Aerophysics of double-skin facades: simulation-based determination of pressure coefficients for multi-story double-skin facades

PDF

Download

Open document

Transcript (if available)

Content AEROPHYSICS OF DOUBLE-SKIN FACADES
SIMULATION-BASED DETERMINATION OF PRESSURE COEFFICIENTS
Copyright 2015
FOR MULTI-STORY DOUBLE-SKIN FACADES
by
Jeffrey Craig Vaglio
A Dissertation Presented to the
FACULTY OF THE SCHOOL OF ARCHITECTURE
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY IN ARCHITECTURE
December 2015
Jeffrey Craig Vaglio
Acknowledgements
The research product covered in the following pages is more than the outcome of the author's
curiosity and investigations and would not be possible without the technical aegis,
encouragement and support of many generous individuals. Gratitude is due for many, including:
• Professor Douglas Noble, FAIA, Ph.D. for his welcoming of me into the doctoral program,
continued advisement, selfless backing and thorough review of the written work
• Dissertation committee members - Professors Marc Schiler, LC, FASES; Anders
Carlson, SE, Ph.D.; and H.L. Wong, Ph.D. - for the time they've invested in listening and
advising me over the years on this dissertation effort
• Fellow Ph.D. in Architecture cohort members at the University of Southern California
School of Architecture
• The University of Southern California, Graduate School, Oakley Fellowship
• Enclos Corp and the Enclos Advanced Technology Studio: the support of academic
endeavor alongside professional practice is a testament to the company's support for
advancing the future knowledge of the facade industry as a whole
• Academic and professional colleague Mic Patterson for his reassurance, intellectual
debate and positive illustration of life-long learning
• All the family, friends and loved ones who provided support throughout my decade plus of
higher-learning, especially my parents, siblings and - last but not least - my wife, who all
know more about double-skin facades than they ever wished.
To all of you my most sincere thanks,
Jeffrey Craig Vaglio
~11
1
1/r
VI
ii
Contents
Acknowledgements ......................................................................................................................... ii
List of Tables .................................................................................................................................. xii
List of Figures ................................................................................................................................ xiii
Abbreviations ............................................................................................................................... xxix
Abstract ........................................................................................................................................ xxx
Preface ........................................................................................................................................ xxxi
1. Introduction to Structural Considerations for Double-Skin Facades .......................................... 1
1.1 Wind Pressure Distribution Characteristics in a Double-Skin Facade 4
1.1.1 Single Surface Distribution .................................................................................... 4
1.1.2 Multi-Layered Distribution ..................................................................................... 5
1.1.3 Multi-Layered Load Sharing .................................................................................. 5
1.1.4 Importance of Pressure Coefficients in Double-Skin Facade Design ................... 7
1.2 How Current Standards Address Wind Load on Double-Skin Facades 11
1.2.1 USCodes-ASCE/SEl ........................................................................................ 11
1.2.2 EUROCODE ........................................................................................................ 12
1.2.3 Glass Codes ........................................................................................................ 13
1.2.4 Limitations to Codes ............................................................................................ 13
1.3 Current Engineering Practice to Determine Pressure Coefficients 14
1.3.1 Full-Scale Tests .................................................................................................. 14
1.3.2 WindTunnelTests .............................................................................................. 15
1.3.3 Computational Fluid Dynamics ........................................................................... 19
1.3.4 Databases ........................................................................................................... 22
1.4 New Approaches to Dynamic Wind-Loading on Double-Skin Facade Structures 23
1.4.1 Fluid-Structure Interaction ................................................................................... 23
1.4.2 Multiphysics Coupling .......................................................................................... 23
1.5 Definition of Terms 25
1.5.1
1.5.2
1.5.3
1.5.4
1.5.5
1.5.6
1.5.7
1.5.8
1.5.9
1.5.10
1.5.11
Double-Skin Facade (DSF) ................................................................................. 25
Transparent Ventilated Facade (TVF) ................................................................. 25
Advanced Integrated Facade (AIF) ..................................................................... 25
Intelligent Facade ................................................................................................ 25
Building Energy Simulation (BES) ....................................................................... 26
Aerophysics ......................................................................................................... 26
Cavity Partitioning ............................................................................................... 26
Ventilation Type ................................................................................................... 26
Ventilation Mode .................................................................................................. 26
Fluid-Structure Interaction (FSI) .......................................................................... 26
Atmospheric Boundary Layer (ABL) .................................................................... 27
iii
1.5.12 Wind Pressure Coefficients (Cp) .......................................................................... 27
1.5.13 Mean Pressure .................................................................................................... 27
1.5.14 Peak Pressure ..................................................................................................... 28
1.5.15 Net Pressure Coefficients (Cp,nerl ........................................................................ 28
1.5.16 Outer Skin Pressure Coefficient Ratio ................................................................ 28
1.5.17 Opening Area Ratio ............................................................................................. 28
1. 6 Hypotheses 28
1. 7 Scope of Study 29
1. 7.1 Multi-Story Configuration with Outdoor Air Curtain ............................................. 29
1. 7.2 Description of Six Airflow Intake Configurations ................................................. 30
1. 7.3 Description of Four Airflow Exhaust Configurations ............................................ 31
1.8 Research Methodology for Evaluating 32
1.8.1 Calibration of a Virtual Simulation ....................................................................... 32
1.8.2 Steady-State Analysis ......................................................................................... 33
1.8.3 Transient Analysis ............................................................................................... 33
1.8.4 Comparative Studies ........................................................................................... 33
1.9 Organization of Dissertation 34
Chapter 1 Endnotes 36
2. Background Research in Modeling Pressure on Double-Skin Facades .................................. 37
2.1 Overview of DSF Modeling Focuses 37
2.2 Research Related to Pressure Distribution Modeling 37
2.2.1 Single-Skin Facade Pressure Distribution ........................................................... 38
2.2.2 Pressure Equalized Rain-Screen (PER) ............................................................. 40
2.3 Double-Skin Facade Wind Interaction 44
2.3.1 Multi-Story Pressure Distribution ......................................................................... 45
2.3.2 Multi-Story Load Sharing ..................................................................................... 56
2.3.3 Box-Window Load Sharing .................................................................................. 60
2.3.4 Box-Window Pressure Distribution (Field Measurement) ................................... 64
2.4 Discussion and Correlations Amongst DSF Wind Engineering Research 69
2.4.1 Modes of Evaluation ............................................................................................ 69
2.4.2 The Unsheltered Tower or Single-Skin Facade as a Baseline ........................... 71
2.4.3 Effects of Sealing or Communicating Between Adjacent Building Faces ........... 72
2.4.4 Influence of Opening Area Ratio ......................................................................... 73
2.4.5 Load Reduction ................................................................................................... 73
2.4.6 Cavity Pressure ................................................................................................... 7 4
2.4.7 Peak Pressures Compared to Mean Coefficients of Pressure ............................ 75
2.4.8 Net Pressure Coefficient ..................................................................................... 75
2.5 Implementing Lessons Learned from Prior Research 76
Chapter 2 Endnotes 77
iv
3. Overview of Double-Skin Facade Applications ........................................................................ 80
3.1 A Clear Definition of Double-Skin Facades 82
3.1.1 Review of Common Terminology ........................................................................ 82
3.1.2 Review of Researchers' Definitions .................................................................... 84
3.1.3 Definition for the Purposes of This Research ..................................................... 89
3.2 Typology and Classification Hierarchy 89
3.2.1 Ventilation Type ................................................................................................... 91
3.2.2 Ventilation Mode .................................................................................................. 91
3.2.3 Cavity Partitioning ............................................................................................... 93
3.2.4 Solar Control Strategy ....................................................................................... 113
3.3 Benefits 114
3.4 Drawbacks 120
3.5 Planning and Feasibility 123
Chapter 3 Endnotes 127
4. Emergence of Double-Skin Facades 129
4.1 Historic Evolution of Double-Skin Facades 129
4.1.1 Environmental Origins of Transparency ............................................................ 130
4.1.2 Industrial Revolution .......................................................................................... 132
4.1.3 European Origins .............................................................................................. 133
4.1.4 Modern Architecture and Glass ......................................................................... 137
4.1.5 Ventilated Facade ............................................................................................. 141
4.1.6 Environmental Awareness ................................................................................ 143
4.1.7 Advent of the Intelligent Skin ............................................................................. 146
4.1.8 Sustainability and the Built Environment ........................................................... 149
4.2 Global Emergence of Double-Skin Facades 156
4.2.1 Occidental Chemical Center ............................................................................. 156
4.2.2 Super Energy Conservation Building ................................................................ 158
4.2.3 Briarcliff House .................................................................................................. 159
4.2.4 Caixa Geral de Dep6sitos, Av. Da RepOblica ................................................... 160
4.2.5 EAL, School of Architecture of Lyon ................................................................. 160
4.3 Europe Takes Leadership 161
4.3.1 Business Promotion Centre .............................................................................. 162
4.3.2 Bibliotheque nationale de France ...................................................................... 164
4.3.3 RWE Headquarters ........................................................................................... 165
4.3.4 Stadttor Dosseldorf (City Gate) ......................................................................... 167
4.3.5 Deutsche Post ................................................................................................... 169
4.3.6 30 St Mary's Axe .............................................................................................. 171
4.3.7 KIW Westarkade ............................................................................................... 173
4.3.8 Roche Diagnostics International AG ................................................................. 175
4.3.9 London Bridge Tower - The Shard .................................................................... 177
v
4.4 Modern Proliferation Outside of Europe 179
4.4.1 Aurora Place ...................................................................................................... 180
4.4.2 Telus William Farrell Building ............................................................................ 182
4.4.3 Sendai Mediatheque ......................................................................................... 184
4.4.4 Manitoba Hydro Place ....................................................................................... 186
4.4.5 1 Bligh Street ..................................................................................................... 188
4.4.6 Shanghai Tower ................................................................................................ 190
4.5 Typological Trends in Global Applications of Double-Skin Facades 192
4.5.1 Construction Evolution ...................................................................................... 193
4.5.2 Programmatic Application ................................................................................. 193
4.5.3 Geographic Distribution ..................................................................................... 193
4.5.4 Typological Distribution ..................................................................................... 196
4.5.5 Global Lessons .................................................................................................. 197
Chapter 4 Endnotes 201
5. Double-Skin Facades in the United States ............................................................................. 203
5.1 Barriers to Implementation 203
5.2 Development Model and Construction Industry 206
5.3 Double-Skin Facade Applications in the United States 207
5.3.1 Yazaki North America ....................................................................................... 209
5.3.2 Seattle Justice Center ....................................................................................... 211
5.3.3 Loyola University Chicago Richard J. Klarcheck Information Commons .......... 213
5.3.4 Cambridge Public Library .................................................................................. 215
5.3.5 USC Eli and Edith Broad Center ....................................................................... 217
5.3.6 New York Presbyterian Hospital ....................................................................... 221
5.3.7 University of Baltimore Angelos Law School ..................................................... 224
5.3.8 Cedar-Sinai Advanced Health Sciences Pavilion ............................................. 226
5.3.9 Columbia University Jerome L. Greene Science Center .................................. 228
5.4 Emerging Trends of Double-Skin Facades in the United States 231
5.4.1 Divergence of Scale .......................................................................................... 231
5.4.2 Structural Aesthetic ........................................................................................... 232
5.4.3 Life-Cycle Assessment ...................................................................................... 232
5.4.4 Post Occupancy Evaluation .............................................................................. 233
5.4.5 Adaptive Re-Use I Retrofit ................................................................................ 233
5.4.6 Translucency ..................................................................................................... 233
5.4.7 Sealed, No Access, Pressurized Box-Windows ................................................ 235
5.4.8 Increased Permeability ...................................................................................... 236
5.5 Design Considerations Going Forward 237
5.5.1 Evaluation and Simulation ................................................................................. 237
5.5.2 Sensibility and Marketability .............................................................................. 238
5.5.3 Integration and Optimization ............................................................................. 240
Chapter 5 Endnotes 241
vi
6. Methodology ...................................................................................................... 242
6.1 Hypotheses 242
6.2 Objectives 242
6.3 Research Design 244
6.3.1 Identification of Problem .................................................................................... 244
6.3.2 Development of an Analytical Prototype ........................................................... 245
6.3.3 Calibration of Analytical Process ....................................................................... 245
6.3.4 Simulations of Multi-Story Double-Skin Facades .............................................. 246
6.3.5 Analysis of Results ............................................................................................ 247
6.3.6 Reflect on the Role of CFO ............................................................................... 249
6.3.7 Develop Guidelines ........................................................................................... 249
6.4 Procedures 251
6.5 Data Analysis 252
Chapter 6 Endnotes 253
7. Development of a Multi-Story Prototype ................................................................................. 254
7.1 Identifying the Most Common Configuration 256
7.2 Dimensional Trends of Multi-Story Double-Skin Facades in the United States 257
7.3 Multi-Story Double-Skin Facade Prototype 259
7.4 Airflow and Structural Variables 260
7.4.1 Airflow Intake Configuration .............................................................................. 260
7.4.2 Airflow Exhaust Configuration ........................................................................... 263
7.4.3 Permeability ...................................................................................................... 264
7.5 Summary of Analytical Prototype 265
Chapter 7 Endnotes 266
8. Calibration of CFO Model Using Multi-Story DSF Wind Tunnel Test Data ............................ 267
8.1 Introduction 267
8.2 Description of Wind Tunnel Models 268
8.2.1 Dimensional Configuration ................................................................................ 268
8.2.2 Boundary Conditions ......................................................................................... 268
8.2.3 Velocity Profile ................................................................................................... 269
8.2.4 Test Configurations ........................................................................................... 269
8.3 Considerations for Turbulent Flow 270
8.3.1 Turbulence Intensity .......................................................................................... 270
8.3.2 Roughness Classification and Length ............................................................... 272
vii
8.3.3 Equivalent Sand-grain Rougness Height .......................................................... 27 4
8.4 CFO Model Description - Simple Cube 275
8.41 Introduction ........................................................................................................ 275
8.42 Test Configuration ............................................................................................. 279
8.43 Comparison of CFO Results to Wind Tunnel Data - Simple Cube ................... 279
8.4.4 Two-Dimensional Steady-State Simulation ....................................................... 285
8.5 CFO Model Description - Unsheltered Tower 287
8.5.1 Introduction ........................................................................................................ 287
8.5.2 Test Configuration ............................................................................................. 287
8.5.3 Comparison of CFO Results to Wind Tunnel Data - Unsheltered Tower ......... 288
8.6 CFO Model Description - Sheltered Tower 291
8.6.1 Introduction ........................................................................................................ 291
8.6.2 Test Configuration ............................................................................................. 291
8.6.3 Comparison of CFO Results to Wind Tunnel Data - Sheltered Tower A ......... 292
8.7 CFO Model Description - Laterally Sealed Sheltered Tower 295
8.7.1 Introduction ........................................................................................................ 295
8.7.2 Test Configuration ............................................................................................. 295
8.7.3 Comparison of CFO Results to Wind Tunnel Data - Sheltered Tower D ......... 296
8.8 Evaluation Criteria moving Forward 298
Chapter 8 Endnotes 300
9. Steady-State Simulation to Determine Wind Pressure Distribution on Multi-Story DSF ........ 302
9.1 Introduction to Steady-State Simulations 302
9.2 Description of Experimental Model 302
9.2.1 Geometry ........................................................................................................... 303
9.2.2 Boundary Conditions ......................................................................................... 305
9.2.3 Loading .............................................................................................................. 306
9.2.4 Turbulence Model Selection .............................................................................. 306
9.2.5 Data Reporting Locations .................................................................................. 307
9.3 Single-Skin Condition 309
9.3.1 Single-Skin Condition ........................................................................................ 309
9.4 Face Inlet 311
9.4.1 Face - Forward ................................................................................................. 311
9.4.2 Face - Through ................................................................................................. 313
9.4.3 Face - Return .................................................................................................... 315
9.5 Trench Inlet 317
9.5.1 Trench - Forward .............................................................................................. 317
9.5.2 Trench - Through .............................................................................................. 319
9.5.3 Trench - Return ................................................................................................ 321
9.6 Raised Inlet 323
9.6.1 Raised - Forward .............................................................................................. 323
9.6.2 Raised - Through .............................................................................................. 325
viii
9.6.3 Raised - Return ................................................................................................ 327
9.7 Shingled Inlet 329
9.7.1 Shingled - Forward ........................................................................................... 329
9.7.2 Shingled - Through ........................................................................................... 331
9.7.3 Shingled - Return .............................................................................................. 333
9.8 Discussion of Results 335
9.8.1 Comparison of Inlet Configurations ................................................................... 335
9.8.2 Comparison of Exhaust Configurations ............................................................ 350
9.9 Conclusions Regarding Steady-State Analysis 358
9.9.1 Summary of Findings from 30 Steady-State Analysis ...................................... 358
9.9.2 Comparison of 30 Steady-State to 20 Transient Analysis ............................... 359
Chapter 9 Endnotes 360
10. Discussion and Conclusions ................................................................................................. 361
10.1 Experimental Summary 362
10.2 Research Limitations 362
10.2.1
10.2.2
10.2.3
10.2.4
10.2.5
10.2.6
Geometric ........................................................................................................ 362
Approaching Flow ........................................................................................... 362
Airflow Configurations and Openings .............................................................. 363
Cavity Contents ............................................................................................... 364
Structure .......................................................................................................... 364
Analytical Method ............................................................................................ 364
10.3 Recommended Design Guidelines for Multi-Story Double-Skin Facades 365
10.3.1 Preliminary Planning Phase ............................................................................ 365
10.3.2 Schematic Design ........................................................................................... 367
10.3.3 Design Development ....................................................................................... 370
10.3.4 Construction Documents ................................................................................. 372
10.3.5 On the Construction Site ................................................................................. 374
10.3.6 Post-Occupancy .............................................................................................. 376
10.3.7 Responsibility Matrix ....................................................................................... 377
10.3.8 Summary of Guidelines ................................................................................... 380
10.4 Recommendations for the Role of CFO in the Design Process 381
10.5 Recommended Pressure Coefficient Determination Methodology 383
10.6 Conclusions 389
10. 7 Suggestions for Future Research 391
Chapter 10 Endnotes 393
Bibliography ................................................................................................................................. 394
ix
Appendices
A International Double-Skin Facade Case Studies ................................................................... 415
A.1 Occidental Chemical Center ................................................................................ 415
A.2 Super Energy Conservation Building ................................................................... 417
A.3 Briarcliff House ..................................................................................................... 418
A.4 Caixa Geral de Dep6sitos, Av. Da RepOblica ...................................................... 419
A.5 EAL, School of Architecture of Lyon .................................................................... 419
A.6 Business Promotion Centre ................................................................................. 420
A.7 Bibliotheque nationale de France ........................................................................ 422
A.8 Gatz Office Building ............................................................................................. 423
A.9 Commerzbank ..................................................................................................... 425
A 10 RWE Headquarters .............................................................................................. 426
A 11 Stadttor Dosseldorf (City Gate) ............................................................................ 429
A 12 debis Headquarters ............................................................................................. 431
A 13 UCB Centre .......................................................................................................... 433
A 14 GSW Headquarters Tower .................................................................................. 434
A 15 Deutsche Messe AG (Trade-Fair Tower) ............................................................. 436
A.16 ARAG ................................................................................................................... 438
A 17 Administration Building in Kronberg ..................................................................... 440
A 18 Aurora Place ........................................................................................................ 442
A 19 Telus William Farrell Building .............................................................................. 444
A.20 Sendai Mediatheque ............................................................................................ 446
A.21 Coeur Defense ..................................................................................................... 448
A.22 Munich Re Offices ................................................................................................ 449
A.23 ING House ........................................................................................................... 451
A.24 Deutsche Post ...................................................................................................... 452
A.25 Kista Science Tower ............................................................................................ 454
A 26 30 St Mary's Axe ................................................................................................. 455
A.27 TiFS lngegneria ................................................................................................... 457
A.28 AZ Medienhaus .................................................................................................... 459
A.29 LTD_1 Office Building .......................................................................................... 461
A.30 Shanghai Tobacco (Group) Corp ......................................................................... 463
A.31 Kraanspoor .......................................................................................................... 465
A.32 Manitoba Hydro Place .......................................................................................... 467
A.33 KfWWestarkade .................................................................................................. 469
A.34 1 Bligh Street ....................................................................................................... 471
A.35 Roche Diagnostics International AG .................................................................... 473
A.36 London Bridge Tower - The Shard ...................................................................... 475
A.37 Pearl River Tower ................................................................................................ 477
A.38 Cleveland Clinic ................................................................................................... 479
A.39 Shanghai Tower ................................................................................................... 481
A.40 Twofour54° Zone 1 ............................................................................................... 483
Appendix A Endnotes 485
B. Geographic Distribution of Double-Skin Facade Applications per Perino (2007) .................. 486
x
C. United States Double-Skin Facade Case Studies ................................................................. 488
C.1 Warren Petroleum Executive Headquarters ........................................................ 489
C.2 Occidental Chemical Center ................................................................................ 491
C.3 Yazaki North America .......................................................................................... 493
C.4 Levine Hall, Pennsylvania School of Engineering ............................................... 495
C.5 Seattle Justice Center ......................................................................................... 497
C.6 Manulife Financial (601 Congress Street) ........................................................... 499
C. 7 Genzyme Center ................................................................................................. 501
C.8 Foundry Square ................................................................................................... 503
C.9 University of Massachusetts Medical School ...................................................... 505
C 10 Loblolly House ..................................................................................................... 507
C.11 University of Michigan Biomedical Science Research Building .......................... 509
C.12 Museum of Contemporary Art ............................................................................. 511
C.13 ComcastTower ................................................................................................... 513
C.14 Riverhouse - One River Terrace ......................................................................... 515
C.15 Loyola University Chicago Richard J. Klarcheck Information Commons ............ 517
C.16 The Walter Cronkite School of Journalism .......................................................... 519
C 17 Whatcom Building ................................................................................................ 521
C.18 Art Institute of Chicago - Modern Wing ............................................................... 523
C.19 Cleveland Museum of Art - East Wing ................................................................ 525
C.20 Cambridge Public Library .................................................................................... 527
C.21 100 Park Avenue ................................................................................................. 529
C.22 USC Eli and Edith Broad Center ......................................................................... 531
C.23 New York Presbyterian Hospital .......................................................................... 535
C.24 University of Oregon Jaqua Center ..................................................................... 538
C.25 New World Center ............................................................................................... 541
C.26 University of Baltimore Angelos Law School ....................................................... 543
C.27 Harvard Business School Tata Hall ..................................................................... 545
C.28 Cedar-Sinai Advanced Health Sciences Pavilion ................................................ 547
C.29 Weill Cornell Medical College - Beller Research Building .................................. 549
C.30 University of Kansas School of Architecture - The Forum .................................. 551
C.31 Northwestern University Recital Hall ................................................................... 553
C.32 Rodino Federal Office Building Moderniziation ................................................... 554
C.33 AJ. Celebreeze Federal Building ........................................................................ 555
C.34 PNC Plaza Tower ................................................................................................ 556
C.35 Columbia University Jerome L. Greene Science Center ..................................... 559
C.36 New Stanford Hospital ......................................................................................... 562
Appendix C Endnotes 564
D. Supplemental Documentation for Calibration Models ........................................................... 565
E. Summary of Preliminary 2D Transient Studies ...................................................................... 570
xi
List of Tables
Table 1-1: Wind Loading Code Wind Speeds and Pressures ................................................... 13
Table 5-1: Double-Skin Facades in the United States ............................................................ 243
Table 5-2: Forthcoming Double-Skin Facades in the United States ....................................... 244
Table 6-1: Summary of Research Methodology and Procedures ........................................... 250
Table 7-1: Summary of Double-Skin Facades in the United States ........................................ 253
Table 7-2: Dimensional Summary of Multi-Story DSFs in the United States .......................... 256
Table 8-1: Terrain Type, Roughness and Surface Drag Coefficient (Holmes 2001) .............. 270
Table 8-2 Davenport Classification of Effective Terrain Roughness (ASCE/SEI 2010) ......... 271
Table 10-1: Responsibility Matrix of a Sample Construction Project ........................................ 378
Table B-1: DSF/AIF Geographical Distribution ........................................................................ 486
Table D-1: CP Profile Values and Coordinates Interpretation for Comparative Analysis ......... 569
xii
List of Figures
Figure 1-1: Double-skin facade multi-story configuration diagram ............................................... 3
Figure 1-2: CP distribution on single envelope and double-skin inner layer ................................. 5
Figure 1-3: Components of a boundary-layer wind tunnel test section ...................................... 15
Figure 1-4: Shanghai Tower scaled wind tunnel model .............................................................. 16
Figure 1-5: Comparison of CP data results of different wind tunnel and full-scale tests ............. 18
Figure 1-6: CFO pressure distribution (Cp) diagrams for wind normal to cube and wind 45°
to cube ...................................................................................................................... 19
Figure 1-7: 20 FSI model of a simple DSF subjected to a sinusoidal 3-second gust velocity ... 24
Figure 1-8: Concept diagram of multi-story DSF with and outdoor air curtain ........................... 29
Figure 1-9: Airflow intake configurations occurring near the ground plane ................................ 30
Figure 1-10: Airflow exhaust configurations occurring near the roof plane .................................. 31
Figure 2-1: CFO pressure distribution (Cp) diagram for wind normal to cube ............................ 39
Figure 2-2: Pressure reduction factor vs. ventilation opening area ratio on curtain wall
spandrel panel .......................................................................................................... 41
Figure 2-3: Influence of gap flow resistance on the net pressure coefficient ............................. 43
Figure 2-4: Influence of oncoming flow on net pressure coefficient derived from the peak
factor approach ........................................................................................................ 43
Figure 2-5: Pressure tap locations, wind incidences and DSF layouts evaluated in wind
tunnel testing ............................................................................................................ 46
Figure 2-6: CP values measured on models for frontal wind incidences for single facade ......... 47
Figure 2-7: CP values measured on models for frontal wind incidences for open double-skin
with 0.8 m (equivalent) depth ................................................................................... 48
Figure 2-8: CP values measured on models for frontal wind incidences for sealed double-skin
facade with 1.6 m (equivalent) depth ....................................................................... 49
Figure 2-9: Pressure and wind coefficients on the outside skin and the cavity of a building-
high double-skin facade ....................................................................................... 51-52
Figure 2-1 O: Net CP distribution for four double-skin facade layouts (A, B, C and D) at a wind
incidence of 0° .......................................................................................................... 53
Figure 2-11: CP distribution on single envelope and a laterally sealed double-skin inner layer
with a wind incidence of 0° and cavity depth of 0.8 m ............................................. 55
xiii
Figure 2-12: Wind flow on a permeable facade schematic diagram ............................................ 57
Figure 2-13: Code proposal for wind loads on a back-vented facade system ............................. 58
Figure 2-14: Wind loads for DSF with sealed exterior and permeable wall .................................. 59
Figure 2-15: Wind pressure test apparatus: front elevation and vertical section ......................... 61
Figure 2-16: Box-window DSF opening position configurations ................................................... 62
Figure 2-17: Box-window distribution load ratio for sealed cavity ................................................ 63
Figure 2-18: Box-window distribution load ratio for cavity with openings to outside .................... 63
Figure 2-19: Full-scale experimental model configuration with four double-skin facade box-
windows; two windows on each of the west and east elevations respectively ........ 65
Figure 2-20: Records of wind speed and wind direction at the experimental field ....................... 66
Figure 2-21: Records of pressure on outer and inner skin and pressure inside testing room ..... 67
Figure 2-22: Peak wind force coefficients on a single-skin, and the outer and inner skins of a
windward double-skin box window ........................................................................... 68
Figure 3-1: Double-skin facades classification scheme ............................................................. 90
Figure 3-2: Five modes of double-skin facade ventilation .......................................................... 91
Figure 3-3: Conceptual diagrams of cavity partitioning .............................................................. 94
Figure 3-4: Conceptual diagram of a Baffle configuration .......................................................... 95
Figure 3-5: Conceptual diagram of a Box-Window configuration ............................................... 97
Figure 3-6: Conceptual diagram of an Alternating Facade configuration ................................... 99
Figure 3-7: Conceptual diagram of a Shaft-Box configuration ................................................. 101
Figure 3-8: Conceptual diagram of a Story-Height (Corridor) configuration ............................. 103
Figure 3-9: Conceptual diagram of a Story-Height (Juxtaposed Modules) configuration ......... 105
Figure 3-10: Conceptual diagram of a Multi-Story configuration ................................................ 107
Figure 3-11: Conceptual diagram of a Multi-Story (Shingled) configuration .............................. 109
Figure 3-12: Conceptual diagram of a Multi-Story (Controllable) configuration ......................... 111
Figure 4-1: Box-window after restoration (kastenfenster nach der restaurierung) ................... 130
Figure 4-2: Palm House at Bicton Gardens (1825), Devon, England ...................................... 131
Figure 4-3 Paxton's Crystal Palace (1851 ), London, England ................................................ 133
xiv
Figure 4-4 Steiff Factory (1903), Giengen, Germany .............................................................. 134
Figure 4-5: Fagus Works shoe factory (1911-1925), Saxony, Germany, Gropius & Meyer .... 134
Figure 4-6 Post Office Savings Bank (1904-1912), Vienna, Austria, otto Wagner ................. 135
Figure 4-7: Fondation Beyeler Museum (1997), Riehen, Switzerland, Renzo Piano
Building Workshop ................................................................................................. 136
Figure 4-8 Halladie Building (1918), San Francisco, Willis Polk ............................................. 138
Figure 4-9: Le Mur Neutralisant for La Cite de Refuge (1930), Paris, Le Corbusier ................ 139
Figure 4-10 Natural History Faculty Library (1964-1968), Cambridge, James Stirling .............. 142
Figure 4-11 House of Culture (1968-1971 ), Stockholm, Celsin & Henriksson .......................... 142
Figure 4-12 PolyvalentWall (1981), Mike Davies, Richard Rogers Partnership ....................... 146
Figure 4-13: Lloyd's Insurance Company (1976-1986), London, Richard Rogers Partnership . 148
Figure 4-14 Institute du Arab Monde (1987), Paris, Jean Nouvel ............................................. 148
Figure 4-15 Green Building (1990), Future Systems ................................................................. 150
Figure 4-16: Annual carbon emissions per capita for the 20 most populated countries ............ 151
Figure 4-17 Eden Project (2001) ............................................................................................... 155
Figure 4-18: Occidental Chemical Center - Overall view ........................................................... 156
Figure 4-19: Super Energy Conservation Building - Overall view ............................................. 158
Figure 4-20: Briarcliff House -Overall view ............................................................................... 159
Figure 4-21: Caixa Geral de Dep6sitos - Overall street view .................................................... 160
Figure 4-22: EAL, School of Architecture of Lyon - Overall street view .................................... 160
Figure 4-23: Business Promotion Centre - Overall view ............................................................ 162
Figure 4-24: Business Promotion Centre - Cavity drawing and exterior photo ......................... 163
Figure 4-25: Bibliotheque nationale de France - Tower view .................................................... 164
Figure 4-26: RWE Headquarters - Overall view ........................................................................ 165
Figure 4-27: RWE Headquarters - Diagram and photo of box-window with Fish Mouth air
openings ................................................................................................................. 166
Figure 4-28: Stadttor DOsseldorf - Overall view ......................................................................... 167
Figure 4-29: Stadttor DOsseldorf - DSF elevation and cavity ................................................... 168
xv
Figure 4-30: Post Tower-Overall view ..................................................................................... 169
Figure 4-31: Post Tower-Shingled west cavity and flush east facades ................................... 170
Figure 4-32: 30 St Mary's Axe - Overall view ........................................................................... 171
Figure 4-33: 30 St Mary's Axe - Installation of outer skin and DSF cavity space ..................... 172
Figure 4-34: KIW Westarkade - Overall view ............................................................................. 173
Figure 4-35: KIW - Sawtooth DSF construction and ventilation diagram .................................. 17 4
Figure 4-36: Roche Diagnostics International AG -Overall view .............................................. 175
Figure 4-37: Roche Diagnostics International AG -Corner elevation and ventilation diagram. 176
Figure 4-38: The Shard - Overall view ....................................................................................... 177
Figure 4-39: The Shard - DSF elevation at corner and intermediate cavity .............................. 178
Figure 4-40: Aurora Place - Overall view ................................................................................... 180
Figure 4-41: Aurora Place - Louvered outer skin of winter garden DSF ................................... 181
Figure 4-42: Telus - Overall view ............................................................................................... 182
Figure 4-43: Telus -West elevation and DSF airflow diagram .................................................. 183
Figure 4-44: Sendai Mediatheque - Overall view ...................................................................... 184
Figure 4-45: Sendai Mediatheque - DSF cavity drawing and constructed structure ................. 185
Figure 4-46: Manitoba Hydro Place - Overall view .................................................................... 186
Figure 4-47: Manitoba Hydro Place - Northwest elevation and operable units on DSF ............ 187
Figure 4-48 1 Bligh Street - Overall view .................................................................................. 188
Figure 4-49: 1 Bligh Street-View through DSF and intake/extract louvers .............................. 189
Figure 4-50: Shanghai Tower-Overall view ............................................................................. 190
Figure 4-51: Shanghai Tower-Atrium DSF rendering and construction photo ........................ 191
Figure 4-52: DSF/AIF -Year of construction or retrofit .............................................................. 194
Figure 4-53 DSF/AIF -Geographic distribution map ................................................................ 195
Figure 4-54 Facade refurbishment in Stuttgart by Behnisch Sabatke Behnisch (1996) ........... 199
Figure 5-1: Yazaki North America - Exterior view of atria ....................................................... 209
Figure 5-2: Yazaki North America -Atria mast truss and overall building ............................... 210
xvi
Figure 5-3: Seattle Justice Center - Overall street view .......................................................... 211
Figure 5-4: Seattle Justice Center - DSF elevation and cavity space ..................................... 212
Figure 5-5: Klarcheck Information Commons - Horizontal cable and DSF end condition ....... 213
Figure 5-6: Klarcheck Information Commons - Elevation and cavity photo ............................. 214
Figure 5-7: Cambridge Public Library- Exterior DSF view ...................................................... 215
Figure 5-8: Cambridge Public Library- Overall exterior and interior views ............................. 216
Figure 5-9 USC CIRM Center - DSF cavity view .................................................................... 217
Figure 5-10: USC CIRM Center - Overall exterior view during construction ............................. 218
Figure 5-11: USC CIRM Center - Exterior and interior views of translucent frit ........................ 219
Figure 5-12: USC CIRM Center - Installation of vertical cables and glass ................................ 220
Figure 5-13: New York Presbyterian Hospital - Exterior DSF view of air inlet ........................... 221
Figure 5-14: New York Presbyterian Hospital - Overall view ..................................................... 222
Figure 5-15: New York Presbyterian Hospital - Detail view ....................................................... 223
Figure 5-16: Angelos Law School - Overall view ....................................................................... 224
Figure 5-17: Angelos Law School -Baffle DSF and external blinds .......................................... 225
Figure 5-18: Cedar-Sinai Advanced Health Sciences Pavilion-Overall view ........................... 226
Figure 5-19: Cedar-Sinai Advanced Health Sciences Pavilion -DSF detail drawing & photo .. 227
Figure 5-20: Jerome L. Greene Science Center - Overall construction progress photo ........... 228
Figure 5-21: Jerome L. Greene Science Center - Train adjacent to the east and southeast
corner of Wall Type 3 DSF ..................................................................................... 229
Figure 5-22: Jerome L. Greene Science Center - DSF's indoor air inlet at third floor and interior
of DSF's cavity at the top exhaust .......................................................................... 230
Figure 5-23: Nelson Atkins Museum - Translucent lenses with thermal buffer wall .................. 234
Figure 6-1: Overview of research methodology ....................................................................... 248
Figure 7-1: Typological summary of identified buildings with DSFs in the United States ........ 256
Figure 7-2: Multi-story double-skin facade prototype variations: raised/entry and trench ........ 259
Figure 8-1: Simulated atmospheric boundary layer wind tunnel configuration ......................... 268
Figure 8-2 Velocity profile at inlet plane .................................................................................. 269
xvii
Figure 8-3: Turbulence intensities of wind tunnels participation in the Fritz et al. (2008)
comparison (Simiu 2009); open exposure and suburban exposure ...................... 271
Figure 8-4: Range of surface pressures for a wind direction of 0° as modeled by twelve
different wind tunnels ............................................................................................. 276
Figure 8-5: Maximum pressure coefficients on the windward face for a wind direction of 0°
as modeled by twelve different wind tunnels ......................................................... 277
Figure 8-6: Comparison of Silsoe Cube (full-scale and wind tunnel) surface pressures to
the aforementioned range of twelve different wind tunnel results .......................... 278
Figure 8-7: CP profile comparison; a varies (U
0
=7mis, IT,inler=0.15, a:varies, z
0
=1.0 m) ......... 280
Figure 8-8: Approaching IT profiles (U
0
=7mis, IT,inler=0.15, a:varies, z
0
=1.0 m) ..................... 280
Figure 8-9: CP profile comparison; z
0
varies (U
0
=7mis, IT,inler=0.15, a=0.22, z
0
:varies) .......... 282
Figure 8-10: Approaching IT profiles (U
0
=7mis, IT,inler=0.15, a=0.22, z
0
:varies) ....................... 282
Figure 8-11: CP profile comparison; IT,inler varies (U
0
=7 mis, IT,inler:varies, a=0.22, z
0
=1.0 m). 284
Figure 8-12: Approaching IT profiles (U
0
=7mis, IT,inler:varies, a=0.22, z
0
=1.0 m) .................... 284
Figure 8-13: CP profile comparison; z
0
varies (U
0
=10.5 mis, IT,inler=0.15, a=0.18, z
0
:varies) .... 289
Figure 8-14: CP comparison; IT,inler varies (U
0
=10.5 mis, IT,inler:varies, a=0.18, z
0
=2.0m) ....... 289
Figure 8-15: Comparitive elevations of CP distribution for Unsheltered Tower ........................... 290
Figure 8-16: Sheltered Tower (Layout A) configuration and section cut locations ..................... 291
Figure 8-17: CP profiles at mid (simulation) and 1i3 (wind tunnel) points with varied roughness
length (U
0
=10.5 mis, IT,inler=0.15, a=0.18, z
0
:varies) ............................................. 293
Figure 8-18: CP profiles at mid (simulation) and 1i3 (wind tunnel) points with varied inlet
turbulence intensity (U
0
= 10.5 mis, IT,inler: varies, a= 0.18, z
0
= 2.0 m) ............. 293
Figure 8-19: Comparitive building-face elevations of elevations of Cp,int distribution for
Sheltered Tower ..................................................................................................... 294
Figure 8-20: Laterally Sealed Sheltered Tower (Layout D) configuration and section cuts ....... 295
Figure 8-21: CP profiles at mid (simulation) and 1i3 (wind tunnel) points with varied inlet
turbulence intensity (U
0
=10.5 mis, a=0.18, z
0
=20 m, s={0.8 m, 1.6 m}) ............... 296
Figure 8-22: Comparitive building-face elevations of elevations of Cp,int distribution for
Laterally Sealed Sheltered Tower .......................................................................... 297
Figure 8-23: CP profiles at x/a = 0.35 and 0.67 for simulations compared to wind tunnel
studies (U
0
=10.5 mis, IT,inler=0.15, a=0.18, z
0
=2.0 m, s=0.8 m) ........................... 299
xviii
Figure 9-1: Simulated wind tunnel configuration with single-skin building configuration ......... 303
Figure 9-2: Opening configurations for each inlet and exhaust type ........................................ 304
Figure 9-3: Key edge locations of user-controlled, forced mesh subdivision ........................... 305
Figure 9-4: Mesh layers of a Trench-Forward configuration with the exterior skin, mesh
subdivision and interior skin ................................................................................... 307
Figure 9-5a: Vertical reporting locations x/a = {0.50, 0.67, 0.90) ............................................... 308
Figure 9-5b Horizontal reporting locations zih = {0.10, 0.50, 0.67, 0.90) ................................. 308
Figure 9-6: Single-skin configuration ........................................................................................ 309
Figure 9-7: cp.Smgfr elevation, vertical cp.Smgfr profiles and horizontal cp.Smgfr profiles for a
single-skin sealed configuration ............................................................................. 310
Figure 9-8: Face-Forward configuration ................................................................................... 311
Figure 9-9: CP elevations interior and exterior, CP profiles and CP·"" profiles for the Face-
Forward configuration ............................................................................................. 312
Figure 9-1 O: Face- Through configuration ................................................................................... 313
Figure 9-11: CP elevations interior and exterior, CP profiles and CP·"" profiles for the Face-
Through configuration ............................................................................................ 314
Figure 9-12: Face-Return configuration ...................................................................................... 315
Figure 9-13: CP elevations interior and exterior, CP profiles and CP·"" profiles for the Face-
Return configuration ............................................................................................... 316
Figure 9-14: Trench-Forward configuration ................................................................................ 317
Figure 9-15: CP elevations interior and exterior, CP profiles and CP·"" profiles for the Trench-
Forward configuration ............................................................................................. 318
Figure 9-16: Trench-Through configuration ................................................................................ 319
Figure 9-17: CP elevations interior and exterior, CP profiles and CP·"" profiles for the Trench-
Through configuration ............................................................................................ 320
Figure 9-18: Trench-Return configuration .................................................................................. 321
Figure 9-19: CP elevations interior and exterior, CP profiles and CP·"" profiles for the Trench-
Return configuration ............................................................................................... 322
Figure 9-20: Raised-Forward configuration ................................................................................ 323
Figure 9-21: CP elevations interior and exterior, CP profiles and CP·"" profiles for the Raised-
Forward configuration ............................................................................................. 324
Figure 9-22: Raised-Through configuration ................................................................................ 325
xix
Figure 9-23: CP elevations interior and exterior, CP profiles and CP,"" profiles for the Raised
Through configuration ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, 326
Figure 9-24: Raised-Return configuration , , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, 327
Figure 9-25: CP elevations interior and exterior, CP profiles and CP,"" profiles for the Raised
Return configuration ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, 328
Figure 9-26: Shingled-Forward configuration ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, 329
Figure 9-27: CP elevations interior and exterior, CP profiles and CP,"" profiles for the Shingled
Forward configuration,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 330
Figure 9-28 Shingled-Through configuration ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, 331
Figure 9-29: CP elevations interior and exterior, CP profiles and CP,"" profiles for the Shingled
Through configuration ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, 332
Figure 9-30: Shingled-Return configuration,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 333
Figure 9-31: CP elevations interior and exterior, CP profiles and CP,"" profiles for the Shingled
Return configuration ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, 334
Figure 9-32: Comparison of CP values at mid-span (x/a = 0,5) for Face inlet configurations
with varied exhaust configurations,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 337
Figure 9-33: Comparison of CP values at mid-span (x/a = 0,5) for Trench inlet configurations
with varied exhaust configurations,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 337
Figure 9-34: Comparison of CP values at mid-span (x/a = 0,5) for Raised inlet configurations
with varied exhaust configurations,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 338
Figure 9-35: Comparison of CP values at mid-span (x/a = 0,5) for Shingled inlet configurations
with varied exhaust configurations,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 338
Figure 9-36: Comparison of outer skins' pressure ratio vertical profiles for Face inlet
configurations with varied exhaust configurations ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 341
Figure 9-37: Comparison of outer skins' pressure ratio vertical profiles for Trench inlet
configurations with varied exhaust configurations ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 341
Figure 9-38: Comparison of outer skins' pressure ratio vertical profiles for Raised inlet
configurations with varied exhaust configurations ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 342
Figure 9-39: Comparison of outer skins' pressure ratio vertical profiles for Shingled inlet
configurations with varied exhaust configurations ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 342
Figure 9-40: Comparison of outer skins' pressure ratio horizontal profiles for Face inlet
configurations with varied exhaust configurations ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 345
Figure 9-41: Comparison of outer skins' pressure ratio horizontal profiles for Trench inlet
configurations with varied exhaust configurations ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 345
xx
Figure 9-42: Comparison of outer skins' pressure ratio horizontal profiles for Raised inlet
configurations with varied exhaust configurations ................................................. 346
Figure 9-43: Comparison of outer skins' pressure ratio horizontal profiles for Shingled inlet
configurations with varied exhaust configurations ................................................. 346
Figure 9-44: Comparison of CP·"" values (atxla = {0.50, 0.67, 0.90)) for Face inlet
configurations with varied exhaust configurations ................................................. 348
Figure 9-45: Comparison of CP·"" values (atxla = {0.50, 0.67, 0.90)) for Trench inlet
configurations with varied exhaust configurations ................................................. 348
Figure 9-46: Comparison of CP·"" values (at x/a = {0.50, 0.67, 0.90)) for Raised inlet
configurations with varied exhaust configurations ................................................. 349
Figure 9-47 Comparison of CP·"" values (atxla = {0.50, 0.67, 0.90)) for Shingled inlet
configurations with varied exhaust configurations ................................................. 349
Figure 9-48: Comparison of CP values at mid-span (x/a = 0.50) for Forward exhaust
configurations with varied inlet configurations ....................................................... 350
Figure 9-49: Comparison of CP values at mid-span (x/a = 0.50) for Through exhaust
configurations with varied inlet configurations ....................................................... 351
Figure 9-50: Comparison of CP values at mid-span (x/a = 0.50) for Return exhaust
configurations with varied inlet configurations ....................................................... 351
Figure 9-51: Comparison of outer skins' pressure ratio vertical profiles for Forward exhaust
configurations with varied inlet configurations ....................................................... 352
Figure 9-52: Comparison of outer skins' pressure ratio vertical profiles for Through exhaust
configurations with varied inlet configurations ....................................................... 353
Figure 9-53: Comparison of outer skins' pressure ratio vertical profiles for Return exhaust
configurations with varied inlet configurations ....................................................... 353
Figure 9-54: Comparison of outer skins' pressure ratio horizontal profiles for Forward exhaust
configurations with varied inlet configurations ........................................................ 354
Figure 9-55: Comparison of outer skins' pressure ratio horizontal profiles for Through exhaust
configurations with varied inlet configurations ........................................................ 355
Figure 9-56: Comparison of outer skins' pressure ratio horizontal profiles for Return exhaust
configurations with varied inlet configurations ....................................................... 355
Figure 9-57: Comparison of CP·"" values (atxla = {0.50, 0.67, 0.90)) for Forward exhaust
configurations with varied inlet configurations ....................................................... 356
Figure 9-58: Comparison of CP·"" values (atxla = {0.50, 0.67, 0.90)) for Through exhaust
configurations with varied inlet configurations ....................................................... 357
Figure 9-59: Comparison of CP·"" values (atxla = {0.50, 0.67, 0.90)) for Return exhaust
configurations with varied inlet configurations ....................................................... 357
xxi
Figure 10-1: Relative comparison of acoustic buffering by ventilation mode ............................. 365
Figure 10-2: Relative comparison of cavity pressure volatility by ventilation mode ................... 367
Figure 10-3: Conceptual design construction process related to developing the building skin .. 379
Figure 10-4: Suggested pressure coefficient determination methodology ................................. 383
Figure 10-5: Process of determining procedure for wind load determination ............................. 385
Figure 10-6: Relative comparison of inner skin pressures of twelve multi-story double-skin
facade configurations ............................................................................................. 387
Figure 10-7: Relative comparison of net pressure across outer skin of twelve multi-story
double-skin facade configurations .......................................................................... 387
Figure A-1: Occidental Chemical Center - Overall view ........................................................... 415
Figure A-2: Super Energy Conservation Building - Overall view ............................................. 417
Figure A-3: Briarcliff House - Overall view ............................................................................... 418
Figure A-4: Caixa Geral de Dep6sitos -Overall street view .................................................... 419
Figure A-5: EAL, School of Architecture of Lyon -Overall street view .................................... 419
Figure A-6: Business Promotion Centre - Overall view ............................................................ 420
Figure A-7: Business Promotion Centre - Cavity drawing and exterior photo ......................... 421
Figure A-8: Bibliotheque nationale de France - Tower view .................................................... 422
Figure A-9: Gatz Office Building - Corner view with fans in the cavity ..................................... 423
Figure A-10 Gatz Office Building - DSF section and detail ....................................................... 424
Figure A-11: Gatz Office Building - Entrance elevation ............................................................. 424
Figure A-12: Commerzbank- Overall view ................................................................................ 425
Figure A-13: Commerzbank-DSF airflow diagram and interior photo ...................................... 426
Figure A-14: RWE Headquarters - Overall view ........................................................................ 427
Figure A-15: RWE Headquarters - Diagram and photo of box-window with Fish Mouth air
openings ................................................................................................................. 428
Figure A-16: Stadttor DOsseldorf - Overall view ......................................................................... 429
Figure A-17: Stadttor DOsseldorf - DSF elevation and cavity ................................................... 430
Figure A-18: debis Headquarters - Overall view ........................................................................ 431
xxii
Figure A-19: debis Headquarters - Exterior louvers photo and diagrams .................................. 432
Figure A-20: UCB Centre - Overall view .................................................................................... 433
Figure A-21: GSW Headquarters Tower - Overall view ............................................................. 434
Figure A-22: GSW Headquarters - West facade multi-story DSF configuration ........................ 435
Figure A-23: Trade-Fair Tower - Overall view ............................................................................ 436
Figure A-24: Deutsche Messe AG (Trade-Fair Tower) - DSF story-height corridor cavity and
interior office ........................................................................................................... 437
Figure A-25: ARAG - Overall view ............................................................................................. 438
Figure A-26: ARAG -Air inlet and elevation ............................................................................... 439
Figure A-27: Building in Kronberg - Overall view ....................................................................... 440
Figure A-28: Building in Kronberg - DSF closed and open ........................................................ 441
Figure A-29: Kronberg - Naturally ventilated DSF concept diagram .......................................... 441
Figure A-30: Aurora Place - Overall view ................................................................................... 442
Figure A-31: Aurora Place - Louvered outer skin of winter garden DSF ................................... 443
Figure A-32: Telus - Overall view ............................................................................................... 444
Figure A-33: Telus -West elevation and DSF airflow diagram .................................................. 445
Figure A-34: Sendai Mediatheque - Overall view ...................................................................... 446
Figure A-35: Sendai Mediatheque - DSF cavity drawing and constructed structure ................. 447
Figure A-36: Coeur Defense - Overall view ............................................................................... 448
Figure A-37: Munich Re Offices - Overall view ......................................................................... 449
Figure A-38: Munich Re Offices - DSF close-up ........................................................................ 450
Figure A-39: Munich Re Offices - Plan view of DSF .................................................................. 450
Figure A-40: ING House - Overall view ...................................................................................... 451
Figure A-41: Post Tower - Overall view ..................................................................................... 452
Figure A-42: Post Tower - Shingled west cavity and flush east facades ................................... 453
Figure A-43: Kista Science Tower - Overall view ....................................................................... 454
Figure A-44 30 St Mary's Axe - Overall view ........................................................................... 455
xx iii
Figure A-45: 30 St Mary's Axe - Installation of outer skin and DSF cavity space ..................... 456
Figure A-46: TifS lngegneria - Overall view ............................................................................... 457
Figure A-47: TifS lngegneria - Drawing and photo of DSF cavity space ................................... 458
Figure A-48: AZ Medienhaus - Overall view .............................................................................. 459
Figure A-49: AZ Medienhaus - DSF elevation and section drawing .......................................... 460
Figure A-50 LTD_ 1 Office Building - Overall view .................................................................... 461
Figure A-51: LTD_ 1 Office Building - DSF elevation and stack joint section drawing ............... 462
Figure A-52: LTD_ 1 Office Building - View up from interior courtyard ....................................... 462
Figure A-53: Shanghai Tobacco - Overall view ......................................................................... 463
Figure A-54: Shanghai Tobacco - Interior view .......................................................................... 464
Figure A-55: Shanghai Tobacco - DSF section detail ................................................................ 464
Figure A-56: Kraanspoor - Overall view ..................................................................................... 465
Figure A-57: Kraanspoor - DSF cavity and louvered outer skin ................................................. 466
Figure A-58: Kraanspoor - Interior view adjacent to DSF .......................................................... 466
Figure A-59: Manitoba Hydro Place - Overall view .................................................................... 467
Figure A-60: Manitoba Hydro Place - Northwest elevation and operable units on DSF ............ 468
Figure A-61: KIW Westarkade - Overall view ............................................................................. 469
Figure A-62: KIW- Sawtooth DSF construction and ventilation diagram .................................. 470
Figure A-63 1 Bligh Street - Overall view .................................................................................. 471
Figure A-64: 1 Bligh Street - View through DSF and intake/extract louvers .............................. 472
Figure A-65: Roche Diagnostics International AG - Overall view .............................................. 473
Figure A-66: Roche Diagnostics International AG - Corner elevation and ventilation diagram . 47 4
Figure A-67: The Shard - Overall view ....................................................................................... 475
Figure A-68: The Shard - DSF elevation at corner and intermediate cavity .............................. 476
Figure A-69: Pearl River Tower - Overall view ........................................................................... 477
Figure A-70: Pearl River Tower - Detail view through DSF ....................................................... 478
Figure A-71: Cleveland Clinic - Overall view .............................................................................. 479
xx iv
Figure A-72: Cleveland Clinic - Exterior rendering of point-supported diagrid DSF .................. 480
Figure A-73: Cleveland Clinic - Interior rendering ...................................................................... 480
Figure A-7 4: Shanghai Tower - Overall view ............................................................................. 481
Figure A-75: Shanghai Tower - Atrium DSF rendering and construction photo ........................ 482
Figure A-76: Twofour54 Media Zone - Overall view .................................................................. 483
Figure A-77: Twofour54 Media Zone - DSF detail view ............................................................. 484
Figure C-1: Warren Petroleum Executive Headquarters - Overall view ................................... 489
Figure C-2: Warren Petroleum Executive Headquarters - Interior office and cafeteria views . 490
Figure C-3: Occidental Chemical Center - Overall view ........................................................... 491
Figure C-4: Occidental Chemical Center - Cavity view ............................................................ 492
Figure C-5: Yazaki North America - Exterior view of atria ........................................................ 493
Figure C-6: Yazaki North America -Atria mast truss and overall building ............................... 494
Figure C-7: Levine Hall - Overall courtyard view ...................................................................... 495
Figure C-8: Levine Hall - Exterior photo and interior corner view ............................................ 496
Figure C-9: Seattle Justice Center - Overall street view .......................................................... 497
Figure C-10: Seattle Justice Center- DSF elevation and cavity space ..................................... 498
Figure C-11: Manulife Financial - Upper tower view .................................................................. 499
Figure C-12: Manulife Financial - Overall view and double-skin facade .................................... 500
Figure C-13: Genzyme Center - Overall view ............................................................................ 501
Figure C-14: Genzyme Center - DSF "loggia" section and photo .............................................. 502
Figure C-15: Foundry Square - Overall view .............................................................................. 503
Figure C-16: Foundry Square - Corner of multi-story DSF's point-supported glass wall ........... 504
Figure C-17: UMass Medical School -Over-clad multi-story DSF cavity ................................... 505
Figure C-18: UMass Medical School -Completed retrofit (top) and conditions prior to reclad .. 506
Figure C-19: Loblolly House - Exterior view ............................................................................... 507
Figure C-20: Loblolly House - Overall exterior (top) and interior (bottom) views ....................... 508
Figure C-21: University of Michigan - Multi-story DSF view ....................................................... 509
xxv
Figure C-22: University of Michigan - Overall view of south curvilinear multi-story DSF ........... 510
Figure C-23: MCA Denver - Exterior DSF view .......................................................................... 511
Figure C-24: MCA Denver - Overall building and interior corridor views ................................... 512
Figure C-25: Comcast Center - South entrance winter garden with multi-story DSF ................ 513
Figure C-26: Comcast Center -Atrium interior and winter-mode diagram ................................. 514
Figure C-27: Riverhouse - Overall view ..................................................................................... 515
Figure C-28: Riverhouse - Unitized box-window DSF exterior views ........................................ 516
Figure C-29: Klarcheck Information Commons - Horizontal cable and DSF end condition ....... 517
Figure C-30: Klarcheck Information Commons - Elevation and cavity photo ............................. 518
Figure C-31: "Aldis" at Arizona State University - DSF exterior view ......................................... 519
Figure C-32: "Aldis" at Arizona State - South Elevation and stairwell interior ............................ 520
Figure C-33: Whatcom Building - Interior space behind Lightcatcher DSF ............................... 521
Figure C-34 Whatcom Building - Lightcatcher day (top) and dusk (bottom) exterior views ...... 522
Figure C-35: Art Institute Modern Wing - Exterior view of multi-story DSF ................................ 523
Figure C-36: Art Institute Modern Wing - Close-up of underside air inlet dampers in closed
position and overall north elevation ........................................................................ 524
Figure C-37: Cleveland Museum of Art - Sculpture gallery interior ............................................ 525
Figure C-38: Cleveland Museum of Art - South elevation and interior ....................................... 526
Figure C-39: Cambridge Public Library- Exterior DSF view ...................................................... 527
Figure C-40: Cambridge Public Library- Overall exterior and interior views ............................. 528
Figure C-41: 100 Park Avenue - Overall exterior view of east facade over-clad ....................... 529
Figure C-42: 100 Park Avenue - East elevation and close-up ................................................... 530
Figure C-43 USC CIRM Center - DSF cavity view .................................................................... 531
Figure C-44: USC CIRM Center - Overall exterior view during construction ............................. 532
Figure C-45: USC CIRM Center - Exterior and interior views of translucent frit ........................ 533
Figure C-46: USC CIRM Center - Installation of vertical cables and glass ................................ 534
Figure C-47: New York Presbyterian Hospital - Exterior DSF view of air inlet ........................... 535
Figure C-48: New York Presbyterian Hospital - Overall view ..................................................... 536
xxvi
Figure C-49: New York Presbyterian Hospital - Detail view ....................................................... 537
Figure C-50: Jaqua Center - Exterior DSF view ......................................................................... 538
Figure C-51: Jaqua Center - Interior view .................................................................................. 539
Figure C-52: Jaqua Center - View through multi-story DSF cavity and solar-control screen ..... 540
Figure C-53: New World Center - Interior daytime view ............................................................. 541
Figure C-54: New World Center - North exterior and interior nighttime views ........................... 542
Figure C-55: Angelos Law School - Overall view ....................................................................... 543
Figure C-56: Angelos Law School - Baffle DSF and external blinds .......................................... 544
Figure C-57 Tata Hall - Multi-story DSF detailed view .............................................................. 545
Figure C-58: Tata Hall - Exterior view of concave double-skin facade ...................................... 546
Figure C-59: Cedars-Sinai Advanced Health Science Pavilion - Overall view ........................... 547
Figure C-60: Cedars-Sinai Advanced Health Science Pavilion - DSF detail drawing & photo .. 548
Figure C-61: Wiell Cornell - Overall view of faceted multi-story DSF ......................................... 549
Figure C-62: Wiell Cornell - Subdivision of elevation and cavity interior .................................... 550
Figure C-63: The Forum - Southeast corner view and air intakes ............................................. 551
Figure C-64: The Forum - Multi-story DSF south elevation at night .......................................... 552
Figure C-65: Northwestern University Recital Hall - Interior rendering of multi-story DSF ........ 553
Figure C-66: The Peter W. Rodino Federal Building - Exploded view of over-clad DSF ........... 554
Figure C-67: A.J. Celebreeze Federal Building - Southeast corner rendering of DSF .............. 555
Figure C-68: PNC Plaza Tower - Overall view ........................................................................... 556
Figure C-69: PNC Plaza Tower-WT-1 DSF with operables closed and open .......................... 557
Figure C-70: PNC Plaza Tower - Section axonometric of natural ventilation and solar
chimney .................................................................................................................. 558
Figure C-71: Jerome L. Greene Science Center - Overall construction progress photo ........... 559
Figure C-72: Jerome L. Greene Science Center - Train adjacent to the east and
southeast corner of Wall Type 3 DSF .................................................................... 560
Figure C-73: Jerome L. Greene Science Center - DSF's indoor air inlet at third floor and
interior of DSF's cavity at the top exhaust ............................................................. 561
Figure C-74: Nw Stanford Hospital - Interior rendering of multi-story DSF ................................ 562
xxvii
Figure C-75: New Stanford Hospital - Preliminary design of OSF unit interface ........................ 563
Figure 0-1: 20 Steady-State Simulation -Simple Cube: Sensitivity to Power Law
(U = 10.5 mis, Ir, mfr,= 0.15, a= varies, Zo: 1.0 m) ................................................. 566
Figure 0-2: 20 Steady-State Simulation - Simple Cube: Sensitivity to Roughness
(U = 10.5 mis, Ir, mfr'= 0.15, a= 0.22, z
0
: varies) ................................................... 567
Figure 0-3: 20 Steady-State Simulation - Simple Cube: Sensitivity to Inlet Turbulence
Intensity (U = 10.5 mis, Ir, mfr, : varies, a= 0.22, z
0
= 1.0 m) ................................. 568
Figure E-1: 20 Transient Simulations: Sensitivity to aspect ratio (at t ~ 1.5 s) ............................ 571
Figure E-2: 20 Transient Simulations: Sensitivity to opening area ratio (at t ~ 1.5 s) ................. 571
Figure E-3: 20 Transient Simulations: Sensitivity to airflow configurations (at t ~ 1.5 s) ............ 572
xxviii
Abbreviations
20 Two-Dimensional
30 Three-Dimensional
A/EiC Architectu re/Engineering/Construction
ABL Atmospheric Boundary Layer
ABS Acrylonitrile Butadiene Styrene
AFN Airflow Network
AIF Advanced Integrated Facade
ALE Arbitrary Lagrangian-Eulerian
BES Building Energy Simulation
BMS Building Management System
C&C Components and Cladding
CFO Computational Fluid Dynamics
CGF Conventionally-Glazed Facades
Cp Coefficient of Pressure
Cp,net Net Pressure Coefficient
DSF Double-Skin Facade
FSI Fluid-Structure Interaction
FT Fully Tempered
HS Heat-Strengthened
IGU Insulated Glass Unit
IEA International Energy Agency
LCCA Lifecycle Costing Analysis
MWFRS Main Wind-Force Resisting Structures
PVB Polyvinyl Butyral
SHGC Solar Heat Gain Coefficient
TVF Transparent Ventilated Facade
xx ix
Abstract
Double-skin facades (DSF) have been implemented to harness benefits of increased energy
efficiency, acoustic isolation, and access to natural ventilation. The continued innovation and
implementation of these systems presents a need for revised structural standards that account for
the differing airflow inlet and outlet configurations. Recent multi-story double-skin applications in
North America are evaluated to document the current airflow and structural design methodologies
utilized in practice. Multi-story double-skin configurations are the focus due to their prevalence in
the United States, as learned following a review of 30 existing projects of which 21 had multi-story
cavity partitioning. Preliminary simulations are performed using a two-dimensional transient
simulation model where the variables considered included cavity aspect ratio (height/depth),
ventilation geometry and permeability. These simulations revealed a prominent sensitivity to
airflow inlet and outlet configurations, leading the research to extensive three-dimensional
simulations with varied airflow configurations while holding other variables constant. To evaluate
the structural response of multi-story double-skin facades, the effects of various airflow
configurations on the pressure coefficient distribution are examined by using computational fluid
dynamics to simulate a wind tunnel testing environment with scaled building geometries. The
simulation method is first calibrated and reviewed against existing wind tunnel studies. A
prototypical geometry is developed and three-dimensional steady-state computational fluid
dynamic simulations are run for twelve configurations of airflow inlet and outlet combinations. The
interior skin's coefficient of pressure profile fluctuates considerably depending on the varying inlet
and outlet configurations. Computational fluid dynamics with multi physics software is evaluated to
determine its potential effectiveness as a design tool for the determination of coefficients of
pressure for double-skin facades. A set of guidelines is provided for 1) the evaluation process of
coefficients of pressure for double-skin facades and 2) the design of various multi-story double
skin facades.
xxx
Preface
A double-skin facade is a response intended to preserve transparent portions of a building while
pragmatically meeting the insulating needs of an envelope subject to varying modalities from
environmental (climate) changes. Double-skin facades are not the only approach that may
achieve these goals, but they are part of a whole ecology of high-performance facade solutions
that seek heightened performance (e.g., acoustic, thermal comfort, daylight and glare control -
just to name a few) with a balance of aesthetics, acting as the interface between interior and
exterior. These performance challenges exist globally with unique solutions specific to individual
projects. Outside of Europe, the emergence and evolution of high-performance facades has
occurred with very little regulation, few guidelines, and has lagged in implementation. Seeing that
the facade is the intersection of many performance and aesthetic elements, it has become the
focus of many research efforts, and is often a focal point within architectural design practice; it is
a design element full of freedom, opportunity, expression, technicality, precision, innovation and
enriched performance. In a time when environmental concerns are prevalent, the double-skin
facade is regularly acclaimed as a sustainable enclosure solution. This can be a shallow
assessment of a double-skin facade's role in a larger whole building performance strategy, and
the architecture/engineering/construction (A/EiC) community's dialogue of such systems must
exercise a more critical evaluation of its adoption and a deeper understanding of its potential in a
wide spectrum of high-performance facade solutions.
As double-skin facades are emerging into more widespread application, the lack of a mature
infrastructure for evaluating the performance and suitability of applications has both benefitted
and damaged the reputation of the concept's merits as a relevant facade solution. As with many
innovative approaches, a lack of controls and recognized conventions allows for innovation, as
well as misappropriation. In an ideal evolution, there is a temporal sensibility that steers the
technology towards a common understanding of known performance advantages and a larger
sample of built precedents that serve as a foundation for more optimal future incarnations of the
xxxi
technology. With a larger sample base, lessons learned provide a platform for new
interpretations of the technology that respond either more specifically to established performance
demands, or new, previously unaddressed criterion. Over the last several decades, double-skin
facades have evolved similarly to other technologies; there are established concepts - including
many tested - but no universally accepted or time-tested, variations. In Europe, there are
different fiscal and cultural dynamics that impact the feasibility of double-skin facades compared
to the United States. As a result of these differences, the evaluation of a double-skin facade in
the United States must factor in different constraints in a cost-benefit analysis.
Though it is generally agreed that the first modern double-skin facade was constructed in Niagara
Falls, New York in the early 1980s, the integration of these systems in the United States went
dormant for nearly two decades before resurfacing near 2000 (see Appendix C.3 to C.5).
Between 1980 and 2000, many European applications were built, often taking the form of box
window solutions in commercial office building towers. After more than a decade of regular
implementation in the United States starting in 2000, there are now dozens of built examples that
establish a basis for assessment and comparison. The most notable difference is that the multi
story configuration is applied at a greater frequency in the United States than compared to the
European patterns. With the scale of multi-story configurations being much larger than box
window designs, the variability in the performance is greater, potentially leading to overheating,
and thermal and acoustical detriments. Additionally, cheaper energy in the United States than in
Europe makes it more difficult to economically justify a double-skin facade. In lieu of the thermal
and acoustic conditions, the preference for a singular, non-compartmentalized, multi-story
configuration may be more substantially attributed to their grandeur and iconicism than envelope
performance. Within certain programs and climates these configurations do have added value,
but the tendency towards the multi-story as a podium or entrance solution within a larger program
rather than a whole enclosure solution points towards the double-skin facade serving as an
identifiable "sustainable" feature. Of course there are many examples of well-thought-out
xxxii
applications of multi-story double-skins, but the multi-story configuration presents a number of
sensitive performance issues that warrant further research, including energy performance,
thermal comfort and structural response.
Considering the modality of double-skin facades in recent years presents a number of challenges
and opportunities. The cautionary tone that surfaces through this dissertation can be attributed to
conveying a mindset that aims to avoid profligate investments into facades (not just double-skin
facades) that fail to achieve their design intent. With clearly articulated design goals,
performance priorities and technologically sawy individuals involved, high-performance facades
can flourish in a sensible response to a site's constraints and environmental conditions. When
the root advantages of any given technology are revisited, the door reopens to undercut what has
become the norm to re-invent and re-envision new solutions that progress technology further.
One of best ways to advance the A/EiC community's collective understanding is to encourage
research, as-built commissioning, on-going monitoring, post-occupancy evaluation, and
disclosure of these findings. With this information, the feedback loop can be closed, reinforcing
the design intent of future facade solutions.
An interesting ramification of the multi-story double-skin facade is that it can be more sensitive to
thermal, structural and acoustical issues. The items often associated with double-skin facades -
energy performance and thermal comfort - are both the most frequently researched and the most
controversial in practice. Often overlooked are the unique structural responses presented in
double-skin facades. With multi-story double-skin facades, the potential for longer-span
structures and larger air cavities is introduced into the building envelope, challenging the
appropriateness of design methodologies that currently exist for more traditional, single-skin
facades. The structural and airflow dynamics are targeted with an aim to raise awareness of the
fundamental issues and provide a series of guidelines for evaluation and design process moving
forward.
xxxiii
1. Introduction to Structural Considerations for Double-Skin Facades
The effort to improve the indoor environment of building spaces while simultaneously reducing
energy consumption and operating costs has been a continual quest for architects and engineers.
While originating in passive strategies, the pursuit has evolved in recent decades to intelligent
facade concepts. Modern glazed building envelopes first implemented simple and proven
passive controls such as shutters, blinds and natural ventilation to provide adaptations to variant
climate conditions. However, these early control strategies did not provide adequate versatility to
address occupant comfort, especially in large buildings with high occupancies. As a result,
innovative glazing systems have been the focus of many designers and researchers, and
ultimately spawned the emergence of the facade engineer as a building systems specialist with
increasing significance. This is in large part due to the building envelope's prominence in the
architectural statement, occupant comfort and overall building efficiency. The high-performance
facade is a continually evolving building system which is incredibly nuclear to the architectural
aesthetic of a structure; mediating between the outdoors and interior spaces, interacting as a thin,
fragile sheath between comfort and harsh environments, with infinite opportunities for systems
integrations with advancing mechanical and energy technologies.
As the building industry strives to transform buildings from energy consumers to energy
producers, the importance of transparency in the building enclosure plays a critical role in
establishing a comfortable interior environment. A transparent facade creates a strong visual
relationship between the building interior and the surrounding environment. The glazed envelope
of modern commercial projects provides a visual transition between these two climates, which
would otherwise be disconnected by an opaque enclosure. Aware of the pleasure a transparent
building creates, but faced with heightened performance expectations, the facade community
began to explore the concept of an advanced integrated facade - one that equilibrates the
external climate with the comfort requirements of the indoor environment - while providing the
1
core functions of shelter, security and minimizing energy consumption. The facade industry is
now driven by priorities of environmental sustainability, climatic sensitivity, programmatic
response, and economic sensibility. With energy and sustainability prevalent, the facade has
transformed into more than just a barrier between interior and exterior, and with that comes new
considerations for facade engineers. Amongst them, leading researchers and practitioners have
identified that "natural ventilation, double layered wall systems, and even integrated wind turbines
are beginning to present yet further exciting challenges for wind engineers" (Irwin 2009).
The increased performance expectations on the glazed envelope resulted in a number of
innovations, including one approach which utilized multiple-layers to tailor the facade to dynamic
climatic conditions. A double-skin facade is characterized by the addition of a second glazing
either behind or in front of the insulated building facade. This layering introduces a greater depth
to the facade system and results in a cavity which may be used to introduce airflow or shelter any
number of control systems, such as shading devices.
Double-skin facades have emerged gradually over the last century, but are considered a
relatively new technology in modern applications and systems integration. The double-skin
facade presents a unique condition compared to conventional single-skin facades. The additional
layer of glazing, and any systems which may occur in the intermediate space, present a complex
interaction of enclosure with natural forces. The introduction of an intentional airflow cavity, air
inlets and air outlets create dynamic behavior which are not integral to the evaluation of single
skin facades. A conceptual diagram of a double-skin facade of the multi-story configuration is
presented in Figure 1.1 as an example of the system layers, depth, airflow functions, and how it
may all be assembled for a building project.
1
Understanding the pressure dynamics of the
double-skin cavity is essential for optimizing the design of both the outer and interior facades of a
double-skin system, as well as accurately inputting wind pressure coefficients for building energy
simulations (BES) and airflow network (AFN) modeling.
2
Double-Skin Facade
Multi-Story Configuration C oncept
Figure 1-1: Double-skin facade multi-story configuration diagram.
1.1 Wind Pressure Distribution Characteristics in a Double-Skin Facade
The interaction of wind forces and the building envelope is characterized by the building
envelope's pressure distribution. Wind pressure always includes static and dynamic components.
The static component is the result of large-scale air movements and is subject only to gradual
changes in magnitude. The dynamic component is the result of gusts, or short-term changes in
wind velocity and direction. Characterizing a facade's pressure distribution is integral in
engineering the glass and structural components, as well as understanding its effects on natural
ventilation strategies. Building shape is a primary influence on pressure distribution. The wind
pressure distribution of a double-skin facade configuration can vary substantially compared to a
geometrically similar conventional single-skin facade.
1.1.1 Single Surface Distribution
The wind loading effect on a local facade area is described by the pressure coefficient ( Cp)· For
conventional facades, building codes and analytical procedures, such as ASCE/SEI 7 Chapters
26 - 31,
2
may be used to determine the wind loading for a structure. The pressure coefficients
provided within the codes are conservative for a wide range of applications, primarily of simple
rectangular configurations. The coefficients available in building codes are based on extensive
wind tunnel testing performed primarily in the 1970's (Geurts, van Staalduinen and de Wit 2004,
17 4). For buildings with non-rectangular geometry, large roofs, complex roof detailing, canopies
or sheltering effects with neighboring buildings, no local coefficients are available in codes.
3
For a simple, rectilinear geometry that is included in the codes with a single surface facade, the
local pressure coefficients are generalized dramatically to minimize variation. This simplification
may include intermediate, edge and corner zones - with larger coefficients most often associated
with the two latter regions. The actual pressure distribution across a single layer facade is more
gradient and dynamic. When studying pressures on a facade, it is not only the pressure values at
a given point that is of interest, but also the variation across the entire surface.
4
1.1.2 Multi-Layered Distribution
In a double-skin facade, the introduction of a second glazed layer and an intermediate cavity
create additional surfaces to which the wind pressure distribution is to be considered. Even if a
building has a rectilinear geometry, there is no clear method outlined to address the pressure
distribution gradient unique to each of the inner and outer skins respectively. For the purposes of
this research, the pressure coefficient for each skin in a multi-layered system is of interest and
evaluated as independent surfaces with a unique pressure distribution gradient, each contributing
to the combined behavior of a double-skin facade in a built application.
Single-Skin Double-Skin Inner Envelope
:: 0.5
N
o_,_,_"-",-~.,---T-~~~~,------,-'---,~"-o-"----r
0 0.5
x/L
0
:: 0.5
N
0
0.5
x/L
-0.4
Figure 1-2: CP distribution on single envelope (left) and double-skin inner layer (right).
Information for this figure based on Effects of different multi-storey double skin facade
configurations on surface pressures (Marques da Silva and Gomes 2005).
1.1.3 Multi-Layered Load Sharing
In a double-skin facade configuration, the intermediate cavity space between the two glazing
systems can be sealed or introduce a ventilated airflow path. Most modern double-skin facades
use an operable mechanism, such as motorized louvers, at the inlet and outlet airflow openings to
5
provide some variable control of the ventilation rate. When the airflow rate varies through the
cavity, the wind pressures on each of the inner and outer skins vary.
In a sealed cavity condition, the exterior facade will generally encounter a greater fraction of the
overall impinging wind pressure than the interior facade; though this ratio is sensitive to the depth
of the cavity. In a case where the cavity is sealed to the interior skin and possesses openings to
the outside, thus creating an outdoor air curtain, the interior facade will receive a greater fraction
of the overall impinging wind pressure in relation to the exterior facade as the opening size
increases (Ishida 2003, 254). The load applied to each skin is necessary to size glass, glazing
attachments, structural components and anchoring systems of each. Additionally, through a
greater understanding of the ratio of load, or distribution load ratio, the facade structures may be
optimized to reflect the systems' load sharing characteristics rather than conservatively designing
both systems to take the entire wind pressure. This impacts and has the potential to reduce
costs, materials and the energy used to harvest and transport those resources.
With a ventilated double-skin facade, the distribution of wind loads between the two facade layers
becomes significant. Additionally, if the interior facade provides operable units for natural
ventilation, the characteristics of how the wind pressures enter the internal spaces must be
considered to maintain comfortable airflow conditions. In a double-skin facade with inner-surface
ventilation, wind tunnel tests have shown that the distribution of loads occurs over both the inner
and outer facade layers (Oesterle 2001, 112-113). The outer facade is primarily affected by the
dynamic components of wind - gusts - and is minimally affected by constant static pressure
elements. The steady components of wind pressure penetrate the ventilated cavity and affect the
inner skin, which is minimally affected by wind gusts. The distribution of the wind loads between
the two skins is also affected by other project specific design variables, including open area ratio,
opening configurations, cavity depth and communication with other skins.
6
1.1.4 Importance of Pressure Coefficients in Double-Skin Facade Design
The pressure coefficient (Cp) is a non-dimensional ratio between the pressure at a point in the
building facade and the dynamic wind pressure in the free stream region. Wind pressure
coefficients are influenced by many parameters, including wind speed, wind velocity, building
geometry, facade detailing, position on the facade, and exposure/sheltering effects of surrounding
structures. The wind pressure coefficients (Cp) on a building envelope is a non-dimensional ratio
generally defined as follows:
Eq. 1.1.1
where Px is the wind induced static pressure at a point x in the building facade, P
0
is the static
reference pressure, and Pd is the upstream dynamic wind pressure measured at the building
height (Etheridge and Sandberg 1996). The upstream dynamic wind pressure, Pd, is the force
per area unit due to the impinging wind normal to the building face, generally defined as follows:
Eq. 1.1.2
where p is the air density and Vr,f is the wind speed. By substituting Equation 1.1.2 into
Equation 1.1.1, the local pressure coefficient ( Cp) becomes,
C _ Px-Po
P-(')'
2 p·Vref
Eq. 1.1.3
Pressure coefficients are integral to the design of all building envelopes, but are especially
important in the engineering of double-skin facades for two primary reasons; 1) the pressure
coefficients are used to translate the dynamic pressure into a pressure load or force onto the
envelope which is used in the design and structural engineering of all glazing system
components, and 2) the pressure coefficients are fundamental to calculate ventilation rates by
7
airflow network models, which is the standard model for ventilation rate calculation in many
building energy simulation tools - of which many double-skin facade energy consumption
improvements and thermal comfort are perceived advantages and driving motivations for
implementation.
Structural Load Distribution
Distribution of wind pressures over the multiple layers in a double-skin facade is a critical
component in engineering and selecting appropriate materials and dimensions for glass, glazing
attachments, structural components, and anchoring systems of each layer. Being able to identify
and appropriately load areas of anomaly, such as near edges, corners or at the roof parapet
where vortex shedding can be expected, is necessary to implement a safe solution with
confidence. Obtaining accurate pressure coefficients for double-skin facade systems most
certainly will require advanced simulation or physical testing beyond the simplified or analytical
methods presented within structural codes, especially considering the potential for dynamic
gradients across each layer.
The wind pressure distribution load ratio depends on a number of factors including, but not limited
to: cavity depth, opening ratio, ventilation geometry and structural rigidity. The distribution
between skins is significant with respect to structural design because it presents opportunity to
introduce different support systems for each skin. In new construction of double-skin facades the
opportunity to use either a shared or independent structure to both facades may depend on
program, acoustics, installation, design aesthetic or budgetary constraints. In a retrofit
application, where a double-skin may be applied to modernize an existing structure and improve
the envelope performance, structural constraints can be more stringent and require an
independent structure for the additional exterior skin or significant upgrades to the existing
structure. In either scenario, the pressure coefficients are a fundamental element that affects
every component from glass to anchor.
8
Building Energy Modeling
A challenging component of double-skin facade energy modeling is accurately simulating the
effects of ventilation through the cavity. Airflow through the building enclosure is an imperative
factor in any heat loss and moisture transfer characteristics of an envelope (Hagentoft 1998), and
internal environment quality and energy demand of a building are greatly influenced by air
infiltration and ventilation (Liddament 1986). The complex interaction of multiple physics across
the double-skin facade makes it computationally overwhelming to tie energy simulation tools
directly to advanced computational fluid dynamics or heat-air-moisture transfer models (HAM).
Instead, two separate models are coupled using common output data from one source as an
input loop for the other. This is true of the BES-AFN programs which consider the effects of
ventilation in the energy simulation.
Natural ventilation is common to many double-skin facade configurations, sometimes providing
ventilation from exterior to interior spaces, but most often a ventilated cavity space to create an
airflow buffer. Wind is an important driving force for infiltration and ventilation, which plays a
significant role in an airflow network when considering possible natural ventilation approaches
(Clarke 2001, 283). Many building energy simulation (BES) applications use an airflow network
(AFN) for ventilation rate calculations. At the core of the AFN are two inputs: wind pressure
coefficient (Cp) and discharge coefficient (Cz). The validity of the CP source data is extremely
influential in the accuracy of the building energy simulation. As it is practically impossible to take
into account the full complexity of CP variation, BES-AFN programs generally incorporate it in a
simplified way. Reliable CP data is challenging to obtain for BES-AFN simulations because of the
complex interaction and interrelationships of turbulent elements (wind direction and speed) and
site-specific conditions (sheltering effects, building geometry, and facade design). other reasons
for the high uncertainty associated with CP values are due to critical issues such as thermal
comfort, energy consumption and mold, and are sensitive to the air change rate, which depends
on CP (Costola, Blacken and Hensen 2009, 2027). As a result, the CP data is often identified as a
9
primary source of uncertainty in BES-AFN models (de Wit 2001, 59-60) and remains a research
focus amongst academics and practitioners.
The BES and AFN programs use a wide range of sources for CP data which can be classified as
primary or secondary sources. Primary data sources are considered the most reliable CP sources
(Costola, Blacken and Hensen 2009, 2028) and include full-scale measurements, wind tunnel
measurements, and computational fluid dynamics (CFO). Secondary data sources are
considered less accurate but are readily available and more economical than primary sources.
The most common secondary sources are CP databases such as tables provided in wind load
standards (i.e. ASCE/SEI 7-10 Minimum Design Loads for Buildings and Other Structures).
These databases are often derived from wind tunnel experiments and applicable to generic
conditions for simple geometries of unsheltered buildings. Two other examples of CP databases
include the Air Infiltration and Ventilation Center (AIVC) references and the ASHRAE Handbook
of Fundamentals; however, these too are limited to simple building geometries and surrounding
conditions.
The uncertainty associated with the pressure coefficient values being input into building energy
simulations raises skepticism about the accuracy of energy modeling. Developing consistent
methods to obtain accurate CP data for ventilated facades is essential since current data sources
produce significantly varied results for even a simple geometry with the same conditions (Costa la,
Blacken and Hensen 2009, 2035). For double-skin facades, which rely heavily on improved
energy efficiency as a validation or justification of higher initial costs, it is vital that more accurate
and available methods of project-specific CP data generation are developed. Additionally, post
occupancy data on actual energy consumption is an excellent comparative data source to review
the accuracy of preliminary projections and refine the simulation process. The focus of this
research is not in the energy performance of double-skin facades, but the significance that the
pressure coefficient (Cp) has at the input level of building energy simulation.
10
1.2 How Current Standards Address Wind Load on Double-Skin Facades
The effects of multi-layer enclosures and permeable enclosure layers on pressure coefficients are
addressed in few codes. When a building envelope consists of multiple layers with an
impermeable (air-tight seal) inner skin and a permeable outer skin, the outer skin may function as
a rain-screen. This leads to pressure equalization across the outer skin with pressure inside the
cavity that relates to the exterior pressure rather than that of the building interior. When the
exterior pressure has a gradient across the facade height, this will generate a pressure gradient
in the cavity, creating cavity airflow. Obtaining appropriate wind pressure coefficients for multi
layer systems is complicated because currently there is no standardized method or accepted
procedure. Additionally, when using scaled wind tunnel experiments, placing pressure
measurement devices on both the outer and inner skins without significantly disrupting the airflow
within the cavity space can become challenging. Ultimately, building codes -which do not
account for the reducing effects of wind load associated with pressure equalization - may lead to
the design of less-economical but safe double-skin facades.
Standards for the structural design of double-skin facade systems to resist wind load do not
account for the geometric orientation of ventilation orifices with respect to the prevailing wind. In
the proposed research the pressure distribution of multi-story DSF systems normal to prevailing
winds is evaluated for multiple ventilation orifice configurations.
1.2.1 US Codes - ASCE/SEI
In ASCE/SEI 7-10 Minimum Design Loads for Buildings and Other Structures, Chapters 26-31,
wind loads for double-skin facades are not addressed.
4
It is unclear how the definitions of
partially enclosed building enclosure and air permeable cladding may relate to the determination
of internal pressure coefficients for a double-skin that includes natural ventilation by operable
units on the inner skin. The configuration of airflow openings is not addressed. In the determining
of wind loads for components and cladding (C&C), the Analytical Procedure (Chapter 30)
11
stipulates that the design wind loads determined for single-layer cladding "shall be used for air
permeable cladding unless approved test data or recognized literature demonstrates lower loads
for the type of air-permeable cladding being considered" (ASCE/SEI 7-10, Section 30.1.5). In the
commentary amendment of the standard, there is reference to a common belief that the
pressures derived from the Analytical Procedure can overestimate the load on air-permeable
cladding (ASCE/SEI 7-10, Section C30.1.5). The commentary proceeds to state that full-scale
pressure measurements or reference to recognized literature should be used to determine the
pressure differential across the permeable cladding. In Chapter 31 it is stated that the Wind
Tunnel Procedure "may always be used for determining wind pressures for the MWFRS and/or
for C&C of any building or structure," also citing that "this method is considered to produce the
most accurate wind pressures of any method specified in this Standard," (ASCE/SEI 7-10,
Section 31.1).
1.2.2 EUROCODE
Eurocode EN-1991-1-4:2005 begins to present guidelines in Section 7.2.10 Pressure on walls or
roofs with more than one skin (CEN/TC250 2005). This section specifies that the wind force on
each skin is to be calculated separately. Section 7.2.10.3 identifies that the wind force on each
skin is dependent on 1) the relative rigidity of the skins, 2) external and internal pressures, 3)
distance between skins, 4) permeability of the skins, and 5) the openings (CEN/TC250 2005). A
series of rules are then presented for determining the pressure on each skin, but is restricted to a
distance between skins less than 100 mm and where the "extremities of the layer between skins
is closed." The presented rules are not applicable to configurations which put air into
communication with other faces, excluding many double-skin facade airflow configurations.
Researchers have concluded that the wind load guidelines established in Eurocode EN-1991-1-
4:2005 require revision (Marques da Silva and Gomes 2008) and (Wellserhoff and Hortmanns
1999).
12
1.2.3 Glass Codes
Current glass codes can be difficult to adapt to emerging design technologies, such as double-
skin facades, and generally use large factors of safety. Additionally, glass codes of different
countries often disagree. Glass codes used in practice are similar to the aforementioned wind-
loading codes in that they are often restricted to specific, simple design conditions. There is
opportunity for the facade industry to develop a basis of design which bridges wind engineering
and glass design (Zammit, Overend and Hargreaves 2010, 656).
1.2.4 Limitations to Codes
Modern facades continue to evolve towards increasingly intricate and diverse solutions frequently
attributable to increased energy efficiency expectations. Codes used in practice to determine
wind loading are limited to simple building geometries and rarely address complex facades. The
designer must proceed cautiously when applying simple codes to a complex design in order to
avoid unsafe designs. On the other hand, applying simple code procedures to determine wind
loads for a complex facade, such as a double-skin facade, may result in overdesigned and costly
solutions because code-derived pressures are often considerably higher than those from wind
tunnel testing. Table 1-1 presents a comparison of wind speeds and pressures determined for a
simple condition using different code procedures, CFO and physical measurements.
Table 1-1: Wind Loading Code Wind Speeds & Pressures (Overend and Zammit 2006, 656)
Result ASCE Eurocode BS 6399 ESDU AS/NZS Measured Pressures CFO
Design Gust Speed
[ms-
1
]
37.12 41.15 40.42 38.39 36.31 38.39 38.39
Maximum External Pressure
-1064.3 -1453.3 -1301.9 -1174.9 -1050.8 -858.55 -813.37
(Pa)
The determination of pressure coefficients and wind loads for energy simulation and structural
design of double-skin facades is a challenge and requires a project-specific analysis.
13
1.3 Current Engineering Practice to Determine Pressure Coefficients
Modern and dynamic architectural forms are being constructed, advancing the limits of existing
engineering knowledge. However, these exciting forms often fall outside of the constraints of
existing building costs, which are in large part based on past industry and research experience.
One building-to-nature interaction which is increasingly volatile with geometrically complex
structures is the flow of wind around and through these structures. The simplified or analytical
methods outlined in wind loading codes of practice are not applicable, and prescribe wind tunnel
investigations to determine a facade's wind pressures. As facades have become increasingly
intricate, pressured to achieve greater energy efficiency, and an iconic representation of many
clients, projects have seen dramatic increases in facade costs (Zammit, Overend and Hargreaves
2010, 656). In large value facade contracts, wind tunnel testing is routine and justifiable due to its
relative cost benefit Seeing that wind tunnel experimentation can be expensive and regularly
occurs far along in the project development, additional data sources are used in practice to
determine pressure coefficients, especially during design development. Similar to the variations
produced by different building codes (see Section 1.2.4), the various CP data sources used in
engineering and energy modeling practice show large variations, even for simple configurations
like unsheltered cubic buildings (Costola, Blacken and Hensen 2009, 2035). The following
sections include a general outline of key CP data sources implemented in practice.
1.3.1 Full-Scale Tests
Full-scale measurements of constructed facades on the project site provide the most reliable CP
data representation. They are most valuable in urban environments for which the analysis of
pressure coefficient data is particularly demanding. The dynamic interactions of many variables
in the double-skin facade cavity may require a full-scale mock-up test before implementation.
This may be costly, but allows for system assurance and improvement of performance before
application. Studying and visualizing the behavior of airflow through openings in a mock-up is
very helpful in making design adjustments or verifying simulations. This method is appropriate in
14
unitized, box-window, double-skin products where there is repetition and mass production of the
design being evaluated. Full-scale tests are generally used for validation purposes because of
their inherent complexity and cost, and are not integral through early design development.
1.3.2 Wind Tunnel Tests
Wind tunnel testing to determine building pressures due to wind-loading is widely practiced
(Reinhold 1982). Scaled-model tests can be used to study the airflows around a building and
within the cavity space of multi-layered facade. Wind tunnel studies allow for scaled visualization
and measurement of external pressures. Structural engineering uses wind tunnel experiments to
evaluate the wind loading on and around a building, taking geometry, surroundings, wind speed
and turbulence into consideration. Wind tunnel testing requires large facilities, expensive
equipment, and most significantly, expertise to produce reliable data. Boundary-layer wind
tunnels for structural model evaluation are on the order of 2.4 m (8ft) wide, 1.8 m (6 ft) tall, and 10
m (33 ft) long. The building geometry model is often three-dimensional printed in an ABS plastic
material with up to many hundreds of pressure taps connected by means of tubing to electronic
pressure transducers (Irwin, Denoon and Scott 2013, 24-25). These trials can be valuable for
unique building forms or complex cavity configurations. Additionally, for buildings where the
along-wind response governs the wind loading, wind tunnel produced loads are likely to be similar
or even less than those calculated using building code procedures (Irwin, Denoon and Scott
2013, 35).
Flow-Straightening
Screens
Wind
Velocity
Profile
Longitudinal Cross-Section View
Velocity
Profile
Figure 1-3: Components of a boundary-layer wind tunnel test section.
Boun~2'. ~~r
Turntable
15
While the code approach is most common for nominally sealed buildings, wind tunnel testing is
the more common method of load determination for buildings with openings in the exterior facade
(Irwin, Denoon and Scott 2013, 24-25). Such is the case with double-skin facades. However, few
project-specific wind tunnel tests to date have taken into account the unique pressure
distributions of the outer and inner layers. Local pressure dynamics, such as turbulent flow
through openings, may be difficult to characterize at a whole-building scale given that a limited
number of pressure taps over the whole facade area may collect data. Larger scale and finer
placement of pressure taps for detailed local analysis is recommended.
Figure 1-4: Shanghai Tower scaled wind tunnel model. Copyright Marshall Strabala (2011).
This is an exceptionally large wind tunnel and scaled model. The project utilizes a double-facade
strategy, though the cavity depth is on the scale of an atrium or winter garden.
16
The use of wind tunnel experimentation has been applied to building applications since the first
half of the 20
1
" century (lrminger and N0kkentved 1936) and translated knowledge early on from
aerospace research. This translation from other fields contributed to years of modeling with use
of laminar approach flows, without consideration of the atmospheric boundary layer (ABL) wind
profile which buildings encounter as they rise above ground. Additionally, the laminar approach
flows did not account for surrounding topographic effects and sheltering conditions. After
researchers compared full-scale measurements to wind tunnel results, the importance of the
atmospheric boundary layer in determining wind loads on structures was revealed and drove the
manifestation of boundary-layer wind tunnels (Jensen and Franck 1965). Nowadays, wind tunnel
testing specific to architectural structures is widely practiced and heavily researched.
The complexity associated with wind tunnel testing is similar to other laboratory measurements in
the sense that it demands special care, precise measurement collectors, standardized methods,
calibration and expertise. Issues of coordination and quality assurance are important (Irwin,
Denoon and Scott 2013, 33-34). Uncertainty in the results of wind tunnel testing compared to full
scale testing are the most evident in the imperfect science of determining wind pressure on
buildings, though great improvements have and continue to be made. Inconsistencies between
12 different institutions and facilities using the same conditions for a simplified cube at three wind
indices exemplifies the importance of quality assurance in testing procedures (Holscher and
Niemann 1998, 599). The variation in this comparative study revealed greater uncertainty on the
roof and leeward exposures than the windward surface. Holscher and Niemann describe several
possible explanations for the variation amongst the different laboratories, including: variability
from the measuring devices; variability in simulated turbulence; judgments made in time and
geometric scales; imperfections in the model, pressure tapping and tubing; variation in data
analysis; and lack of accuracy due to human error. Atmospheric boundary-layer wind tunnels
with simulated full-scale flows can provide realistic wind loads and effects if some fundamental
requirements are considered (Holscher and Niemann 1998, 607).
17
1.0
- Mean Result
0.5
Range of Results
c.
0
"i
Q)
0
0
. C3
~
0
0
Q)
::;
-0.5
{j)
{j)
Q)
ct
-1.0
-1 .5
Distance Along Trajectory [0-1-2-3)
Figure 1-5: Comparison of Cv data results of different wind tunnel and full-scale tests.
(Holscher and Niemann 1998), (Richards, et al. 2007). Information for this figure based on
Ovetview of Pressure Coefficient Data in Building Energy Simulation and Airflow Network
(Costola, Blacken and Hensen 2009, 2029).
The application of wind tunnel studies for double-skin facade simulations experiences the same
challenges discussed above, with several heightened issues. In theory, to measure the pressure
distribution on both the inner and outer skins requires twice as many pressure taps and tubing
across the double-skin area. This scenario would require tubing to travel through the cavity
space in order to provide pressure taps on the exterior skin. Tubing in the cavity creates an
obstruction to airflow which would render the simulation inaccurate. The deviations resulting from
tubing in the cavity may be minimized by scaling the model specimen to provide a larger model
cavity space; however, the scale of the building may impose on the tunnel's geometric limits and
proper clearances required to maintain appropriate boundary conditions. A larger wind tunnel or
a smaller test sample may be appropriate. Another possibility is two models: one that measures
the exterior skin's pressure as if the cavity was sealed, and a second that has openings for airflow
through the cavity that measures the inner skin's pressures.
18
1.3.3 Computational Fluid Dynamics
Computational Fluid Dynamics (CFO) simulations exploring wind pressure on the building facade
have been implemented increasingly over the last two decades. The primary resistance to CFO
in practice has been due to the availability, costs of software, and computational demand required
to conduct complex simulation. The use of CFO in practice continues to become more prevalent
as computer performance improves, costs reduce, and researchers or practitioners have
increased access to the tools. An advantage of CFO to study double-skin facades, or any
building system, is that it can be used earlier in the design process to inform decisions rather than
validate them after the fact. The ability to evaluate designs iteratively, or perform parametric
evaluations, is a trait not common in wind tunnel or full-scale testing. Despite the benefits CFO
introduces into the design process, the user must be aware of the common downfalls and
limitations specific to the analysis model.
-0.2
-0.3
-0.3
=
-0.5
0
-0.7
-0.5
-0.5
-0.3
/e = 45°
Figure 1-6: CFO pressure distribution (Cp) diagrams for wind normal to cube (left) and wind
45° to cube (right). Information for this figure based on Improved Computational Methods for
Determining Wind Pressures and Glass Thickness in Facades (Zammit, Overend and
Hargreaves 2010, 659).
19
It is not common practice to use CFO as a source of custom CP data primarily because the
expertise, costs and computational resources required often outweigh those of the BES
simulation itself (Costola, Blacken and Hensen 2009, 2030). One approach to supersede such
limitations is to integrate the pre- and post-processing between the CFO and BES tools. Current
attempts to incorporate pre- and post-processing show promise, but are limited in practice
because of minimal grid refinement flexibility. The grid density of a mesh is significant in that a
finer mesh increases accuracy - especially of local phenomena - and reduces the likelihood of
divergence; however, a greater density of data points requires additional computing power.
Being able to control the grid density gives the user control to create finer regions of data
collection where variations may be anticipated, while minimizing the density of large regions
(which will not encounter the same anomalies in results) is one of the quickest methods to reduce
computational demand while preserving the fidelity of the results. The ability to perform local grid
(or mesh) refinement is a significant component of high-end CFO software packages, but not
necessarily tools which combine the analysis with energy simulation. This level of detail is
important when designing the cavity of double-skin facades. The type of interior and exterior
openings, size and the geometry of intermediate placed shading devices, and type of ventilation
strategy can influence the type of airflow. Thus, CFO simulations can provide useful details,
decreasing the possibilities of unpredicted mistakes during the design stage. However, the airflow
simulations are still difficult (Poirazis 2006, 222) and require the attention of a mechanical
engineer. ASCE currently doesn't allow CFO analysis for the determination of loads. Accuracy of
models utilized in CFO should be reviewed and confirmed to the engineer's best knowledge.
Similar to wind tunnel testing, uncertainty in CFO simulations can be attributed to the sensitivity of
boundary layer conditions. As discussed in the wind tunnel section, the wind speed alone is not
adequate enough to perform simulations. The gradient of wind speed near the ground surface
varies with height and is affected by surrounding topographic conditions. The atmospheric
boundary layer in wind tunnels is well documented and has been validated when compared to
20
full-scale measurements. The same standardization and disclosure of boundary layer modeling
is not as established in CFO; however, collaborative research efforts are underway to improve the
accuracy and reliability of CFO modeling for wind flow in urban environments (Franke, et al.
2007).
In CFO analysis, the boundary layer conditions are characterized by formulae at the inlet
conditions that define flow (Blacken, Stathopoulos and Carmeliet 2007, 239). Several examples
of turbulence models used in CFO include Reynolds-Averaged Navier-Stokes (RANS), Detached
Eddy Simulations (DES), and Large Eddy Simulations (LES). The RANS models were originally
developed for streamlined flow applicable to the aerospace, ship and automotive industries.
When applied to surface mounted bluff bodies, the RANS models do not perform well in
predicting flow separation and reattachment. This is especially true of low-rise buildings where
flow-separation around corners and roofs within the boundary layer are frequent. Researchers
acknowledge that RANS models are useful to specific applications, but not all - including wind
engineering (Hanjalic 2005). Instead, the future of CFO modeling will use unsteady simulations to
predict structural loads in wind engineering. Examples include the aforementioned DES and LES
models, as well as the Unsteady Reynolds-Averaged Navier-Stokes (URANS). It is accepted that
LES can provide better results for CP compared to RANS simulations. The use of unsteady
models in transient simulations requires more sophisticated definition of the inlet profile, which
results in increase strain on computational resources. An advantage of utilizing CFO to perform
transient analyses is that the effects of both steady and dynamic loading can be modeled to
evaluate the effects of fluctuating wind loads. Though CFO is rarely used as a source of CP data
currently, it might become an important source either on its own, or integrated within BES-AFN
programs (Costola, Blacken and Hensen 2009, 2035). For enclosures intended to provide natural
ventilation, measured wind pressures from wind tunnel models may be used as input boundary
conditions for unsteady CFO modeling of internal flows (Irwin, Denoon and Scott 2013, 25). The
reference to CFO playing a supplemental role to wind tunnel testing is increasing.
21
The potential of CFO modeling to predict for CP data for building facades is promising, but
requires further standardization and refinement. The most significant advantage that CFO has
compared to wind tunnel or full-scale testing is that it is not limited by scale. Additionally,
implementation of CFO modeling during design development could inform crucial early decisions
rather than playing a validation role. The reality is that for CFO modeling to become accepted it
must become more accurate, and this can only happen by working with wind tunnel and full-scale
testing to justify its reliability. By using physical testing to calibrate the pressure distribution, CFO
could be used to predict pressures on more complex facades, including externally ventilated
double-skin facades.
1.3.4 Databases
Wind pressure coefficient databases are collections of CP data from a single or multiple sources,
usually classified by parameters such as geometry and orientation relative to the wind attack
vector. CP databases, such as tables provided in wind load standards (i.e. ASCE/SEI 7-10
Minimum Design Loads for Buildings and Other Structures), are the most common form of data
used in BES-AFN models. These databases are often derived from wind tunnel experiments and
applicable to generic conditions for simple geometries of unsheltered buildings. Two other
examples of CP databases include the Air Infiltration and Ventilation Center (AIVC) references
and the ASHRAE Handbook of Fundamentals; however, these too are limited to simple building
geometries and surrounding conditions. An additional source of CP data is the web-based
application Cp Generator that "predicts the dimensionless static wind pressure coefficients, CP, on
the facades and roofs of blocked shaped buildings" (TNO Building Reseach). The values
generated are based on wind tunnel and field-measured tests. For simple geometries, tools like
Cp Generator serve as a means to determining coefficients of pressure without the expense of
wind tunnel studies. Though fit and validated for simple conditions, Cp Generator should not be
used to determine coefficients of pressure on complex building structure geometries or systems
with small openings.
22
1.4 New Approaches to Dynamic Wind-Loading on Double-Skin Facade
Structures
Increasing accuracy demands in today's simulation tasks involve simultaneously considering
multiple physical effects - resulting in a multiphysics problem (Bungartz and Schafer 2006, v).
For example, fluid flow can continuously deform surrounding structures. If the deformations are
large, they will in turn affect the fluid flow. This behavior is present in double-skin facades,
especially those with a more flexible outer skin (i.e. cable-net facade - see Section 5.3.3). The
magnitude of the structural deformation's effect on the fluid flow through the cavity space is
unknown and requires further investigation under a series of loading conditions.
1.4.1 Fluid-Structure Interaction
The structural deformation interaction with fluid flow for a double-skin facade described above is
an intricate behavior. Fluid-structure interaction (FSI) problems are complex because they
consist of structural nonlinear boundary conditions imposed on fluid moving boundaries where the
position is part of the solution. Because the moving position of the structure prescribes part of the
fluid boundary, it becomes necessary to perform the integration of the Eulerian fluid equations on
a moving mesh (Vazquez 2008, 149). This iterative computation requires substantial
computational power on the magnitude of a non-linear structural finite-element analysis.
1.4.2 Multiphysics Coupling
The multi-faceted analysis of both structure dynamics and fluid flow require a linkage between the
two. This link is often referred to as a couple. The combination of these analyses result in a
multiphysics solution. Fluid-structure interaction problems usually are viewed as a two-field
coupled problem, however the moving mesh common to both physics-models can be viewed as
another structural problem and therefore the complete coupled problem can be formulated as a
three-field system: the structure, the fluid and the moving mesh. Structural formulations are
commonly based on Lagrangian descriptions. Fluid formulations are commonly Eulerian
23
descriptions. In FSI problems, the Eulerian mesh of the fluid is not fixed in space anymore since
the motion of structure moves the fluid mesh. The Arbitrary Lagrangian-Eulerian (ALE)
formulation is a hybrid technique to combine the advantages of the Lagrangian and Eulerian
methods while minimizing the disadvantages (Vazquez 2008, 152). In Figure 1-7, a simplified
two-dimensional model of a double-skin facade subject to a sinusoidal gust in a fluid-structure
interaction domain is shown. Notice the location of the mesh nodes is changing from frame to
frame, as is the pressure magnitude (evident by the color change).
t =O s t= 0.5 s t= 1.0 s
t = 1.5 s t = 2.0 s t = 2.5 s
0
·l
-2
-3
-4
Pa
Figure 1-7: 20 FSI model of a simple DSF subjected to a sinusoidal 3-second gust velocity
profile. Moving mesh and pressure magnitudes shown fort= {O, 0.5, 1.0, 1.5, 2.0, 2.5}.
24
1.5 Definition of Terms
1.5.1 Double-Skin Facade (DSF)
A double-skin facade is a strategy for improving building envelope performance through the
introduction of a second transparent glazed layer, resulting in an airflow cavity ranging from 10
cm (4 in) to 2 m (6.6 ft) in depth.
1.5.2 Transparent Ventilated Facade (TVF)
A transparent ventilated facade uses an air gap between two glazed panes to reduce the thermal
impact on the building environment (Perino 2007, 12).
A double-skin facade is often a transparent ventilated facade, but the term TVF does not include
opaque ventilated facades, also thought of as Trombe walls.
1.5.3 Advanced Integrated Facade (AIF)
An advanced integrated facade is a building envelope that possesses adaptive qualities in
response to both local climatic context and indoor environment to provide basic functions while
minimizing energy consumption (Perino 2007, 11).
A transparent ventilated facade is one type of advanced integrated facade. other types include
opaque ventilated facades (i.e. Trombe walls) and operable windows tied into the HVAC system
1.5.4 Intelligent Facade
An intelligent facade is "a composition of construction elements confined to the outer, weather
protecting zone of a building, which performs functions that can be individually or cumulatively
adjusted to respond predictably to environment variations, to maintain comfort with the least use
of energy" (Wigginton and Harris 2002). The concept of intelligence associated with double-skin
facades represents a change from a static to a dynamic envelope that is capable of adapting to
25
changes in outdoor conditions in order to achieve indoor comfort requirements and reduce energy
consumption.
1.5.5 Building Energy Simulation (BES)
Building Energy Simulation (BES) is the use of computer simulation methods for the analysis of
energy efficiency and building load consumption profiles.
1.5.6 Aerophysics
Aerophysics is the study of airflow on, around and within a building.
The investigation of aerophysics associated with double-skin facades includes several interesting
areas of focus across the entire envelope and at localized regions within the cavity. These
include pressure due to wind action, pressure differences due to thermal buoyancy, airflow
through inlets and outlets, airflow through the cavity, airflow around shading, ventilation and load
sharing.
1.5.7 Cavity Partitioning
Cavity partitioning is the subdivision of the double-skin cavity geometry.
1.5.8 Ventilation Type
The driving airflow forces at the origin of the cavity ventilation: natural, mechanical or hybrid.
1.5.9 Ventilation Mode
Ventilation mode is a characteristic of the cavity airflow path from entry to exhaust orifices.
1.5.10 Fluid-Structure Interaction (FSI)
Fluid-structure interaction (FSI) is the interdependent interaction of a deformable structure with a
fluid flow.
5
26
1.5.11 Atmospheric Boundary Layer (ABL)
The Atmospheric Boundary Layer (ABL) is the bottom layer of the Earth's atmosphere, the
troposphere, which is in contact with the surface of the earth. This zone above the earth's
surface is often turbulent, may extend a height of 100 meters or so, and is also known as the
friction layer, or boundary layer. Wind speed increases with increasing height above the ground
until reaching a constant - gradient wind speed - at a height where the effects of surface friction
are negligible. In engineering practice, a power law function is used to characterize the varying
vertical velocity profile.
The atmospheric boundary layer affects the design of all buildings with larger wind velocities
applied at higher elevations of the structure. Additionally, the replication of the atmospheric
boundary layer is at the core of uncertainty associated with wind tunnel testing and, even more
so, computational fluid dynamics.
1.5.12 Wind Pressure Coefficients (Cp)
Wind pressure coefficients (Cp) is a derived quantity which requires data for pressure on the
facade (Px), reference wind speed (Vref), and - technically speaking - air temperature and
atmospheric pressure to determine the air density (p). The relationship between the
dimensionless coefficient and the dimensional numbers is,
1.5.13 Mean Pressure
Mean pressure is the average pressure over a sampling period or area is known as the mean
pressure (p). Similarly, the average wind pressure coefficient across a region is known as the
mean wind pressure coefficient (Cp).
27
1.5.14 Peak Pressure
The largest pressure from a sampling period or area is known as the peak pressure (p). Similarly,
the largest wind pressure coefficient magnitude across a region or from a time period is known as
the peak wind pressure coefficient (Cp)·
1.5.15 Net Pressure Coefficients (Cp,netl
The difference in pressure coefficients of the exterior and outer skins, CP·"" (or !1CP = Cp.n,- Cp.m,).
1.5.16 Outer Skin Pressure Coefficient Ratio
A comparative ratio of a DSFs outer skin pressure coefficient to a single-skin's, cp.DSFcmm I cp.smgfr.
1.5.17 Opening Area Ratio
The opening area ratio (OAR)
6
is the area of the openings (A,) divided by the total wall/cavity
surface area (a* hrnv):
1.6 Hypotheses
This research focuses on the following two primary hypotheses. Additional objectives are outlined
in Section 6.2.
Hypothesis 1: Computational fluid dynamics with multiphysics software is an essential
design tool in the determination of pressure coefficients for multi-story double
skin facades.
Hypothesis 2: Airflow openings in multi-story double-skin facades have significant effects on
the design pressure requirements.
28
1.7 Scope of Study
This research focuses on the structural response of multi-story double-skin facades by evaluating
the wind pressure distribution upon the exterior and interior layers. The variables considered in
this evaluation include the cavity aspect ratio (height/depth), ventilation geometry, and opening
area ratio. Multiphysics software is used to model the steady-state and transient CFO simulations.
The results will be compared against one another, as well as against a base case single-glazed
facade. Fluid-structure interaction, which combines the effects of structural deformation with fluid
flow and considers the outer skin's structural flexibility as a variable, is occasionally referenced,
but outside the scope of this study.
1. 7 .1 Multi-Story Configuration with Outdoor Air Curtain
In Chapter 3 it will be evident that there are many possible configurations and airflow solutions a
double-skin facade can possess. Despite the numerous classifications and options for designers,
one configuration and one airflow strategy are present in a majority of the double-skin facades in
the United States. Chapter 7 evaluates 30 double-skin facades in the United States to find that of
the built DSFs sampled, 70% are of a multi-story configuration and 70% use an outdoor air
curtain airflow mode.
7
Considering the emergence of a preferred strategy, the simulation
component of this research will concentrate on a multi-story DSF with an outdoor air curtain.
Figure 1-8: concept diagram of multi-story DSF with an outdoor air curtain.
29
1.7.2 Description of Six Airflow Intake Configurations
The airflow mode of a double-skin facade is the result of an entry and exit strategy. The airflow
intake configurations considered in this research include 1) sealed, 2) face (flush), 3) trench
(recessed), 4) raised (entry), 5) louvered and 6) shingled, as diagrammed in Figure 1-9.
EXT INT EXT INT
Sealed Face I Flush
EXT INT EXT INT
Recess I Trench Raised I Entry
Louvered Shingled
Figure 1-9: Airflow intake configurations occurring near the ground plane.
30
1.7 .3 Description of Four Airflow Exhaust Configurations
The airflow exhaust configurations considered in this research include 1) sealed, 2) forward, 3)
return and 4) through, as diagrammed in Figure1-10. These configurations may be combined
with the aforementioned airflow intake configurations to create a number of ventilation modes.
8
EXT INT EXT INT
Sealed Forward
EXT INT EXT INT
Return Through
Figure 1-10: Airflow exhaust configurations occurring near the roof plane.
31
1.8 Research Methodology for Evaluating
The method employed by this research is computational fluid dynamics of steady-state and
transient problems representative of wind tunnel test sections with scaled building geometries.
Calibration of the CFO analysis is performed using previous wind tunnel studies. The use of
simulation for this research requires special attention be given to tuning the modeling process in
an effort to preserve the relevancy of the results. Researchers have identified the potential of
CFO modeling, when used in conjunction with wind tunnel studies, to tackle challenging design
situations. Once standard modeling procedures are established, using wind tunnel testing to
calibrate CFO modeling may be used to predict accurate pressure distributions on facade
structures (Zammit, Overend and Hargreaves 2010, 659-660). The following sections describe
important features of the methodology. A more elaborate outline is presented in Chapter 6 as well
as in the individual sections preceding each primary analysis type.
1.8.1 Calibration of a Virtual Simulation
Without readily available wind tunnel facilities as an option, two calibration models will be
adopted: one of a simple cube and the other with an unsheltered tower that is also examined with
second-skin configurations. The simple cube calibrations occur at a model scale of ,l = 1i100 and
reference velocity of U
0
= 7 mis. The second calibration (unsheltered and sheltered towers) is the
most similar physical testing of multi-story double-skin facades to date: Marques da Silva and
Gomes' testing of Gap inner pressures in multi-storey double skin facades (2008) in an open
circuit boundary layer wind tunnel. The research of Marques da Silva and Gomes is modeled
using an unsheltered tower with model dimensions of 32.5 cm wide, 20 cm deep and 70 cm tall,
and the same tower with a second-skin with the outer skin beginning 7.5 cm above ground level.
All calibration models herein maintain the study parameters, including: model scale of\= 1i40;
wind tunnel section 9 m x 3 m x 2 m; inlet velocity of Urer= 10.5 mis; a velocity profile power law
equation exponent of a= 0.18: and a suburban exposure (Marques da Silva and Gomes 2008,
1554-1555). A detailed description of this research is outlined in Section 2.3.1.
32
1.8.2 Steady-State Analysis
All simulations are computational fluid dynamics models of wind tunnel test sections with scaled
building geometries contained within. All of the double-skin configurations' simulations that make
up this research rely on the parameters outlined in Section 7.5. The 30 steady-state simulations,
representative of the wind tunnel test section, incorporate a Reynolds Averaged Navier Stokes
(RANS) turbulence flow model type with a k-B turbulence model with incompressible flow. The
simulations are conducted in the CFO module of COMSOL Multiphysics 5.0.0.244. Thirteen
models are simulated in total: one of a single-skin configuration and twelve of various airflow
opening configurations. The process for the simulations is further outlined in Section 9.2.
1.8.3 Transient Analysis
Preliminary 20 transient simulations with a k-B turbulence model are conducted to
investigate the sensitivity of the analytical prototype to 1) different dimensional variables
(height, depth and permeability) and 2) various airflow intake (face, trench, raised) and
exhaust (forward, through and return). See Appendix E for these simulations' findings.
1.8.4 Comparative Studies
The simulation models' wind pressure distribution characteristics are evaluated by
extracting 1) contour elevations for both the interior and exterior skins, and 2)
characteristic profile curves at key vertical and horizontal cut locations. The local
pressure coefficients for each of the inner and outer skins are used to determine the net
pressure coefficients occurring across the outer skin. These results of each airflow
intake/exhaust combination are compared to one another as well as to the results of a
single-skin facade. The comparison process is carried out in Section 9.8. Additionally,
qualitative comparisons are made between the findings from the 30 steady-state analysis
to those of the preliminary 20 transient studies to identify any supporting affirmations or
conflicting findings. See Section 9.9.
33
1.9 Organization of Dissertation
Chapter 1 introduces the importance of wind pressure distribution characteristics in the structural
engineering and energy simulation fields. The current codes, practices and emerging analytical
techniques are briefly reviewed. Additionally, terminology and abbreviations that are integral to
this document are defined.
A comprehensive review of previous research in modeling wind pressure on double-skin facades
is presented in Chapter 2. The review begins by introducing the various interests of researchers:
typology, thermal comfort, airflow modeling, energy simulation, fire/life safety and life cycle
assessment. Research on pressure equalization, rain-screen principles and pressure distribution
on facades are presented as antecedent technologies. The last part of Chapter 2 presents
research specifically focused on wind interaction of double-skin facades - the most pertinent to
the focus of this research.
Chapters 3, 4 and 5 establish a common discourse and historical context of double-skin facades.
Chapter 3 establishes a clear definition, reviews existing typologies and proposes a classification
scheme. A summary of advantages, disadvantages and project approach considerations is
included. Chapter 4 reviews the evolution of facades towards the double-skin and presents
international case studies. Chapter 5 then evaluates double-skin facade applications in the United
States and defines a set of emerging trends.
Chapter 6 and 7 serve as a transitory bridge between the background and the analytical chapters.
In Chapter 6 the research design and methods are reviewed, including an outline of hypotheses,
objectives, procedures and data analysis. Next, Chapter 7 identifies a prototype configuration for
simulation of a multi-story double-skin facade. The multi-story is evaluated and several
classifications specific to the type, primarily related to airflow and structure characteristics, are
offered.
34
Chapter 8 presents a calibration of the three-dimensional steady-state analysis to existing wind
tunnel experimental results of scaled simple cubes and towers with a second skin. The
calibrations seek to determine the wind tunnel experiment's parameters that develop the tunnel's
characteristic turbulence, particularly the unknowns of inlet turbulence intensity (Ir, mlee) and
roughness length (z
0
).
Chapter 9 outlines the steady-state analysis models for a single-skin facade and twelve double
skin facades with different combinations of airflow intake and exhaust openings. The chapter
also includes results and comparisons for each of the experimental models to one another, as
well as to the single-skin facade's pressure distribution.
Chapter 10 discusses the primary results for multi-story double-skin facade pressure profiles as
well as the role of computational fluid dynamics as a supplemental tool in the project delivery
process. The chapter also includes a set of multi-story double-skin facade guidelines and
suggests a methodology for the determination of multi-story double-skin facade pressure
coefficients.
Appendix A presents 40 case studies of double-skin facades from around the globe.
Appendix B presents the statistical breakdown of 215 DSF project locations.
9
Appendix C presents 30 built and six forthcoming double-skin facades from the United States.
Appendix D presents supplemental documentation for the calibration models shown in Chapter 8.
Appendix E presents the preliminary two-dimensional transient analysis models, procedure and
results that informed the research to focus on the primary variable of airflow opening
configurations.
35
Chapter 1 Endnotes
2
3
This image was created by the author in October 2011 for the Advanced Technology Studio of
Enclos' Insight 02 publication. It is a diagrammatic rendering of the multi-story double-skin
facade at the Columbia University's Jerome L. Greene Science Center. The Insight 02
publication may be accessed at http://www.enclos.com/library/publications/insight-02>.
ASCE/SEI 7-10, a revision of ASCE/SEI 7-05, presents a complete reorganization of the wind
load provisions, expanding from one chapter (Chapter 6 in ASCE/SEI 7-05) to six chapters
(Chapters 26 -31 in ASCE/SEI 7-10). The commentary chapters have also seen an expansion
of commentary items related to the wind loading provisions.
In ASCE/SEI 7-10, Section 30.1.3 Limitations, the loads on buildings that have "unusual shapes
or response characteristics, shall be determined using recognized literature documenting such
wind load effects or shall use the wind tunnel procedure specified in Chapter 31," (ASCE/SEI
2010, 315).
4
Though double-skin facades are not specifically addressed, in ASCE/SEI 7-10 Chapter C30
there is increased commentary (relative to what was contained within ASCE/SEI 7-05)
regarding air permeable cladding. The tone of some of the guidelines is beginning to sound
similar to what we see in the Eurocode regarding communication between adjacent skins' air
cavities.
5
6
In double-skin facades, fluid flow can continuously deform the exterior structure, airflow
openings and shading. If the deformations are large, they will in turn affect the fluid flow. This
behavior may be increasingly present in double-skin facades with multi-story configurations,
long-span structures, and those with a more flexible outer skin.
Some studies in this document characterize the opening area ratio by using wall permeability
(s). In their research Wind Loads on Wind Permeable Facades, Gerhardt and Janser utilized a
boundary layer wind tunnel to perform a parametric study that systematically evaluated porous
facade systems with wall permeability (s). The wall assembly in this research had a gap
dimension of 75 mm (3 in).
7
The statistical figures in this document differ slightly than those presented in (Vaglio and
Patterson 2011, 35) and (Vaglio 2011, 398) because the number of projects in the case study
sample base has increased: 23 built projects in the previous articles, 30 built projects herein.
8
See ventilation mode definitions and diagrams in Section 3.2.2.
9
The breakdown of country locations for the 215 double-skin facade projects from around the
globe are extracted from the International Energy Agency's, Annex 44: Integrating
Environmentally Responsive Elements in Buildings, State of the Art Review Vol. 2A:
Responsive Building Elements (Perino 2007, 23).
36
2. Background Research in Modeling Pressure on Double-Skin
Facades
2.1 Overview of DSF Modeling Focuses
The primary emphasis of recent research is on the thermal performance (Tanaka, et al. 2009),
(Jiru and Haghighat 2008), (Gratia and Herde 2007), (Manz and Frank 2005) and energy
modeling (Manz 2004), (Saelens, Carmeliet and Hens 2003), (Hensen, Barlak and Frantisek
2002) of double-skin facades. other pertinent research focuses include acoustics, shading
optimization, fire spread, and life-cycle assessment (LCA). Investigation into the structural
performance and pressure distributions of double-skin facade systems requires further
exploration. Past research conducted in the area of pressure equalized rain-screen (PER) design
and pressure equalized cavities provides a precept for analysis of double-skin facades; however,
the geometric scale and modes of ventilation associated with DSF configurations are more
expansive and affect the pressure distributions across internal and external layers. Additionally,
the structural flexibility of the interior and exterior skins affects the load sharing distribution of a
double-skin system.
This Chapter presents an overview of common research interests pertaining to multi-layered
facade system evaluation, followed by an expanded review of literature focused on wind
interaction and pressure distribution characteristics on double-skin facades.
2.2 Research Related to Pressure Distribution Modeling
Two relevant areas of research that inform double-skin facade wind interaction research include
the pressure distribution on single-skin facades and pressure equalized rain-screens. Key
research efforts into both of these areas are presented herein to establish a foundation and
discourse from which double-skin facade research has evolved.
37
2.2.1 Single-Skin Facade Pressure Distribution
The determination of pressure coefficients on building structures is integral in all structural codes.
These generalized guidelines utilize simple building geometries with single layer enclosures.
Zammit, Overend and Hargreaves (2010)
The research paper Improved computational methods for determining wind pressures and glass
thickness in facades acknowledged that the wind flow effects on the multifarious geometries and
structures often fall "outside the scope of wind loading codes of practice" (Zammit, Overend and
Hargreaves 2010, 655). Since wind loading codes are generalized to simple building geometries,
wind tunnel testing is often employed in large projects, projects with complex facades, or
expensive facade scopes. Zammit, Overend and Hargreaves presented an evaluation of five
wind loading codes,
1
a wind tunnel study and a computational fluid dynamics values for a
simplified geometry (see Table 1-1). This comparison showed that 1) there is variance between
the wind loading codes, and 2) the predicted facade pressures from the wind loading codes were
noticeably higher than those measured in wind tunnel tests (Zammit, Overend and Hargreaves
2010, 656).
The paper proceeded to discuss the current and future roles of computational fluid dynamic
(CFO) modeling and how CFO may be used to complement wind tunnel data. The authors
mentioned the physical limitations of wind tunnel testing, the absence of scalar limitations to CFO
models, and the prospects of using wind tunnel testing to calibrate a CFO model and progress
towards predicting accurate pressure distributions via simulation. Amongst the data presented
was a pressure distribution mapping for a cubed-shaped single-facade building' subject to normal
indices of wind (see Figure 2-1). The research also included a transient model simulation to
characterize the effects of fluctuating wind load on the structural engineering of glass. The
investigation utilized existing full-scale test data and transient simulations to calculate the
equivalent uniform glass stresses for equivalent load duration.
38
-0.2
-0.3
-0.3
=
-0.3
-0.5
-0.5
-0.5
-0.7
-0.7 eA! .1
-0.7
O_
Figure 2-1: CFO pressure distribution (Cp) diagram for wind normal to cube.
Information for this figure based on Improved Computational Methods for Determining Wind
Pressures and Glass Thickness in Facades (Zammit, Overend and Hargreaves 2010, 659).
Within the conclusion of Improved Computational Methods for Determining Wind Pressures and
Glass Thickness in Facades, the authors suggested "advances in CFO make it possible to give
further information on pressure distribution over facades, particularly over intricate geometries
such as brise soleils and double skin facades". Additionally, within the future work section, the
authors emphasize "CFO analysis and wind tunnel testing will be used together to enable the
accurate design of double skin facades" through (Zammit, Overend and Hargreaves 2010, 666).
39
2.2.2 Pressure Equalized Rain-Screen (PER)
A pressure equalized rain-screen (PER) consists of a protective, porous outer cladding that
permits extraction of moisture, an airflow cavity, and an impermeable building wall. In Europe, a
comparable system is often referred to as a back-vented rain-screen (BVR). The wind loading on
the exterior, protective layer is the pressure differential between the exterior environment and the
gap pressures.
Past research conducted in the area of pressure equalized rain-screen (PER) design (Suresh
Kumar and Wisse 2001 ), (Quirouette and Rousseau 1998), (Choi and Wang 1998), (Baskaran
1994), (lnculet and Davenport 1994) and pressure equalized cavities (Sharma 2007), (Xie,
Schuyler and Resar 1992), (van Schijndel and Schols 1998) provides a guideline for analysis of
multi-layered systems; however, the geometric scale and modes of ventilation associated with
double-skin facade configurations are more extensive and influence the load sharing between the
internal and external skins. Pressure equalization is a principle that has been implemented with
facade rain-screens for many decades. However, "true pressure-equalized rain-screen (PER)
walls are rare," (Brock 2005, 31). The objective of pressure equalization of a cavity within a
facade system is to prevent water penetration into the cavity. In reality, most existing curtain wall
types "fall somewhere between a PER and a drainage cavity wall," (Brock 2005, 31 ). By
equilibrating the cavity with the exterior pressure, water will not migrate into the intermediate gap.
A fundamental difference between the traditional curtain wall rain-screen principle and its
application to double-skin facades is that the latter may include interior facade openings, where
the rain-screen principle restricts openings in the back, impermeable surface. Additionally, a rain
screen is panelized, introducing airflow passages frequently and regularly across the height of a
building envelope. Contrastingly, a double-skin facade may consist of a wide array of opening
configurations as described in Section 1.6.2. For instance, a multi-story double-skin facade often
has an airflow inlet at the bottom and exhaust at the top - unlike a PER
40
Xie, Schuyler and Resar (1992)
The net pressure on pressure equalized cavities was studied using full-scale physical tests that
were then used as a baseline comparison for a predictive numerical model. The full-scale panel
studied, representative of a curtain wall spandrel panel, was 1.5 m (5 ft) wide by 1.2 m (4 ft) tall
with an initial cavity volume of 0.14 m
3
(5 ft\ The inner skin was composed of galvanized sheet
steel while the exterior skin was 5 mm (3/16 in) aluminum. An opening placed at the bottom of
the exterior skin resulted in a 3 mm (1/8 in) gap along the full panel width (1992, 2450).
The full-scale tests substantiate that load reductions are possible with suitable pressure equalized
curtain wall design and that the reduction factor is a function of the ventilation opening area ratio
as shown in Figure 2-2, where Pel P
0
is a ratio of peak external pressure to the mean pressure
(Xie, Schuyler and Resar 1992, 2453). The pressure reduction factor below considers the net
peak pressure across the exterior cladding and its relationship to the peak imposed pressure.
1.0
0.9
0.8
0
0.7
t5
(lj
LL
0.6
c
0
-~
:J
0.5
""C
Q)
0:::
Q)
:; 0.4
(f)
(f)
Q)
ct
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Ah/A[%]
( Vent Hole Area I Panel Area )
Figure 2-2: Pressure reduction factor vs. ventilation opening area ratio on curtain wall
spandrel panel. Information for this figure based on Prediction of Net Pressure on Pressure
Equalized Cavities (Xie, Schuyler and Resar 1992, 2458).
41
A pointed insight was provided by the authors in the closing paragraph that may serve as a
precept in determining double-skin facade pressure distributions. A cautionary message about
the applicability of the results found in the paper recognizes that other factors (not considered in
their study) affect the net pressure across the exterior skin of a pressure equalized curtain wall -
including communication between the cavities. The advisory warns against communication
between one building face to another and encourages compartmentalization of cavities. In the
words of the authors, "there should probably be no air movement from floor to floor, around
corners, or greater than ten feet on a flat face" (Xie, Schuyler and Resar 1992, 2456).
Gerhardt and Janser ( 1994)
In their research Wind Loads on Wind Permeable Facades, Gerhardt and Janser utilized a
boundary layer wind tunnel to perform a parametric study that systematically evaluated porous
facade systems and their sensitivities to building dimensions (height-to-width, h/a, and depth-to
width ratios, (b/a), wall permeability (s), relative gap width (gap-to-width ratio, s/a), and oncoming
wind flow profile and direction (a). The wind tunnel model scale was 1 :350 (Gerhardt and Janser
1994, 39) and the system configurations studied were representative of a pressure-equalized
rain-screen, though the notions of convective facades and double-skin facades' were alluded to
(37). This paper astutely recognized two key factors
4
: (1) the role of the net pressure coefficient
(Cp,nerl across the exterior skin (see Figure 2-3), and (2) the net pressure coefficients are
effectively independent of the oncoming flow (see Figure 2-4).
The latter portion of the paper submitted a comparison between full-scale measurements of a
back-vented rain-screen facade with a cavity depth of 75 mm at ETERN IT AG, Berlin, to wind
tunnel tests of a comparable configuration at a scale of 1 :150 (Gerhardt and Janser 1994, 45).
Since the configuration common to the full-scale and wind tunnel did not possess cavity airflow
communication with other building faces, the authors concluded that the pressure equalization
occurs primarily across the facade panels, resulting in very low net pressure coefficients (47).
42
-1.4
b/a = 2, aP = 0.1 , E = 1 O/o
-1.2
-1
~
I()
~
:J
-0.8
"'
"'
~
0...
0
c -0.6
....
Q)
~
"(3
ti=
I
Q)
0
()
-0.4
.. :\.
-0.2 .....
0
0 2 3
h/a
cp,int
~~ ._.,.. s/a = 0.005
J ,,,_ Q.Q10
------T
4 5
Figure 2-3: Influence of gap flow resistance on the net pressure coefficient. Information for
this figure based on Wind Loads on Wind Permeable Facades (Gerhardt and Janser 1994, 41 ).
-6
b/a = 2, E = 1 % , s/a = 0.00251
-5
. . ~
~A-
-4
~=·2~k d~
a =0.1 I ._
I
p ....
+
~
()
-3
-
~
:J
"'
"'
~
-2 -
0...
0
/\
c
cp,oet
Q)
-1 -
"(3
ti=
Q)
0
()
0 -
..>t.
----
OI
Q)
/\
0...
I
~
!
I
1
c
p, int
?
0
2 -
3L------------------------------'L
0 2 3 4
h/a
Figure 2-4: Influence of oncoming flow on net pressure coefficient derived from the peak
factor approach. Information for this figure based on Wind Loads on Wind Permeable Facades
(Gerhardt and Janser 1994, 44).
43
2.3 Double-Skin Facade Wind Interaction
The ramifications of the double-skin facade on the wind pressures across the inner skin and outer
skin have been investigated sporadically. Box-window configurations, often unitized, have been
the focus of a number of early pressure distribution and load sharing investigations since they are
readily mocked-up at full-scale, and lend themselves more to a manufactured product approach
than multi-story one-off configurations. The research of Asahi Glass of Japan (Ishida 2003)
evaluated the influence of openings, depth and airflow paths on the distribution load ratio of
pressure on the inner and outer skins of a box-window DSF at full-scale. Another full-scale field
measurement of box-windows, in this case located on opposing building facades allowing for
cross ventilation, each with operable units for natural ventilation, was performed more recently by
Kawai, Nishimura, Suzuki and Oura (2009) to characterize the peak pressure coefficients and
mean pressure coefficients from measurements collected over a 15 month period. The box
window configuration has also been the emphasis of thermal, energy and comfort research
investigations for their innate capacity to be studied at various scales and in multiple modalities -
theoretical, numerical, simulation (computational fluid dynamics), scaled experiments (wind
tunnel) and full-scale experiments, such as mock-ups.
The most pertinent research of pressure distributions on multi-story double-skin facades includes
Marques da Silva and Gomes (2008) investigations of inner-face pressure distributions of multi
story DSF models obtained through wind tunnel tests for five wind incidences applied to four
different systems with three gap depths and Wellershoff and Hortmanns (1999) wind tunnel study
on DSF systems with gaps larger than 15 cm. The results of these investigations demonstrate
the sensitivity of pressure distribution in double-skin facades to project specific site conditions and
design configurations. Recent research on the ballooning of flexible membranes (Shi and Burnett
2008) also has potential application to double-skin cavities with an exterior skin subject to large
deformations, such as a cable-net facade. There is a need for a standardized approach for the
determination of wind loading and pressure coefficients on double-skin facade systems.
44
2.3.1 Multi-Story Pressure Distribution
Marques da Silva and Gomes (2008)
One recent wind tunnel evaluation of double-skin facade pressure distribution is the most similar
to the research objective of this dissertation. Gap Inner Pressures in Multi-Storey Double-Skin
Facades by Marques da Silva and Gomes (2008) used boundary layer wind tunnel testing to
determine mean pressure coefficients (Cp) on the inner skin of a series of model configurations.
This research referenced several key considerations of why the pressure within the double-skin
cavity is significant: glazing structural safety, natural ventilation strategies, and the safety of
maintenance staff that may occupy the cavity or breach the inner skin seal to gain cavity access.
The results support the conclusion that wind load evaluation established in codes requires
revision (Marques da Silva and Gomes 2008, 1558).
This research focused solely on the multi-story condition where there is no vertical or horizontal
partitioning between the two skins. A series of wind tunnel tests under a boundary layer velocity
profile were performed for four different DSF layouts, three cavity depths, at five wind incidences.
Each layout included openings at the base and top allowing airflow to occur within the cavity. The
testing occurred in an open-circuit wind tunnel with a 3 m width, 2 m height, 9 m length and a
contraction ratio
5
of 4.5:1.
6
All experiments implemented a boundary layer incident flow with a
velocity of-10.5 mis with a velocity profile represented by the power law function with an
exponent aP = 0.18 (Marques da Silva and Gomes 2008, 1554-1555).
The authors described the 1 :40 scaled building model used in the wind tunnel experiments in
dimensional terms. In an effort to characterize the building models used in various experiments,
this review will translate the dimensional values provided into relative, dimensionless terms.
The tested specimen was a rectangular geometry with a relative building height, hf a= 2.15,
relative aspect ratio bf a= 0.615, and varied relative gap width of sf a= 0.025, 0.037 and 0.049,
45
where a is the building width. The tested depths are also characterized by the cavity aspect ratio
(height:depth), where h/s = 87.50, 58.33, and 43.75. A second-skin was placed on two sides of
the model starting approximately 0.107 h above the ground level and extending the full height of
the building. The experimental approach is best summarized in Figure 2-5 where the various
layouts, wind indices and pressure tap locations are diagramed. The four layouts included A) a
cavity open on all sides, B) continuous corner with open edges, C) continuous corner with sealed
edges, and D) fully lateral closure.
-'- rna•Y
13;"
0.9
go•>-
0.8
45°-1 a· A
0.7 Layout A
0.6
I
......
0.5
N
0.4
Layout B
0.3
0.2
Layout C
0.1
E9 E9
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0. 7 0.8 0.9
x/L
Layout D
- Measured Facade
Figure 2-5: Pressure tap locations (left), wind incidences (top right) and DSF layouts
evaluated in wind tunnel testing (right). Information for this figure based on Gap inner
pressures in multi-storey double skin facades (Marques da Silva and Gomes 2008, 1554).
46
The results presented in the paper for configurations exposed to wind incidences normal to the
front facade exhibit a negative CP pressure distribution on the inner facade in regions behind the
second skin (see Figures 2-7 and 2-8) compared to the single-skin facade (see Figure 2-6). The
single-skin region, which exists near the ground, exhibited both positive and negative pressures
where the wind acts directly. Compared to the single-skin facade, the double-skin configurations
exhibited a more variant and complex pressure distribution.
0.9
0.8
0.7
0.8 0.8
0.6
0.6
0.6
0.4
0.5
::c:
......
0.4 N
0.4
0.8
0.2
0.2
0.3
0.2
0.1
D
Front
t
0 oo
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
x/L
Figure 2-6: CP values measured on models for frontal wind incidences for single facade.
Information for this figure based on Gap inner pressures in multi-storey double skin facades
(Marques da Silva and Gomes 2008, 1555).
47
The experimental data for the multiple double-skin configurations illustrated a significant
difference in the pressure distribution compared to the single-skin configuration. The interior skin
pressure distribution was layout sensitive, but always negative regardless of the incident wind
direction (Marques da Silva and Gomes 2008, 1558). The consistent negative interior pressure
has also been identified in other double-skin research (Potangaroa and Aynsley 2003) where the
system was open at the top but closed at the bottom.
0.9
0.8
0.7
0.6
-1
-1
:::c
0.5 ......
N
0.4
0
0.3
-1
0.2
0.1
IDO.Sm
Front
i
0
oo
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
x/L
Figure 2-7: CP values measured on models for frontal wind incidences for open double-
skin with 0.8 m (equivalent) depth. Information for this figure based on Gap inner pressures in
multi-storey double skin facades (Marques da Silva and Gomes 2008, 1555).
48
The lower area of the front elevation that did not include a second layer had a strong pressure
gradient between the ground and start of the second skin. Higher cP values were measured at
the lower edge of the raised double-skin facade as the cavity depth narrowed (1557). This
sensitive region is important in the design of double-skin facades, serving as an entrance into the
building interior. Other sensitive regions with heightened magnitudes of pressure were identified
in configurations with open vertical edges (see Figure 2-7) as compared to sealed conditions (see
Figure 2-8).
0 0.1 0.2 0.3 0.4 0.5
x/L
0.6
0
0
0.7
0.9
0.8
0.7
0.6
:::c
0.5 .......
N
0.4
0.3
0.2
0.1
Front
0
i
oo
0.8 0.9
Figure 2-8: CP values measured on models for frontal wind incidences for sealed double-
skin facade with 1.6 m (equivalent) depth. Information for this figure based on Gap inner
pressures in multi-storey double skin facades (Marques da Silva and Gomes 2008, 1555).
49
These results were useful in understanding the pressure dynamics associated with double-skin
facades and identify the opportunity to modify existing codes to address double-skin systems as
opposed to applying single-skin pressure coefficients.
One of the limitations of wind tunnel studies in evaluating double-skin facades discussed in
Section 1.3.2 is the difficulty in making measurements on both the interior and exterior skins. The
challenge of placing pressure taps on both surfaces is complicated by the tubing in the air gap
creating an obstruction to airflow. Marques da Silva and Gomes (2008) did not attempt to collect
pressure data for both skins - just the inner facade. This strategy was valuable in that the inner
skin acts as the primary weather barrier and was readily compared to the conventional single-skin
approach. If an accurate CFO model can be calibrated to produce comparable results to the wind
tunnel experiments outlined in this research, than the advantage of identifying CP data for both
skins, as well as airflow velocity and temperature, would be incredibly useful.
Kawai (2006)
The research presented by Kawai in Pressure in a cavity of building-high double facade
presented a numerical method to approximate the cavity pressure of a multi-story double-skin
facade. Within the introductory remarks, the double-skin facade landscape was described as
being comprised of two types story-high and building-high (Kawai 2006, 7 41 ). Before isolating
the evaluation on the building-high or multi-story configuration, Kawai explained that the exterior
and cavity pressures on a story-high double-skin facade are commonly balanced and very small,
except near the facade's corners. This pattern of behavior is similar to the pressure-equalized
rain-screen approach, but may include increased cavity depths for performance or to house
shading devices. In contrast, the multi-story pressure characteristics were described, therein, as
having different pressures at the top and bottom cavity openings that produce in an imbalance of
pressures between the exterior and the cavity, resulting in amplified coefficients of pressure on
the outer skin (741), as seen in Figure 2-9b.
50
The development of a numerical method to approximate the cavity pressure of a double-skin
facade was derived from a combination of experimental wind tunnel tests on a single-skin and an
application of Bernoulli's theory, wherein the cavity pressure consists of two components: friction
loss and velocity pressure (Kawai 2006, 741-742). The building face was subdivided into a grid
of points for calculation of the friction loss pressure component, velocity pressure component,
total pressure at the point, and finally, the equivalent coefficient of pressure that was normalized
by the velocity pressure of the upstream flow. In evaluating the numerical method of summing
the friction loss and velocity pressure, Kawai found the pressure distribution to be akin to the
experimental results across the majority of the cavity, with discrepancy occurring in the numerical
model's tendency to under estimate the flow separation in the bottom region where flow
separation occurs in the region of openings, generating large negative pressure (744).
Ultimately, numerical methods rooted in experimental data would be incredibly valuable for
estimating Cr distributions, especially during design development prior to wind tunnel testing.
External pressure on the outside skin
Figure 2-9a: Pressure and wind coefficients on the outside skin and the cavity of a
building-high double-skin facade (Kawai 2006, 742).
51
Pressure in the cavity
o.(>
0.8
0.7
Net wind force on the outside skin
Figure 2-9b: Pressure and wind coefficients on the outsides skin and the cavity of a
building-high double-skin facade. Information for this figure based on Pressure in a cavity of
building-high double facade (Kawai 2006, 742).
52
Marques da Silva and Gomes (2005)
The aforementioned Gap Inner Pressures in Multi-Storey Double-Skin Facades by Marques da
Silva and Gomes (2008) represented an evolutionary progression from earlier boundary layer
wind tunnel testing (2005) to determine mean pressure coefficients (Cp) on the inner skin of a
series of model configurations. In Effects of different multi-storey double skin facade
configurations on surface pressures, a base case of a single envelope (unsheltered) was used to
compare a multi-story double-skin facade with the bottom opening of the second (sheltered) skin
beginning 3 m (full scale equivalent) above ground and a cavity depth of 0.8 m (Perino 2007, 31 ).
The measurements were reported as the net difference, liCP = cp,DsF - cp,unsht• between the local
CP value for the respective DSF layout and the correspondent value for an unsheltered envelope
(Marques da Silva and Gomes 2005, 355). Data lines were presented for each configuration in
both the horizontal and vertical direction.
0.9
0.8
0.7
0.6
I
-.. 0.5
N
0.4
0.3
0.2
0.1
0
-3 -2 0
Figure 2-10: Net cp distribution, /iCP = cp,DSF - cp,unsht• for four double-skin facade layouts
(A, B, C and D) at a wind incidence of 0°. Information for this figure based on Effects of
different multi-storey double skin facade configurations on surface pressures (Marques da Silva
and Gomes 2005, 356).
53
The tested specimen was a rectangular geometry with a relative building height, hf a= 2.15, and
relative aspect ratio b/a = 0.615. The only tested depth, 0.8 m (full scale equivalent), was
characterized by a cavity aspect ratio (height:depth) of h/s = 87.50. A second-skin was placed
on two sides of the \=1/40 scaled model starting approximately 0.107 h above the ground level
and rising the full height of the building. Five wind directions (0°, 45°, 90°, 135° and 180°) were
tested for each configuration, as well as four layouts including: A) a cavity open on all sides, B)
continuous corner with open edges, C) continuous corner with sealed edges, and D) fully lateral
closure (Marques da Silva and Gomes 2005, 354).
In comparison to the latter studies (2008), the coefficients of pressure for both the single-skin
and double-skin inner envelope appear to vary slightly; however, this was not the case.
7
As seen
in Figure 2-11, the 2005 single-skin values showed a maximum CP value of 0.9 at the building
midsection, at approximately 2/3 the building height, where as the 2008 single-skin values
showed a maximum CP value of 0.8 in the respective region (see Figure 2-6). The variation
between the 2005 and 2008 tests for the unsheltered skin were simply in the level of reporting
and were consistent data values constant in both research papers.
8
For the double-skin facade
configurations with 0.8 m (full scale equivalent) gap widths, common to both research papers, the
net pressure coefficients on the inner skin presented in 2005 can be compared to the 2008
measured pressure coefficients by subtracting the unsheltered condition's corresponding values
for each local point. For example, in Figure 2-10 the curve A:x/L =0.35/0.67 at z/H = 0.5 has a
;:,cP =-1.8. In the 2008 studies, the same value can be construed as the difference of CP values
in Figures 2-6 and 2-7 (herein) for x/L = 0.35 at z/H = 0.5 where (-1) - (0.8) = -1.8.
The authors concluded that the CP values on the inner skin of double-skin facade configurations
demonstrated considerable divergence from the single facade, or unsheltered, building,
especially for wind incidences between 0° and 45°, and particularly in DSF layouts A and B where
the cavity possessed some open lateral edges (Marques da Silva and Gomes 2005, 358).
54
Single-Skin
I
~ 0.5
N
o+-'-_,__,,_~--'T-~.,__~~~~~-----~-.--'----'-r~-+
I
...... 0.5
N
0
0
0
0.5
x/ L
Double-Skin Inner Envelope
0
0.5
x/ L
-0.4
Figure 2-11: CP distribution on single envelope (top) and a laterally sealed double-skin
inner layer (bottom) with a wind incidence of 0° and cavity depth of 0.8 m. Information for
this figure based on Effects of different multi-storey double skin facade configurations on surface
pressures (Marques da Silva and Gomes 2005) as presented in (Perino 2007, 30).
55
2.3.2 Multi-Story Load Sharing
The double-skin facade with airflow introduced to the cavity spaces creates an interesting
dynamic of load distribution between the inner and outer skins. Generally speaking, when the
exterior skin is sealed, it will receive the majority of wind pressure. When the exterior skin is
permeable, both skins receive wind pressure, with a lower pressure occurring at the outer skin
compared to the sealed condition.
9
The benefits of an air permeable facade may include lower
panel loads as the result of load sharing. Adhering to codes, however, will apply a single-skin
methodology to load derivation and may result in a conservatively designed double-skin facade.
10
Gerhardt and Kruger (1997)
The determination of pressure coefficients across a back-vented facade system were compared
to wind tunnel loads for a case study of a double-skin facade - the Stadttor Dusseldorf office
building (Gerhardt and Kruger 1997, 337-339).
A code proposal for back-vented rain-screens was summarized: much of it based on earlier
research where measurements were collected in a boundary layer wind tunnel with a test section
1. 75 m (5. 7ft) wide x 0.9 m (3.0 ft) tall x 2.0 m (6.6 ft) long with a fetch section of 6.0 m (20 ft)
(Gerhardt and Janser 1994, 38). The earlier research evaluated three boundary layers with
profiles having power law exponents exponent aP = 0.1, 0.2 and 0.3. The authors described the
1 :350 scaled building models used in the wind tunnel experiments and a series of variables
considered in the analysis: relative building height, h/a = 0.5, 1.0, 2.0 and 4.0, relative aspect
ratio b/a = 1.0, 2.0 and 4.0, relative gap width of s/a = 0.0025, 0.005 and 0.01, where a= 100
mm (4 in) was the building width, and relative permeability values of E = 0.5%, 0.75% and 1 %.
A schematic of variable definitions for a permeable facade are presented in Figure 2-11, and the
time-averaged pressure coefficients recommended for code integration are shown in Figure 2-12.
56
-c
p
f =:--t--
..,_____
s
cp,inl
)I
A
~
gap flow
)I
t
)I
A
_!_)
)I
l
through flow
)I
porosity,£
)I
...
-c
p
Figure 2-12: Wind flow on a permeable facade schematic diagram. Information for this figure
based on Doubles Skin Glass Facades: Investigations into the Load Sharing Possibilities
(Gerhardt and Kruger 1997, 336).
57
b/a ~ 2 cp = - 0.40
bla>2 CP = -0.50
0.1 h
h
b/a~2 CP= -0.30
bla>2 CP= -0.35
Figure 2-13: Code proposal for wind loads on a back-vented facade system. Information for
this figure based on Doubles Skin Glass Facades: Investigations into the Load Sharing
Possibilities (Gerhardt and Kruger 1997, 336).
The second part of the research evaluates a built case study of an office tower enclosed with a
convective double-skin glass facade. The motivations for implementing a double-skin facade in
the Stadttor Dusseldorf project included natural venting for the non-air conditioned space, and
better noise protection. Wind tunnel study validation tests performed after the design ascertained
the structural wind loads but underestimated the panel loads due to external suction (Gerhardt
and Kruger 1997, 337).
To ensure proper flow through the cavity space a 1 :250 model reflective of the facade's geometric
and permeability properties was studied through wind tunnel testing. A full lateral closure
between adjacent building elevations was integrated. As seen in Figure 2-13, the wind loads on
the exterior skin were reduced in the permeable condition compared to the sealed exterior
condition.
58
w = 1.5 kN/m
2
w = 0 kN/m
2
w = 0.85 kN/m
2
w = 1.0 kN/m
2
pressure pressure pressure pressure
.. .. .. ..
... ... ... ...
wsuction
= 2.0 kN/m
2
wsuction
= 0 kN/m
2
wsuction
= 1.28 kN/m
2
w
suction
= 1.1 kN/m
2
Outside Office Space Outside Office Space
Closed Exterior Facade Permeable Exterior Facade
Figure 2-14: Wind loads for DSF with sealed exterior (left) and permeable wall (right).
Information for this figure based on Doubles Skin Glass Facades: Investigations into the Load
Sharing Possibilities (Gerhardt and Kruger 1997, 339).
It is evident that the net pressure across the permeable exterior skin was smaller than on an
impermeable enclosure (Gerhardt and Kruger 1997, 336).
The permeable facade experienced a 43% reduction in pressure and 36% reduction in suction
compared to the sealed condition. However, the inner skin experienced load in the permeable
facade, and no wind pressure or suction in the sealed configuration. The work of Gerhardt and
Kruger (1997) was a pivotal shift in testing the aptness of back-ventilated rain-screen (BVR) and
pressure equilibration rain-screen (PER) research and theory to double-skin facade systems.
Though the data is project specific, it established the notion of load sharing in double-skin
facades as an advantage and potential enabler of naturally ventilated high-rise applications.
59
The research presented thus far of Marques da Silva and Gomes (2008) and Gerhardt and
Kruger (1997) have both studied double-skin facades with a multi-story cavity. Both
investigations utilized wind tunnel testing to characterize the wind pressure coefficients
associated with the double-skins. It is significantly more challenging to perform field-testing on
multi-story DSF structures to validate the wind tunnel testing due to the sheer number of data
points that would be required, not to mention costs. other researchers have focused their
investigations on the pressure distribution across double-skin facades of the box-window
configuration.
2.3.3 Box-Window Load Sharing
The box-window configuration is more feasible to mock-up than a multi-story configuration and
can be tested on-site at full-scale rather than in a laboratory wind tunnel. Additionally, compared
to multi-story approaches, the box-window is more-likely to be a mass-produced solution that has
provoked intense industry research to seek honed DSF products.
11
Primary differences between
a box-window and a multi-story double-skin facade include: cavity depth is usually smaller in a
box-window, smaller air openings occur more frequently across a facade with box-windows along
an elevation, and box-windows are more likely to experience load sharing between glass skins.
Ishida (2003)
The wind pressure interaction on a box-window double-skin facade system was rigorously
evaluated and well-documented by the Asahi Glass of Japan (Ishida 2003). The use of room
height double-skin glazing had emerged as a trend in the Japanese commercial building. A lack
of experimental data motivated a series of full-scale tests to evaluate the influence of openings,
depth and airflow paths on the distribution load ratio of pressure on the inner and outer skins of a
box-window DSF. The test cell included 1.8 m (5.9 ft) x 1.8 m (5.9 ft) tempered glass lites, 10
mm (3/8 in) thick, for both the inner and outer skins (see Figure 2-14).
60
E
E
0
0
"'
-
~
,
'
,,.
·~
~
\ I
CD
I \
~ ~
1800 mm
Front Elevation
~
\
'
,
~
~
®
Vertical Section
CD 1.8 m x 1.8 m Glass
® Inner Glass
@ Cavity
® Opening
® Outer Glass ® Pressure box
Figure 2-15: Wind pressure test apparatus: front elevation (left) and vertical section (right).
Information for this figure based on Doubles-Skin Glass Facade - Design Method of Wind
Pressure (Ishida 2003, 253).
The distribution load ratio was measured for two cavity depths and four opening configurations: 1)
no openings, 2) openings on the outside, 3) openings on the inside and 4) openings on both
sides. The types of opening positions are diagrammed in Figure 2-15. Though Ishida tested four
opening configurations, the two most insightful positions for this research are the sealed condition
and the openings to the outside. Studying these two prototypes in particular eases comparison to
the aforementioned research efforts which isolate the sealed and permeable exterior skin
conditions.
61
®
2 @ 3
®
No Openings Openings on Outside Openings on Inside Openings on Both Sides
CD 1.8 m x 1.8 m Glass
@ Inner Glass
@ Cavity
® Opening
@ Outer Glass ® Pressure box
Figure 2-16: Box-window DSF opening position configurations. Information for this figure
based on Doubles-Skin Glass Facade - Design Method of Wind Pressure (Ishida 2003, 253).
In the case of the sealed cavity, the two layers behaved similar to an insulated glass unit. When
the air gap between the two glass lites was small, the deformed outer panel subject to pressure
changed the air volume within the cavity, transferring load to the inner lite. As the depth of the
cavity increased, the ratio of air change due to deformation relative to the overall cavity volume
became less. Therefore, as the depth increased in a sealed cavity, the greater the distribution
load ratio was on the outer skin. For instance, a 90 cm (35 in) sealed cavity was expected to
attract in excess of 80% of the pressure to the outer surface (Ishida 2003, 254). The distribution
load ratio of both the inner and outer skins of a sealed box-window experiment and analysis are
summarized in Figure 2-16.
62
0.9 Outer Skin
0.8
?-- 0.7
0
& 0.6 --()-- Outer Skin (Analysis)
1J - • - Inner Skin (Analysis)
_g 0.5 "-"""-----------'----------------!
Outer Skin (Experimental)
Inner Skin (Experimental)
0.2
----'---...±_- ·
0.1 Inner Skin
o~-----~--~--~---------~--~-----~
0 200 400 600 800 1000
Depth of Cavity [mm]
Figure 2-17: Box-window distribution load ratio for sealed cavity. Information for this figure
based on Doubles-Skin Glass Facade - Design Method of Wind Pressure (Ishida 2003, 254).
0.9
0.8
-;:::. 0.7
0
& 0.6
"C
ro
_g 0.5
c
0
·-s 0.4
.c
·;::
1i5
Ci 0.3
0.2
0.1
0
- _:::I ..:::-
V !!
~
y·
•
,,
,,
..&.:,·
, . ,,
, . ,,
,,,
~ ,,
~;
iti
,, .
"1:
,,, .
II \'
~ ·
IL
I I'
·~ ~ ~
I '~
6
I
0 5
-
I
I
I
LS'
I
Inner Skin
I
I
I
--()--
Outer Skin 600 mm
-
-·-
Inner Skin 600 mm
- -·- -
Outer Skin 300 mm '"""
. - -1:::..- - .
Inner Skin 300 mm
-
t
-- -- -- --
--+--
l
- - -
I
Outer Skin
I
II
I
'.:A ""
I I
-
10 15 20
Opening Area Ratio ( p) [cm
2
I m
3
]
Figure 2-18: Box-window distribution load ratio for cavity with openings to outside.
Information for this figure based on Doubles-Skin Glass Facade - Design Method of Wind
Pressure (Ishida 2003, 254).
63
When the cavity possessed openings to the outside, the distribution load ratio was influenced
more by the opening area ratio than the depth of the cavity. Figure 2-17 shows the experimental
and analytical data for two cavity depths each at five opening area ratios. When the cavity was
sealed, the outer skin had a distribution load ratio of y
0
"' ~ 0.78, and the inner skin a ratio of Yin~
0.22. As the cavity area ratio increased, the outer skin received less of a distribution ratio while
the outer skin took the majority of the load. This suggests that when a double-skin facade has
openings to the outside, and a larger opening area ratio, it may be possible to use thinner glass
on the outer skin (Ishida 2003, 254). The investigation proceeded to evaluate the inside and both
opening configuration types before drawing several conclusions.
A primary difference between the work of Ishida (2003) and the work of Marques da Silva and
Gomes (2008) or Gerhardt and Kruger (1997) is the use of the distribution load ratio, y, to
characterize the pressure associated with each skin. The distribution load ratio may be suitable
for box-window configurations where the local CP distribution is not as variant: but how might the
ratio be applied to the multi-story configuration?
12
2.3.4 Box-Window Pressure Distribution (Field Measurement)
Kawai, Nishimura, Suzuki and Oura (2009)
Recognizing an increase of double-skin facade applications across Europe, Asia and the United
States, the research presented by Kawai et al, Field Measurement of Wind Pressure on a Double
Skin with a Ventilator, investigated four full-scale box-window configurations with ventilators
located in a natural wind environment in an experimental field at Kyoto University. The research
built off earlier investigations using a simulated wind environment, Kawai et all 2005 and 2007,
where wind tunnel studies were used to evaluate various double-skin configurations and
proposed wind loads for the outer and inner skins. One key conclusion from the research was
that the peak wind loads on the full-scale box-window double-skin facade were significantly
affected by the window opening (Kawai, Nishimura, et al. 2009, 232).
64
The full-scale experimental model had a total width of 5.20 m (17 ft), total depth of 2.52 m (8.3 ft)
and a total height of 5.825 m (19 ft) as seen below in Figure 2-18 (Kawai, Nishimura, et al. 2009,
230). The test bed contained four box-windows; two located on the west and two located on the
east elevations. All four box-windows were 1.61 m (5.3 ft) wide, 4.205 m (13.8 ft) tall, had a
cavity depth of 0.20 m (0.66 ft) between the two skins, and had a 0.10 m (0.33 ft) wide by 2.052
m (6.7 ft) tall ventilator installed on the inner skin creating an operable opening capable of
delivering 300 m
3
/hr (177 CFM) of air from the natural environment to the room interior (230).
LO LO
N 0
Cl) N
LO ~
CD 1400 x 2052 DSF Box-Window
@ 1 00 x 2052 Ventilator
@ Airflow Opening
r
5200
1610
3 )::::::::::::::
West I East Elevation
r
2520
, .
I I
I I
I I
I I
I I
I I
I I
I I
• •
I
I
~
I I
I I
I I
,
.
North I South Elevation
Figure 2-19: Full-scale experimental model configuration with four double-skin facade of
box-windows; two windows on each of the west and east elevations respectively.
Information for this figure based on Field Measurement of Wind Pressure on a Double Skin with a
Ventilator, (Kawai, Nishimura, et al. 2009, 230).
65
Over a 15 month period, data measurements recorded velocity levels using a sonic anemometer
at pressure hole locations on each of the outer skin, inner skin (cavity side) and interior (room
side) of the inner skin. The research paper focused on a 39-hour period of data collection to
evaluate the wind speeds, peak velocities, mean velocities and coefficients of pressure specific to
the experimental model and environment. During the selected evaluation period, the full-scale
model was subjected to peak gusts of 29.6 m/s (66.2 mph) and a mean wind speed in excess of 6
mis (13.4 mph) (Kawai, Nishimura, et al. 2009, 230). Multiple operable unit combinations were
modeled over this duration by varying the opposing elevation's window position.
The first configuration summarized in-depth the results section of Field Measurement of Wind
Pressure on a Double Skin with a Ventilator was a 10 minute evaluation period where the full
scale model was exposed to prevailing winds from 277 degrees, a nearly direct west wind
13
(see
Figure 2-19 for a 10-second excerpt from the evaluation period). During the evaluation period, the
wind velocity "fluctuated very violently" and recorded a turbulence intensity of 34% (Kawai,
Nishimura, et al. 2009, 231).
14
35 360
- W ind Direc1ion
30
'iii'
._
25
s
"C
20 QJ
QJ
~ 15
300
c:
- Wind Speed
240
0
·~
~
180
0
"C
"C
c: 10
~
5
120
c:
~
60
0 0
320 321 3 22 323 3 24 325 3 26 3 27 328 329 330
Time [s)
Figure 2-20: Records of wind speed and wind direction at the experimental field.
Information for this figure based on Field Measurement of Wind Pressure on a Double Skin with a
Ventilator, (Kawai, Nishimura, et al. 2009, 231).
66
00. 4
-E 3
Ql
·13
:t: 2
Ql
8
Ql
~ 0 ~=------~--.....,_,c____~c----:------------~4
"'
Ql
ct -1
-2~-------------------------~
320 321 322 323 324 325 326 327 329 329 330
Time [s)
- Outer Skin
- cavity
-Room
Figure 2-21: Records of pressure on outer and inner skin and pressure inside the testing
room. Information for this figure based on Field Measurement of Wind Pressure on a Double
Skin with a Ventilator, (Kawai, Nishimura, et al. 2009, 231).
For the case where the full-scale model's west elevation was exposed to nearly normal prevailing
wind directions and the inner skin's operable windows were in the closed position, the west
elevation's box-window DSF units, with air intake and exhaust to the exterior environment,
experienced positive pressures on the outer skin and within the cavity space (see Figure 2-20).
At the point of maximum gusts at the experimental field, the peak pressure at the west elevation's
box windows' outer skins experienced pressure coefficients larger than 4.0 (at t:::: 324.8 s). The
pressure coefficient within the airflow cavity demonstrated a similar fluctuation pattern as the
outer skin but at approximately 80% of the magnitude. The cavity space experienced a maximum
peak pressure coefficient around 3.0 at several points during the 10-second record period. At
other data collection points on the west facade's outer skin, peak pressure coefficients exceeded
8.0 at times; however, the primary range of mean pressure coefficients for the same data points
were between 0.6 and 1.1 (Kawai, Nishimura, et al. 2009, 231).
The research proceeded to characterize the peak wind force coefficient for the same condition
with a closed ventilator on the inner skin and a windward exposure, and related the mean values
across each skin to that of a single-skin.
67
1.4
0 Single Skin
1.2
e Outer Skin
1.0
() Inner Skin
c
CD
"(3
~ 0.8
8
CD
~
0.6
ii:
-0
0
c::
~
0.4
• • , .
fl'
• • •
•
0.2
·- ---~
•• •
•
250 260 270 280 290 300 310 320 330 340
Wind Direction [de~rees]
Figure 2-22: Peak wind force coefficients on a single-skin, and the outer and inner skins of
a windward double-skin box window. Information for this figure based on Field Measurement
of Wind Pressure on a Double Skin with a Ventilator, (Kawai, Nishimura, et al. 2009, 232).
As seen in Figure 2-21, for the single-skin condition, the mean peak wind force coefficient were
recorded as approximately 0.80. In comparison, for the double-skin condition when the operable
windows were closed , the mean peak wind force coefficients were 0.23 and 0.69, respectively, for
the outer and inner skins.
15
Based on these findings , one of the paper's conclusions was that for
these box-windows with no opening to the interior room and subjected to nearly-normal wind
attack, the mean peak wind loads on the outer skin was 29%, and inner skin was 86%, when
compared to the mean peak wind loads on a single-skin facade (Kawai, Nishimura, et al. 2009,
232).
Consistent with many of the aforementioned research efforts, the full-scale field experiments led
the researchers to corroborate that window openings have an integral effect on peak wind loads.
68
2.4 Discussion and Correlations Amongst DSF Wind Engineering Research
The study of pressure distribution characteristics on multi-layered enclosure systems and
specifically double-skin facades is a topic that has been, and is currently, a multi-faceted research
investigation with full-scale experimental models, scaled wind tunnel studies, simulation and
numerical models all playing an integral role. As wide as the spectrum of possible double-skin
facade configurations is, so is the number of various configurations implemented in testing. From
box-window mock-ups placed in a natural wind environment in Japan, to scaled, building-high,
multi-story studies in a boundary layer wind tunnel at Portuguese laboratory, the description of
relevant research efforts pertaining to double-skin facades is a sampling of this diversity. Going
forward, how could understanding the relationships between the findings of these assorted
studies inform future research endeavors, as well as applications of double-skin facades in the
built environment?
This section begins to identify several key correlations and distinctions between the background
research efforts that have informed the development of this dissertation research.
2.4.1 Modes of Evaluation
For double-skin facades of a box-window configuration, as well as the pressure equalized rain
screen precedent research, full-scale testing was used to understand pressure characteristics. In
the cases of Xie et al. (1992) and Ishida (2003), controlled full-scale tests were implemented.
These analyses have value when the configuration being tested is repeated frequently on a
project, a representation of a box-window product system, or a common configuration. In the
research of Xie et al. (1992), the specimen was of one panel assembly 1.5 m (5 ft) wide by 1.2 (4
ft) tall with a fixed cavity volume of 0.14 m
3
(4.9 ft'). This means the cavity depth was fixed at a
depth around 78 mm (3 in). The variable of interest within this experiment had more to do with
the opening area size than anything else. In the research of Ishida (2003), the 1.8 m (5.9 ft)
square box-window specimen was setup to have a cavity depth that ranged from 0 to 1 m (3.3 ft),
69
able to achieve a cavity volume of 3.24 m
3
(114 fl'). Similar to the research of Xie et al. (1992),
lshida's evaluation varied opening area. A key difference in the work of Kawai et al. (2009) was
that the specimen was placed on-site with field measurements taken over time. The advantages
of full-scale testing in-situ include the transient nature of the wind conditions to which the
specimen is exposed. Additionally, the data collection must also be taken with time at a
frequency that lends it to a greater volume of information. Kawa i's test specimen also
differentiates itself in the fact that it was more than one unit; instead, four 1400 mm (4.6 ft) wide
by 2052 mm (6. 7 ft) tall - two on opposing elevations - and part of a larger 5200 mm (17 ft) wide
by 2520 mm (8.3 ft) deep by 5825 mm (19.1 ft) tall assembly placed in the field. In the review of
these research efforts utilizing field measurement, two categories emerge to classify the
simulation of wind effects: controlled and natural. Though a controlled study can simulate
transient conditions, the field-controlled natural evaluation only has transient environmental
conditions.
For double-skin facades of multi-story configurations, full-scale specimens to evaluate the wind
effects and pressure coefficients are impractical. Instead, scaled wind tunnel evaluations,
numerical analysis and simulation can provide insight. The work of Marques da Silva and Gomes
(2008) and Gerhardt and Kruger (1997) have both studied double-skin facades with a multi-story
cavity utilizing boundary-layer wind tunnel testing to characterize wind pressure coefficients.
Gerhardt and Kruger used a 1 :250 scaled model of the Stadttor Dusseldorf office building facade
while Marques da Silva and Gomes specimen was a \=1/40 scaled model. When afforded the
opportunity to, a larger model scale has benefits of acquiring measurements at more data points
and is an easier scale to coordinate with equipment such as pilot tubes. In the case of Gerhardt
and Kruger, they were able to complement their wind tunnel analysis with field measurements
from the actual project. In both wind tunnel testing and computational fluid dynamics, the
atmospheric boundary layer is defined by the boundary conditions, which can be sensitive
70
(Zammit, Overend and Hargreaves 2010). A clear argument is often made that the merit of such
evaluations -wind tunnel or CFO - lies in the validation from full-scale measurements.
Numerical modeling was exhibited in the work of Xie et al. (1992) on pressure equalized cavities
and Kawai (2006) on multi-story double-skin configurations. Ultimately, numerical models are
tremendously valuable if they reflect accurate conditions because they can avoid costly full-scale
experiments, timely simulations and readily inform design decisions - all of which would be
useful, in particular for multi-story double-skin configurations that are rarely tested at full-scale.
Though Kawai found experimental and numerical data to match in the majority of the surface, the
numerical model had a tendency to under estimate the flow separation at the bottom in regions of
openings. Potentially, numerical models can serve as a platform for typical design conditions with
full-scale modeling or computational fluid dynamics playing a supplementary role in the evaluating
local effects. This is a key balance, especially in practice, because, with the exception of
numerical, all other evaluation modes are time-consuming.
The range of evaluations exemplified that attaining detailed information about the pressure
performance of box-window systems is far easier than that of multi-story double-skin facades.
This is evident in the ability of Kawai et al. (2009) to collect data over a 15 month period for a full
scale field-measured specimen that included four box-windows, a room, operable windows and
the ability to simulate natural ventilation, including cross ventilation. Seeing that box-windows are
just one of numerous configuration possibilities, attention must be paid to those that are not as
readily modeled or evaluated -the multi-story double-skin facades.
2.4.2 The Unsheltered Tower or Single-Skin Facade as a Baseline
The use of an unsheltered facade is commonly used as a calibration tool for boundary layer wind
tunnels and computational fluid dynamics. Calibrating the pressure coefficients on an
unsheltered element serves as a baseline of comparison for distribution of pressure coefficients
for layered systems. One example of the unsheltered structure in analysis is the use of a study
71
cube in both CFO and field measurements by Zammit, Overend and Hargreaves (2010). Though
not a cube, the work of Marques da Silva and Gomes (2008) utilized an unsheltered tower in their
wind tunnel analysis as a baseline to multi-story double-skin facade configurations with adjacent
sides, ranging from no lateral closure to fully sealed. Both of these evaluations also reviewed the
pressure distributions when subject to a corner wind incidence 45° from normal. Historically,
evaluation of an unsheltered cube has served as a method of comparing the performance of wind
tunnels from varying institutes around the world.
2.4.3 Effects of Sealing or Communicating Between Adjacent Building Faces
One stern advisory that leapt out of several of the discussion and conclusion sections of the
research review was to avoid communication of cavity air between multiple building faces. The
work of Gerhardt and Janser (1994) used a configuration without cavity airflow configurations with
other building faces in both their full-scale and wind tunnel models. The advisory was also clearly
stated by Xie, Schuyler and Resar (1992). They discouraged air movement between floors,
around corners and even in flat cavities exceeding ten feet. This would exclude the application of
multi-story applications. For the purposes of this research, the multi-story facade will be
evaluated, but only in configurations that possess full vertical edge closure between adjacent
building walls will be considered. Some researchers, such as of Marques da Silva and Gomes
(2008), were inclusive of configurations that were open to airflow communication with double-skin
facade cavities on adjacent faces, as well as configurations with full lateral closure. All together,
these consistent findings and tones of caution are reflected in the exclusion of configurations
which put air into communication with other faces in the select guidelines present in Eurocode,
Section 7.2.10 (CEN!TC250 2005). This point is also echoed in ASCE/SEI Minimum Design
Loads for Buildings and Other Structures where there is supplemental commentary in Section
C30.1.5 Air Permeable Cladding that advises that the air space between cladding and the next
adjacent building envelope surface be compartmentalized to avoid communication between
different pressure zones of a building's surfaces (2010, 570). Now this is most likely in reference
72
to a traditional rain-screen cladding assembly but the air-permeable cladding commentary is one
of the few parts of ASCE/SEI 7-10 that even remotely relates to the considerations for double
skin facades. The majority of built case studies that are presented in Chapters 4 and 5 and
Appendices A and C follow this guideline to avoid communication between adjacent building
faces, often by isolating each building elevation's cavity from the adjacent, or limiting the scope of
the double-skin facade to a single elevation all together.
2.4.4 Influence of Opening Area Ratio
One variable that was credited with having a critical impact on pressure distributions and peak
loads was openings and the permeability of skins. Opening area ratio (p) describes the area of
opening compared to the area of panel or enclosure. Ishida suggested that with a larger opening
are ratio it is possible to use thinner glass on the outer skin (2003, 254). Additionally, when
exposed to outside air, the opening area ratio has a greater influence than cavity depth on the
load distribution. Xie, Schuler and Resar found the load reduction factor to be a function of the
opening area ratio (1992, 2453). Kawai, Nishimura, et al. identified that window openings have
an integral effect on peak wind loads (2009, 232). All of these references to the influences of
openings and opening area ratio are consistent with the Eurocode, Section 7.2.10.3, that
identifies the wind force on each skin being dependent on the permeability of the skins and the
openings (CEN/TC250 2005).
2.4.5 Load Reduction
The possibility for load reduction is commonly referenced when load sharing considerations are
applied to double-skin facade configurations. The reality is specific to a project's design and local
climate. Contrastingly, sometimes a double-skin facade is associated with increased suction
loads in the cavity. In the review of research there are several evaluations the address the load
reduction factors associated with a double-skin facade as compared to a sealed or single-skin
facade. The first research noted is Xie, Schuyler and Resar's (1992) evaluations of spandrel
panels in suitable pressure equalized curtain wall design. The findings include a rapidly
73
exponential decay of the load reduction factor as a function of opening area ratio. As seen in
Figure 2-2, for p,/P
0
= 7, the reduction factor drops to an apparent bottom limit of approximately
0.12 when the opening area ratio reaches 2.0%.
As for double-skin facades specific mentions of load reduction, Gerhardt and Kruger (1997) found
reductions of 43% positive pressure and 36% suction compared to a sealed condition for their
project case study.
16
In the full-scale tests of box-window configurations, Ishida (2003) used
distribution load ratio to characterize the relationship between loads received by each of the inner
and outer skin. Figure 2-17 exemplifies that the distribution load ratio is more sensitive to opening
area ratio than cavity depth as seen in Figure 2-16. When sealed, the outer skin, regardless of
depth, receives the majority of pressure on the order of 70-80%. As the opening area ratio
approaches 3% the outer skin load distribution starts to fall to 0.10. The work of Xie, Schuyler
and Resar (1992) and Ishida (2003) both show that an even distribution of load between surfaces
occurring around 0.5% opening area ratio, however, beyond that, the inner skin receives the
majority of the full pressure. Load reduction is only an opportunity appropriate across the outer
skin in a double-skin facade ventilated to exterior air.
2.4.6 Cavity Pressure
Cavity pressure, consisting of friction loss and velocity pressure components (Kawai 2006, 7 41-
742), is magnified in multi-story double-skin facades as compared to box-window assemblies.
The cavity pressure is essential for determining inner skin loads and net pressure across the
outer skin. In a sealed cavity the pressure is far less, or nonexistent, as compared to a
configuration that contains an open air inlet
2.4.7 Peak Pressures Compared to Mean Coefficients of Pressure
Multiple experiments found the peak pressure coefficients to be quite high; in excess of 8.0 at
times in the full-scale natural field measurements by Kawai, Nishimura, et al. (2009, 231) and up
to 11 in edge regions in the 1 :250 scaled wind tunnel study of Stattor DOsseldorf by Gerhardt and
74
Kruger (1997, 338). Though high peak pressures were measured in the work of Kawai,
Nishimura, et al., the mean pressure coefficients ranged from 0.6 to 1.1; much closer to values of
a single-skin facade. The work of Xie, Schuyler and Resar (1992, 2453) was presented and
organized with the dimensionless relationship of peak external pressure to the mean pressure,
p,/P
0
, as a key analytical component. This is a useful practice to convey the range in magnitude
of peak coefficients experienced in different double-skin facades.
2.4.8 Net Pressure Coefficient
At any given instant the total net pressure across a building envelope assembly is the sum of the
partial pressures across the individual layers (ASCE/SEI 2010, 569). The net pressure coefficient
across the outer skin is used to characterize the combined effects of impinging wind loading and
cavity suction occurring simultaneously. If the net pressure coefficient is less, it is possible to
reduce the outer skin glass thickness and structure. As Gerhardt and Kruger noted in from their
scaled wind tunnel study and field measurements of Stattor DOsseldorf, the wind suction forces
on an impermeable facade are greater than the net wind loading acting on the outer permeable
sheeting (1997, 336). The net peak suction coefficients for the same project ranged from 2.56 to
3.20 (Gerhardt and Kruger 1997, 338). The CP·"" values across the outer skin must be evaluated
unique to each design, especially in the case of multi-story DSFs. Net pressure coefficients may
be heightened in corners, especially when cavity airflow is exposed to areas of vortex shedding.
On the outer skin the greatest impinging wind loads are most often near the top. When the cavity
is exposed to the roof level, as in a forward exhaust outdoor air curtain, large suction loads may
occur near the top of the cavity. When the external positive pressure and the internal cavity
negative pressure are combined, the net pressure coefficients will have a greater magnitude.
75
2.5 Implementing Lessons Learned from Prior Research
An idealistic approach to determining wind pressure coefficients for building envelopes is to use
full-scale experiments in advance of wide-spread construction of a design. In reality, this is rarely
an option, primarily due to costs and a lack of refined design detail. When a design could be
refined in design detail, full-scale testing is primarily feasible for box-window applications and not
practical for the evaluation of multi-story double-skin facades. Even when full-scale testing is
conducted, the results are specific designs under specific local conditions, and the results are not
generally applicable. As Zammit, Overend and Hargreaves (2010) have noted, the use of
simulation and wind tunnel testing are applicable for addressing double-skin facade wind
pressure effects. In order to provide pressure coefficients earlier in the design process,
simulation is more conducive to iterative design evaluations.
If a reasonably accurate CFO model can be calibrated to produce comparable results to the wind
tunnel experiments of multi-story double-skin facades outlined in this chapter, than the advantage
of identifying CP data for skins, as well as airflow velocity, would be incredibly useful in double
skin facade design development. Whether it is simulation, wind-tunnel testing, or some
combination thereof, the treatment of wind loads of multi-story double-skin facade systems is
highly variant and requires development of an evaluation approach instead of a prescriptive code
approach.
Considering the overwhelming recommendations in the research reviewed, as well as the
recommendations of Eurocode, Section 7.2.10 (CEN!TC250 2005), configurations with
communication between skins will be avoided. These evaluations carried out in this research will
focus solely on multi-story double-skin facades with full lateral closure.
76
Chapter 2 Endnotes
The five wind loading codes considered in this evaluation (Zammit, Overend and Hargreaves
2010, 656) include 1) BS 6399-21997, Loading for buildings-Part 2 Code of practice for
wind loads, (BSI 1997); 2) EN 1991-1-42005, Eurocode 1 Actions on structures - Part 1-4
General actions - wind actions (CEN!TC250 2005) in conjunction with NA to BS EN 1991-1-4,
Draft 4.1 Oa, 2005 U. K. National Annex to Eurocode 1: Actions on structures - Part 1-4: General
actions - wind actions (BSI 2005); 3) ESDU Data Item 82026, Revision E, Strong winds in the
atmospheric boundary layer, Part 1: hourly-mean wind speeds (ESDU International 2002) with
ESDU Data Item 83045, Revision C, Strong winds in the atmospheric boundary layer, Part 2:
discrete gust speeds (ESDU International 2002); 4) ASCE 7-05. Minimum Design Loads for
Buildings and other Structures (ASCE/SEI); and 5) AS/NZS 1170.22002. Structural design
actions, Part 2 Wind actions (AS/NZS 2002).
2
The usage of single-facade building herein is synonymous with unsheltered tower and single
skin facade.
3
The authors, Gerhardt and Janser, do not use the term double-skin facade, but acknowledge a
recent (at the time, 1994) trend of the convective facade with a second glazed envelope in their
research paper.
4
The two "key factors" are in the eyes of this author, and though the 1994 research of Gerhardt
and Janser was not primarily specific to double-skin facades, later research (Gerhardt and
Kruger 1997) reveals a natural progression from pressure-equalized rain-screens as the
permeable, or porous, facade to serve as a basis for double-skin facade coefficient of pressure
evaluation.
5
A contraction in a wind tunnel is established to increase the airflow velocity through a low-
speed region - at the inlet where honeycomb screens and airflow straighteners exist - into the
smaller test section while minimizing pressure losses. The contraction ratio is the area of the
air inlet compared to the area of the test section in a wind tunnel. Contraction ratios in small
wind tunnels are usually between 6 to 9 (Mehta and Bradshaw 1979, 448).
6
The wind-tunnel contraction ratio of 4.5:1 was clarified by Fernando Marques da Silva via email
correspondence on August 16, 2010. The test section dimensions of the wind tunnel were 3 m
(9.8 ft) wide by 2 m (6.6 ft) tall with airflow inlet dimensions of 5.2 m (17 ft) x 5.2 m (17 ft).
7
The difference between the data presented in Effects of different multi-storey double skin
facade configurations on surface pressures (2005) and Gap inner pressures in multi-storey
double skin facades (2008) was clarified by Fernando Marques da Silva via email
correspondence on October 13, 2011. There is no difference in the data common to the two
papers. The only differences are the extent presented (in 2008, three cavity depths were
studied compared to just one, 0.8 m, in 2005) and the method of presenting the results. The
results for the double-skin configurations in Effects of different multi-storey double skin facade
configurations on surface pressures (2005) were communicated as the net ;:,cP = cp,osF -
Cp,unsht, where as the same configuration results in Gap inner pressures in multi-storey double
skin facades (2008) were presented as the measured values themselves, cp,osF·
8
Ibid.
77
9
This statement refers to a permeable facade with a consistent frequency of airflow openings
along the height of the building. This may be similar to a pressure equalized rain-screen, box
window double-skin facade, or louvered multi-story double-skin facade. This statement does
not globally apply to all double-skin facade configurations, especially multi-story double-skin
facades with a single inlet at the bottom and single exhaust at the top. A lower pressure on the
exterior skin does not globally apply to all double-skin facade configurations either, because in
double-skin facades that have communication with other building faces, such as the roof, the
net pressure differential across the exterior skin may heightened the pressure coefficients used
in facade design.
10
This statement refers to a permeable facade with a consistent frequency of openings along the
height of the building, such as a pressure equalized rain-screen, box-window double-skin
facade, or louvered multi-story double-skin facade.
11
Global facade contractor, Permasteelisa Group and its subsidiaries, Permasteelisa, Josef
Gartner GmbH and Scheldebouw, have developed an approach towards facades that are
integrated into the building's environmental system that they have trademarked as Blue
Technology since 2001. These systems utilize double skin facades including Active Wall®,
Interactive Wall®, and the MFREE-S Closed Cavity Facade. Though not standardized
products, these systems, design strategies and philosophies provide a framework for
technological advancement while possessing the flexibility to apply to numerous geometric
configurations.
12
The distribution load ratio for a multi-story double-skin facade will vary with height. Describing
distribution load ratios at equal heights on the outer and inner skin will provide explicit point
values.
13
A wind incidence, or wind direction, nearly normal to the west elevation of the full-scale model
was the most comparable configuration in this study to the normal wind indices, often studied in
wind tunnels, which are often described as wind attack angles of 0° (Kawai, Nishimura, et al.
2009).
14
The turbulence intensity value of 34% is not stated in the Fifth European & African Conference
on Wind Engineering, Volume 51 of Proceedings e-report version of Field Measurement of
Wind Pressure on a Double Skin with a Ventilator (Kawai et al., 2009), but is stated in what
appears to be the expanded paper available at the International Association for Wind
Engineering's 5EACVVE Proceedings website. See the following web address:
http/twww.iawe.org/Proceedings/5EACWE/168.pdf>
15
The mean peak wind force coefficients of 0.23 for the outer skin and 0.69 for the inner skin
were not stated in the Fifth European & African Conference on Wind Engineering, Volume 51 of
Proceedings e-report version of Field Measurement of Wind Pressure on a Double Skin with a
Ventilator (Kawai et al., 2009), but were stated in what appears to be the expanded paper
available at the International Association for Wind Engineering's 5EACVVE Proceedings
website. See the web address in the previous endnote.
78
16
Presumably this configuration adheres to the code proposal requirement outlined within the
paper for porosities E ~ 0.75.
79
3. Overview of Double-Skin Facade Applications
Facade design is rooted in the basic human need for protection from external natural forces such
as wind, rain, and extreme temperatures. The utilization of glass within the facade is a strategy
which enhances human comfort by maintaining a visual connection and acting as an interface
between architecture and nature. In the pursuit of transforming buildings from energy consumers
to energy neutral, the glazed envelope has come under considerable scrutiny
1
while
environmental pressures have yielded exciting new technologies, solutions and ideas for high
performance facades.
Over the last three decades, no other building material has experienced as great a transformation
or experienced as much innovation as glass. It has become a high-tech material that can be
combined with countless structural possibilities to achieve traditional building enclosures and to
create sleek, bold, filigree, sculptural constructions. Modern architecture has embraced the
glazed building for its airy, transparent and light-filled designs which provide greater access to
daylight than typical office buildings. Such spaces possess a sense of openness, connection to
the environment and an impression of future. Depending on the internal function, complete
transparency can communicate a company or institution's openness towards society. The
aesthetics of the transparent facade are often desired by owners and architects, but future
solutions will have to meet higher performance metrics. The advances in coming decades will
transform the transparent facade even further.
In the 1990's, material advances such as low-emissivity and spectrally selective coatings became
available, allowing the transparent enclosure to achieve considerably high levels of thermal
insulation and solar protection. A common criticism of these coatings is the need to incorporate
additional shading devices. The glass envelope requires layering of shading systems to cope
with seasonal changes. Shading solutions may include internal blinds or external louvered
80
devices, but the latter become problematic when exposed to high wind pressures seen on high
rise buildings. There is an opportunity to develop solutions which collapse separate performance
layers into one, enhancing the aesthetic of the glass facade and truly becoming high-tech.
A possible resolution for facade solutions with protected solar controls is the double-skin facade.
These solutions are characterized by the addition of a secondary glazing either behind or in front
of the insulated building facade. The resultant cavity creates a shelter for shading and other
systems. This protection from the wind is especially valuable in high-rise applications. The cavity
air connects either with the air inside the room, known as internally ventilated cavity systems, or
with the external air, known as external ventilated cavity systems. This strategy can be
implemented to provide natural ventilation in regions of high wind speed, shield from noise or air
pollution, or give an old structure a new modern skin. As American coastal cities become
denser,
2
the issues of air pollution, acoustic comfort and historic preservation will continue to
present the double-skin facade as a viable solution to architects and owners.
The double-skin facades have emerged gradually over the last century, but are a relatively new
technology in its modern applications and systems integration. They are most commonly found in
Europe where the increased initial costs offset quicker by energy savings due to the high cost of
energy. Common reasons for the implementation of a double-skin facade solution include
improved indoor comfort, acoustical protection, a reduced energy consumption profile, and the
aesthetic of increased transparency. There are many types of double-skin facades, and
numerous opportunities for integration of technologies and active systems. For those reasons,
double-skin facades constitute a research field for everyone involved in the building process. The
complex range of performances facades are being pushed to achieve - and their sensitive
relation with comfort, place and the building perception - often make double-skin facade
implementation a unique experimental solution. The site-specific facade solutions are exciting
and require collaboration between various experts in different fields. In an era of heightened
81
dialogue regarding sustainability, the potential for reducing energy consumption, optimizing
interior conditions and potential for new aesthetics, makes the double-skin facade one of the most
interesting, and debatable, recent developments in the facade industry.
3.1 A Clear Definition of Double-Skin Facades
The design approach for a double-skin facade can vary substantially depending on climate,
existing conditions, desired performance, aesthetics, level of control and more. The approaches
in a rehabilitation project or preservation of an existing building skin will undoubtedly possess
different priorities than a new construction. Existing concepts include the double-skin facade with
controllable ventilation (through controls or manually), and other solutions where ventilation is not
controllable. Solutions can include indoor and outdoor skins that may be operable, thus not
airtight. The cavity may be occupied by shading devices, fan controls, or renewable energy
sources. There are many combinations of systems which make the double-skin facade typology
layered and complex. The primary difference between an entirely sealed buffer multi-glazing
facade and a double-skin facade is the intentionally controlled ventilation of the cavity in a double
skin facade.
A definition of double-skin facades for the purposes of this research is outlined in Section 3.1.3
following a review of existing definitions and descriptions to date from a survey of other DSF
researchers.
3.1.1 Review of Common Terminology
The term Double-Skin Facade (DSF) is often referred to in other terms that have been adopted by
different regions, institutes, or focus on a specific typology. There tends to be a lack of
consistency in the idiom of double-skin facades making it challenging to understand the scope of
systems included within this typology. Some other terms that may be used to communicate the
ideas of the double-skin facade include, but not limited to the following:
82
•
Cavity facade
•
Double facade
•
Double envelope
•
Double-leaf facade
•
Dual-wall
•
Dual-layered glass facade
•
Environmental facade
•
Environmental second skin
•
Intelligent glass facade
•
Multiple-skin facades
•
Twin-skin facade
•
Ventilated facade
•
Ventilated double-skin
other common terms that appear in literature characterizing the airflow strategy of double-skin
facades include, but are not limited to the following:
• Active facade (mechanically ventilated)
• Airflow window
• Exhaust window/facade
• Hybrid facade (mechanically and naturally ventilated)
• Passive facade (naturally ventilated)
• Supply air window
• Ventilated facade
83
3.1.2 Review of Researchers' Definitions
It should be noted that in Double-Skin Facade: Integrated Planning (Oesterle 2001, 12), a
landmark book on DSFs, it is stated that the term 'double-skin facade' is yet to be clearly defined.
This makes it difficult for individuals to classify multiple-skin facade types due to a lack of
definition of the typological limits. Over the last two decades, many double-skin facades were
constructed feverishly across Europe and seen increased implementation in the United States,
China, Japan, and other regions. Numerous researchers and institutes have proposed their
interpretation of a double-skin facade, but still to this day, there is a lack of consistent
understanding.
3
The established researchers and research institutes tend to be European based.
Here is chronological evolution of double-skin facade definitions proposed by leading researchers
of the subject matter.
Arons (2000, 14) in Properties and Application of Double-Skin Building Facades states:
DSFs are characterized by having at least two membranes between the interior, occupied space
and the exterior environment.
Oesterle et al. (2001, 12) in Double-Skin Facade: Integrated Planning:
Double-skin facades are based on a multilayer principle. They consist of an external facade, an
intermediate space and an inner facade. The outer facade layer provides protection against the
weather and improved acoustic insulation against external noise. It also contains openings that
allow the ventilation of the intermediate space and the internal rooms. The flow of air through the
intermediate space is activated by solar-induced thermal buoyancy and by the effects of the wind.
To achieve greater adaptability in reacting to environmental conditions, it may be possible to
close the openings in the outer facade layer
Uuttu (2001, 12) describes the Double-Skin Facade as:
84
... a pair of glass skins separated by an air corridor (also called cavity or intermediate space)
ranging in width from 20 cm to several meters. The glass skins may stretch over an entire
structure or a portion of it. The main layer of glass, usually insulating, serves as part of a
conventional structural wall or a curtain wall, while the additional layer, usually single glazing, is
placed either in front of or behind the main glazing. The layers make the air space between them
work to the building's advantage primarily as insulation against temperature extremes and sound
Selkowitz (2001, 55) in Integrating Advanced Facades into High-Performance Buildings
comments on the double envelope:
[The double envelope designs] introduces at least two primary glazed layers in the facade, often
in conjunction with sun control systems (e.g., shades, blinds, louvers}, light redirection systems
and ventilation systems. The characteristic "thickness" of such multilayer, multifunctional systems
is on the order of 1 meter
Compagno, (2002, 118) describes the Double-Skin Facade as:
" ... an arrangement with a glass skin in front of the actual building facade. Solar control devices
are placed in the cavity between these two skins, which protects them from the influences of the
weather and air pollution a factor of particular importance in high rise buildings or ones situated in
the vicinity of busy roads."
Lee et al. (2002, 18) in High-Performance Commercial Building Facades (LBNL) states
A second layer of glass placed in front of a conventional facade reduces sound levels at
particularly loud locations, such as airports or high-traffic urban areas. Operable windows behind
this all-glass layer compromise this acoustic benefit, particularly if openings in the exterior layer
are sufficiently large to enable sufficient natural ventilation.
85
Saelens (2002, 5) in Energy Performance Assessment of Multiple-Skin Facades defines the
multiple-skin facade:
A multiple-skin facade is an envelope construction, which consists of two transparent surfaces
separated by a cavity, which is used as an air channel. This definition includes three main
elements: (1) the envelope construction, (2) the transparency of the bounding surfaces and (3)
the cavity airflow"
Wigginton & Harris (2002, 41) in Intelligent Skins
The double skin is a system involving the addition of a second glazed envelope which can create
opportunities for maximizing daylight and improving energy performance.
Harrison and Meyer-Boake (2003, 2) in the Tectonics of the Environmental Skin:
[DSFs are] essentially a pair of glass "skins" separated by an air corridor The main layer of glass
is usually insulating. The air space between the layers of glass acts as insulation against
temperature extremes, winds, and sound. Sun-shading devices are often located between the
two skins. All elements can be arranged differently into numbers of permutations and
combinations of both solid and diaphanous membranes.
Poirazis (2004, 175) in the Double Skin Facades for Office Buildings - Literature Review:
The Double Skin Facade is a system consisting of two glass skins placed in such a way that
airflows in the intermediate cavity The ventilation of the cavity can be natural, fan supported or
mechanical. Apart from the type of the ventilation inside the cavity, the origin and destination of
the air can differ depending mostly on climatic conditions.
86
Lancour et al. (2004, 5) in Ventilated Double Facades: Classification & illustration of facade
concepts (BBRI) states
A ventilated double facade can be defined as a tradition single facade doubled inside or outside
by a second, essentially glazed facade. Each of these two facades is commonly called a skin A
ventilated cavity - having a width which can range from several centimeters at the narrowest to
several meters for the widest accessible cavities - is located between these two skins.
Knaack et al. (2007, 29) in Facades: Principles of Construction, states:
A double facade is obtained by adding an extra layer of glazing outside the facade to provide the
building with ventilation or additional sound-proofing. This system may be realized in various
ways, depending on the functions desired and the requirements made of the facade.
BESTFACADE (2007, 19) in WP5 Best Practice Guidelines describes the ventilated double
facade:
A ventilated double facade can be defined as a traditional single facade doubled inside or outside
by a second, essentially glazed facade. Each of these two facades is commonly called a skin
(hence the widely-used name "ventilated double-skin facade'). A ventilated cavity - having a
depth which can range from about 10 centimeters at the narrowest to 2 meters for the deepest
accessible cavities - is located between these two skins.
Hausladen et al. (2008, 88) in ClimateSkin: Building-skin Concepts that Can Do More with Less
Energy:
With double-skinned facades the primary facade has a second glass plane in front. [. . .] The
functional elements are placed one behind the other, with the result that they may adversely
87
influence on another An unrestricted relationship with the outside world in respect of ventilation
and outside view is not possible.
Some of these definitions vary from general description to specific constraints. One variable
frequently alluded to is the cavity depth, often by defining a minimum and maximum value. This
may be because ventilation mode and ventilation type are not readily visible; air sources and
exhausts are often concealed out of sight. In the survey of definitions, the minimum cavity depth
presented is 'a couple centimeters', 10 cm (4 in), and 20 cm (8 in) in various reports. The
maximum ranges from 1 m (3 ft) to 'several' meters. A strong definition of double-skin facades
must incorporate a minimum and maximum value for the cavity depth to distinguish between a
deep insulated glazed unit (on the low end) and atria spaces (on the high end). For the purposes
of this research, the range selected is in-line with the 10 cm (4 in) to 2 m (6 ft) as stated in
BESTFACADE (2007, 19).
A curious observation from the survey of definitions is how authors address (or don't address)
shading devices, or solar control, as an integral element in double-skin facades. A few notable
inclusions are Harrison and Meyer-Boake (2003, 2) and Selkowitz (2001, 55), who mention that
sun-control devices "often" occupy the cavity space between the two glazed skins. Compagno
(2002, 118) includes solar control devices explicitly in a description of double-skin facade. A
strong argument may be made for including solar control strategies as a requirement, or at least a
classification tier, in double-skin applications since it is so commonly recognized as one of the
two clear advantages as evidenced in Green Building: Guidebook for Sustainable Architecture:
"For new buildings, double-skin facades really only offer two advantages: decrease of wind
influence on solar protection devices and window ventilation; lowering the level of sound
penetration" (Bauer, Measle and Schwarz 2009, 88).
88
3.1.3 Definition for the Purposes of This Research
A double-skin facade is a strategy for improving building envelope performance through the
introduction of a second transparent glazed layer and solar control strategies, resulting in an
airflow cavity ranging from 10 cm (4 in) to 2 m (6.6 ft) in depth.
3.2 Typology and Classification Hierarchy
There are different ways to classify double-skin facades. The criterion with the greatest
discrepancy amongst researchers is facade partitioning. The three common criteria common to
many researchers include:
• Cavity Partitioning
• Ventilation Type
• Ventilation Mode
To remain consistent with the above definition of double-skin facades, a second transparent
glazed layer .. .resulting in an airflow cavity, the typology is inclusive of unsealed systems such as
louvered outer skins and acoustical barriers. The proposed classification system is a fusion of
BESTFACADE (2007, 19-24) and ClimateSkin (Hausladen, Sa Idanha and Lied I 2008, 102-114)
and introduces a fourth tier in the hierarchy of solar control strategies. Within the solar control
level, there are a series of sublevels for three variable classifications: solar control type,
operability and controls. These variables alone do not fully detail a solar control strategy. other
factors include the position of the shading elements (exterior, intermediate cavity or interior),
position within the cavity (front 1/3 near exterior skin, middle or back 1/3 near the interior skin),
shading depth and material type. These variables are not part of the classification scheme solely
in an effort to maintain simplicity since some of them present many possibilities. Most double
skin facade solutions can be described within this typology, but of course, there are exceptions
that exist outside this typological framework and the aforementioned definition.
89
co
0
1. Cavity Partitioning
Alternating Story-Height
Story-Height
Baffle Box-Window Shaft-Box Juxtaposed Multi-Story
Facade Corridor
Modules
I I I I
I I
2. Ventilation Type Natural Hybrid Mechanical
I I
Outdoor Indoor
Air Air Buffer
3. Ventilation Mode Air Air
Supply Exhaust Zone
Curtain Curtain
4. Solar Control Strategy
Solar Control Type
Overhangs
Fins Louvers Blinds Screens
(exterior)
I I
Adjustable
Retractable
Operability Fixed Retractable and
(tillable)
Adjustable
I I
I I
Controls Manual Automated
Automated w/
Override
Figure 3-1: Double-skin facades classification scheme.
Multi-Story
Multi-Story
Controllable
Shingled
{Louvers)
I I
3.2.1 Ventilation Type
The ventilation type refers to the driving airflow forces within the cavity between two glazed skins.
All double-skin facade concepts possess a single ventilation type. The three possible ventilation
types are:
• Natural
• Mechanical
• Hybrid (mix of natural and mechanical)
3.2.2 Ventilation Mode
The different ventilation modes possible are presented below. It is possible for several ventilation
modes to co-exist within a single double-skin facade. This is most common with naturally
ventilated facades where operable windows exist. The ventilation modes are:
1 r
~ / /
IT IT lY
EXT
I
INT EXT
I
INT EXT
I
INT EXT
I
INT EXT
I
INT
I
Jt
Outdoor Outdoor Air Air
Buffer
Air Curtain Air Curtain Supply Exhaust
Figure 3-2: Five modes of double-skin facade ventilation.
91
• Outdoor Air Curtain
Air enters the cavity from outside and immediately returns towards the outside. An air
curtain forms along the outer skin.
• Indoor Air Curtain
Air enters the cavity from inside the room and returns to the room directly, or
mechanically. An air curtain forms along the interior skin.
• Air Supply
Air enters the cavity from the outside and penetrates into the building.
This can be the building's supply air.
• Air Exhaust
Air enters the cavity from inside the room and exits to the outside.
• Buffer Zone
Each skin is airtight. No ventilation of the cavity is possible.
It is possible for the first four ventilation modes in Figure 3-2 to be reversed to create a downward
airflow within the cavity, however, this is uncommon.
In either air curtain ventilation mode, an insulated glass unit is most commonly placed on the
sealed, non-source skin. In other words, for an outdoor air curtain the insulated glass unit is
closest to the interior while closest to the exterior for an indoor air curtain. The second skin is
most commonly monolithic and laminated glazing. The insulated glass unit is generally placed on
the skin serving as the primary air and weather barrier.
92
3.2.3 Cavity Partitioning
The primary classification characteristic of double-skin facades is the cavity partitioning. There
are nine possible partitioning configurations presented in detail in this section. The primary
attributes unique to each system and information about materials, assembly, erection and
durability are presented in this section. The configurations described herein, include:
•
Baffle
•
Box-Window
•
Alternating Facade
•
Shaft-Box
•
Story-Height: Corridor
•
Story-Height: Juxtaposed Modules
•
Multi-Story
•
Multi-Story: Shingled
•
Multi-Story: Controllable
The first three configurations (baffle, box-window and alternating facade) are all variations that
are conceivable unitized and occur within the dimensions of a conventional facade module. The
next three configurations (shaft-box, story-height corridor, and story-height juxtaposed modules)
have cavities that extrude in one direction beyond the limits of a single module: the shaft-box
extrudes vertically across multiple stories while the story-height configurations span multiple
modules horizontally but occur between floor slabs. The last set of configurations is multi-story
partitioning schemes that propagate in two directions, vertically and horizontally, to create an
airflow cavity that spans beyond the unitized module of the primary enclosure system. Any given
project may possess multiple configuration types. The selection of a cavity partitioning type is a
balancing of programmatic function of the interior space, performance requirements of the
composite enclosure, and desired operations and maintenance plans over the expected life-span
of the structure.
93
Baffle Box-Window
Shaft-Box Story-Height: Corridor
Multi-Story Multi-Story: Shingled
Figure 3-3: Conceptual diagrams of cavity partitioning.
'
'
'
'
'
'
kb:- --
'
'
'
'
'
'
~--
- t
I\
' ' I ..
I I ~ i
i i : I
~--'.:.-- - : -- J
Alternating Facade
- =-
Story-Height: Juxtaposed Modules
Multi-Story: Controllable
94
Baffle
A baffle panel, or system, is an additional glazed layer designed to deflect or regulate the
passage of fluid, light, or sound. The panel typically hovers a short distance in front of an
insulated glazed interior layer. It is a method of minimizing the disadvantageous of single-layer
facades with respect to acoustical insulation and ventilation. Baffle panels may be used to
provide exterior solar screening or protect an operable window from high wind conditions. Other
advantages of this system include its reliable protection against weather and a security layer
during night flushing of the building. Disadvantages of this system may include restricted views,
reduced solar gain, decreased daylight penetration and access for cleaning.
r---
1 T---r---
1 I -1---
I I I
I I I
I I I
I I I
I I I
I I I
I I I
I
---·---
~---._, __ _
I I
I I
I I
I I
I I
I I
I I
I I
I
Figure 34: Conceptual diagram of a Baffle configuration.
,_ ____ I
I I
I I
I I
I I
I I
I I
I I
I I
I I
-- -1
I
I
I
I
I
I
I
I
I
- -1
I
I
I
I
I
I
I
I
I
- .J
95
Applications of baffle panels can include any scale, from residential buildings to administration
buildings. These systems regularly occur in urban environments where higher wind-speeds and
noise-pollution exists. Common applications include high-rise buildings to allow for natural
ventilation, and are a viable retrofit on existing structures.
These panels can be appended in front of window wall or curtain wall facades. Panel size,
module, operability and offset distance depend on the desired function of the outer layer. Baffle
elements usually occur within 0.3 m (1 ft) of the facade but may be greater. This system can be
used to preserve the durability of an interior facade.
The energy demand profile of baffle panel systems is dependent on the panel offset and the
combined insolation of the glazed layers, but is unlikely to provide significant improvement over a
single-skin facade. The glazed systems can range from off-the shelf components to customized
systems combining shading functions. These systems may be operable, and may be selected for
the aesthetic they achieve without the complex integration required for more sophisticated
double-skin facades.
Summary:
• Possible ventilation
• Sound insulation
• Security feature allows for simple night cooling
• Cost-effective
• Weather protection
• Difficult to clean exterior face
96
Box-Window
The box window is the oldest form of double-layer facades. It includes a second glass layer
placed in front of an operable window that generally opens inwards to avoid interference and
allow for cleaning. The box cavity can be ventilated. When arranged adjacent to other box
windows, continuous vertical and horizontal divisions prevent sounds or odor transmission from
adjacent rooms. Other advantages include an external skin that may be fixed or variable
depending on the climatic conditions and desired function. Box-windows are often prefabricated
and part of a repetitive facade system. Disadvantages of this system may be restricted view,
increased construction costs, possible cavity overheating and increased consideration for
maintenance access to clean cavity-based solar control systems.
r--
I
I
I
I
I
I
I
I
1-- -
Figure 3-5: Conceptual diagram of a Box-Window configuration.
97
Typical applications of box-windows can include any scale, from residential buildings to
administration buildings. They may be arranged as a single unit in a punched-opening building
skin, or aggregate in a curtain wall facade. These systems tend to occur in urban environments
where higher wind-speeds and noise-pollution exist. Common applications include high-rise
buildings to allow for natural ventilation, and the possibility of night cooling. Box-windows may
also be appropriate for facade renovation projects.
Box-windows are typically prefabricated in a shop environment. They usually consist of an
insulated glazed interior unit and a single-pane outer skin. The cavity depth is a product of the
desired function and ranges from 10 cm (4 in) to 50 cm (20 in).
The energy demand profile of a box-window system depends on the airflow openings into the
cavity and the resultant insolation of the multiple layers. In general, box-windows take advantage
of solar gain and have improved heating energy performance than a single-skinned facade.
During the summer, box-windows may overheat if the ventilation and operability functions fail to
perform correctly, thus resulting in increased internal room temperatures. Standardized
prefabricated units may be available, but customized systems are costly and most appropriate in
a repetitious design. Increased installation and mock-up testing may be incurred.
Summary:
• Natural ventilation
• Sound insulation, outward and laterally
• High construction costs
• Weather protection
• Night cooling
• Overheating of cavity
98
Alternating Facade
The alternating facade is a combination of both single- and double-skinned facade systems. The
desired performance includes the advantages of both systems. Either the single-skin or double-
skin element may be used to provide ventilation to the interior space. This decision is a function
of the climatic conditions and airflow strategies. An alternating facade typically includes at both
systems available in each room. The combined facade performance may be difficult to predict
with regards to ventilation, noise insulation and solar gain. The composite performance is
dependent on the ratio of each system to the overall facade.
I
I
I
I
I
I
I
I
L
Figure 3-6: Conceptual diagram of an Alternating Facade configuration.
99
Typical applications of an alternating facade can include any scale, frorn residential buildings to
administration buildings. These systems generally occur in urban environments where higher
wind-speeds and noise-pollution exists. Cornrnon applications include high-rise buildings to allow
for natural ventilation, and the possibility of night cooling.
The construction of an alternating facade increases in complexity with the introduction of multiple
systems. The box window element rnay be prefabricated and contain a cavity depth frorn 10 crn
(4 in) to 50 crn (20 in). The cavity space rnay include operable shading devices. The final
proportion of each systern is dependent on the desired function and performance. Attention rnust
be paid to resolving any differential depths or faces of glass that rnay exist between the single
and double-skinned systems.
The energy dernand profile for an alternating facade depends on the proportion of the double-skin
facade and the solar radiation penetration. These systems can yield a slight irnprovernent in
heating dernand when compared to single-skinned facades, but at increased construction costs.
The box-window elements are subject to cavity overheating in the surnrner which will produce
greater interior surface and roorn temperatures.
Surnrnary:
• Many ventilation options
• High construction costs
• Improved sound insulation (depends on proportion)
• Weather protection
• Very high user-acceptance
• Provides interior comfort
• Overheating of cavity
100
Shaft-Box
The objective of the shaft-box configuration is to increase the stack effect to provide natural
ventilation of the cavity space. The facade configuration may be referred to as a twin-face.
Natural air enters each module at floor level and extracts from each cavity to a vertical shaft
which ejects the air at outlets several levels above. Mechanical extraction may be incorporated to
the vertical shafts to accelerate thermal uplift. The primary advantages of this system are thermal
comfort and acoustical insulation. The main disadvantages include complex construction and
fire/smoke requirements. It is common for the shaft-box to include adjacent branches to story-
height or box-window configurations from which air is exhausts.
I
I
I
I
I
I
I
I
f---
1
I
I
I
I
I
I
I
f---
I
I
I
I
I
I
I
I I
I I
+---- t- I
- -- 1- _ _
I
I
I
I
I
I
I
I
.1... _ __ ..._ I
--- .....
~
I\
--...., __ _
I
I
I
I
I
I
I
I
---.., _ _
, __ __
Figure 3-7: Conceptual diagram of a Shaft-Box configuration.
-y- ----.,
I
I
I
I
I
I
I
I I
I
-t-- - 1
I
I
I
I
I
I
I
I
I
-1
I
I
I
I
I
I
I
I
I
~
101
Applications of a shaft-box double-skin facade can include any scale, from low-rise buildings,
administration buildings and high-rises (segmentation may be necessary). Since the height of
stacks has limits in practice, the shaft configuration is most suitable for low-rises. These systems
tend to be located in urban environments where high noise-pollution exists. These systems often
occur in a naturally ventilated strategy but may utilize mechanical extraction.
Shaft-box system constructions contain box-windows with inlets adjacent to a vertical shaft runs.
The cavity space in multi-skin double-skin facades may range from 0.3 m (1 ft) to 1.2 m (4 ft) in
the shaft and 10 cm (4 in) to 20 cm (8 in) for box windows. The box-windows may be
prefabricated units. Mockup tests and technical costs can be high.
The large uninterrupted shaft forms a climatic buffer zone which reduces the winter heating
demand, but can magnify the summer cooling demand. It is possible to circulate solar gain in the
cavity to other areas of the building; however, the buffer zone usually extracts air by natural or
mechanical means. During summer months, increasing airflow by accelerating the extraction rate
may prevent overheating. Modifying extraction is achieved by varying inlet and outlet openings.
Summary:
• Mechanical ventilation
• High construction costs
• Weather protection
• Improved sound insulation from exterior
• Lateral and vertical transmission of sound and odor
• Overheating of cavity in summer
• Consistent aesthetic
• Many structural opportunities
102
Story-Height- Corridor
The corridor facade consists of an airflow cavity which closes off at each floor level. The
horizontal extent of the corridor typically encompasses multiple rooms and allows sound and
odors to be transmitted to adjacent spaces. The ventilation of corridor facades may occur
vertically, horizontally or both. Magnified pressure effects must be addressed at the horizontal
extents of the subdivision. On the exterior facade, air intake and extract openings stagger to
avoid direct recycling of exhaust air from a lower cavity. Primary advantages of the corridor
facade include easy cleaning, integration of other functions, opportunity to control cavity pressure
and a homogeneous aesthetic. Disadvantages may include restricted views and increased fire-
safety requirements to mitigate fire/smoke spread.
-·----·---..., ___ ..., __ _
1---.,
I
I
I
I
I
I
I
I
I
Figure 3-8: Conceptual diagram of a Story-Height (Corridor) configuration.
103
Typical applications of corridor facades occur in high-rise and administration buildings. These
systems are primarily located in urban environments where higher wind-speeds and noise
pollution exists. Common applications include high-rise buildings to allow for natural ventilation
where it would not be possible with a conventional single facade.
A corridor facade construction typically uses an external curtain wall system. The interior facade
systems may be made of wood or materials sensitive to weather protection. The corridor cavity
depth ranges from 0.3 m (1 ft) to 1.2 m (4 ft) and is dependent on the performance strategies and
a combination of functions. Operable lites on the interior skin, often internally opening or sliding
doors, provide access into the cavity space. The cavity depth may be occupiable beyond
maintenance events, but rarely functions as such.
The configuration provides improved heating energy demand compared to single-skinned
facades and has the possibility of circulating solar gain to other areas of the building. Solar heat
gain may be harvested efficiently during the winter, but presents the possibility of overheating in
the corridor cavity. Corridor systems usually do not require exceptional materials or assemblies
but will have high construction costs.
Summary:
• Ventilation options and pressure control
• High construction costs
• Weather protection
• Improved sound insulation from exterior
• Lateral transmission of sound and odor
• Overheating of cavity in summer
• Uniform aesthetic
104
Story-Height- Juxtaposed Modules
The story-height facade with juxtaposed modules is similar to the corridor facade including an
airflow cavity that seals off at each floor level. The horizontal extent of the module typically
separates rooms (or more frequently at a structural module) and prevents sound and odors to be
transmitted to adjacent modules' spaces.
4
The ventilation of corridor facades may occur
vertically. This module can be looked at as a horizontally extended box-window. Primary
advantages of the story-height facade with juxtaposed modules include easy cleaning, integration
of other functions and a homogeneous aesthetic. Disadvantages may include restricted views
and overheating of the cavity in the summer.
Figure 3-9: Conceptual diagram of a Story-Height (Juxtaposed Modules) configuration.
105
Typical applications of story-height double-skin facades with juxtaposed modules tend to occur in
high-rise and administration buildings. These systems can emerge in urban environments where
higher wind-speeds and noise-pollution exists. Common applications include high-rise buildings
to allow for natural ventilation where it would not be possible with a conventional single facade.
A story-height facade with juxtaposed modules construction usually contains an external curtain
wall system. The interior facade systems may be made of wood or materials sensitive to weather
protection. The cavity depth ranges from 0.3 m (1 ft) to 1.2 m (4 ft) and is dependent on
performance strategies and functions. Operable lites on the interior skin, often internally opening
or sliding doors, provide access into the cavity space. The cavity depth may be occupiable
beyond maintenance events, but rarely functions as such.
The configuration provides improved heating energy demand compared to single-skinned
facades and prevents room to room transmissions. Solar heat gain may be harvested efficiently
during the winter, but presents the possibility of overheating the cavity in summer. Story-height
facades with juxtaposed modules systems usually do not require exceptional materials or
assemblies but will have high construction costs.
Summary:
• Natural ventilation
• High construction costs
• Weather protection
• Improved sound insulation
• Overheating of cavity in summer
• Uniform aesthetic
106
Multi-Story
The multi-story double-skin facade has a cavity which has no horizontal or vertical divisions. The
space may encompass multiple rooms, or even an entire building facade. The outer and inner
skins of a multi-story facade are typically independent of each other. Airflow opening locations
only occur at the bottom and top of the cavity, magnifying overheating problems. Air exhausts at
the top of the facade by thermal buoyancy and/or mechanical means. The primary advantages of
this system are a consistent, uninterrupted appearance, and exterior noise reduction. The
primary disadvantages include lateral and vertical transmission of sound and odor via cavity, strict
fire-safety requirements and overheating. The possibility of designs is greater in the multi-story
classification than any other category.
Figure 3-10: Conceptual diagram of a Multi-Story configuration.
107
Typical applications of a multi-story double-skin facade can include any scale, from residential
buildings to administration buildings. These systems appear in urban environments where higher
wind-speeds and noise-pollution exists. Common applications include building renovations and
preservation of historic buildings. Most applications of these systems occur as part of a
mechanically ventilated strategy.
The multi-story skin can be constructed in front of any number of new or existing building types.
The external support skin can be conventional or high-tech and is independent of the interior
skin's structure, with few exceptions. The cavity space in multi-skin double-skin facades may
range from 0.6 m (2 ft) to 2.0 m (6 ft); greater than other classifications. The cavity depth may be
larger, but a maximum depth differentiates the multi-story double-skin from climate halls and atria.
The large uninterrupted cavity forms a climatic buffer zone which reduces the winter heating
demand, but can magnify the summer cooling demand. It is possible to circulate solar gain in the
cavity to other areas of the building; however, the buffer zone usually extracts air by mechanical
means. During summer months, increasing airflow by accelerating the extraction rate may
prevent overheating. The material availability for multi-story DSFs depends on the complexity of
the exterior structure. Long span conventional structural systems can be implemented, or
conversely, lightweight structures with minimal material and visual intrusions.
Summary:
•
Mechanical ventilation
•
Lateral and vertical transmission of
•
High construction costs sound and odor
•
Weather protection
•
Overheating of cavity in summer
•
Improved sound insulation from the
•
Consistent aesthetic
exterior
•
Many structural opportunities
108
Multi-Story: Shingled
The multi-story shingled double-skin facade has a cavity which has no horizontal or vertical
divisions. The space may encompass multiple rooms, or even an entire building facade. The
outer and inner skins of a multi-story shingled facade are typically independent of each other.
Airflow openings are frequent along the cavity, usually at story or unit heights, helping to mitigate
overheating problems with greater air permeability compared to the aforementioned multi-story.
Air extracts from the top of the facade by thermal buoyancy and/or mechanical means. The
primary advantages of this system are more evenly distributed airflow patterns within the cavity
space and exterior noise reduction. The main disadvantages include a non-planar exterior
aesthetic.
Figure 3-11: Conceptual diagram of a Multi-Story (Shingled) configuration.
109
Typical applications of a multi-story shingled double-skin facade can include any scale, from
residential buildings to administration buildings. These systems may appear in urban
environments where higher wind-speeds and noise-pollution exists. Common applications include
building renovations and preservation of historic buildings. Most applications of these systems
occur as part of a mechanically ventilated strategy.
The construction of multi-story shingled double-skin facade is more complex to design and
implement. The multi-story shingled skin can be constructed in front of any number of new or
existing building types. The external support skin can be conventional or high-tech and is
independent of the interior skin's structure, with few exceptions. The cavity space in multi-skin
double-skin facades ranges from 0.3 m (1 ft) and 1.2 m (4 ft). When planning a shingled facade
the adjacent spaces and corner conditions must account for the angled profile interface.
The large uninterrupted cavity forms a climatic buffer zone which reduces the winter heating
demand, but can magnify the summer cooling demand. The regular introduction of airflow across
the height of the facade, as opposed to a single inlet at the bottom, provides some relief to the
concerns of overheating present in a traditional multi-story double-skin facade. The airflow inlets
may be mechanically controlled to regulate the amount of airflow introduced at each inlet.
Summary:
• Natural ventilation
• Higher costs: technical design, construction and maintenance
• Improved sound insulation
• Weather protection
• Reduced overheating
• Non-planar exterior face
110
Multi-Story: Controllable
This variable system is similar to a multi-story facade, except that the outer skin includes pivoting
louvers and is not airtight. The system may be segmented at each floor level or extend across
the entire outer skin. The cavity usually contains metal floors at each floor to provide access for
maintenance and cleaning. The primary advantage of this system is its variable settings which
respond to climatic conditions. Disadvantages of this system include lateral and vertical
transmission of sound and odor when the cavity closes. Additional disadvantages include the
construction and maintenance costs.
Figure 3-12: Conceptual diagram of a Multi-Story (Controllable) configuration.
111
Typical applications of a controllable multi-story double-skin facade can include administration
buildings, high-rise structures and building refurbishment. These systems may materialize in
urban environments where higher wind-speeds and noise-pollution exists. Common applications
include building renovations and preservation of historic buildings. These systems are typically
part of a natural ventilation envelope strategy.
The construction a controllable multi-story double-skin facade is the most complex to design and
implement. The outer skin includes an array of glass louvers which can rotate to vary the area of
opening. To achieve this level of responsiveness with controls requires many moving parts which
require commissioning and maintenance. The louvered skin is offset between 0.3 m (1 ft) and 1.2
m (4 ft). When planning a louvered facade the adjacent spaces must account for the louver
profile when occupied to avoid interferences.
The louvered outer skin creates a climatic buffer zone which harnesses solar gain during winter
months, thus reducing the heating demand. During the summer, overheating can be avoided by
opening the outer skin. The materials required for a controllable skin are available, but the
system design, fabrication, installation, commissioning and maintenance are costly.
Summary:
• Natural ventilation
• Controls adapt skin to climate
• Very high technical design, construction and maintenance costs
• Improved sound insulation (depending on louver position)
• Weather protection
• Less visual interruption
• Reduce or eliminate overheating
112
3.2.4 Solar Control Strategy
The solar control type, operability, and controls describe the systems integral to a double-skin
facade that reduce solar heat gain when desired. Often the cavity space houses and protects the
shading devices. The most common systems include horizontal louvers that retract and roller
blinds of fabric materials. It is possible for double-skin facades to have multiple solar control
system types. In deep cavity spaces, it is beneficial to locate solar control systems not to close to
either skin (.:':_20 cm if possible) and ideally in the front 1/3 of the cavity depth. Material type is
also a key consideration. The solar control strategy is classification includes:
Solar Control Type:
• Overhangs (exterior)
• Fins
• Louvers
• Blinds
• Screens
Operability
• Fixed
• Retractable
• Adjustable (tillable)
• Retractable and Adjustable
Controls:
• Manual
• Automated
• Automated with Manual Override
113
3.3 Benefits
Advanced facade concepts are more relevant today than due to growing environmental
consciousness and a demand for improved indoor air quality. Access to daylight, natural
ventilation and comfortable indoor climates are requirements of most modern constructions.
Increased performance expectations coupled with decreased energy consumption present
planners, architects and engineers with a new challenge for the future. Ultimately, new building
typologies and solutions will emerge to transform the building skin as we know it - double-skins
being one of them. The most frequently cited benefits of double-skin facades are acoustical
insulation (Lee, et al. 2002, 18), protection of solar control devices and restoration/retrofit.
Double-skin facades can have the following potential advantages when compared to conventional
single-glazed facade.
Acoustical Insulation
Sound insulation is one of the best reasons to implement a double-skin facade (Oesterle 2001,
34). Multiple layers restrict the penetration of sound from the exterior to the building interior.
Depending on the double-skin configuration, internal room to room sound transmission can be
prevented.
Renovation
The potential of double-skin facades in retrofit projects is an effective and economical alternative
to new construction. This approach has many benefits including preservation of existing building
stock, an aesthetic upgrade (modernization), improved acoustical insulation, potential for reduced
energy consumption and a perception of sustainable initiative. Buildings that fit the profile for a
facade retrofit are typically structures that remain in the possession of one entity for an extended
period of time - government, institutional, or historically landmarks.
114
Winter Thermal Insulation
The addition of an external skin, combined with the cavity space, increases the external heat
transfer resistance. The cavity space acts as a buffer zone that is warrner than the exterior, thus
reducing the rate of heat transfer at the exterior glass. The rnain benefit of double-skin facades in
colder climates is their ability to use the solar heat gain. The air located between the two glass
skins inside the air cavity acts as a thermal buffer. This buffer facade insulates the building
interior frorn losing heat and significantly improves the U-value of the building. Lower heat losses
rneans a reduced heating consumption profile. The use of low-e glass in double-skin facades
permits solar heat gain and daylight to penetrate the envelope while simultaneously preventing
heat loss frorn the space.
Summer Thermal Insulation
Double-skin facades can only cool the air inside the structure by several degrees lower than the
actual external temperature. It would be ideal if naturally ventilated DSFs could provide the sarne
(or better) cooling action as air-conditioners. However, the combination of both natural and
mechanical ventilation will offer better results. Of course, mechanical rneans energy consumption,
but this rnay be offset as internal temperatures are already lower than outside temperatures.
Heat inside a building is develops with the penetration of the sun through transparent surfaces. In
office buildings, the primary heat loads rnay be internal heat gain: computers and office
equipment, artificial lighting, body heat, etc.
Cooling loads can be reduced by using a double-skin systern to night cool a building. The basic
principle of this systern is to open and ventilate the facade at night when the temperature is lower
than in daytime. Overheating of the cavity air space during surnrner rnonths can be avoided by
natural or mechanical ventilation. Air extraction occurs at the top of the cavity space, removing
warrn air transported by the natural stack effect. The exterior skin of a double-layer envelope can
115
provide security to the interior skin during night ventilation. Eliminating the need for an air
conditioning system is unlikely but it is possible to reduce overheating by regularly ventilating at
night, providing adequate shading in the summer, and using minimizing the internal heat loads.
Night Cooling
During the summer, reduced energy demand can be achieved by night cooling a building. This
process of ventilating the building, and pre-cooling for the next work day reduces indoor
temperatures during the early morning. The addition of a second skin allows DSF systems to be
night ventilated while preserving a layer of security in the outer skin.
Energy Savings
Double-skin facades can save energy when located and designed correctly. The reduced
dependency (or elimination) of a mechanical system will lower electricity costs. Warm air trapped
in the cavity space may be redirected to other areas of the building during winter months. Energy
savings are greatest when there are levels of variability and control of the ventilation strategy.
Natural Ventilation
Since dual-facades use a second glass skin, building users can also leave windows open during
various climatic conditions. The first exterior skin shields the entire building from wind and rain,
allowing natural ventilation to occur through air corridors between the skins. By introducing the
second skin facade, a 24-hour natural ventilation program is possible without compromising
interior comfort Exposure to high-wind loads pose a threat to the feasibility of natural ventilation;
however, some researchers remain optimistic that there is a moderate zone to be harnessed: in
addition to saving energy for air conditioning, the facade can introduce fresh air into an interior
room under moderate strong wind condition (Kawai, Nishimura, et al. 2009, 2).
116
Thermal Comfort
In theory, the inside temperature of the glass in a dual-wall system will be closer to room
temperature than a traditional curtain wall system. The comfort of the overall space aims to be
more comfortable than with a traditional facade system. Improved thermal comfort may be
achieved parts of the year while succumbing to cavity overheating during the hotter summer
months. This is attributed to seasonal climatic variations.
Daylighting
Openings in the building envelope provide access to daylight and views. These openings are
appropriate at eye level, allowing deep penetration of sunlight into the depth of the room. The
addition of a second glass layer reduces the daylight factor in the depths of the interior spaces.
Additionally, any shading devices or other systems with reflective materials can negatively
contribute to issues of glare. The effect of glare can result from high contrast of luminance or
illuminance in three forms; direct glare, contrast glare, or reflection glare. The magnitude of glare
relates to the surface properties and colors of materials. This discomfort can be avoided by
simulation and through glare control strategies within the facade - the intermediate cavity.
Reduction of Wind Pressure
The additional glazed layer shields the interior layer from high winds. This strategy permits
natural ventilation through operable interior windows, even in high-rise applications. The system
depths and glass makeups of the interior and exterior skin are a function of the direct pressure
encountering each layer. Lower wind pressure across taller building heights may permit
protected solar control devices and operable windows where they could not exist in a single-skin
alternative.
117
Transparency- Aesthetic
The use of dual-facade systems often exposes a structural mechanical aesthetic visible to both
the interior and exterior. This look is often attractive to designers and allows for tremendous
transparency outwards from interior spaces. These high transparency facades can achieve an
ethereal, fragility that often contrasts the larger steel and concrete building structure members.
Double-skin facades are not only aesthetically beautiful, but they also promote greener
technology with the opportunity to integrate alternative energy generation, thus, reducing fossil
fuel usage and carbon emissions into the atmosphere. An example of such a system is one
designed in such a way that, in addition to permitting natural air circulation, it also uses solar
energy converted into electrical energy
Security
Security improves with the addition of a second glazed layer. Baffle panels and louvers even
provide protection during night-ventilation.
Improved Protection of Shading Devices
Placing shading devices within the cavity space can reduce or eliminate exposure to wind and
moisture. This preserves the shading system and improves accessibility for cleaning and
maintenance.
Fire Escape
The cavity space in a corridor or multi-story double-skin configuration can function as a circulation
or fire escape route.
118
Lower Construction Costs (Compared to other intelligent technologies)
Intelligent facade technologies such as variochromic panes (properties adapt to climatic
conditions) are more expensive. Though variochromic solutions utilize less material, double-skin
system variability can be achieved with a combination of components and controls in a layered
enclosure approach.
The notion that lower construction costs is an advantage with double-skin facades is not widely
stated in literature or practice, but it is a relative comparison and raises the question of what is a
reasonable baseline comparison for a double-skin? It is the author's opinion that a conventional
single-skin facade is not a fair comparison, but instead should be evaluated against a high
performing single-skin facade that provides a similar level of solar control with exterior sunshade
devices. Each system must possess equivalent performance metrics to fairly assess the costs.
Additionally, the operations and maintenance costs of each system over time should be factored
into a life-cycle cost evaluation (e.g., external sunshades on a high-performance single-skin
facade likely require more frequent cleaning and replacement than those housed in a double-skin
facade).
LowSHGC
The solar heat gain coefficient (SHGC) of the entire assembly can be lower than a conventional
single-glazed facade. When SHGC refers to the center-of-glass solar energy transmittance it is
analogous to the g-value coefficient commonly used in Europe.
5
A lower SHGC reduces the
transmission of heat through the building facade, resulting in reduced peak loads and potential
energy savings.
119
3.4 Drawbacks
Double-skin facades can have the following disadvantages when compared to conventional
single-glazed facade.
Higher Investment Costs
Double-skin facades require more glass, more structure, more fabrication, more installation, and
more money than conventional facades.
Maintenance Costs
The double-skin facade has twice the number of surfaces which require cleaning compared to
single-skin facades (four to two). The accessibility to the glazed surfaces can be a factor in the
costs of cleaning. Ultimately, the most labor intensive and costly maintenance is the cleaning of
the shading elements. These will govern the total cleaning facade cost (Bodart and Gratia 2002).
Operational Costs
The problem of overheating in the cavity space requires additional cooling to maintain interior
comfort levels. This equates to additional operations costs.
Overheating
This can be due to the integration of blinds or solar deflectors which become the hottest element
in the system. Air movement through the cavity can alleviate some of this heat gain, but some
re radiates in the infrared spectrum to the glass surfaces on either side of the cavity. The inner
surface of glass ends up hotter than if there were no blinds.
120
Increased Airflow
In large multi-story cavity spaces, considerable pressure differences may cause undesirable
drafts for interior-skin natural ventilation.
Increased Structural Weight
The additional material required in for a multi-layered facade results in greater construction
weight Ultimately this increases material, transport and installation costs associated with the
building envelope.
Daylight Quality
The introduction of an additional external skin reduces the visible transmittance and quality of
daylight penetrating into the interior. The additional depth as a result of the cavity can also
detract from daylight factors, in particularly with solar control devices in a shading position as
opposed to a light penetration position.
Increased Electrical Loads by Lighting
If the double-skin facade reduces daylight levels below comfort levels, increased use of artificial
lighting systems is probable.
Acoustic lnsulation;Cavity Sound Transmission
Depending on the cavity configuration, sound transmission can occur between rooms or floors.
The cavity can magnify minor noise disturbances. An operable exterior facade also has less
sound insulation when it is in an open position.
121
Fire Protection
The cavity of a double-skin facade may function as a vessel for smoke or fire transmission
between spaces.
Reduced Floor Area
The cavity depth may subtract floor area from the usable space - not an appealing idea for any
building developer or client The depth of the cavity should be selected to optimize performance
and maximize interior space. Additionally, a firm understanding of local building codes' regulation
of the enclosure's relationship to property setback lines is critical, specifically in retrofit
applications where appending the second skin to the exterior of the existing building face may be
desired.
Overestimated Energy Savings
Inaccurate modeling and a lack of envelope commissioning can implement systems which never
reach their intended design performance levels. This yields increased peak heating and cooling
loads. Ensuring early energy modeling assumptions carry through implementation, or, revisiting
the energy model and updating to reflect any post-value-engineered as-built conditions would aid
in improving the accuracy. Precaution should be taken when basing the justification for a double
skin facade primarily on energy performance or savings.
Surface Condensation
Condensation occurs when the ambient air humidity outside a surface is greater than or equal to
the saturation humidity at the surface.
122
3.5 Planning and Feasibility
Successful implementation of double-skin facades requires early planning, evaluation, sensible
problem-solving and ideation amongst a multi-disciplinary project team. As the facade takes on
more layers and heightened performance expectations, the coordination, execution and
communication becomes more critical to making a double-skin facade concept a reality. The
following presents a number of facets integral to the planning and feasibility of projects that
include double-skin facades.
Clear Performance Priorities
There are many reasons for considering a double-skin facade, but sole-dependency on improved
energy performance may not be a justifiable basis for implementation. Sound insolation - on the
other hand - could frame a strong enough argument to see a double-skin facade application.
Before selecting a double-skin facade strategy, and even before considering the alternatives, the
project team must clearly articulate the performance priorities, which are primary, what their
weighting and hierarchy are, and which performance priorities are local conditions and which are
global requirements. After establishing priorities, assigning goals for performance metrics can
steer early design iterations and ultimately become integral to the project specifications for a
specialty facade subcontractor to implement.
Support of Ownership
The involvement of client, ownership or ownership representation is critical when looking to
introduce a facade solution beyond the common conventional single-skin facade. There must be
a commitment by ownership to the aforementioned performance priorities so that budget
conversations become more about the value gained than the greater initial costs. The
formulation of clear performance goals is key to establishing what the baseline enclosure would
be without a double-skin, so budget pricing and feasibility exercises are in the same arena of the
123
final performance. In return, the project team must work to provide regular cost information to
ownership so that there is an understanding of what gains may be possible with the introduction
of a double-skin facade. The need for a double-skin facade should rise from a combination of site
issues, programmatic demands, performance goals and ownership's desire for a certain quality of
occupant experience and comfort. When continuously gauging costs of early design strategies, a
clear understanding of an owner's expectations for comfort performance is essential. An active
dialogue with, and support for the exploration of a double-skin facade are instrumental in
effectively implementing such a system if deemed appropriate.
Design Team Coordination
Uniting the parties affected by an advanced facade strategy early in a project's development is
beneficial to accurate upfront budgeting, as well as the long-term outcome of a system's quality in
design and through execution. This may include, but is not limited to, designers, general
contractor, facade consultant, mechanical engineer, specialty consultants (i.e. acoustical,
structural and security) and facade subcontractor. Front-end involvement from the builder is
quintessential if opportunities for developing efficient field-erection techniques are to be identified.
With many players involved early, it is critical to leverage one another's knowledge, create a
forum for collaboration, identify each party's limits, outline scope of responsibilities and establish
methodologies for decision making. The interdisciplinary dynamic is unique to each project.
Building Orientations
A sensible application of a double-skin facade first begins with determining if the macro-climate
presents opportunities for improved performance followed by a micro-scale assessment of what
opportunities there are to respond to local conditions including solar gain, prevailing winds,
acoustical sources and views. Given the primary advantage of protected solar control devices
within the cavity, a double-skin facade is most effective at an orientation that harnesses solar
124
heat gain in the winter. In the northern hemisphere, a southern exposure permits solar gain
during winter and mitigates overheating in surnrner rnonths.
Leveraging Installation Methodologies'
Assembly and installation issues with double-skin facades range as widely as the systern
variations. Unitized double-skin curtain wall systems can be complicated by the need for panel
operability to accornrnodate maintenance needs. Prefabrication rnay include the installation of
shading devices and controllers as part of the unit assembly process. Once the unit assembly is
complete, however, installation proceeds similarly to conventional units, except the units are
typically heavier, which rnay preclude lifting several units simultaneously.
Multi-story double-skin facades present quite another scenario. Because of the long spans
typically involved, these applications will often have exposed structural systems, sornetirnes
requiring architecturally exposed structural steel (AESS) standards. This type of work is often
unfamiliar to glazing contractors and steel fabricators and is rightly regarded as a specialty itern
In fact, rnany of the rnulti-story DSFs referenced above rnake use of structural glass facade
technology, including the use of frarneless glass systems, as a support strategy for the exterior
skin. The interior skin is often a conventional curtain wall or storefront type systern The issue is
with the exterior skin, its rneans of support, and the required cavity work. The cavity often
incorporates maintenance platforms, shading devices, and potentially other mechanical
components such as operable vents. These rnay or rnay not be included in the facade
contractor's scope of work. An issue of particular concern is the cavity depth: the deeper the
cavity the easier it is for workmen to operate with all the required equipment. Cavity depths less
than 0. 75 rn (30") seriously constrain ease of rnovernent for the workmen.
Access is a consideration on any facade, with little difference here. If installation of maintenance
platforms in the cavity occurs before the outboard skin, they can be instrumental in placing the
125
second skin. If not, temporary platforms may be required within the cavity. Depending upon the
glazing system design, it is possible to have workers on both sides of the skin.
It is beneficial to bring facade subcontractors who will construct the enclosure into discussions as
early as possible in the design process. This accelerates their understanding of implementation
strategies to improve the installation means and methods that minimize costs and uphold the
integrity of the system's design.
Maintenance and Operations
It is always prudent to include those who will operate a building during its life in planning and
design discussions, but is so often overlooked. For projects that include a double-skin facade a
facilities management team's input is relevant to an enclosure's development, more so than for a
conventional enclosure. Such integration allows those who will operate the building and
enclosure over time to understand and contribute to design conversations. This can lead to a
better system design, improve functionality and ease maintenance requirements. This should
include discussions about maintenance frequencies of systems within the cavity, replacement
strategies for components over time and access to the cavity (if any is to be provided).
Integration with Building Management System (BMS)
If the double-skin facade utilizes automated elements (i.e., mechanical systems, operable vents
and operable shading), it is critical to consider how these controls tie into a larger building
management system, and include those responsible for the implementation and use in pertinent
discussions. Additionally, commissioning of these systems is quintessential before project
handoff to a building owner and occupants.
126
Chapter 3 Endnotes
Recent efforts by the American Society of Heating, Refrigerant and Air Conditioning Engineers
(ASHRAE) 90.1 Committee have proposed reducing the allowable window-to-wall ratio (Wl/VR)
from 40% to 30% for the prescriptive path, along with more restrictive U-factors and SHGC
values and the addition of a new minimum for VT/SHGC. The glazing industry was successful
in overturning the Wl/VR reduction through an ASHRAE appeals panel, preserving the
prescriptive Wl/VR of 40% for the time being (GANA 2010). Part of the grounds for appeal was
that the "Project Committee's decisions were technically flawed since they lacked sound
estimates of likely energy saving". Though there is mounting pressure in restricting the use of
glass as part of the development of prescriptive energy codes, there is hardly a sentiment on
the part of building owners and leading architects to take this into account So often the "cutting
edge" architectural visions use glass rampantly, a vision often desired by building owners and
driven by aesthetics. To work towards more efficient, and ultimately energy neutral buildings,
both communities must expand the dialogue. For energy code regulators, there must be an
emphasis on performance-based design to avoid black-box prescriptive thinking, stifling
potential innovations in smaller-scale commercial projects. For architectural designers and
owners, they must embrace energy performance as a driver for decision making and balance
with the aesthetic goals.
2
Projected population change in Coastal Shoreline Counties (as defined by NOAA) from 201 Oto
2020 will be an increase of 8% with an increase in density from 37 pers/mi
2
over the same time
frame (National Oceanic and Atmospheric Administration (NOAA) 2013). The density of Coastal
Shoreline Counties exceeded six times greater than the corresponding inland counties in 2010.
3
This statement is in large-part based on a series of AIA/CES courses, conducted by the author,
with dozens of architectural firms between 2011 and 2013. These sessions often included
discussions around classification of systems and where the bounds relate to multi-layered
(scrim) systems, non-transparent systems (often rain-screens examples) and materials other
than glass.
4
To the author's best knowledge, the use of "story-height with juxtaposed modules" differs
herein from the meaning of "ventilated double facade partitioned by storey [sic] with juxtaposed
modules" outlined in Ventilated double facades classification and illustration of facade concepts
(Lancour, et al. 2004, 7-10). The usage herein refers to a corridor-like story-height configuration
with subdivisions, but not to the point that the subdivision equals that of the primary skin's
module. The believed usage in Lancour, et al. (2004, 10) would be encompassed in the "box
window" definition herein. The paper's description limits the "ventilated double window" (aka
box-window) to "a filling element in the wall" while the "storey [sic] with juxtaposed modules"
configuration would account for a unitized system equal in dimension to the primary skin
module.
Based on forthcoming applications of double-skin facades that have corridor-like story-height
facades with vertical subdivisions at every primary structural line (approximately 10 m (33 ft)
on-center), the term juxtaposed modules applies differently. For the purposes of this research,
a "box-window" includes both infill and unitized configurations up to one module wide, and
"story-height with juxtaposed modules" include cavities greater than one module wide but less
than the width of the elevation.
127
5
In the United States, SHGC is used to describe the solar energy transmittance of a whole
window or wall assembly. When used solely to describe the center-of-glass solar energy
transmittance, the SHGC is analogous to g-value, the coefficient used in Europe. Both SHGC
and g-value range from 0 to 1.
6
Narrative describing comparing installation issues of unitized and non-unitized multi-story
double-skin facades is a partial extraction from Seeing double: applications of double-skin
facade (DSF) systems on the rise (Vaglio and Patterson 2011 ).
128
4. Emergence of Double-Skin Facades
Double-skin facades are not a new concept. In fact, considering James Stirling's Natural History
Faculty Library at Cambridge University (1964-1968) as an early double-skin example, there is
half a century of experience, application and debate that form the context from which modern
systems have evolved. Even prior to Stirling's project, ancestral applications to the double-skin
facade were implemented not by world-renowned architects, but in residential applications by
homeowners in cold European climates. This chapter begins with these novel systems,
progresses towards the modern double-skin facade, emphasizes the leadership of Europe in
advancing the understanding and application of these systems, and ends with trends that have
been observed by researchers and in practice. Some of the systems break the typological
definitions outlined in the previous chapter and are identified where possible. However, these
systems remain in the discussion because widespread acceptance of these projects as influential
developments exists in the current literature. The common thread over centuries of evolution is
the fundamental desire for transparency in combination with occupant comfort; thermal, daylight
access, views, solar control and access to ventilation.
4.1 Historic Evolution of Double-Skin Facades
Transparent windows that have the ability to adapt to different seasons -trapping heat during the
winter to enhance internal comfort - have been around for centuries and are the foundation of the
double-skin facade emergence. Over many years and across multiple continents, the quest for
increased transparency while controlling thermal performance has come in the form of both
simple solutions and high-performing intelligent skins. This section outlines the historical
progress of facade design's implementation of transparent systems, including precursor systems
to double-skin facades, notable architectural and research movements, and seminal projects that
use transparency as part of a built-response to environmental conditions.
129
4.1.1 Environmental Origins of Transparency
The historical evolution of multiple-skin facades correlates with increasing application of
transparent building enclosures first beginning in the late 1 J1h century with the invention of the
cast glass process in 1687 (Staib 2007, 11 ). This process allowed the quality and consistency of
flat glass production to improve. The increasing use of glass in facades soon revealed the
inherent weaknesses of heat loss in winter and overheating in summer. These disadvantages
were offset by the double window construction with an additional window outside, referred to as a
Kastenfenster, which even today are added in front of existing windows during winter to create an
air buffer space that increases the insulation compared to single window constructions
(Compagno 2003, 244). These early inceptions in cold climates used the double window to
improve thermal comfort and performance by sealing the outer layer during the winter months to
trap heat. These "storm windows" have been common in central Europe since the 18th century
and can still be observed there today (Herzog et al. 2004, 233). The use of box-type windows
also continues today. These windows are common on old farmhouses and were built with a
removable outer glazing taken off during the summer (Oesterle 2001, 8). These examples were
permitted with glass manufacturing developments, but leveraged by individuals who looked to
layering of the transparent material to enhance comfort in dynamic climatic conditions.
Figure 4-1: Box-window after restoration (kastenfenster nach der restaurierung).
Source: http://commons.wikimedia.org/wiki/File:Kastenfenster_nach_der_Restaurierung.jpg>;
©Handwerker (accessed August 23, 2013).
130
The development of clear glass presented new opportunities for designers and builders. Once
society became familiar with the transparency of clear glass, it was favored over stained or
colored glass because it was used to "enhance the beauty of other elements rather than being a
beautiful material itself," (Elkadi 2006, 19). The popularity of the transparent aesthetic created a
clearer connection between the interior and exterior. Nature could be visible while occupants
protected from inclement weather. This acceptance fed the expanding demand for daylit spaces.
The concept of using glass to control climate, rather than simply protecting against weather,
originates in horticulture. As world travel became more accessible, the displacement of exotic
plants back to populated civilizations was considered a luxury. It was a means to educate locals
on the climates and natural beauty that exists beyond their region. However, to maintain exotic
plant life required the creation of climate zones not inherent to the colder climates of Europe. The
greenhouses and conservatories accommodated wild plant life with long-span structures and
climate imitation while serving as a building typology for innovation, not by architects, but by
gardeners and engineers like Claudius Loudon and Joseph Paxton. These elegant structures
were free of ornamentation, and prioritized achieving maximum sunlight and reducing material.
Such large fragile structures were made possible by the interaction of iron and glass.
Figure 4-2: Palm House at Bicton Gardens (1825), Devon, England. Source: Addis (2007,
340), courtesy Bicton Gardens Archive.
131
Constructed by the Bailey brothers, the Palm House's glass skin acts like a membrane
subdivided by structure only to attain the necessary bracing, resulting in a monocoque structure.
4.1.2 Industrial Revolution
The advent of new materials and production methods presented new opportunities within
architecture. Iron and glass had become widely accepted in the 19
1
" century, and the new-found
limits of transparency were pushed even further. Architecture sought structural efficiency which
maximized transparency and minimized material. The iron mullions in many designs were
spaced at a module that would permit the most slender profile possible. "The process of
dissolving the building skin - its de-materialization - is directly linked to the progressive
independence from its load-bearing function," (Schittich 2001, 12). Elegantly glazed designs
became particularly common due to improved efficiency in their fabrication and erection.
The climax of the transparent buildings of the time is undoubtedly Paxton's monumental Crystal
Palace (1851) design for the World Fair in London. Paxton, a gardener, had explored iron and
glass structures in his development of greenhouses and conservatories. This structure
addressed programmatic conditions, issues of scale and span, costs, prefabrication and
assembly times. The entire structure covered 70,000 m
2
(750,000 sf) and was constructed in just
twenty-seven weeks (Addis 2007, 359). This was achieved through a strict modular approach,
the use of rigid connections and a structural frame that could be extended in two directions. Such
a structure erected in such a short-time had not been seen before. Paxton's problem solving was
innovative and the fame of the Crystal Palace spread globally.
The Crystal Palace taught the world how to build great spaces of structure and light The
fascination with iron and glass translated into many large, long-span, train stations around Europe
designed by engineers. These designs were driven by structural efficiency and economy and
benefited from the exhibition Paxton displayed in his monumental construct.
132
Figure 4-3: Paxton's Crystal Palace (1851), London, England. Source: Addis (2007, 358).
4.1.3 European Origins
Around the same time of Paxton's Crystal Palace, ideas of multiple-skin envelopes began to
surface around Europe. In 1849, Jean-Baptiste Jobard described one of the first mechanically
ventilated multiple-skin envelope concepts. Serving as the director of the Industrial Museum in
Brussels at that time, Jobard described this idea as circulation of hot air between two glazings in
the winter and circulation of cool air during the summer (Saelens 2002, 7).
The building credited as being the first double-skin facade is the Margarete SteifF Factory (1903)
in Giengen, Germany. The structure includes the first-known story-height, fully glazed buffer
facade (Herzog et al. 2004, 241 ). The second skin spans the entire three-story eastern wing of
the toy factory. This additional glazed layer improves the thermal and acoustical insulation
properties of the building while still maximizing daylight. Structural steel elements are positioned
within the air cavity and rest on a concrete base. The building was a success and two additions
were built in 1904 and 1908 with the same double-skin system, but swapping timber in lieu of
steel for budget reasons. The buildings are still in use today. Buildings like the Steiff Factory, and
AEG Turbine Hall (1908-1909) by Peter Behrens, are frequently overlooked in architectural
history because they are considered engineering solutions rather than architectural works.
133
Figure 4-4: Steiff Factory (1903), Giengen, Germany.
Source: Herzog et al. (2004, 241 ); © Achim Bednorz.
Figure 4-5: Fagus Works shoe factory (1911-1925), Saxony, Germany, Gropius & Meyer.
Source: Addis (2007, 407).
134
The principle of "curtain wall" was first expressed in Walter Gropius' (in collaboration with Adolf
Meyer) design for Fagus Works shoe factory (1911-1925). A three-story transparent skin was
suspended in front of the industrial hall, free of all load-bearing function. This freedom from load
bearing was articulated by a fully-glazed corner spanning three floors, void of corner structure.
In 1903, Otto Wagner designed a double-skin skylight in the main hall of his competition winning
Post Office Savings Bank in Vienna, Austria - later constructed from 1904-1912. A similar
solution nearly a century later is Renzo Piano Building Workshop's roof design for Fondation
Beyeler Museum (1997) in Riehen, Switzerland. A double glazed roof provides natural
illumination of the art objects while a mechanically ventilated buffer zone protects against harmful
temperature and humidity variances.
Figure 4-6: Post Office Savings Bank (1904-1912), Vienna, Austria, Otto Wagner
Source: http://www. fotopedia. com/items/flickr-1368980440#context=e37 a32efaf9>;
© kymtyr (accessed August 25, 2013).
135
- 7 - - -
Figure 4-7: Fondation Beyeler Museum (1997), Riehen, Switzerland, Renzo Piano Building
Workshop. Source: Compagno (2002, 126); © Fondation Beyeler I Thomas Dix (top).
Though not all of these works are strictly double-skin facades, their significant contribution to the
evolution of facade technology is integral to understanding the methods and solutions
implemented in a successful double-skin system. These early origins of the double-skin facade
were founded on the same principles that drive the high-tech designs of today's best architects -
comfort, performance, adaptability, and aesthetic.
136
4.1.4 Modern Architecture and Glass
To understand the double-skin facade solutions of the last several decades, it is necessary to
establish a context from which modern DSF's emerged by reviewing the introduction of curtain
wall and large expanses of glazing in architectural applications. The curtain wall first occurred on
commercial buildings near the end of the industrial revolution (i.e. Fagus Works by Walter
Gropius with Adolf Meyer in Germany, 1911-1913; see Figure 4.5). The use of iron and steel in
the building envelope made larger windows and greater proportions of the glazing-to-opaque area
in a building's facade possible. It was the implementation of the skeletal structural frame that
allowed the external facade to be supported floor-by-floor, reducing undesired structural elements
from the occupants' sightlines. It was the architects of the Bauhaus movement, most notably
Walter Gropius and Adolf Meyer, who conceived, advanced and implemented the curtain wall
concept into practice.
Modern architecture would be dramatically different without the global proliferation of glass that
occurred. The use of vast glass expanses, however, can lead to undesirable energy loss, solar
gain, glare and overheating that must be compensated for by energy consuming building
services. The emergence of transparency - and specifically curtain wall technologies - would
present challenges and criticism, but ultimately, the desire for transparency persisted architects of
the modern movement towards a non-load bearing enclosure. This paradigm shift would present
new architectural possibilities as well as challenges and opportunities related to occupant
comfort
The first urban commercial office building to incorporate a suspended curtain wall was the
Halladie Building (1918) in San Francisco designed by William (Willis) Jefferson Polk. A four
story curtain wall was suspended external to the primary load-bearing structure. This facade
prioritized functionality, including a wrapping of the iron fire escapes around the glass and metal
wall as seen in Figure 4.8.
137
Figure 4-8: Halladie Building (1918), San Francisco, Willis Polk. Source: Schittich (2006, 15);
©Christian Schittich.
The swing towards the large glazed expanses provided by the curtain wall created over-heating
and comfort problems in many buildings. These systems were streamlined for economics,
efficient fabrication and ease of construction; however, they did not possess an ability to respond
to changing climatic conditions. The curtain walls' short-comings were quickly hidden by the
advent of air conditioning. The long-seen success of the lightweight curtain wall benefited
considerably from Willis Carrier's introduction of the first fan coil dehumidifying system in 1902.
Carrier's invention enabled an abundance of skyscrapers to be erected in urban areas across the
world with flawed, single-glazed, poorly performing facades. These new systems ensured
thermal interior comfort, but not without the associated energy consumption.
138
One of the first to embrace the new air-conditioning technology was Le Corbusier. In 1930 Le
Corbusier proposed Le Mur Neutralisant, a double-skin system for La Cite de Refuge. With the
integration of air conditioning and a sealed interior, Le Corbusier developed the principle of the
double wall as the carrier of air to change the interior temperature of the building. The cavity
volume provided warm air in winter and used air condition technology to cool air within the cavity
space during summer. In the built design, the air conditioning was omitted for economic reasons
and the building was extremely uncomfortable (Wigginton 2007, 29).
Some scholars do not perceive La Cite de Refuge as an early example of a double-skin facade.
Instead, Arons argues "rather than being the first double-skin facade, this is truly an early version
of the sealed, mechanically controlled building" (Arons 2000, 17). The sealed envelope,
mechanically treated interior prototype was repeated many times including many notable high-
rises of the International Style. This included the New York towers, the Lever Building (1951-
1952) by SOM and the Seagram Building (1954-1958) by Mies van der Rohe,
Figure 4-9: Le Mur Neutralisant for La Cite de Refuge (1930), Paris, Le Corbusier.
Source Knaack et al. (2007, 88); ©VG Bild Kunst
139
Building off Le Corbusier's concept of Le Mur Neutralisant, or neutralizing wall, Bruce Graham of
Skidmore, Oivings and Merrill (SOM) began to investigate double-skin curtain wall strategies in
the United States. Believed to be "the first building to incorporate a double glass wall," (Smith
2007, 11), the Warren Petroleum Executive Headquarters (1957) in Tulsa, Oklahoma applied two
layers of single-glazing, separated by a balcony space. Within this interstitial cavity, there is no
means of mechanical ventilation, making it "ineffective as a device for controlling heat gain into
the occupied spaces of the building" (Smith 2007, 11 ). The balconies that form the cavity space
include sunshades at the head condition comprised of gray-tinted heat absorbing glass. The
application of the heat-absorbing sunshade may have been "intuitive and successful", but even in
the words of Graham himself, "even with what we learned from Le Corbusier's experience, we
knew very little about how to deal with inclement weather," (Tigerman 1989, 14). In the end, the
DSF was ineffective at controlling heat gain in office spaces and steered soon-to-follow SOM
projects, such as the Inland Steel Building in Chicago (1957), away from the double-skin curtain
wall and towards double-glazed insulated units.
The curtain wall enabled many aspects of Modernism and the International Style, but material
and manufacturing advances in the 1950s would establish the curtain wall as an economically
viable solution for decades to come. With the development of laminated glazing, and soon after
the insulated-glazed unit (IGU), the building envelope continued to improve performance (relative
to preceding technologies) in the window pane. These new products were more expensive than
single-glazed panes and did not serve as the primary economic driver. It was the advent of the
float process for glass, patented by Alastair Pilkington in 1959, which improved the industrial
production of glass while driving down prices (Compagno 2003, 244). The advances made in
glass during the modern movement transformed the built environment, especially in the United
States, from a place of stone and opacity to cityscapes of glass, light, transparency and
connection between interior spaces and their exterior environment.
140
4.1.5 Ventilated Facade
In contrast to the sealed envelope approaches of the 1930-1950's, the 1960's marked the
beginning of a movement towards breathable skin explorations, also described as building
envelope strategies that integrated air movement. The first patent for an internally ventilated
facade was rewarded to Lunden and Sbdergren of Sweden. The system was first realized in the
cold climate of the Scandinavian countries, and shortly followed by Germany. The research,
development and execution of ventilated facades remained focused within European countries.
The first DSFs were designed in the 1960's. One of the early examples was the Natural History
Faculty Library at Cambridge University 1964-1968 designed by James Stirling (Compagno 2003,
245). The building is referred to as one of Stirling's avant-garde masterpieces by many in the
architectural community. Compagno also references the Gerri! Rietveld Academie in Amsterdam,
1959-67, realized by the architects GT Rietveld, van Dillen and van Tricht as another early
double-skin construction.
Another early example of a ventilated facade system is The House of Culture in Stockholm (1968-
71) by Celsin and Henriksson. The interior facade is floor height glazing, with an external skin
made of insulated glass. The air cavity is 0.7 m (2.3 ft) deep and contains mechanical convectors
to ensure the circulation air temperature is not too cool in the cold climate. Single-glazed doors
on the interior surface provide access to the cavity space (Compagno 2003, 245).
Fundamentally, ventilation can be integrated within the building facade using either mechanical or
natural sources. Using natural ventilation, a facade can provide passive cooling strategies in
some climates and temperate seasons, resulting in energy consumption avoidance. To respond
to the varying seasons experienced in most climates, ventilated facades often rely on mixed
mode, or hybrid, ventilation strategies that combine natural (passive) ventilation with mechanical
(active) ventilation and cooling.
141
Figure 4-10: Natural History Faculty Library (1964-1968), Cambridge, James Stirling.
Source: http://www.flickr.com/photos/seier/>; © seier+seier (accessed July 18, 2013).
Figure 4-11: House of Culture (1968-1971 ), Stockholm, Celsin & Henriksson.
Source: http://www.flickr.com/photos/2840231O@N06/6662520395/in/set-2157626091522551 >;
©Hans Nerstu (accessed July 18, 2013).
142
4.1.6 Environmental Awareness
The 1960's was a period characterized by counter-culture with increased social activism including
powerful anti-war and civil right movements. During this time, few advances were made with
respect to glass as a material, but the increased implementation of laminated glass and IGU's
became common and economically feasible. Additionally, plastics evolved greatly during this
decade and became well integrated into construction practices. Amongst all these developments
sealed, air-conditioned, curtain wall clad high-rises continued to be built across the United States,
often neglecting their local climatic conditions nor providing much responsive adaptability to
various seasons.
In 1969, Reyner Ban ham was critical of the curtain wall in his book The Architecture of the Wei/
Tempered Environment. He believed there was room for the building skin to reach a greater level
of environmental responsiveness. In his text Banham proceeds to quote the futurist Peter
Scheervbart, author of Glasarchitektur (1914) as follows:
Since air is one of the worst conductors of heat, all glass architecture need this double wall. The
two glass skins can be a metre, or even farther, apart. Lights between these walls shine both
inwards and outward, and the outer, as well as the inner glass can have coloured ornamentation
If too much light is then lost in the colors, the outer skin can be left clear, and all that is needed is
a coloured glass screen
Banham found Scheerbart's vision had an acute consciousness of mechanical and environmental
realities. Furthermore, he believed and understood how double-glazing works, and how it could
be used aesthetically as a responsive skin (Banham 1984, 127). The foresight that Scheerbart
presented is remarkable, and that nearly a century later the cavity depth dimension of 1.0 m (3.3
ft) he alluded to is in agreement with many of the definitions and dimensions suggested by
current researchers and facade practitioners.
143
Following the first oil crisis of 1973, the public at large became aware of the energy crisis and
Earth's dwindling resources. There was a sense of re-evaluating the methods of design and
construction which had necessitated energy guzzling air-conditioning plants for heating and
cooling. It was around this time that the differentiation between natural and mechanical
ventilation received was emphasized, making distinctions between active and passive solar
designs (Braham 2005, 3). Discussion about the benefits of solar heat gain in commercial office
buildings was considered. Many of the reflective, solar-controlled glasses that had been used to
limit sunlight from entering office spaces were now viewed with a negative perception because
they required greater amounts of electricity to power artificial lighting compensating for reduced
visible transmittance. A growing awareness of building energy consumption placed the facade
into a very critical light. This opportunity led architects, engineers and the glass industry to
reconsider the methodologies that had long been known to harness solar energy.
Since the early 1970's, many new glass products designed to mitigate excessive heat gain during
the summer and prevent undesirable heat loss in the winter while maximizing daylight have been
introduced into the marketplace. Examples of such products that improve the glazing's thermal
performance include insulated glazed units, gas-filled air spaces, selective coatings, light diffusing
glazing and more. As these products have infiltrated architectural design, architects and
engineers have aimed to create building systems that respond to, and are sensitive to, solar
energy usage, providing natural ventilation and access to daylight (Andreotti 2005, 1 ). The
increased social awareness of environmental concerns has elevated the building facade, along
with mechanical systems, to be the most influential parts of a building's energy performance.
In search of smarter envelope solutions researchers in the late 1970's began exploring the
possibilities of improved exterior performance through multiple layers. The re-emergence of solar
architecture led researchers to test new construction approaches first at a residential scale.
Several relevant works of the time included: Lee Porter Butler's case studies of double-envelope
144
houses (Smith and Butler 1978); Don Booth's The Double Shell Solar House (1980); and William
A. Shurcliff's Superinsulated Houses and Double-Envelope Houses: A Preliminary Survey of
Principles and Practice (1980) and Superinsulated Houses and Double-Envelope Houses: A
Survey of Principles and Practice (1981 ). In Shurcliff's investigations, he became in favor of
super-insulated solutions over the double-envelope house. He felt that the super-insulated
strategy was already a proven commodity with many built examples performing well. It was
acknowledged in his work that humidity can be a minor issue with the super-insulated home, but
can be combated by opening a few windows. To the contrary, Shurcliff found double-envelope
houses to be a more expensive solution, yet inconclusive whether it was an effective strategy. It
was stated that designers do not have the data to understand which features make the greatest
improvement on performance. He acknowledges that there is a lack of research, measurement
and computer simulation at that time to understand the double envelope strategy (Shurcliff 1980,
12.02).
The research conducted on double-envelope houses differs in several ways from the double-skin
facade condition. The residences studied had small wall opening ratios compared to a fully
glazed facade. The scale of the projects is also different since the double-skin facade is most
applicable to commercial buildings, specifically high-rises. Additionally, the internal functions of a
residence compared to an office space affect the envelope performance tremendously. In an
office setting, the space can be dominated by internal loads generated from equipment. Despite
fundamental differences, these decades-old studies express many of the same arguments that
exist today related to the DSF. Is the long-term performance significant enough to justify the
initial investment in the system? Where is the statistical data to support double-skin facades?
A common urgency exists today that may be similar to the feeling experienced during the energy
crisis of 1970's. As Shurcliff stated, "in an era of energy crisis and cold homes, we cannot afford
to spend much time planning for excitement and fun" (Shurcliff 1980, 12.01). Innovative facade
145
solutions and credible research investigation for multiple climates are required to justify double-
skin facades as a consumable energy-efficient strategy.
4.1.7 Advent of the Intelligent Skin
Continuing on the momentum of the environmental and consumption concerns in buildings raised
in the previous decade, the early 1980s was a period that included attempts at experimental
facade solutions that responded to climatic conditions. Until this point, many facade solutions
had responded to climatic conditions by an addition-of-parts mentality; mechanical systems,
shading components, glass coatings, etc. It was in 1981 when Mike Davies, an English architect
who helped establish Richard Rogers Partnership (known as Rogers Stirk Harbour+ Partners
since 2007), proposed that all levels of climatic and energy regulation between interior and
exterior spaces should be achieved in a multi-layered compound element. This idea of a
"polyvalent wall" which is climatically responsive drives many product developers and research
programs still to this day (Compagno 2003, 244).
-_.,
I
I
I
I
I
I
I
I
I
_)
Figure 4-12: Polyvalent Wall (1981), Mike Davies, Richard Rogers Partnership.
Source: Knaack et al. (2007, 89); ©Mike Davies I Richard Rogers Partnership.
146
Investigation into innovative solutions was not solely in the hands of researchers and product
developers. Practitioners began to test new design strategies and embraced opportunities to
create facades which used natural, renewable energy sources (e.g., solar energy) to alleviate a
building's heating, cooling and lighting demands. Early discourse and built explorations
eventually led to the emergence of the term intelligent skin, which is contained within the
intelligent building. The phrase is most commonly attributed to responsive enclosure systems
with an adaptability based on climatic conditions. A couple notable explanations from facade
researchers include:
The idea of the 'intelligent building' has achieved a certain currency in the past few decades.
With concepts such as 'smart material' it represents the introduction into design principles of
ideas related to self-adjustment and responsiveness, made possible by new technologies in
general, and information technology in particular (Wigginton & Harris 2002, 3).
The word 'intelligent"' indicates the dynamic, almost living capability of a facade to adapt to
changing daily or seasonal conditions in order to achieve a reduction in a building's consumption
of primary energy (Compagno 2002, 129).
An example of efforts towards an intelligent facade is the Lloyd's Insurance Company (1978-86)
in London by Richard Roger's Partnership. The facade principles strive to achieve Le
Courbusier's "mur neutralisant" first attempted at Le Cite de Refuge fifty years earlier (Wigginton
2007, 30). A mechanically ventilated cavity with down-ward airflow improves the thermal comfort
near the internal surfaces of the glass facade. Warm air is mechanically extracted from office
level rooms via air ducts that feed the narrow wall cavity at transom height The air circulates
downward to the bottom of the cavity before moving to the building's mechanical system. The
active envelope is both performative and uniquely aesthetic with visible exterior mechanical
ductwork. This machine-like quality imposes a high-tech persona on its surroundings.
147
Figure 4-13: Lloyd's Insurance Company (1976-1986), London, Richard Rogers
Partnership. Source: Compagno (2002, 113).
Figure 4-14: Institute du Arab Monde (1987), Paris, Jean Nouvel.
Source: author photo;© Jeffrey Vaglio.
148
The Institute du Arab Monde (1987) in Paris is one of Jean Nouvel's early facade experiments
which ultimately catapulted him into architecture allure. The most striking feature of his design is
a pneumatic brise soleil implemented on the south facade. Along this face, high-tech
photosensitive mechanical devices control light and transparency levels by varying the position of
metallic screens within each glass module, reminiscent of traditional Arab latticework. The
computer controlled operability of the 240 modules was intended to vary light levels by
responding to climatic conditions, similar to the dilation of an eye. With innovation and use of
'smart' systems, Nouvel's scheme was ultimately over budget and fell short in execution;
underestimating the internal heat generated by the mechanical controls caused cavity
overheating in the layered facade. The system no longer operates, though the facade's intent
remains a beautiful feature, even in its static state.
These examples of early 'intelligent' systems were complimented by a majority of economic
driven facade constructions across the globe. During the 1990's, the glass industry made
advances in developing numerous insulating and solar control glazing units. The introduction of
low-emission and selective coatings immediately permitted considerably improved levels of
thermal insulation and solar protection. Despite these advances, the conventional facade had yet
to integrate adaptive technologies sensitive to varying solar conditions - which is still lacking to
this day.
4.1.8 Sustainability and the Built Environment
It isn't always easy to draw the line between a useful skin and ornamental packaging . .Between
these extremes, lies a third, equally contemporary path: the building skin as a responsive skin, as
one component of a sustainable low-energy concept. This begins with simple folding and sliding
shutters or with the popular moveable louvers and culminates in multi-layered glass facades
equipped with a multitude of devices for shading and glare protection, light deflection, heat- and
energy gain (Schittich 2001, 9).
149
Familiarity with double-skin facades became more common in the 1990's as a result of increasing
environmental concerns, political influence and the evolution of 'green'. In a very short period of
time, it became fashionable and marketable for a corporation to grasp a sustainable image. The
focus on re-designing the way buildings are designed was in great force and the double-skin
concept was embraced for its thermal buffer and natural ventilation possibilities.
An example of sustainable strategies combined with a completely glazed structure is the 'Green
Building' (1990) research project developed by Future Systems. This design is an early attempt
to reduce dependency on artificial (mechanical) air supply by utilizing natural ventilation. The
egg-shaped form includes adjustable light shelves to redistribute light deep into the office space.
The multi-skin enclosure provided acoustical insulation but was selected primarily for its natural
ventilation and thermal uplift characteristics. The design was never realized but arguably
influenced many later forms such as London City Hall.
Figure 4-15: Green Building (1990), Future Systems.
Source: Compagno (2002, 130); © Future Systems, Jan Kaplicky and Amanda Levete.
150
Increased environmental and climatic awareness did not come without the formation of
leadership, organization and representation. The most significant gathering of world leaders to
discuss sustainable concerns was Earth Summit of 1992 (formally known as the United Nations
Conference on Environment and Development) located in Rio de Janeiro, Brazil (Haase 2005,
763). At this gathering, world leaders pledged to reduce carbon dioxide emissions by avoiding
increased consumption in order to preserve the planet's biodiversity. The bi-products of this
gathering included several documents and legally binding agreements. The Rio Declaration on
Environment and Development
1
consists of 27 principles aimed at future sustainable
development around the world, including Principles 7) State Cooperation to Protect Ecosystem,
8) Reduction of Unsustainable Patterns of Production and Consumption, and 9) Capacity Building
for Sustainable Development. More significantly is the legally binding agreement in the
Framework Convention on Climate Change (UNFCCC/ that encourages a commitment from
each party to provide "calculations of emissions by sources and removals by sinks of greenhouse
gases" (United Nations 1992, 12) in Article 4.2(c). The UNFCCC initiated in 1994 and was later
supplemented by the Kyoto Protocol to create commitments for emission reduction targets.
5
~
"@-
0
Q;
0..
-0
2
E
E
w
3
c
0
_Q
a;
0
0
(/)
l>
c
c
0
I-
c;;
1
:J
c
c
l:
0
c
C1l
0..
C1l
--,
C1l
.!:
..c
0
-0
c
..!!!
·c;;
..c
I-
0
()
-~
~
-~
Ill
C1l
·c;;
l>
c
0
-0
.E
E
C1l
c
Q)
>
(/)
l>
c
·c..
~
..c
[J_
C1l
-~
Ol
z
..c
(/)
l>
al
m
c
C1l
Ill
-~
J5 0
5.. Ol
l> c
Cl:'. 0
.0
E
l> 0
0
Figure 4-16: Annual carbon emissions per capita for the 20 most populated countries.
3
151
One factor nurturing this embryonic interest in high-performance facades in the United States was
Leadership in Energy & Environmental Design (LEED). This organization began forming in 1994
under the leadership of Robert Watson. With the U.S. Green Buildings Council, the organization
developed a performance-based rating system and design standard to evaluate the
environmental impact of buildings. Since its inception in 1998, LEED has grown to encompass
more than 44,998 projects in the United States and another 6700+ in other countries across the
globe (as of April 2013).
4
LEED has evolved from one standard to a suite of six interrelated
standards covering different aspects of development. In the United States, numerous local
governments - and a handful of states - have, or are considering, incentive programs to support
LEED-certified development. Additionally, the USGBC announced the LEED Earth campaign in
2013 where it will refund certification fees for the first LEED-certified project in a country that is
currently without one.
5
Into the late Nineties discussions of escalating environmental concern continued amongst world
leaders. The most significant of the commitments made by governments towards an
environmentally-conscious future was the signing of the Kyoto Protocol in 1997. This agreement
legally binds countries to reduce greenhouse gases by an aggregate 5.2% (Wigginton and Harris
2002, 9). The Kyoto protocol is against global warming and ozone layer depletion, and is
representative of expanding awareness around the globe. The Kyoto Protocol is a valuable
commitment, but its weakness is in permitting emissions trading that has enabled countries to
purchase quotas of emitting harmful gases in an open market. It will be interesting to see how
governments are held to the standards set forth in this agreement. At this juncture, both the
energy efficient design of new buildings and increasing the energy efficiency in existing building
stock became a primary matter of the construction sector. Nearing the third millennium in a more
energy sensitive design and construction market with terms like sustainability and green so
prevalent (but with a broad spectrum of understanding), it was essential to develop a consistent
understanding and clear definition to the A/E/C community.
152
One such term that wrestled with ambiguity is high-performance, in particular with how it relates
to buildings. The U.S. Department of Energy established a 20-year industry plan in its document
High-Performance Commercial Buildings: A Technology Roadmap. The document is a self
proclaimed "ambitious vision" that addresses a "whole-buildings approach", outlines a vision for
the year 2020 and defines four strategies to support the vision's direction (US Department of
Energy, 2). Included in these four strategies is a top priority of "establishing key definitions and
metrics for high-performance commercial buildings". The emphasis of the Roadmap is identifying
current trends, identifying barriers, formulating the vision and identifying steps to progress
towards the vision. In addition to identifying smart, integrated building controls as a mechanism
for progress, the report acknowledges a need for market transformation that must overcome
existing financial barriers. The roadmap intentionally avoids prescribed implementation
approaches and instead points to a joint agreement of strategies between industry and
government while soliciting feedback. The document communicated the state of the challenges
at the time and served as a framework for moving forward; however, many terms and definitions
remained obscure.
In a similar fashion, the U.S. Department of Energy also established a 20-year industry for the
building enclosure in its document Building Envelope: Technology Roadmap. This document also
identified a 2020 vision, including enclosures becoming: energy-positive, adaptable, affordable,
durable, environmental, healthy and comfortable, and intelligent" (US Department of Energy, 2).
The primary technology strategies identified were categorized into materials, systems,
process/design and performance evaluation. In a list of ten potential systems to prioritize
development of were the double envelope and intelligent envelope. Additionally, the results of an
industry survey identified the double envelope as an energy-positive system, low-risk
advancement with few technical unknowns, and within the respondents' research and
development budgets (US Department of Energy, 23).
6
The roadmap was established as a living
text to take the complex long-term challenges and deconstruct them into manageable portions.
153
In an effort to bring focus to high-performance building terminology, the US Energy Independence
and Security Act of 2007 defined a high-performance building as one that "integrates and
optimizes on a life cycle basis all major high-performance attributes, including energy
conservation, environment, safety, security, durability, accessibility, cost-benefit, productivity,
sustainability, functionality, and operation considerations" (H.R. 6--11oth Congress: Energy
Independence and Security Act of 2007). In Sec.401.Definitions, there is a differentiation
between high-performance buildings and high-performance green buildings that is recognition
that high-performance is not unequivocally "green".
Looking forward, and beyond the United States, a technology roadmap for Energy-efficient
Building Envelopes
7
by the International Energy Agency (IEA) is being developed with an
expected release in 2014. The vision for this roadmap is driven by cutting energy-related carbon
dioxide emissions by more than half by 2050 (as compared to 2009 levels) and is part of 22 low
carbon technology roadmaps that are now available or will be in the near future. The
organization aims to provide affordable and clean energy to its 28 member countries, including
the United States. The Agency was initiated in response to the oil crisis of 1973 and initially
addressed forming and refining strategies for managing oil supply disruptions. The Agency has
evolved significantly since its founding and is now at the core of global energy discourse with
focused priorities on energy security, economic development, environmental awareness and
engagement worldwide. It will be interesting to compare the anticipated roadmap for 2050 to the
preceding roadmap to 2020 developed by the U.S. Department of Energy.
Working continuously through a continuous sea change of environmental awareness and
sustainable building, architects, engineers and builders have developed a tremendous diversity of
projects that represent their balancing of performance requirements and project needs. These
solutions range from high-tech to passive, complex to simple and known material application to
exploration of emerging material technologies.
154
Figure 4-17: Eden Project (2001). Source: Schittich (2006, 113). ©Simon Burt I Apex .
A momentous project at the turn of the millennium was the completion of the Eden Project (2001 )
in Cornwall, United Kingdom. The project, designed by architect Nicholas Grimshaw, is a visitor
attraction that includes two large artificial biomes comprised of a lightweight and long-span
enclosure of hexagonal, inflated ethylene tetrafluoroethylene (ETFE) pillows supported by tubular
steel. Within the two biomes are artificial climates and vegetation replicating Tropical and
Mediterranean climates. The building is a synergistic reflection of the heightened environmental
awareness as well as advancements in material technology over the previous century.
Reminiscent of the early botanical houses, the Eden Project represents a departure from the
static, rigid and fragile nature of iron and glass and signifies a direction towards dynamic, flexible
and lightweight solutions. The solution continues to dissolve the amount of structure disrupting
the interior from the exterior environments permitting increased solar gain when compared to the
early botanical structures; addressing the same performance priorities of a century earlier.
155
4.2 Global Emergence of Double-Skin Facades
Double-skin facades started to emerge across Europe, Asia and the United States beginning in
the 1980's. Many early constructions were low-rise applications, shortly followed by high-rise.
These early examples sought to reduce energy consumption and protect otherwise exposed
external shading. In some applications, natural ventilation alleviates the energy required for
mechanical ventilation. The primary objectives of these double-skin solutions are the same
approaches required today. The following case studies provide seminal works in the lineage of
DSFs. One note in particular: the ventilation modes listed for each project are in addition to a
buffer zone mode. Most DSFs can be sealed during cooler months to trap solar heat gain.
Additionally, projects and regions not covered here exist and are the focus of other researchers.
8
4.2.1 Occidental Chemical Center
Completed: 1980
Location: Niagara Falls, NY, USA
Architect: Canon Design
Cavity Partition: Multi-Story
Ventilation Type: Natural
Ventilation Mode: Outdoor air curtain
Cavity Depth: 1.2 m (3.9 ft)
Shading: Louvers, automated with override
Sources: Boake, Harrison & Chatham (2001, 1-5);
Wigginton & Harris (2002, 163-168); Compagno (2003,
245); Poirazis(2006, 205-206); Wright (2005, (4))
Figure 4-18: Occidental Chemical Center- Overall view.
Source: Wigginton & Harris (2002, 165); ©Barbara Elliott Martin I Cannon Design.
156
This building is widely recognized as the first modern double-skin facade. The 18,580 m
2
(200,000 sf) building consists of a nine-story square plan around a central core resulting in
column free office space for approximately 500 employees at its time of completion. The nine
story double-skin facade was adopted to protect solar control devices at the highly wind-exposed
site, which would have made use of exterior solar control devices with a single-skin facade
challenging (Compagno 2003, 245). Additionally, during operations the envelope remains
transparent giving views on out towards Niagara Falls. When the building is completely vacant,
the louvered shading system is shut to provide insulation, retaining internal heat that was
generated during the day to reduce heating cost. The double-skin reduces the impact of severe
external temperatures by minimizing infiltration from the cavity to the interior conditioned space.
During the winter, the cavity acts as a thermal buffer. A mockup done at Arizona State University
found the double-skin to have a U-value of 0.27 Btui°F-ft
2
-h (Wigginton and Harris 2002, 164).
The 1.2 m (3.9 ft) cavity space is encompassed by a blue-green tinted IGU exterior with a visible
transmittance up to 80% and single clear glazing for the interior. Within the cavity depth are
operable white-painted louvers spaced 200 mm (8 in) apart. The louver position varies between
horizontal to 45° angle based on sunlight hitting a single sensor placed at the center of each
elevation. The white louvered surface reflects daylight reaching over half the usable floor space.
The original system relied on the BMS to control the louver position and airflow dampers at the
cavity inlet and exhaust. Occupant override was only provided for the louver position at corner
offices.
The original design was projected to perform at 39,000 Btu/sf, but has actually performed as low
as 32,000 Btu/sf (Wright 2005, (4)). One complaint about the operations is that the louvers are
not utilized as originally intended; they typically remain open at night except on occasional
weekends. The building has been only partially occupied since 2000 and has fallen into dire
states including extended periods of time with failed, inoperable controls.
157
4.2. 2 Super Energy Conservation Building
Completed: 1982
Location: Tokyo, Japan
Architect: Ohbayashi-Gumi
Cavity Partition: Multi-story
Ventilation Type: Hybrid
Ventilation Mode: Air supply,
outdoor air curtain
Cavity Depth: n/a
Shading: Blinds, automated
Sources: Wigginton & Harris (2002,
159-162); Wong (2008, 87-89)
Figure 4-19: Super Energy Conservation Building - Overall view.
Source: V\figginton & Harris (2002, 160); © Ohbayashi Corporation.
This is the first double-skin facade application in Japan. This four-story building was originally
designed to be an exhibit building shCM'casing innovative technologies that can be used in
generating a low-consumption office building for one of the country's largest general contractors,
Ohbayashi Corporation. The building uses an inclined double-skin facade on the south exposure
with automated controls to vary between two seasonal operation modes. During the winter
months, the cavity preheats incoming air before directing it to the air handling units for distribution
within the building. In the summer, air intakes located ICM' in the wall and air-extracts at the top of
the cavity are opened to create natural ventilation to help reduce peak cooling loads. The
building uses night cooling to reduce cooling loads during the follawing day. This approach is
improved by lowering insulating blinds outside the inner facade (within the cavity) during night
cooling to minimize losses.
158
4.2.3 Briarcliff House
Completed: 1984
Location: Farnborough, England
Architect: Arup Associates
Cavity Partition: Multi-Story
Ventilation Type: Natural
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 1.2 m (3.9 ft)
Shading: Blinds, automated
Sources: Compagno (2002, 119);
Poirazis (2006, 191-192)
Figure 4-20: Briarcliff House - Overall view.
Source: Compagno (2002, 119).
This double-skin facade was selected for it acoustic protection from traffic and aircraft and its
ability to give the interior offices solar control. The intermediate cavity space is 1.2 m (3.9 ft)
deep encompassed by an exterior skin with 10 mm (3/8 in) heat-absorbing single glazing and an
inner skin of insulated glass. The cavity space contains sensor-controlled louver blinds at the
interior skin and exposed ventilation ductwork. During the winter months, outdoor air enters at
ground level and is lifted up the cavity space before being recovered by a heat exchanger.
During the summer months, the heat exchanger is bypassed and the air exits the cavity at
louvered extract openings atop the roof return. Poirazis describes the primary facade function as
an active wall or climate wall (2006, 191 ). The cavity space provides access for cleaning and
contains the ventilation ducts.
159
4.2.4 Caixa Geral de Dep6sitos, Av. Da Republica
Completed: 1983-1987
Location: Lisbon, Portugal
Sources: BESTFACADE (2005, 54 and A 11);
Streicher (2007, 1-28)
Figure 4-21: Caixa Geral de Dep6sitos - Overall Street View.
Source: maps.google.com.
This is the first known double-skin facade constructed in Portugal. Most double-skin facades in
Portugal are located in Lisbon. These designs are usually taller than five stories tall. Common
typologies are corridor facades and multistory facades. Primary reasons for double-skin facades
in this region are aesthetics and energy conservation (BESTFACADE 2005, 54). Further
information regarding this facade was not available.
4.2.5 EAL, School of Architecture of Lyon
Completed: 1987
Location: Vaulx en Velin, France
Sources: BESTFACADE (2005, 55 and A 7)
Cavity Partion: Story-Height Corridor
Width: 119 m (390 ft)
Height: 12 m (39 ft)
Figure 4-22: EAL, School of Architecture of Lyon - Overall Street View.
Source: BESTFACADE (2005, A 7).
160
4.3 Europe Takes Leadership
The proliferation of double-skin facades in Europe can partially be credited to regional climatic,
cultural, legislative and technical features. The major development and implementation of the
double-skin facade technology took place in Northern Europe through the 1990s and 2000s, with
numerous completed works of great variety, driven by legislative mandates for improved energy
efficiency in buildings, daylight access and natural ventilation. The focused exploration of double
skin facades has since grown widespread, often applying similar techniques used in Europe, but
in a fundamentally different climate, market and context.
The predominant areas of the double-skin facade development have occurred in the warm
summer continental (Dfb) and maritime temperate (Cfb) climates.
9
The warm summer
continental climate can be found in central Scandinavia and eastern Central Europe and is
characterized by summer high temperatures between 21-28 cc (70-82 cF) and winter lows below
-3 cc (27 cF). This climate type is typified by long cold winters, mild summers and less
precipitation than the Hot Summer subtype. It is not rare for Dfb climates to experience short
stints of extreme heat. Areas in North America with the same classification include the northern
United States and southern Canada, from about 44cN to 50cN latitude. In Europe, the maritime
temperate climates occur in coastal areas of Western Europe, including most of France, Belgium,
The Netherlands and the United Kingdom, and even small patches of Denmark and Norway.
These climates are typified by mild winters (when compared to other climates at similar latitudes),
cool summers, overcast weather and subjected to the polar front year round. Areas in North
America with the Cfb classification are limited but do include the Olympic Peninsula (west of
Seattle, Washington), the Alaska Panhandle and Vancouver, British Columbia. Both the Dfb and
Cfb climates in Europe are heavily influenced by their proximity to water, including the Atlantic
Ocean, North Sea and Baltic Sea.
This section provides a select sampling of European case studies over the span of 30 years.
10
161
4.3.1 Business Promotion Centre
Completed: 1993
Location: Duisburg, Germany
Architect: Foster and Partners w/ Kaiser
Bautechnik
Cavity Partition: Multi-Story
Ventilation Type: Mechanical
Ventilation Mode: Indoor air curtain
Cavity Depth: 0.2 m (8 in)
Shading: Blinds, automated with override
Sources: Wigginton & Harris (2002, 125-128);
Compagno (2002, 120)
Figure 4-23: Business Promotion Center- Overall view.
Source: Wigginton & Harris (2002, 125); ©Dennis Gilbert I VIEW.
This eight-story, 27 m (86 ft), lens-shaped building includes rentable office space and galleries.
The entire perimeter of the building is surrounded with a curved, full-height, multi-story double
skin facade. The exterior layer of glazing follows a 1.5 x 3.3 m (4.9 x 10.8 ft) module and is a
faceted representation of a 46 m ( 151 ft) radius curve comprised of 12 mm (1 /2 in) tempered
single glazing 200 mm (8 in) in front of the inner skin. The outer layer is point-fixed using a
Pilkington Planar™ glass bolt system tied back to aluminum mullions that are hung at roof level
and transfer lateral loads back at each floor slab. The vertical mullions are hung above from a
steel ring beam along the curvature of the building profile. The interior skin is a thermally broken
aluminum glazed IGU comprised of 6 mm (1/4 in) outer lite, 12 mm (1/2 in) argon-filled air space
and an 8 mm (5/16 in) laminated inner lite with a low-e coating. The inner glazing consists of
side-hung windows that are operable only for maintenance (Wigginton and Harris 2002, 127).
162
- I
--! '
_,,
-'----.e=~iif-~ -
,_
-
~~~~~~t 1 1
li~ 11
Figure 4-24: Business Promotion Centre - Cavity drawing (left) and exterior photo (right).
Source (left): Wigginton & Harris (2002, 126); ©Foster+ Partners. Source (right): Compagno
(2002, 120).
Since the building is situated next to a busy roadway on its east elevation, natural ventilation was
deemed not feasible. Instead, air is supplied to the cavity by a displacement ventilation system
and rises as the result of the stack effect. Issues of overheating in the higher floors have been
reported (Compagno 2002, 121). Within the cavity space, 7% perforated aluminum louver blinds
are located closer to the inner skin. The computer-controlled blinds are integrated into a central
BMS that adjusts tilt angle based on heat and light sensors located in each room. Natural
daylighting is experienced throughout the building due to the high levels of transparency, shallow
depth floor plan and integration of the motorized blind controls with the BMS. The occupant has
manual control to override the blind position, but the BMS frequently checks and repositions for
optimal performance.
163
4.3.2 Bibliotheque nationale de France
Completed: 1995
Location: Paris, France
Architect: Dominique Perrault
Architecture
Cavity Partition: Box-Window
Ventilation Type: Mechanical
Ventilation Mode: Indoor air curtain
Cavity Depth: 90 mm (3.5 in)
Shading: Fins (interior wood panels)
Sources: Compagno (2002, 106-108)
Figure 4-25: Bibliotheque nationale de France - Tower view.
Source: www.archdaily.com; ©Yuri Pal min.
Four 20-story, L-shaped towers facing inwards towards one another create a central courtyard on
the site. Each tower primarily consists of floors of book storage but also has support levels for
building services. The enclosure contains story-height box-window double-skin facade modules
that are pressurized with mechanical air. The pressurized air can be pre-heated or pre-cooled to
respond to the exterior temperatures accordingly and mitigates against condensation on the
interior side. Additionally, the humidity can be controlled in a manner that is best for conserving
the books stored within the towers. Both layers follow a 1.8 x 3.6 m (5.9 x 11.8 ft) facade module.
The outer skin is comprised of laminated lites, either of two 6 mm (1 /4 in) or 8 mm (5/16 in)
panes, structurally glazed with silicone to aluminum frames. The interior skin also uses a
laminated lite of two 10 mm (3/8 in) lites - additionally, this layer is fire-resistant. Placed 0.9 m
(3.0 ft) behind the inner glass lites are pivoting wooden panels to mitigate solar radiation.
164
4.3.3 RWE Headquarters
Completed: 1997
Location: Essen, Germany
Architect: lngenhoven Overdiek Kahlen and Partner
Cavity Partition: Box-Window
Ventilation Type: Hybrid
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 500 mm (20 in)
Shading: Blinds, automated with override
Sources: Herzog et al. (2004, 256-257); Compagno (2002, 138-140);
Wood & Salib (2012, 24-31); Schittich et al. (2007, 290-295);
Company Headquarters Tower in Essen (1997, 355-363)
Figure 4-26: RWE Headquarters - Overall view.
Source: Schittich et al. (2007, 290); © Holger Knauf.
This project is marketed as the first contemporary high-rise tower to be naturally ventilated.
Every office has access to at least one operable window. The approximate percentage of the
year that natural ventilation can be used is 75% (Wood and Salib 2012, 24). Essen's mild climate
can largely be attributed to proximity to the coastline. The 31-story cylindrical tower is wrapped in
a double-skin facade made from prefabricated box-window modules. The exterior facade
consists of tempered 'extra clear' glass supported by drilled glass point supports occurring at four
locations along the each window's 3.5 m (11 .5 ft) height. The interior consists of floor-to-ceiling
sliding glass doors made with thermally insulated glazing. The 500 mm (20 in) cavity houses
centrally controlled perforated aluminum venetian blinds, and the interior space has translucent
textile anti-glare roller blinds along the inner-skin. Alternating sliding doors are operable but are
limited to 150 cm (6 in) opening for safety reasons, except for maintenance and cleaning.
165
Figure 4-27: RWE Headquarters - Diagram and photo of box-window with Fish Mouth air
openings. Source: Wood & Salib (2012, 27-28). © lngenhoven Architects .
The 2.0 m (6.6 ft) wide box-window units follow an alternating pattern that staggers air supply
modules and air exhaust modules. These modules are separated from one another by a vertical
glass fin between units and at each floor level. The alternating pattern prevents recontamination
of exhaust air into an adjacent units supply air. The key feature that permits the type and rate of
airflow in each unit while providing acoustic insulation is the "Fish Mouth". There are air emit and
exhaust variations as well as two different opening sizes; one for higher wind zones in the upper
half of the building and one for the bottom. The design of the Fish Mouth and its openings was
determined using CFD and large scale wind-tunnel testing. The hybrid ventilation mode can
switch between mechanical and natural ventilation as needed, on a daily or seasonal basis.
Post-occupancy evaluation was conducted between 1998-2000 measuring room temperatures,
facade temperatures and air exchange rates.
166
4.3.4 Stadttor Di..isseldorf (City Gate)
Completed: 1997
Location: Di.isseldorf, Germany
Architect: Petzinka Pink und Partner
Cavity Partition: Story-Height Juxtaposed Modules
Ventilation Type: Natural
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 0.9 m (3 ft) or 1 .4 m (4.6 ft)
Shading: Blinds, automated with override
Sources: Oesterle et al. (2001, 21-22); Lee et al. (2002, 99);
Wigginton & Harris (2002, 65-70); Bodart & Gratia (2003, 7);
Compagno (2002, 140-142); Herzog et al. (2004, 252-253)
Figure 4-28: Stadttor Di..isseldorf - Overall view.
Source: Compagno (2002, 140); © Gundelfingen.
This 19-story, 70 m (230 ft) rhomboidal building consists of two-office towers offset and bridged at
the top three stories. The entire building exterior is wrapped in glass. The use of a double-skin
envelope on three sides was implemented to reduce traffic noise and wind pressure on the
exposed site. The outer facade consists of 12 mm (1 /2 in) or 15 mm thick tempered single
glazing with a cavity behind it that is either 0.9 m (3 ft) or 1.4 m (4.6 ft). The outer glazing is
continuously captured along top and bottom edges and has an additional glass bolt point-fixing at
the railing. It is also low-iron to permit maximum transparency. 200 mm (8 in) behind the exterior
skin are venetian blinds that are lowered and raised automatically according to light, insolation
levels and the need for nighttime insulation (Wigginton & Harris 2002, 68). The interior skin is
made of 1.5 x 2.85 m (4.9 x 9.4 ft) modules of IGU's with a low-e coating in a wood frame.
Alternate bays pivot about their center to provide access to the cavity and for natural ventilation .
167
Figure 4-29: Stadttor Dusseldorf - DSF elevation (left) and cavity (right).
Source: Compagno (2002, 141); © Gundelfingen (right).
The story-height DSF cavities are divided into 20 m (66 ft) lengths by an escape staircase, the
atrium or corners. The double-skin facade, in conjunction with other design elements, allows for
natural ventilation 60% of the year (Compagno 2002, 142). Natural ventilation into the cavity
space is through staggered air inlet and outlet vents that occur at the outer skin along the floor
levels. Alternating between the inlet and outlet vents is done to avoid short-circuiting effects. The
ventilation flaps can be 0%, 10% or 100% opened (Wigginton & Harris 2002, 67) and are
automatically closed when wind speed exceeds 9 m/s (20 mph). Within the cavity space,
aluminum louver blinds directly behind the outer skin (and in-line with the railing) provide solar
shading. Both the air inlet and exhaust flaps, as well as blinds within the cavity, are controlled by
the BMS. However, occupants do have the ability to override whether the blinds are up or down
and choose between one of three tilted positions (0°, 45° and 90°).
168
4.3.5 Deutsche Post
Completed: 2003
Location: Bonn, Germany
Architect: Murphy/Jahn Architects
Cavity Partition: Multi-story
Ventilation Type: Hybrid
Ventilation Mode: Outdoor air curtain
Cavity Depth: South - 1.7 m (5.6 ft); North - 1.2 m (3.9 ft)
Shading: Blinds, automated with override
Sources: Compagno (2002, 152-153); Oesterle etal. (2001, 160);
Sobek (2003, 127-130); Wood & Salib (2012, 74-83)
Figure 4-30: Post Tower- Overall view.
Source: Wood & Salib (2012, 74); ©Murphy/Jahn Architects.
Bonn's climate has four distinct seasons and is heavily influenced by its proximity to the Rhine
Valley. The 162.5 m (533 ft), 42-story elliptical shaped tower consists of two offset perimeter arcs
of offices spaces that surround a full-building-height central atrium space and prioritized giving
access to outside air (and controlling how much) to occupants. The DSF treatments on the north
and south facades vary in width, but both are multi-story configurations - separated 32 m (105 ft)
vertically every nine floors (11 for the uppermost) - with a laminated glass outer skin , full-height
operable sunshades located closest to the outer skin, and floor-to-ceiling argon-filled IGU's on the
interior. The outer skin is supported by stainless steel members hung from outriggers every nine
floors and laterally braced back to each floor slab. An 8° angle shingle aesthetic on the south
exterior layer introduces operable flaps between lites. These flaps can vary in opening area to
address cavity overheating and to moderate wind pressure along the inner skin.
169
Figure 4-31: Post Tower- Shingled west cavity (left) and flush east (right) facades.
Source: Wood & Salib (2012, 78); ©Murphy/Jahn Architects.
The north facade utilizes a smooth aesthetic on the outer skin with operable flaps occurring in the
plane of glass. The deeper depth and the shingled effect applied to the south facade are
strategically employed to avoid overheating from its direct solar exposure. The inner facades
have centrally controlled motorized operable windows located every other bay that have an
override option. Fan coils and radiant ceilings supplement the space conditioning during the
extreme seasons. The BMS system controls the outer skin flaps, highly reflective perforated
adjustable blinds within the DSF cavity on both elevations, artificial lighting, radiant slabs and the
operable windows on the interior layer. The project set out to achieve aggressive energy
performance benchmarks that were assessed after completion in 2003 (Wood and Salib 2012,
81). The study found the project consuming more energy than the aggressive design targets, but
performing very efficiently compared to conditioned office buildings in that region.
170
4.3.6 30 St. Mary's Axe
Completed: 2004
Location: London, England
Architect: Foster+ Partner
Cavity Partition: Story-Height Juxtaposed Modules
Ventilation Type: Hybrid
Ventilation Mode: Air exhaust, air supply, indoor air curtain
Cavity Depth: 1.0 - 1.4 m (3.3 - 4.6 ft)
Shading: Blinds, automated
Sources: Wood & Salib (2012, 84-91);
(Compagno, Innovative Tower Block Facades 2003, 820-826);
(Gardner 2013, 60-62)
Figure 4-32: 30 St. Mary's Axe - Overall view.
Source: Wood & Salib (2012, 84).
The 180 m (591 ft), 42-story tower is most distinguishable by its bulging cylindrical form that
begins with a 50 m (164 ft) diameter at base, expanding to 57 m (187 ft) approximately 40% of
the way up the building, and then tapers to the top. The form is aerodynamic and reduces wind
pressures on the facade and structure as well as at the pedestrian environment. A load bearing
structure on the perimeter comprised of a diagonal steel diagrid forms a shell of envelope and
structure that result in a column free interior office environment. Six triangular atria spaces spiral
up and around the building at a rate of 5 degrees per level and are partitioned into two or six floor
volumes (similar to the village strategy used in Commerzbank Tower by the same architect). The
combination of cross-ventilation and stack effect created by the atria spaces assists in providing
natural ventilation to the office space approximately 40% of the year. The double-skin facade
occurs between the atria wedges and encloses the office functions .
171
Figure 4-33: 30 St. Mary's Axe - Installation of outer skin (left) and DSF cavity space (right).
Source: (Compagno, Innovative Tower Block Facades 2003, 820-822). ©Nigel Young I Foster+
Partners.
The double-skin facade has an outer layer of clear, low-e IGUs of a 10 mm (3/8") toughened
outer pane, 16 mm (5/8") air space and 10 mm (3/8") laminated inner pane. On the exterior skin,
air intake and extract is provided via narrow slits between glazing units at each floor level. The
inner skin is 1 O mm (3/8") laminated safety glass and contains sliding casements that are only
opened for cleaning purposes. The cavity space is sealed from the atria spaces and ranges from
1. 0 - 1. 4 m (3. 3 - 4. 6 ft) in depth. Air from the office spaces is exhausted through the cavity
space to mitigate solar heat gain. Venetian blinds are incorporated into the cavity near the inner
skin and are controlled by the BMS. The inner skin does not contain operable windows, thus
limiting occupant control of their environment. The building was initially designed to be owner
occupied. Since its inception, it has changed to a multi-tenant occupied building.
172
4.3.7 KfW Westarkade
11
Completed: 2010
Location: Frankfurt, Germany
Architect: Sauerbruch Hutton
Cavity Partition: Story-Height Corridor
Ventilation Type: Hybrid
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 0.7 m (2.3 ft)
Shading: Blinds, automated
Sources: Wood & Salib (2012, 122-131)
Figure 4-34: KfW Westarkade - Overall view.
Source: Wood & Salib (2012, 74).
The 56 m (184 ft), 14-story tower is an addition to an existing office complex in the temperate
Frankfurt climate and houses 700 employees. Preserving views and access to daylight for both
the new and existing facilities in the complex was a top priority. The building's plan is represented
as an airfoil and was developed in response to prevailing winds and the sun path. The perimeter
offices are naturally ventilated through a continuous story-height corridor DSF that is 0.7 m (2.3 ft)
deep. The outer skin follows a jagged sawtooth pattern with glass lites on the broader surface.
The shorter return leg of the sawtooth is a slim glass panel that varies in color around the
building, alternating between an acoustic panel and a side-hung operable that acts as a fresh air
inlet. The continuous story-height cavity around the building acts as a pressure ring and is
always at a higher pressure than the building interior. This positive pressure ring sources natural
ventilation into the office interiors and then exhausts through the building core by stack effect.
173
Figure 4-35: KfW - Sawtooth DSF construction (left) and ventilation diagram (right).
Sources: Wood & Salib (2012, 126-128). ©Jan Bitter (photo);© Sauerbruch Hutton (diagram).
The interior skin of the DSF is made up of argon-filled IGU's with a low-e coating. The inner
surface also alternates between fixed lites and operable lites that open inwards to allow natural
ventilation and are occupant controlled. The vertical operable panels on the exterior are
controlled by the BMS and used to create constant regulated flow in the cavity and for ventilation,
mitigating negative effects created by high wind speeds. Additionally, the broad glass lites at
strategic locations around the plan (i.e., the tip or point) are also operable and controlled to open
to prevent overheating in hot summer conditions. The design and climate allow natural ventilated
approximately 60% annually (Wood and Sa lib 2012, 122). Solar control is provided by venetian
blinds near the inner skin that redirect light. The story-height partitioning provides fire protection
with steel sheeting at each slab. Additionally, smoke protective curtains in the cavity aid in
creating three partitioned fire-protection zones (Wood and Sa lib 2012, 125).
174
4.3.8 Roche Diagnostics International AG
Completed: 2011
Location: Rotkreuz, Switzerland
Architect: Burckhardt + Partner
Cavity Partition: Box-Window
Ventilation Type: Mechanical
Ventilation Mode: Buffer (pressurized)
Cavity Depth: 190 mm (7.5 in)
Shading: Louver blinds, automated
Sources: Hell & Kaltenbach (2011, 404-412);
De Bleecker, et al. (2012, 310-318)
Figure 4-36: Roche Diagnostics International AG - Overall view.
Source: W Hell & Kaltenbach (2011, 404); ©Frank Kaltenbach.
The 16-story, 68 m (223 ft) tower brings together administrative departments that were disparate
prior. Its structure includes a central core, perimeter diagrid and reinforced concrete slabs. The
three work together to form the structure's stiffness to resist loads, and they create an
uninterrupted workspace. The transparent facade provides expansive views of Lake Zug and the
Alps using a maintenance free, sealed, closed-cavity facade (CCF) developed by Josef Gartner
GmbH. This DSF box-window uses entirely separate and sealed units that are mechanically
pressurized, protect shading elements and the cavity, and do not need to have the cavity
cleaned. Maximum transparency, high thermal insulation and sound protection, reduced cleaning
costs, maintenance free solar shading devices and ease of installation are all significant positive
impacts of the CCF (De Bleecker, et al. 2012, 317). This technology, along with decentralized
facade ventilation units, form an efficient comfort and performance strategy for the building that
operates in a "free cooling" mode most of the year (Hell and Kaltenbach 2011, 405).
175
Figure 4-37: Roche Diagnostics International AG - Corner elevation (left) and ventilation
diagram (right). Source: W Hell & Kaltenbach (2011, 410-411 ); © Daniel Spehr (left).
The typical floors use 1350 x 3780 mm (4.4 x 12.4 ft) facade units ....nth an outer lite of 12 mm (112
in) laminated float glass, intermediate cavity of 190 mm (7 .5 in) dehumidified air, and an inner
triple-glazed lite of5 mm (3/16 in) float glass, 16 mm (5/8 in) air space, 5 mm (3/16 in) float glass,
16 mm (5/8 in) air space and 8 mm (5/16 in) laminated innermost pane. All glass panes are low
iron an do not use coatings of any kind. The sole source of solar control is by electronically
controlled aluminum louvers with 10% perforation located ....nth in the CCF cavity. To prevent
condensation on any of the cavity facing glass surfaces, a steady flow of clean, dehumidified, dry
air is mechanically pumped into each cavity space at a bottom inlet tethered to an air supply line
beneath the raised office floor. Natural ventilation into the office space is provided, but not
integrated into the CCF double-skin box-window. Fresh air enters 200 x 50 mm (7.9 x 2 in) inlet
vents that align ....nth the face of glass located at each floor. The ventilation flaps are then
connected to an air duct located beneath an aluminum grating around each room's perimeter.
176
4.3.9 London Bridge Tower -The Shard
Completed: 2012
Location: London, England
Architect: Renzo Piano Building Workshop
Cavity Partition: Story-Height Corridor
Ventilation Type: Natural
Ventilation Mode: Outdoor air curtain, air supply
Cavity Depth: 20 cm (7.9 in)
Shading: Blinds, automated with override
Sources: (Moazami and Slade 2013); Masera (2012, 46-59);
Bucci (2012, 66-99)
Figure 4-38: The Shard - Overall view.
Source: author photo;© Jeffrey Vaglio.
The 72-story, 310 m (1017 ft) tall mixed-use tower is the tallest in London's skyline (and Europe),
sitting just south of the River Thames in the London Bridge Quarter. The building functions
include offices, retail, restaurants, hotels, dwellings and observatories. The overall mass consists
of eight shard planes leaning in 6° against one another. Each of these shards utilizes a ventilated
DSF used to reduce heat gain, increase comfort levels around the floor plan's perimeter and
permit access to natural daylight and views. The typical facade units are 1.5 m (4.9 ft) wide by
3.8 m (12.5 ft) tall and the project contains over 10,000 of them (Masera 2012, 53). The outer skin
consists of 13 mm (1 /2 in) laminated single glazing with a colorless solar-control coating (24%
reflective coating) while the inner skin has 34 mm (1 5/16 in) IGUs with low-e coatings and an
argon-filled air space. Both skins utilize low-iron glass to maximize transparency. The inner skin
has side-hung operable IGUs to provide access to the 20 cm (7.9 in) wide cavities.
177
Figure 4-39: The Shard - DSF elevation at corner (left) and intermediate cavity (right).
Source (left): detail-online.com>; ©Rob Telford (accessed March 21, 2013).
Source (right): redchalksketch.wordpress.com>; ©Bart Akkerhuis (accessed August 23, 2012).
The initial design called for a mechanically ventilated cavity; however, a passive strategy was
implemented in the end due to changing building regulations over the course of the design (Bucci
2012, 99). At the meeting rooms, natural ventilation is provided via glass doors on the inner skin.
A significant priority of implementing the DSF solutions was to protect shading devices that would
otherwise be exposed to high wind loads at higher building elevations. The primary solar control
comes from a motorized roller blind of woven glass-fiber located in the intermediate cavity that is
semi-transparent when extended. The red roller blinds are controlled by an automated BMS that
lowers them when the solar radiation reaches 200 W/m
2
(Masera 2012, 53) and are capable of
reducing solar radiation by 85% (Bucci 2012, 99). The two facades are supported by separate
aluminum mullion frame systems tied together by a series of narrow bridge bars of cast aluminum
per unit. The frames were preassembled into a single unit prior to arrival and installation on site.
178
4.4 Modern Proliferation Outside of Europe
The number of applications outside of the European continent began to grow around the turn of
the century with new constructs in Australia, Canada and the United States. This initial
adaptation of DSF technology elsewhere was further reinforced soon after by exploration in
Japan and China. Some of these projects were designed by European-based architects and
engineers with familiarity with DSFs and implemented in very similar climatic pockets around the
globe. For instance, the Telus William Farrell Building (see Section 4.4.2) in Vancouver, British
Columbia is a maritime temperate (Cfb), similar to London, Amsterdam and Western Germany.
Many of the other projects, however, are of the humid subtropical Koppen classification (Cfa),
including Sydney, Australia (see Sections 4.41 and 4.4.5); Sendai, Japan (see Section 4.4.3);
Shanghai and Guangzhou, China (see Sections 4.4.6, A.30 and A.37). The humid subtropical
(Cfa) climate is temperate and usually occurs on the eastern coasts of continents. The case
study provided in Section 4.4.4, Manitoba Hydro Place, is located in Winnipeg, Canada, a humid
continental climate (Dfb) with more extremes between winter and summer conditions with short
intermediate seasons. The biggest climatic deviation in applications outside of Europe is in the
deserts of the Middle East Though implementation in this region lagged behind North America,
Australia and East Asia, there is a stock of projects recently completed or nearing completion that
utilize double-skin facade systems to buffer between conditioned interior spaces and the hot
exterior desert climate. Two projects in Abu Dhabi, United Arab Emirates, are presented (see
Sections A.38 and A.40) - one nearing completion and the other a competition design.
12
The
Abu Dhabi's Koppen climate classification is hot desert climate (BWh), characterized by blue
skies, extreme heat and humidity during summer months and hot weather even in cooler
seasons. These examples, along with some of the examples in Section 5.3 Double-Skin Facade
Applications in the United States, show some sensible transfer of technology to similar climate as
well as some fundamentally dissimilar applications with respect to climate.
179
4.4.1 Aurora Place
G:'ompleted; 2000
Location: Sydney. Australia
Architect Renzo Piano Building· Workshop
Cavity Partitian :story-Height Corridor with Juxtaposed Modules
and Louvered Outer Skin
Ventil-atia n Type Natural
Ventilatinn Mode: Air supply. outdoor air curtain
Cavity Depth: > 3 m (> 10 It)
Shading Blinds. automated with override
.~ournes Compagno (2002, 148-149);
figure 440: Aurora Place- Overall view.
Source: Compagno (20o:2, 148-14.9).
The 44-story, 218 m (715 ft) ler1s ,,shaped tower of offices and residences is oriented with its
primary axis running north to south, resulting i!'I offices along the curved east and West elevations.
East and west facades continue past the north · and south facades to create large fins of.glazing
that are exposed on both sides, ¢reate a weather shield and contribute to the aerodynamic Wind
flow patterns by avoiding sharp comer conditions, The east and west facades consist of 1.35 x
2,4 m (4.4 x 7 .9 ft) IGUs of loW-iron glass With a low-e coating and a continuous edge covered in
a white:-fritted dot pattern The use of low-iron glass for both lites of the IGU allowed the quality of
ligt1t i!'I the office spaces to be natural in color Along tt1e narrow north and south elevations, the
flyby fins of the east and west facades create a wind-protected, less turbulent airflow zorne that
Was taken advantage of by implementing a double-slin facade: With an exterior layer of glass
louvers to provide natural ventilation and views, including the harbor and Sydney Opera House.
180
Figure 4-41: Aurora Place - Louvered outer skin of winter garden DSF.
Source: Compagno (2002, 149); ©Michel Denance (left) and Chris Kelly (right) I Renzo Piano
Building Workshop.
The outer skin consists of six 12 mm (1 /2 in) laminated and tempered low-iron glass louvers per
story. Each louver is 1.20 x 0.46 m (3.9 x 1.5 ft) and fixes on both sides to stainless steel
brackets that attach to vertical laminated glass mullions. The top four louvers can rotate open
45° in the commercial zone (90° in the residential zones) while the bottom two remain fixed to
function as a parapet. There is always natural ventilation to the corridor cavities due to the open
gap between louvers, even in the closed position. The cavity space is deep enough to be
occupied and also houses solar control systems: fabric blinds for the north and south and
additional horizontal metal sunscreens for the solar-exposed north. The inner facade consists of
insulated glazing and includes sliding doors for cavity access and bottom-hung operable windows
to provide occupants access to natural ventilation.
181
4.4.2 Telus William Farrell Building
Completed: 2000 (Renovation); 1947 (Original
Construction)
Location: Vancouver, Canada
Architect: Busby + Associates
Cavity Partition: Multi-Story
Ventilation Type: Outdoor air curtain, air supply
Ventilation Mode: Natural
Cavity Depth: 0.9 m (3.0 ft)
Shading: Screen (frit glass) and overhangs
(cavity daylight shelves), fixed
Sources: Boake, Harrison and Chatham (2001,
6-12)
Figure 442: Telus - Overall view.
Source: author photo;© Jeffrey Vaglio.
The eight-story office building is a renovation of an existing concrete structure located in busy
downtown Vancouver, a temperate climate. On the south and west elevations, a multi-story DSF
was implemented to create a thermal and acoustic buffer while permitting preservation of the
existing facade. The multi-story was different from the systems being most commonly applied in
Europe at that time in that it was not compartmentalized and ran the full-height of the building
(Boake, Harrison and Chatham 2001, 11). The new skin is comprised of differentially glazed
aluminum curtain wall with insulated glass units and operable windows. The 0.9 m (3.0 ft) offset
from the single-glazed punched windows within the existing concrete wall creates an intermediate
cavity space that can be accessed for cleaning. Natural ventilation is achieved by motorized
windows on the new exterior skin as well as operable units on the existing inner skin that provide
occupants with access to natural ventilation.
182
mm
Figure 4-43: Telus - West elevation (left) and DSF airflow diagram (right). Source (left):
author photo;© Jeffrey Vaglio. Source (right): Boake, Harrison and Chatham (2001, 8).
Cavity airflow is regulated by inlet louvers at the base and exhaust dampers at the top. Additional
air inlets are provided by mechanized operable exterior windows at each level across the exterior
facade height. During the winter, the DSF acts as a thermal buffer, trapping solar heat gain.
During summer months, the DSF is exhausted and uses the stack effect (and the assistance of
fans) to exhaust heat and prevent overheating. The first level of solar shading is provided by the
frit treatments on the exterior glazing in spandrel zones. On the exterior skin, between spandrel
zones, clear vision glass is used in alignment with the height of the interior windows. Within the
cavity space, daylight reflectors are used to allow a deeper penetration of light into the building
plan. The use of a multi-story established a precursor for other North American commercial
application that is believed to be less expensive than compartmentalized DSF systems (Boake,
Harrison and Chatham 2001, 13).
183
4.4.3 Sendai Mediatheque
Completed: 2001
Location: Sendai, Japan
Architect: Toyo Ito & Associates
Cavity Partition: Multi-Story
Ventilation Type: Natural
Ventilation Mode: Outdoor air curtain
Cavity Depth: 1.0 m (3.3 ft)
Shading: Blinds, retractable
Sources: Media Centre in Sendai (2001, 1263-1277); Herzog
et al. (2004, 202-203);
Figure 4-44: Sendai Mediatheque- Overall view.
Source: Media Centre in Sendai (2001, 1263) ; © H iro Sakaguchi I A to Z .
The six-story Mediatheque is a transparent volume that houses a mix of programmatic functions ,
including library, galleries, multimedia library, seminar rooms and cafe. The varying height floor
levels are supported by thirteen distinct tree-like load-bearing columns, some of which house
stairs and elevators. Around the perimeter, each elevation has a unique facade design. To the
south, a multi-story double-skin is used as a thermal and acoustical buffer from its surrounding
environment. The highly transparent double-skin overlooks a primary roadway and preserves a
visible connection between inside and outside with its views and openness. During winter
months, the double-skin facade cavity is sealed and acts as a thermal buffer. When solar heat
gain builds up in the summer months, a glazed ventilation flap at the roof parapet level is opened
to exhaust the cavity. The outer skin has 19 mm (3/4 in) laminated glass that is corner clamped
and the inner has 10 mm (3/8 in) toughened glass that is continuously supported on four-sides.
184
.. A "
: l~ ·~
an:::--
I
11
t ~
11
~ .. · · ciniJJc
- p -
Figure 4-45: Sendai Mediatheque - DSF cavity drawing and constructed structure.
Source: Media Centre in Sendai (2001 , 1274-1275); © Hiro Sakaguchi I A to Z.
A unique feature of the double-skin facade is a shared structural support for both skins. A
vertically spanning 19 mm (3/4 in) laminated glass fin is located closer to the inner skin, providing
continuous support while stainless steel corner clamps tie back to the fin at intermediate points to
support the outer skin. The stainless steel glass fixings transfer dead load back to steel fin
outriggers at floor levels via a 14 mm (9/16 in) diameter stainless steel rod. Since the cavity
space is continuous and spans multiple floors, all steel members in the cavity space are treated
with intumescent paint. Screen printed frit patterns are applied to the outer glazing to disguise the
floor level behind, essentially acting as a spandrel panel and mitigating some solar heat gain.
Additional daylight control is provided to occupants in the form of adjustable anti-glare roller
blinds that are located on the interior side of the inner skin.
185
4.4.4 Manitoba Hydro Place
13
Completed: 2008
Location: Winnipeg, Canada
Architect: Kuwabara Payne McKenna Blumberg Architects (design
architect); Smith Carter Architects & Engineers (associate architect)
Cavity Partition: Story-Height Corridor
Ventilation Type: Hybrid (mixed-mode)
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 1.3 m (4.3 ft)
Shading: Blinds, automated with override
Sources: Wood & Salib (2012, 112-121).
Figure 4-46: Manitoba Hydro Place - Overall view.
Source: Wood & Salib (2012, 112).
The 115 m (377 ft), 22-story office tower is located in the extreme climate of Winnipeg that
encounters harsh winters and hot summers. The primary building form includes two 18-story
column-free office towers bridged by a service core and winter garden wedge between, forming
an "A" in plan. The shape and orientation of the tower is a response to the prevailing wind
(predominantly from the south) and solar conditions. The use of cross-ventilation and the stack
effect - drawing the air up the core towards a prominent solar chimney to the north - combine to
allow natural ventilation approximately 35% of the year. A primary source of fresh air is the sky
gardens. Three six-story sky gardens, located along the south, form the atria I winter gardens
that precondition the natural air before it enters the main office environment. These sky gardens
act as the building's lungs as well as informal breakout and social spaces for occupants.
Additional ventilation is provided by a corridor DSF along the northeast and west elevations.
186
Figure 4-47: Manitoba Hydro Place - Northwest elevation and operable units on DSF.
Source courtesy KPlvlB. ©Eduard Hueber (left);© Tom Arban (right).
The outer skin is low-iron IGU's followed by a 1.3 m (4 3 ft) air cavity with motorized blinds and a
single-glazed interior curtain wall. The outer skin has a field of motorized glazed lites that open
and close to supply air and avoid condensation respectively On the interior facade, occupants
have the option to open operable windows near their workspace These operable windows are
not tied to the BMS and are solely controlled by manual means. The central BMS controls
operable windows, vents, solar shading, thermal comfort, ventilation rates and radiant ceilings
(Wood and Sa lib 2012, 118) Louver shades are controlled by a central BMS but have means of
override via a custom computer interface. The use of the story-height corridor uouble-skin
facades along the broad elevations of the project is one. facet of a larger, more holistic natural
ventilation strategy that uses the sky garden atria as an air source and the 115 m (377 ft) solar
chimney as an air exhaust by way of the stack effect
187
4.4.5 1 Bligh Street
14
Completed: 2011
Location: Sydney, Australia
Architect: lngenhoven Architects
Cavity Partition: Corridor (continuous around entire perimeter)
Ventilation Type: Hybrid (mixed-mode)
Ventilation Mode: Outdoor air-curtain
Cavity Depth: 0.6 m (2 ft)
Shading: blinds, automated
Sources: Wood & Salib (2012, 132-137)
Figure 4-48: 1 Bligh Street- Overall view.
Source: Wood & Salib (2012, 132).
The 139 m (456 ft), 30-story tower is located in the temperate climate of Sydney, provides access
to daylight and views, but does not rely on natural ventilation in the office spaces. The elliptical
form is skewed to the main city grid to maximize harbor views to the north from the office floors,
additionally enhanced by locating the circulation core tower to the south. Daylight is provided on
two sides of the office plates with a central atrium running the full-height of the tower, daylit from
above via a skylight. The naturally ventilated atrium also houses glass-encased elevators for
circulation. The office space is mechanically sealed from the naturally ventilated atrium space as
well as the exterior environment. The building is enveloped by a corridor DSF that has a
laminated low-iron exterior lite, a 0.6 m (2 ft) air cavity and a low-e IGU interior unit. The inner
skin does not contain operable windows and does not permit natural ventilation. The purpose of
the DSF is to protect the solar control blinds within the cavity. Additionally, there are thermal and
acoustical benefits to implementing the corridor with outdoor air-curtain configuration.
188
t
t
Figure 449: 1 Bligh Street - View through DSF (left) and intake/extract louvers (right).
Source: Wood & Salib (2012, 136-137). © DEXUS Property Group (left);© ingenhoven architects.
The air intake and exhaust on the exterior layer is permitted by a series of fixed louvers that have
an airfoil profile that was designed using CFO and wind-tunnel tests. These tests evaluated the
wind speeds, and cavity air speed, to ensure 1) the blinds would not be disrupted and 2) there
would not be recontamination of exhaust air into the supply air in the above unit. Since the intake
and exhaust are fixed louvers, the BMS system is unable to regulate the opening amount, airflow
rates, or seal the cavity; it is a passive airflow strategy. The blinds, on the other hand, respond to
on-site measurements via sensors and sun-tracking software to mitigate heat gain and glare.
The use of blinds within the cavity also allows the glass to be low-iron, preserving a high purity of
view and high visible light transmission. Natural ventilation is believed to be possible for
approximately 35-40% of the year if the interior skin was retrofitted with operable windows (Wood
and Sa lib 2012, 136). The missed opportunity to harness natural ventilation in such a temperate
climate is unfortunate.
189
4.4.6 Shanghai Tower
Completed: 2014
Location: Shanghai, China
Architect: Gensler
Cavity Partition: Multi-story
Ventilation Type: Mechanical
Ventilation Mode: Indoor air curtain, air exhaust
Cavity Depth: Varies, but> 5 m (16 ft) in many zones
Shading: Fins, fixed
Sources: Gensler (2010, 1-15); Zeljic (2010, 169-187)
Figure 4-50: Shanghai Tower- Overall view.
Source: Gensler (2010, 3); ©Gensler.
The 632 m (2073 ft), 121-story, twisting tower will be the second tallest tower in the world when
complete in 2014 and was designed with the notion of a "vertical city" in mind. The unique
rotating form was optimized by the design team through a series of wind tunnel tests to reduce
the building's wind loads by 24% resulting in a refined, lighter structure with a 32% reduction in
material (Gensler 2010, 6). The tower is subdivided into nine vertical programmatic zones that
vary from 12 to 15 stories high. These zones represent the height of the multi-story DSF atrium
spaces between the independent outer and inner curtain walls. The depth of the atriums varies
and is often greater than the dimensional limits of 2 m (6.5 ft) outlined in Section 3.1.3. The outer
skin is cam-shaped in plan and utilizes a staggered aesthetic in part to prevent harsh reflections
to the surrounding environment. The inner curtain wall is circular in plan. Space between creates
public garden spaces that improve air quality, preserve views and provide an insulating buffer.
190
Figure 4-51: Shanghai Tower-Atrium DSF rendering (left) and construction photo (right).
Source:© Gensler.
The outer skin is supported by a series of 356 mm (14 in) diameter steel pipe rings at the
horizontal stack joints. This steel rings transfer lateral loads back to the primary interior structure
via a series of 219 mm (8.6 in) diameter steel struts that span the full atrium depth. Additionally,
the steel rings ascend at a set incline to create a spiraling effect around the outer skin. The dead
load of the outer skin is transferred by tension rod elements at the perimeter steel ring that carry
the gravity load up to double-belt trusses. The double-belt trusses separate and support each
vertical neighborhood. The outer skin utilizes a 26 mm (1 in) laminated glass fin integral to the
vertical mullions while the inner skin uses a 4-sided aluminum frame. Both the inner and outer
curtain walls have spectrally selective low-coatings. On the outer skin, solar control is provided
by the use of fritted glass as well as metal panel shelves that serve as fixed horizontal fins. There
is a BMS that monitors and adjust systems including lighting, HVAC and power-generation.
191
4.5 Typological Trends in Global Applications of Double-Skin Facades
Double-skin facade technology evolved considerably during the mid-1990's and through the first
decade of the 2000's. The initial concentration of double-skin facades were implemented in
Northern European climates and primarily on office buildings. The primary motivations for
implementing double-skin facades in Europe often included, but were not limited to improved
thermal insulation during both hot and cold seasons, reduced solar gains, providing natural light,
protecting shading devices, extended use of natural ventilation, and adaptability to assist a
reduction in energy consumption. Many of these projects failed to meet all the performance goals
established, often incurring overheating of the cavity, reduced or obstructed daylight, and
disruptive sound transmission through the cavity. The primary design objectives of these double
skin solutions are the same approaches required today; however, the lessons learned from the
technology's pioneers should be understood and inform applications in other cultures and
climates.
An extensive evaluation of built projects was included in the International Energy Agency's,
Annex 44: Integrating Environmentally Responsive Elements in Buildings, State of the Art Review
Vol. 2A: Responsive Building Elements (Perino 2007). This summary collected advanced
integrated facade (AIF) project data on a sample of 215 buildings from 20 different countries. The
authors of this report acknowledged the challenges in classifying advanced integrated facades
stemming from the numerous aspects to be considered, but ultimately used the most common
criteria: ventilation type, ventilation flow path and system configuration (cavity partitioning). The
location, by country, and the building's programmatic application were also considered.
This section of the dissertation relies heavily on the State of the Art Review Vol. 2A: Responsive
Building Elements (Perino 2007) to describe the global evolution of advanced integrated facades.
A similar evaluation is performed in Section 5.4 Emerging Trends of Double-Skin Facades in the
United States, which reviews a database of 26 built DSF applications in the United States.
192
4.5.1 Construction Evolution
The rampant diffusion of double-skin facade systems in Europe truly took off in the early 1990's.
However, earlier investigations date back to the 1960's. The State of the Art Review Vol. 2A:
Responsive Building Elements' collection of 215 projects accounts for one double-skin facade
designed in the 1960's. This is the Natural History Faculty Library at Cambridge University 1964-
1968 designed by James Stirling (Compagno 2003, 245), which some researchers identify as the
first double-skin facade. Furthermore, the State of the Art Review Vol. 2A: Responsive Building
Elements recognizes eight DSF/AIF projects completed during the 1980's. This sample most
likely includes the Occidental Chemical Center (United States, 1980), Super Energy Conservation
(Japan, 1982), Briarcliff House (United Kingdom, 1984), Caixa Geral de Dep6sitos - Av. Da
RepOblica (Portugal, 1987), and EAL- School of Architecture of Lyon (France, 1987). Detailed
information about these select projects is presented herein, Section 4.2.
4.5.2 Programmatic Application
The principle building type that first received double-skin facades was of the office type. Of the
215 projects included in the evaluation, fewer than 20 buildings (9.3%) were not office programs.
other programs included airports, malls, schools, courts, libraries, hotel, hospital and research
buildings (Perino and Serra 2006).
4.5.3 Geographic Distribution
A prevailing portion of the DSF/AIF applications is concentrated in the northern European
countries with 60.9% of the projects located in Germany, Switzerland, Finland, Belgium, Poland,
Netherlands, Austria, Netherlands, Denmark or Luxembourg. The country responsible for the
most DSF/AIF projects was Germany (49 or 22.8%), followed by Japan (28 or 13.0%),
Switzerland (26 or 12.1 %), Finland (16 or 7.4%), France (16 or 7.4%) and Belgium (15 or 7.0%).
The United States had seven projects included in this survey of built DSF/AIF. A complete
interpretation of the geographic distribution and country breakdown is provided in Appendix B.
193
1geo
These eight projects occurred over the
course of the 1980s. Specific years of
completion were not available.
1
1g70
1964-1968
Natural History
Faculty Library at
Cambridge University
designed by James
Stirling (Compagno
2003).
Figure 4-52: DSF/AIF - Year of construction or retrofit.
8
1980
1980
Occidental Chemical Center
(aka Hooker Building)
designed by Cannon Design
for Niagara Falls, New York.
18
1990 2000
Information for this figure based on IEA-ECBCS Annex 44, State of the Art Review Vol. 2A: Responsive Building Elements, (Perino 2007).
CD
C.11
0
1-4
5-9
-
10-14
""
-
15-19
-
-
20-24
25-29
-
30-34
-
35-39
-
40+
Figure 4-53: DSFIAIF - Geographic distribution map.
Information for this figure based on I EA-ECBCS Annex 44, State of the Art Review Vol. 2A: Responsive Building Elements, (Perino 2007).
4.5.4 Typological Distribution
In the State of the Art Review Vol. 2A: Responsive Building Elements, the authors disclose that it
was not possible to obtain information concerning the year of construction, ventilation type,
ventilation flow path and system configuration for all 215 projects. In some cases, no information
was acquired, and in many cases at least one of the typologies is absent (Perino 2007). The
percentages are a decent indicator of the global typological trends.
Ventilation Type
Information concerning each DSF/AIF's ventilation approach was gathered for 191 projects. The
most common ventilation type is natural ventilation (58.1 %), followed by mechanically ventilated
(32.4%). The least common solution is hybrid ventilation (9.4%) per Perino (2007, 23-24).
Ventilation Flow Path
It is unclear how many projects had Information about ventilation airflow path available since the
data includes a significant number of DSF/AIF projects with multiple flow paths. The most
common ventilation flow path strategies include the outdoor air curtain (49.1 %), indoor air curtain
(17.3%), supply air (15.6%) and buffer (15.6%) per Perino (2007, 24).
Cavity Partitioning
Information classifying the system configuration, or cavity partitioning, was available for all 215
projects. The most common configuration is the multi-story (47.0%) double-skin facade. Box
window (14.0%) and corridor (14.0%) configurations accounted for the same amount of projects.
The shaft-box (6.0%) was the least implemented double-skin facade configuration (Perino 2007,
24).
196
4.5.5 Global Lessons
The evolution of double-skin facades, heavily concentrated in northern Europe and diffused
globally, can be partially attributed to the European and global efforts to reduce energy
consumption. The early adopters of ventilated facades vowed it to be a sustainable technology
that would reduce energy consumption while improving thermal comfort for the building's
occupants. Similar to rnany of the criticisms which have surfaced in the United States, these
design goals were often not estimated and validated in a comprehensible rnanner. This led to a
nurnber of built projects with double-skin facades that did not perform to design goals and owner
expectations, leading rnany designers, engineers and clients to be skeptical of the technology as
a sustainable solution. Following an initial surge in Europe, there has been a retraction of double
skin facade applications with a core justification of improved energy performance. Instead, the
emphasis on protection of shading devices, enhanced acoustical insulation and access to natural
ventilation are the primary drivers. In reviewing the forty case studies presented in this Chapter,
several noteworthy patterns seern to reoccur or progress over tirne, including: automated controls
of solar control with occupant override, increasing permeability, the quest to rninirnize
maintenance and the pursuit of super-tall building structures with double-skin facades to protect
shading equipment. Each of these is described further in the following pages.
Occupant Override of Solar Controls
Many of the built DSFs rely primarily on an automated control strategy for deploying, retracting,
tilting and adjusting solar control strategies such as louver or roller blinds, usually located in the
cavity space. Sarne systems even provide an additional layer of glare control in the interior space
to supplement the BMS-controlled system A cornrnon solution with automated systems is to
provide an occupant override strategy that rnay include a panel control switch within the
respective office space to respond to individual needs. A design consideration then becomes the
length of tirne the occupant override remains before reverting back to BMS control.
197
Increasing Permeability
The DSFs that have a box-window cavity partitioning have a regular frequency of openings
across the face of an elevation, usually every floor. In DSFs that have a multi-story geometry, the
frequency of openings can vary substantially. The early examples of DSFs - Occidental
Chemical Center (Section 4.2.1), Super Energy Conservation Building (Section 4.2.2) and
Briarcliff House (Section 4.2.3) - possess a sealed outer skin facade with airflow penetration
solely occurring at the base of the system. This approach increases the risk for overheating
across the height of the cavity and also enhances pressure differential effects that drive the stack
effect. As more and more applications came to light, there was an increase in the diversity of
solutions to airflow permeability in the outer skin, including shingled, louvered and vertical
ventilation strategies. The Deutsch Post Tower (Section 4.3.5) introduced air at every lite/story
across the height of one of its facades by sloped glass planes to create a shingled effect, and
with ventilation flaps occurring horizontally in the offset dimension at the base of each module.
The louvered outer skin effect was embraced at Aurora Place in Sydney (Section 4.4.1 ), de Bis
Headquarters in Berlin (Section A.12) and Kraanspoor in Amsterdam (Section A.31).
Another interesting observation from the case studies is the use of a ventilation inlet that is
vertical in orientation and not normal to the impinging wind pressures. This is the case on the
Administration Building Kronberg (Section A 17), Munich Re Offices (Section A.22), AZ
Medienhaus (Section A.22) and KIW Westarkade (Section 4.3.7). Of these examples, two have
controllable openings and two have fixed gaps for inlets. The controllable solutions occur at the
Building in Kronberg with an outer lite that hinges about one vertical edge and the KIW
Westarkade that has slim ventilation doors on the short side of a sawtooth profile. The fixed
solutions occur in the static outer lite of the Munich Re Offices and the "cranked"-on-one-side
outer lite on the AZ Medienhaus. With regular frequency of openings, there is greater control
over the pressure within the cavity space and exhaust strategies to reduce overheating.
198
Figure 4-54: Facade refurbishment in Stuttgart by Behnisch Sabatke Behnisch (1996).
Source: Herzog, Krippner and Lang (2004, 248); © Ralf Richter.
199
Maintenance Free DSF Cavity Solutions
One strategy that was employed early in France and has seen recent application elsewhere is the
pressurized, sealed box-window strategy. These systems achieve a mechanically pressurized
cavity space by inputting a constant dry airflow into the cavity to prevent dust, dirt and
condensation from surfacing. Dominique Perrault employed this approach on the Bibliotheque
nationale de France (Section 4.3.2) in 1995, followed by Jean-Paul Viguier at Coeur Defense
(Section A.21) in 2001, both in Paris. The system has encountered a recent surge from facade
engineers at Permasteelisa I Josef Garner GmbH I Scheldebouw in what they term the closed
cavity facade. One recent project utilizing a maintenance free closed cavity facade is the Roche
Diagnostics International AG tower (Section 4.3.8) in 2011. The advantages over other, more
common, DSF systems are that the pressurized, closed cavity facade is simple, efficient and
reduces maintenance demands. A similar pressurized box-window approach will be implemented
on the forthcoming New Stanford Hospital project (Appendix C.36).
Supertall Buildings
15
with DSF
Though Commerzbank Tower at 259 m (850 ft) in 1997 (Section A.9) and Deutsche Post Tower
at 162.5 m (533 ft) in 2002 (Section 4.3.5) are examples of skyscrapers with DSFs, they pale in
comparison to the super heights of recently completed and soon to be completed projects. The
recent completion of The Shard in London at 310 m (1017 ft) in 2012 (Section 4.3.9) marked the
completion of one of the largest DSF-clad projects to date as well as becoming the tallest building
in Western Europe and a new icon on the south bank of the River Thames. The Shard is slightly
taller than the Pearl River Tower (Section A.37) at 300 m (984 ft) but is less than half the height of
the mega-tall Shanghai Tower at 632 m (2073 ft) to be completed in 2014 (Section A.39). The
scale of DSF applications continues to expand and intertwine with the explosion of supertall
buildings (particularly in the country of China) and it appears that will continue.
200
Chapter 4 Endnotes
Rio Declaration on Environment and Development (1992) available at http://habitat.igc.org/
agenda21 /rio-dec. him>.
2
The text of the UNFCCC is available at http://unfccc.int/files/essential_ background/
background_publications_htmlpdf/application/pdf/conveng.pdf>.
3
Population data from The World Factbock available at www.cia.gov> with information updated
as of July 2013 (est.). Carbon emissions data is from the 2010 estimates from the Carbon
Dioxide Information Analysis Center available at cdiac.ornl.gov>.
4
The figures of LEED projects are as of April 2013 and based on information provided by the
USGBC at http://www.usgbc.org/articles/infographic-leed-world> (USGBC 2013).
5
Announcement of the LEED Earth campaign is available at the website: http://www.usgbc.org/
arti cles/usg be-off er -f ree-leed-certificatio n-gro undbrea king-projects-new-markets> (USG BC
2013).
6
The "double envelope" is defined in the Building Envelope: Technology Roadmap (US
Department of Energy, 25) as:
.. . glass and opaque double-envelope systems with integral energy collection
and distribution These systems combine two or more layers separate by one or
more cavities for the purpose of collecting and utilizing (winter mode) or
rejecting (summer mode).
7
Technology roadmaps are available at the International Energy Agency's website:
http://www. iea. org/roadmaps/>.
8
One such example is Uuttu's (2001) extensive research and documentation of double-skin
facade structures in Finland).
9
Classifications per Koppen-Geiger climate classification system.
10
The accuracy of the case study information provided within this chapter is to the author's best
knowledge, from review of what other researchers have documented and the information
available through literature, periodicals, project drawings and photographs. Where information
was not attainable, these items are marked "n/a" for not available.
11
This case study was partially informed by Erik Olsen, PE, Managing Director of Transsolar's
presentation CLIMATE ENGINEERING Energy Efficiency and Environmental Quality in
Buildings at the Facade Tectonics 6 conference at the University of Southern California on July
29
1
", 2011.
12
The current status of the Twofour54° Zone 1 project in Abu Dhabi - and whether its developed
design still includes any portions with a double-skin facade enclosure strategy - is unknown to
the author.
201
13
This case study was partially informed by Erik Olsen, PE, Managing Director of Transsolar's
presentation CLIMATE ENGINEERING Energy Efficiency and Environmental Quality in
Buildings at the Facade Tectonics 6 conference at the University of Southern California on July
29
1
", 2011.
14
This case study was informed by Christoph lngenhoven's presentation Vl/hy I Love Skyscrapers
at the National Museum of the American Indian, Alexander Hamilton U.S. Custom House in
New York City on April 1 o'", 2013, as well as his presentation SUPER GREEN at the
facades+PERFORMANCE Symposium on April 11
1
", 2013 at the McGraw Hill Conference
Center.
15
"Supertall" is defined by the Council on Tall Buildings and Urban Habitat (CTBUH) as a building
taller than 300 m (984 ft) in height measured from the lowest entrance level to the "architectural
top" of the building, which does include spires. "Megatall" is defined as 600 m (1968 ft)
(CTBUH 2013).
202
5. Double-Skin Facades in the United States
The application of modern double-skin facades in the United States began with a multi-story
solution on the Occidental Chemical Center (or Hooker Building) in Niagara Falls, New York over
three decades ago in 1980. Nearly twenty years passed before other projects in the United States
followed suit by integrating double-skin facade solutions; however, within the last ten years,
nearly two dozen such projects have been realized. As the sample size grows, the differences in
motives and justifications for double-skin facades in the United States as compared to Europe
continue to reveal themselves. With increased implementation and consideration of double-skin
facades in the United States, a heightened dialogue with respect to life-cycle performance,
maintenance access, and installation has become more common earlier in facade design
development. This chapter presents project case studies as groundwork to address these issues
as well as hurdles to implementation, emerging trends in U.S. double-skin facade applications
and design considerations moving forward.
5.1 Barriers to Implementation
Active facades have been implemented frequently in Europe, but the rest of the world has lagged
in applying, or investing in, high-performance facade technology. The use of double-skin facades
in the United States for commercial buildings presents a variety of challenges and impediments:
• Double-skin facade systems are not common building envelope technology
• Architects, engineers and specialty contractors in the United States are not generally
familiar with radiant systems behavior and design
• The traditional segregation of the U.S. specialty trades operating relatively independent
of each other is counter to the integrated and collaborative design approach that requires
coordination across building design disciplines and trades, though this is improving in
recent years with the development of new project delivery models like design-assist
203
• Early integration of facade engineers is not a common design process for architects,
though this is becoming increasingly more common in recent years
• The cost of the high-performing technologies can be very costly compared to
conventionally-glazed facades (CGFs) or higher-insulating approaches like triple-glazing
• Developers are looking for a quick payback, or return on investment, on building systems;
not a conducive time span for double-skin facade
• The United States building codes do not require access to daylight and fresh air in office
environments as some European countries' codes mandate
• There is no engineering discipline responsible for solar controls such as shading and
blinds, often resulting in neglect for the role and integration of these systems
• There is not as great an emphasis, or value, for healthy work environments compared to
European nations.
Experience and extensive testing show that double-skin facade systems have the ability to adapt
to applications in regions other than central Europe. To date these structures have been
implemented sparsely in Japan, China, Canada and the United States. The principal drawback in
Europe and elsewhere is the increased cost compared to traditional curtain wall facades. Lang
and Herzog (2000) speculated that double-skin facades are about twice the price of conventional
curtain walls in Europe. As for comparable systems in the United States, the relative cost is likely
to be four to five times more expensive. These greater costs are a result of unfamiliarity with
these systems among trades, the additional glass material required, engineering and installation
costs (Lang and Herzog 2000). Notwithstanding the greater investment, double-skin facades are
more common amid European high-rises commercial developments because the cost of energy
is considerably higher in Europe than in the United States. Overtime, the greater energy costs
lend double-skin facades in Europe to a quicker return on upfront investment.
204
A primary hesitation to integrate double-skin facades in the United States is due to an uncertainty
in how the systems' effectiveness will translate to different climates outside of central Europe. A
critical factor in determining if a system will yield comfortable spaces depends on humidity. A
humid environment may require an active facade to address dehumidification and condensation
control. Condensation problems can be detrimental to preserving the quality of installed glazing
systems over time. These problems exist in some of the European buildings, and it is suggested
that these difficulties would be magnified in climates that are particularly humid or cold. In addition
to condensation control, a technical analysis must be conducted to evaluate the potential for
using natural ventilation at each site in specific climates. It is likely that the number of annual
hours that natural ventilation is possible in cities across America is fewer than at the successful
European applications in cities such as Frankfort, London, or Amsterdam. This would mean a
direct translation of the double-skin facades as we know them would not be optimal, but
advancements in the systems would be required on an application-specific basis.
Beyond the financial and climatically hurdles for double-skin facades in North America, a cultural
discrepancy exists between the expectations of European and Americans in the quality of work
environments. In Germany, workers have demanded access to daylight and outside air in
commercial buildings to the point that it is implemented in the building code. Additionally,
European research proves that employee satisfaction and productivity with active facades is
higher, thanks to the many benefits including natural ventilation, daylighting, and greater control
variability over workplace atmosphere. This level of expectation is not seen outside the European
continent, and does not exist in the United States. As sustainable issues and climate change
awareness have escalated in recent years, it is possible that concerns about indoor air quality
and human welfare may provide the impetus for paradigm shifts in the expectations for U.S.
commercial buildings. Change is unlikely for the near future since the dollar will continue to speak
louder than people for some time.
205
5.2 Development Model and Construction Industry
To date, developers in the United States are solely interested in deep-plan buildings evaluated by
their cost per square foot. Developers are also generally not concerned with the economies of
worker satisfaction and productivity associated with views, fresh air, and comfort, because the
common model is to sell a developed property within the first five years following completion.
Double-skin facades with operable features can be very challenging to control energy efficient
levels with the common open floor plan office space. This American building typology is not
conducive to high-performance building skins in the same way as European office towers with
perimeter offices or work spaces. The European model allows for segmented zoning to optimize
efficiency through local control systems. The only real way to overcome this economy of open
floor plan development will be through building code regulation to mandate change beginning at
the local level.
Another obstacle for implementation of double-skin facades in the United States is a result of the
fragmented nature of the construction industry, where trades are segregated. Double-skin
facades require more material, as well as coordination and interaction between glazing,
mechanical and electrical subcontractors. This can be difficult to achieve since each party will
generally look to shed the risk associated with their scope of work to ensure profits. The
involvement of multiple trades tends to inflate costs due to the scheduling impact. Involvement by
specialists is required early in the design process to see the implementation of double-skin
facades through from conceptual development to installation.
The greatest difficulties with double-skin facades in the United States appear to be related to
development regulation, industry fragmentation, and lower social expectation for quality
workspaces. The difficulties related to climate-specific issues can be combated through system
modifications and advancements tailored to each individual project location. The development of
sustainable standards, such as LEED, is the beginning of changing building performance
206
expectations, and as local governments begin to require sustainable building, active facade
systems in American cities will be considered and accornrnodated at a higher frequency.
5.3 Double-Skin Facade Applications in the United States
Double-skin facade applications are often regarded as infrequent or rare, but the reality is there is
a growing sarnple size of built constructions and an increasing rate of application. Are these
applications in response to climatic, comfort or other motivational drivers? It is difficult to access.
Motivations aside, it is important to highlight the diversity of solutions, trends present in current
applications, and strategies for rnore sensible applications in the future.
Table 5-1 presents a database of completed double-skin facade projects within the United States,
several of which are expanded upon in the following sections, with all presented in Appendix C.
This list is comprehensive of the authors understanding of completed double-skin projects as of
January 1 '
1
, 2014. There are projects that have been completed since this project database was
completed and a high likelihood of previously completed projects that the author did not have
information about. This list, and the extraction of trends frorn this list, are intended to be
characteristic of the domestic double-skin facade technology, but are in no way an all
encornpassing surnrnary of the dynamic facade design and engineering industry. Additionally, in
Table 5-2, a list of known forthcoming double-skin facades in the United States is presented. The
five projects are also expanded on as case studies following the built examples in Appendix C.
Despite these successful projects, there are reservations by industry and developers to integrate
dual-facades. Since there are very few examples of such enclosure types in North America, the
building skin typology lacks convincing evidence on how well the dual-skin facades work in
conserving energy in particular climates. Simulating these environments and innovative systems
can be unreliable with the existing modeling software and practices, and private developers are
unlikely to assurne a perceived risk at an inflated prerniurn relative to traditional facade solutions.
207
Table 5-1: Double-Skin Facades in the United States
1
Location
Project Name Year City State Architect
Warren Petroleum Executive HQs 1957 Tulsa OK SOM, Bruce Graham
Occidental Chemical Center 1980 Niagara Falls NY Cannon Design Inc.
Yazaki North American 1999 Canton Ml Plantec
Levine Hall - Pennsylvania School of Engineering 2001 Philadelphia PA Kieran Timberlake Associates
Seattle Justice Center 2002 Seattle WA NBBJ, Kerry Hegedus
Manulife US Headquarters 2003 Boston MA Skidmore, Owings & Merrill, LLP
Genzyme Headquarters 2003 Boston MA Behnisch, Behnisch and Partner
Foundry Square 2003 San Francisco CA STUDIOS A rchitecture
UMass Medical School 2006 Worcester MA Payette Associates
Loblolly House 2006 Taylors Island MD Kieran Timberlake Associates
University of Michigan Biomedical Science Research Building 2006 Ann Arbor Ml Polshek Partnership Architects, LLP
Museum of Contemporary Art 2007 Denver co David Adjaye w/ Davis Partnership
Comcast Tower 2008 Philadelphia PA Robert A.M. Stern Architects
Riverhouse - One River Terrace 2008 New York NY Polshek Partnership Architects, LLP
Loyola Information Commons 2008 Chicago IL Solomon Cordwell Buenz
Walter Cronkite School of Journalism 2008 Phoenix AZ Ehrlich Architects
Whatcom Museum 2009 Bellingham WA Olson Kundig Architects
Art Institute of Chicago - Modern Wing 2009 Chicago IL Renzo Piano Building Workshop
Cleveland Museum of Art - East Wing 2009 Cleveland OH Rafael Vinoly Architects
Cambridge Public Library 2009 Cambridge MA William Rawn Associates
100 Park Avenue 2009 New York NY Meed de Armas & Shannon Architects
USC Eli and Edythe Broad CIRM Center 2010 Los Angeles CA Zimmer Gunsul Frasca Architects, LLP
New York Presbyterian Hospital 2010 New York NY Pei Cobb Freed & Partners
University of Oregon Jaqua Center 2010 Eugene OR Zimmer Gunsul Frasca Architects, LLP
New World Center 2011 Miami FL Gehry Partners
University of Baltimore Angelos Law School 2013 Baltimore MD Behnisch Architekten
Harvard Business School Tata Hall 2013 Boston MA William Rawn Associates
Cedars-Sinai Advanced Health Sciences Pavilion 2014 Los Angeles CA HOK
Weill Cornell Medical College - Belfer Research Building 2014 New York NY Ennead Architects
University of Kansas School of Architecture - The Forum 2014 Lawrence KS Studio 804
Table 5-2: Forthcoming Double-Skin Facades in the United States
Location
Project Name Year City State Architect
Northwestern University Recital Hall 2015 Evanston IL Goettsch Partners
Rodino Federal Office Building 2015 Newark NJ Dattner Architects
A.J. Celebrezze Federal Building 2015 Cleveland OH Interactive Design Eight Architects
PNC Plaza Tower 2015 Pittsburgh PA Gensler
Columbia University Jerome Greene Science Center 2016 New York NY Renzo Piano Building Workshop
New Stanford Hospital 2018 Palo Alto CA Rafael Vinoly Architects
208
5.3.1 Yazaki North America
Completed: 1999
Location: Canton, Michigan
Architect: Plantec Architects, Inc.
Cavity Partition: Multi-Story
Ventilation Type: Natural
Ventilation Mode: Outdoor air curtain
Cavity Depth: 3 m (10 ft)
Shading: n/a
Sources:
Figure 5-1: Yazaki North America - Exterior view of atria.
Source: Advanced Structures Inc.;© Enclos/ASI.
The four-story headquarters building encloses utilizes a series of 12 exterior atria, each 20.5 m
(67.3 ft) wide. Ten of the atria are double-skin systems with the other two being single skin
facades at the primary lobby entrances that include a distinct canopy penetration. The exterior
glass is point-fixed using a Pilkington Planar™ system to support monolithic glass lites that are
1867 mm (6.125 ft) wide by 1905 mm (6.25 ft) tall. The glass bolt fittings attach to a series of 73
mm (2.875 in) diameter horizontal spreader tubes that align with the horizontal glass joints every
1905 mm (6.25 ft). The spreader elements rigidly connect to a 15 m (50 ft) tall, 114 mm ( 4.5 in)
diameter vertical tube mast. The spreader elements separate the front and back 22 mm (0.875
in) diameter galvanized steel cables. Cable elements are also used 149 mm (5.875 in) behind the
face of glass as a vertical dead-load cable and every 3.8 m (12 .5 ft) vertically at the mast tube for
lateral bracing to floor levels. The interior facade is a conventional curtain wall that is not
supported by the structural mast trusses located within the atria cavities.
209
Figure 5-2: Yazaki North America -Atria mast truss (left) and overall building (right).
Source: Advanced Structures Inc.;© Enclos/ASI.
The mast truss assembly is the primary load resisting system for the vertical glass v.tall. The
tensile loads developed by the activation of cable elements under load are resisted internally by
the mast element. The bow trusses are pinned from the top at a series of steel outriggers and
vertically slotted at the base to allow for expansion and contraction or footing settlement. The
advantage of this approach v.tas the accelerated erection of the steel trusses permitted by the
prefabricated mast trusses. The point fixed glazing still required field installation and sealing.
Airflow into the cavity is through a trench condition at ground level. An outdoor air curtain is
created when the seasons permitted, otherwise, the configuration acts in a thermal buffer
ventilation mode. Warm air rises to the top of the cavity space where a glass skylight - slightly
sloped tov.tards the outermost vertical face - caps off the cavity volume using a similar point fixed
attachment system.
210
5 .. 3.2 Seattle Ji.uilic, e CeJlter
2
ColT)pleted: 2002
Loe.ti on; ·s eot lie, Wa Sil ingtorr
llrcltlte.ct: Kerry Hegedus, NB . BJ
f avlty Portlt1on: Mtlltl Story
\YeQtilo\ion Typ~ H)'.bnd
\1en1il~tion M o _de~ Outdoor· air ~urttiR
Cavtty. Depth: Q .7~ m (30 ii'()
'Shading: Blind:i; automated
· source. s: Le.e et ah '('.200;!, 72'17');
Hegetlus. (2001 ·~ Poirozis (2006
1
204 ·205')
figose·S~: Seattle' Juslice'Center- Uveratl'slreet view.
S0~1'Ce; .~ ee,i:e- .bulld\riggreen.rcom >; ©Chtistisin Richters (access. e. d September 9, 2011 ).
Jliis 12-s)Q l.Y. 26,000 IT)
2
(300:,QQQ $fj LEED Silver eertified .~uilciing is swrouncie'tj on thi ee sides
by e)i(j·stJng build.l ngs:· vW.h ~he $Outh1J1.est fa'cad' e exp· dred to su,nUght. The interior pr.o.gram
incl'UUe.s cifY· courts: $ind' police he.adqoarter.s:. Pait pf (he de· s:iie;d 6e:Sttietic ~sto 1 Provide, vie\l\e.
into tne building .to reflect ~.n op· en, in..,;ting end ·tr~n·sp~rent view of government o-t ege· dus-· 2o· oi A
11 ),. O'n the Vies! la!ide e 9-story .. 1672 ln
2
(18,000 st) m utti-stor; double-Skin facade' glazesthe
building. The system uses~ multi-story 0 .76 m (30 in) deep cavity .sp, ace enoompassed by $ngle
lo\'11/~ton glazing to ttie e~etiof :S:rjd instilatin!;J glazing ta the (nterior. Withif1 the. cavity spare there
1s- a· tv1 ecnoS· hade sti·tldiAg system that consists otse1ni-lransparent roBer blind- stha:t tntiv.es: In
u[lison by floor aqcordirrg 1 0.11istoric.al s. eas. onal data. Additionally) th: ere i. s- a :cat\ri.e_ lk loc. ated :at
each floor leVel and an interior li~ht shelf &I 2.4 m (8 ft} above ~nisned flooe. Due lo Cd's!
constraints, the·btin: ds only loV\ertO' 1 :s m (6 1),) above 1ldor level (lee,. e.t at ,~02, 7 .5) . The
Svste1niis not progr-etnm ed to respond to· reaJ-tim e - \l\eather data.
Figure 54: Seattle Justice Center - DSF elevation (left) and cavity space (right).
Source eere buildinggreen com>;© Christian Richters (accessed September 9, 2011)
Th· e intermediate cavity space includes air intake and air extract louvered openings at the bottom
and top of the. wall respectively These louvered openings are regulated by the cavity air
temperature and the building's HVAC operating mode . The louvers· remain closed on cold days
to retain heat and act as an insulating barrier. Durin9 hot days, the louvers are opened to vent
heat from the interior cavity space, extracting thermal uplift at the roof parapet The primary
complaint for this system by the own er is glare .
The project sought to develop an affordable, simple and sensible solution that was driven by
performance and costs . The intent to keep the design simple was so the owner. general
contractor. facade contractor nor budget felt overwhelmed. Keeping true to this mantra. a
standard set of materials and aluminum curtain wall systems was implemented to create a
familiarity or comfort with the layered system
212
5.3.3 Loyola University Chicago Richard J. Klarcheck Information Commons
3
Completed: 2007
Location: Chicago, Illinois
Architect: Solomon Cordwell Buenz
Cavity Partition: Multi-Story
Ventilation Type: Hybrid
Ventilation Mode: Outdoor air curtain
Cavity Depth: 0.9 m + (3 ft+)
Shading: Blinds; automated
Sources: Patterson (2011, 207-221);
Vaglio et al. (2010, 191-199); Kirchoff et al.
(2010, 267-277)
Figure 5-5: Klarcheck Information Commons - Horizontal cable and DSF end condition.
Source: Patterson (2011 , 220); ©Mic Patterson.
This 6,690 m
2
(72,000 sf), three-story, LEED Silver university building is a high-tech library sited
immediate to the Lake Michigan waterfront to the east. Interestingly, it is a completely electronic
resource library- no books! High transparency walls are used on both the east and west
exposures to preserve views of the Lake from the center campus quad. A 46 m (150 ft) wide by
17 m (56 ft) tall multi-story double-skin facade with a two-way tensioned cable-net on the west
facade was implemented to manage heat flow and natural ventilation through the year. The 0.9 m
(3 ft) intermediate cavity space acts primarily as a buffer space to provide thermal insulation
during cold winter months and protecting shading elements that would otherwise be external. The
system is also designed to operate in a hybrid ventilation mode under controlled conditions. The
interior program spaces are exhausted through the air cavity to relieve the interior space of high
internal loads, while outdoor air enters the space at ground level trench through metal louver
dampers and is raised by thermal uplift before extraction at the roof return via operable windows.
213
Figure 5-6: Klarcheck Information Commons - Elevation (left) and cavity photo (right).
Source: Patterson (2011, 219 & 211); ©Mic Patterson (left) and Enclos (right).
The exterior skin is an innovative application of two-way cable-net support structure - the first of
its kind in the United States. The crossing network of tensioned vertical and horizontal stainless
steel cables support the 12 mm (Y2 in) fully-tempered, monolithic, low-iron glass lites at each
corner with stainless steel clamping components. The glazing and cable modules follow the 14
mm (5/8 in) butt-glazed silicone joints, occurring every 1.5 m (5 ft) vertically and 2.4 m (8 ft)
horizontally. The slender cables along with a minimal field-applied silicone seal maximize the
diaphanous appearance of the skin. The interior skin of the DSF is insulated glass units that are
point-supported by patch plate clamps along vertical aluminum extrusions. The joints between the
panels use field-applied silicone as a weather seal. The depth of the cavity space varies as the
inner-facade curves inward to form a vestibule. Within the cavity space are 100 mm (4 in)
horizontal blinds near the outer facade which track the sun's movement to allow natural daylight
into the building when open, but reflecting away excess radiant energy when closed.
214
5.3.4 Cambridge Public Library
4
·
5
Completed: 2009
Location: Cambridge, Massachusetts
Architect: William Rawn Associates
Cavity Partition: Multi-story
Ventilation Type: Natural
Ventilation Mode: Outdoor air curtain, air supply
Cavity Depth: O .9 (3 ft)
Shading: Louvers; automatic
Sources: www.rawnarch.com>; (Gonchar
2010, 76-81 ); Vaglio et al. (2010, 195-198)
Figure 5-7: Cambridge Public Library - Exterior DSF view.
Source: © Edward Lifson.
William Rawn Associates' design for a new 6,500 m
2
(70,000 sf) two-story building alongside the
preservation of the existing 3,250 m
2
(35,000 sf) historic library structure in Cambridge includes a
multi-story double-skin facade to increase comfort and reduce operating costs. The transparent
facade embraces the adjacent park and communicates a symbolic message of openness and
welcoming. The southwest facing double-skin facade by Gartner Steel and Glass GmbH, based
in Wurzburg, Germany, consists of inner and outer glass walls separated by a 0.9 m (3 ft) deep
by 12.3 m ( 40'-6") tall cavity. The two skins are supported by a series of vertical steel frames tied
back to the building's primary structure. The steel framing - common support for both the interior
and exterior skins - supports the point-clamped monolithic exterior skin, large translucent glass
shading canopies on the exterior that act as light shelves, aluminum blinds within the cavity, as
well as the aluminum curtain wall interior skin utilizing insulated glass units. An air inlet located at
the bottom and an exhaust vent at the top are used to introduce airflow into the cavity
215
Figure 5-8: Cambridge Public Library- Overall exterior (left) and interior (right) views.
Source:© Edward Lifson.
During winter months the bottom air inlet and top exhaust vent are closed to create a thermal
buffer to insulate the interior spaces. In summer months, an outdoor air curtain is created by
introducing cooler air through the air inlet into the cavity which rises as it warms and exits the top
vent creating a stack effect. During the more temperate fall and spring seasons, an acceptable
cavity temperature is maintained through operable windows, adjustment of air openings and
variant shading configurations providing natural ventilation to the interior. The design uses
different shading strategies for zones within occupants' height and zones above. Shading within
the cavity is achieved via perforated aluminum venetian blinds at the upper half of each floor
which are used to minimize glare and provide even light levels to the interior library space. The
operable louvers vary in position throughout the day and year by computer controls tied to the
BMS. The shading slats are closed during winter days to reduce glare from low sun angles.
During summer the louvers are configured to provide shading against high-angled sunlight.
216
5.3.5 USC Eli and Edith Broad Center
6
'
7
Completed: 2010
Location: Los Angeles, California
Architect: Zimmer Gunsul Frasca Architects
Cavity Partition: Multi-story
Ventilation Type: Natural
Ventilation Mode: Outdoor air curtain
Cavity Depth: 0.9 (3 ft)
Shading: Fixed; translucent frits
Sources: Patterson (2011, 195-206);
Vaglio etal. (2010, 196-198);
www.zgf.com>
Figure 5-9: USC CIRM Center - DSF cavity view.
Source: Patterson (2011 , 206); ©Mic Patterson.
The multi-story double-skin facade is located on the southeast face of the new four-story stem
cell research facility located on the campus of University of Southern California Keck School of
Medicine in Los Angeles. The double-skin facade, designed by ZGF Architects LLP, was
selected because the project demanded a 'world-class' facade . The original concept used a
shared cable truss to support the interior and exterior walls by drilled point supports (glass bolts)
on both sides. This concept was ultimately scaled back to separate structural systems due to
cost. The final design includes an interior skin with low-e coated insulated glazing supported by a
unitized aluminum curtain wall system. The exterior 19 mm(% in) low-iron laminated glass is
comprised of a 6 mm(% in) fully-tempered outer lite, custom frit on the number two surface, 1.52
mm (0.06 in) SentryGlas® interlayer, and 12 mm ( Y:z in) fully-tempered inner lite. Though the floor
levels are 4.88 m (16 ft) apart, the outer skin's laminated lites are 1.6 m (5.25 ft) wide by 2.44 m
(8 ft) tall. Alternating translucent and clear lites creates a vertically banding along the outer skin.
217
Figure 5-10: USC CIRM Center- Overall exterior view during construction.
Source: Patterson (2011, 195); ©Mic Patterson.
The exterior skin is cable-suspended from the building structure. The vertical cables are
supported by T-shaped outriggers at each floor to minimize cable span and deflections under
loading. The steel outriggers double as support for maintenance grating and fixed shading
devices located within the cavity. Point-fixed bolted laminated glass attaches on a 1.6 m (5.25 ft)
vertical grid to the cables. The two skins enclose a 0.9 m (3 ft) deep by 20.7 m (68 ft) tall cavity
that acts as an outdoor air curtain by bringing fresh air into the lower cavity and using the stack
effect to exhaust warm air through a series of automated, louvered vents at the roof parapet.
These operable louvers allow the cavity to be opened or closed to optimize cavity temperature
and airflow, which will reduce undesirable conductive gains and losses through the interior glass.
The louvers are controlled by comfort temperature sensors located on the interior surface of the
interior facade. Though considered in schematic design, natural ventilation is not introduced to
the interior spaces due to the sensitivity of the laboratory environment.
218
Figure 5-11: USC CIRM Center - Exterior (left) and interior (right) views of translucent frit.
Source: Patterson (2011, 201 ); ©Mic Patterson.
Both the 61 m (200 ft) long east double-skin and the west glazed facade explore translucency
effects to control interior light levels within the laboratory spaces. The southeast double-skin
facade has an alternating pattern of full-coverage translucent frit between the interior low-e
insulated glass and exterior laminated glass. This creates a completely translucent effect normal
to the facade while permitting transparent angular views for the researchers located within. The
use of different glass suppliers for the exterior skin and interior skin presented unique challenges
in matching the appearance of the translucent frit treatment.
Access to the cavity for maintenance and cleaning is possible through end doors at each level.
The interior grating platforms proved instrumental in permitting two-sided access for installation.
219
Figure 5-12: USC CIRM Center - Installation of vertical cables (left) and glass (right).
Source: Patterson (2011, 200 & 204); ©Mic Patterson.
Energy modeling was performed during design development as part of a life-cycle analysis of the
whole-building model. The early concept included a double-skin facade on the west facade as
well. In combination the two double-skin facades contributed toward a 40% reduction compared
to Title 24 standards. Simulations were conducted using eQuest during schematic design and
later evaluated through proprietary software after the deletion of the double wall on the west
elevation. Computational fluid dynamics (CFO) analysis of the cavity was also performed using
TRNSYS. Analysis of the double-skin facade was necessary to understand its contribution in
achieving the LEED energy efficiency credit. The project is on target to achieve LEED Gold
certification
220
5.3.6 New York Presbyterian Hospital
Completed: 2010
Location: New York, New York
Architect: Pei Cobb Freed and Partners
Cavity Partition: Multi-story
Ventilation Type: Mechanical
Ventilation Mode: Air exhaust
Cavity Depth: 0.76 (2.5 ft)
Shading: Fins (fabric); automated
Sources: Patterson (2011, 167-182);
www.pcf-p.com/>
Figure 5-13: New York Presbyterian Hospital - Exterior DSF view of air inlet.
Source: © Nathan Lucero.
The transparent addition of the Vivian and Seymour Milstein Family Heart Center is nestled
between two existing hospital structures that possess predominantly rectilinear, opaque, brick
aesthetic. The new addition includes a four-story curvilinear glass climate wall that seeks to
maximize daylight and views, mitigate solar gain in the summer, and reduce excessive heat loss
in the winter. The four floors vary in height from 4.6 m (15 ft) to 4.9 m (16 ft). Daylight harvesting
and view preservation are achieved through the 0.76 m (2.5 ft) cavity space that acts as a
chimney exhaust for air from the building's plenum system and houses 56 vertical columns of
fabric solar vanes that respond to the movement of the sun. Offset from the vertical datums of
the solar vanes are 28 stainless steel rods, 1.5 m (5 ft) on-center, that create a four-story tension
truss that supports the outer skin, fabric fins, and grating platforms at each floor level. The
stainless steel truss system ties back laterally to each floor slab but the dead load of the outer
wall and cavity elements is hung from a vertical dead load rod located inches behind the outer
221
Figure 5-14: New York Presbyterian Hospital - Overall view.
Source:© Nathan Lucero.
skin and tied up to a steel outrigger at roof level. Essentially the outer skin is structurally
suspended from the roof with a post-tensioning occurring at its base. The exterior skin uses 24.5
mm (1 in) laminated lites that are 1.5 m (5 ft) wide by 4.9 m (16 ft) tall, comprised of a 15 mm
(9/16 in) fully tempered outer lite, 1.52 mm (0.06 in) PVB interlayer and 8 mm (5/16 in) heat
strengthened inner lite. The outer lites are supported at their corners as well as mid-span by
point-supported fixings, or glass bolts. The mid-span point-fixing is tied to the crossing
intersection of the diagonal tension rod truss elements. The inner skin is a 44.5 mm (1.75 in)
laminated insulated glass unit consisting of a 15 mm (9/16 in) fully-tempered outer lite, 16 mm
(5/8 in) air space, and two 6 mm (1/4 in) fully-tempered lites laminated by a 1.52 mm (0.06 in)
PVB interlayer. The inner skin is channel captured between the slabs and spans floor-to-floor
creating a horizontally-uninterrupted view out over the Hudson River. All glass on the double-skin
facade is low-iron Optiwhite by Pilkington.
222
Figure 5-15: New York Presbyterian Hospital - Detail view. Source:© Nathan Lucero.
223
5.3.7 University of Baltimore Angelos Law School
8
Completed: 2013
Location: Baltimore, Maryland
Architect: Behnisch Architekten
Cavity Partition: Baffle
Ventilation Type: Natural
Ventilation Mode: Outdoor air-curtain, air supply
Cavity Depth: approximately 25 cm (10 in)
Shading: Blinds; manual
Sources: Sokol (2013)
Figure 5-16: Angelos Law School - Overall view.
Source: author photo;© Jeffrey Vaglio.
The 12-story new construction for the John and Frances Angelos Law Center is an intersection of
three distinct programmatic elements: 1) office and classroom, 2) library, and 3) clinics, faculty
and administration spaces. Atrium neighborhoods are used between these spaces to resolve any
differential between the floor heights and elevations. Each of the three programmatic elements is
given a unique facade solution, including the office and classrooms' unique DSF-rain-screen
hybrid. The primary goals of the building's facade program were to provide access to natural
ventilation, enhance daylight access and control, and protect the shading devices external to the
primary enclosure layer. Specifically for the double-skin facade system, protection of the shading
devices from wind, insulating from external noise and providing operable windows for occupants
were design objectives; not much more - energy was not a key driver. The result is an interior
skin of punched window vision zones, an approximately 25 cm (10 in) deep air cavity housing
venetian blinds, and outer skin that is point-supported by capture clamps along vertical edges.
224
Figure 5-17: Angelos Law School - Baffle DSF (left) and external blinds (right).
Source: author photos;© Jeffrey Vaglio.
For the double-skin facade around the office and classrooms, the exterior skin is a low-iron 12
mm (1/2 in) fully tempered laminated lite that is point supported by white armatures at corners
and intermediate points along the panel's height. The exterior skin is more or less a baffle, or
rain-screen, since the joints between lites remain open, no sealant, and therefore permits airflow
into the cavity space regularly along the surface of the facade. This is a more evenly distribution
of air inlet/outlet than seen in most of the aforementioned examples. This consistent permeability
aids in preventing overheating but also allows dirt to accumulate within the cavity. The interior
skin has both opaque metal panel spandrel zones and operable window vision zones. The
operable windows are low-iron insulated glass units with a low-e coating. Where a vision lite
occurs, so do exterior venetian sunshades. When prompted about the annual cleaning strategy
for the DSF, Stefan Behnisch described it as "sub-optimal" and a "sore spot" due to the restricted
access and shallow cavity depth.
9
225
5.3.8 Cedars-Sinai Advanced Health Sciences Pavilion
Completed: 2014
Location: Los Angeles, California
Architect: HOK
Cavity Partition: Multi-story
Ventilation Type: Natural
Ventilation Mode: Outdoor air-curtain
Cavity Depth: 0.9 m (3 ft)
Shading: Screens; fixed
Sources: (Medical Construction and Design
2014); www.bensonglobal.com;
www.hok.com>
Figure 5-18: Cedars-Sinai Advanced Health Sciences Pavilion - Overall view.
Source: www.bensonglobal.com>; ©Benson Industries, Inc. (accessed April 21, 2014).
The 11-story, LEED Gold new construction for the Cedars-Sinai Advanced Health Science
Pavilion contains a multitude of functions, including housing the Cedars-Sinai Heart Institute, the
Regenerative Medicine Institute, neurosciences, research laboratories, an education center and
procedural spaces (Medical Construction and Design 2014). The east and west facades utilize a
multi-story double-skin facade, as well as vertical fins, to mitigate solar heat gain and glare on the
broad building elevations exposed to low sun angles. The curving east facade is shaped by San
Vicente Boulevard and is the largest application of the 0.9 m (3 ft) deep multi-story double-skin
with a laminated glass lite to the exterior and 25 mm (1 in) double-glazing on the interior skin. The
interior skin is comprised of a 200 mm (8 in) deep aluminum curtain wall assembly. A structural
outrigger ties the exterior system back to the inner curtain wall and creates a platform for grating
that permits access for maintenance within the cavity. The cavity is more permeable than some
of the other multi-story systems in that it allows air to enter and exhaust at each floor level.
226
Figure 5-19: Cedars-Sinai Advanced Health Sciences Pavilion -DSF detail drawings and
photo. Source: www.bensonglobal.com>; ©Benson Industries, Inc. (accessed April 21, 2014).
227
5.3.9 Columbia University Jerome L. Greene Science Center
Completed: 2016
Location: New York, New York
Architect: Renzo Piano Building
Workshop
Cavity Partition: Multi-story
Ventilation Type: Mechanical
Ventilation Mode: Air exhaust
Cavity Depth: 36 cm (14 in)
Shading: Blinds; automatic
Sources: www.rpbw.com>;
www.enclos.com>
Figure 5-20: Jerome L. Greene Science Center - Overall construction progress photo.
10
Source: enclos.com>; © Enclos (accessed January 4, 2015).
The 10-story structure is the first building on the Manhatanville campus expansion of Columbia
University. The master plan for the campus was also designed by RPBW. The new building will
house Nobel Prize-winning neuroscientists that continue their mind, brain and behavior research
initiatives. The program calls for a high-performance enclosure, but one that is sealed due to the
strict air quality requirements of laboratory research facilities. The curtain wall Type 3 DSF occurs
along the south facade and portions of both the east and west faces. The DSF creates a climate
wall buffer that improves thermal comfort, protects automated sun-shades, and creates an
acoustical buffer much needed to insulate from the nearby elevated train. The outer skin is a
laminated glass panel supported by an extruded aluminum mullion cassette. The interior skin is
an IGU support by aluminum extrusions as well. Both skins' units are 0.8 m (2.6 ft) wide, 4.5 m
( 14. 75 ft) tall and connected by cast bracket armatures that are bolted approximately every 0.9 m
(3 ft), creating a composite truss structure common to the inner and outer skin.
228
Figure 5-21: Jerome L. Greene Science Center - Train adjacent to the east (left) and
southeast corner of Wall Type 3 DSF (right).
11
Source: author photos;© Jeffrey Vaglio.
The curtain wall units were preassembled offsite, interior and exterior skins adjoined. The units
were three modules wide and constructed as double-story height, meaning their overall
dimension they left the assembly shop at and arrived to site in was 2.4 m (7'-10 W') wide by 9 m
(29'-6") tall. A primary driver in the selection of double-skin facade was the acoustic performance
needed to create a comfortable interior environment despite the nearby elevated train. The initial
design of the double-skin called for a passive system that utilized an outdoor air curtain, bringing
natural ventilation into the cavity space via inlet vents. This approach was modified though due to
its degrading effects - serving as a source or noise - on the overall wall's ability to reach desired
acoustic ratings. The final design of the double-skin does not have inlet vents that communicate
with the exterior and instead draws air from mechanical sources within the building's HVAC
mezzanine. The cavity pressure is constantly maintained at a prescribed positive pressure and
continuously exhausts air out the relief louvers on the backside of the parapet assembly.
229
Figure 5-22: Jerome L. Greene Science Center - DSF's indoor air inlet at third floor (left)
and interior of DSF's cavity at the top exhaust (right).
12
Source: author photos;© Jeffrey
Vaglio.
Outdoor air enters the facade through wall Type 5's slotted steel panel soffit that occurs at the
transition between the Type 1 storefront glazing and the opaque zone, just beneath level three,
that conceals level two's mechanical mezzanines. Outdoor air enters air handling units in this
zone and exhaust air is extracted, also through Type 5's opaque zone, but through a separate
sealed plenum. The exhaust air is then fed into the wall Type 3's DSF cavity via a metal spandrel
with fixed vent louvers. The exhaust air then elevates seven stories to its exhaust at the roof
parapet through louvered grills, thus making the DSF a mechanically ventilated, air exhaust
chimney. The 36 cm (14 in) cavity houses automated shades that are concealed into cylindrical
head-boxes hung from the inner skin mullions. Open air grillages occur at each level, just above
the shade head-box, and hinge upwards to allow downward access to maintain/replace the
shades. Cavity access for maintenance is permitted by operable doors that occur intermittently on
the inner skin. Inboard, a second solar control layer exists on the interior in the form of roller
shades that are manually controlled with an automated override that can be controlled remotely.
230
5.4 Emerging Trends of Double-Skin Facades in the United States
Several trends, based on Section 5.3 and Appendix C's review of built and future DSFs are
identified.
5.4.1 Divergence of Scale
An emerging trend in U.S. double-skin facade applications is a divergence of scale. Many
designs can be separated into modular unitized systems, or a single multi-story cavity. Selkowitz
previously acknowledged this divergent trend of miniaturization on one hand and large scale
double-skins on the other (Selkowitz 2001, 54-56).
The primary driver towards unitization of double-skin facade systems is economics. Ad\!_anced
facades are typically customized and expensive, but designers are now seeking products which
can be purchased off-the-shelf, prefabricated or unitized. Examples of this are the use of unitized
products on Levine Hall and the custom unitized system on One River Terrace. Key benefits of
unitized systems compared to large multi-story systems include easier repair and maintenance,
as problems such as leakage or condensation are localized, and repair and replacement can be
accomplished without disrupting the operation of the entire system. These advantages, along
with economic forces favoring mass production as facilitated by the prefabricated systems, will
result in accelerated future development for these system types.
The other end of the divergence spectrum is the trend towards large-scale, multi-story DSF. The
multi-story facades presented herein tend to be geometrically similar with widths of 30 m (100 ft)
to 61 m (200 ft), heights of 12 m ( 40 ft) to 21 m (70 ft), and depths of 1 m (3 ft). Of course,
projects like the University of Michigan Biomedical Science Research Building (Section 5.3.7) and
Seattle Justice Center (Section 5.3.2) present dimensional exceptions outside these general
ranges. The range of cavity height to depth aspect ratio was predominantly from 10 to 30 until
recently. The multi-story configurations on future projects (Jerome L. Greene Science Center -
see Section 5.3.8) are pushing the height-to-depth aspect ratio greater than 80 with tall but
231
relatively shallow cavities. Multi-story single-cavity applications are appealing to designers
because they possess a homogenous aesthetic from both the interior and exterior. These large
systems can also possess a sense of grandeur and make a powerful statement, but require a
more sophisticated understanding of cavity dynamics and systems interactions. These systems
are often praised for their beauty, but criticized for under-performing.
5.4.2 Structural Aesthetic
The use of multi-story cavity spaces creates an opportunity to express a dramatic structural
solution to the support of the two skins. This is evident in the use of a highly flexible two-way
cable-net on the Loyola Klarcheck Information Commons (see Section 5.3.3) and vertical cables
on the USC Eli and Edith Broad CIRM Center (see Section 5.3.5). It is also apparent in the
conceptual quest for a singular structure supporting both facades on the Broad Center (via cable
truss), and the shared steel frame that supports the DSF at the Cambridge Public Library (see
Section 5.3.4), that designers are intrigued by the double-skin as a whole performance envelope.
The use of point supported glazing systems for the exterior skin is present in each of the multi
story case studies presented; corner patch plate clamps on Loyola Commons, bolted fixings on
the Broad CIRM Center, and intermediate patch plate clamps on the Cambridge Public library.
The use of innovative structural systems with double-skin facades will continue to be a symbiotic
balance of transparency and technological iconicism.
5.4.3 Life-Cycle Assessment
The acceptance of double-skin facade systems ultimately depends on the life-cycle assessment
costs. When comparing the cost of a double-skin facade to a single-skin facade system it is
necessary to evaluate not just the investment, but the cost of operations over the structures'
expected life. A standard method of life-cycle assessment of double-skin facade systems is
necessary to consistently evaluate their feasibility of application. As double-skin facades become
more commonplace in buildings, familiarity with the technology will likely result in decreasing
costs. The perceived risk associated with double-skin facades will decline. Design, engineering,
232
fabrication, and installation costs are likely to decrease, and increasing competition will lead to
further cost competitiveness. Finally, the factor that will reduce the payback period of a DSF the
quickest is increasing energy costs. Higher initial costs of material production will increase the
initial investment for all facades, but operations costs and energy performance will be more
critical in reducing the payback period for a double-skin facade.
5.4.4 Post Occupancy Evaluation
An even greater factor that could influence the use of double-skin facade applications is post
occupancy evaluation (and disclosure) of real performance data. This information could be
compared to expected performance to hone in on an accepted life cycle assessment model for
future designs, thus closing the performance gap.
5.4.5 Adaptive Re-Use I Retrofit
Double-skin facades are being considered for the renovation of older buildings. This is one of the
most promising applications for double-skins, but further study is required. With increased
environmental awareness, re-use of existing building stock is often a viable alternative to new
construction. This approach may provide greater economy, modernize the appearance of a
building, and improve energy performance, all while projecting a positive perception of
environmental consciousness. Retrofitting with a DSF also avoids removal and land filling of the
existing skin. The kind of buildings that fit the profile for a facade retrofit are typically structures
that remain in the possession of one entity for a long period of time - government or institution.
5.4.6 Translucency
The use of double-skin facades with a single translucent layer or both skins translucent appears
to be increasingly prevalent. With a translucent inner surface, facilities like museums are able to
achieve a filtered, diffused interior lighting. This is the case in the Museum of Contemporary Art
in Denver (see Appendix C.12) as well as the Nelson Atkins Museum in Kansas City." For the
new addition completed in 2007, Steven Holl Architects introduced a double-layer buffer facade
233
with a translucent, insulated·outer layer separated appr(O(imately 1.2 m (4 ft) from an inner
laminated safety glass, etched on the inside (yl/eller, et al. 20.09, 84-85). An advantage of the
translucent treatments is.the a~ility to transformth'e envelope with lighting·(in oombination with
roller shades) at night into a smoothly lit lantern effect.
Figure 5-23: Nelson Atkins Museum - Translucent lenses with thermal buffer wall.
Spurce: •www.stevenholl.com»; ©Andy Rfan.
The .USC Eli and Edith Bro. ad CIRM Center (see Section 5.3.5) also applied translucency to
create an interesting visual effett. By offsetting Vl!rtical bands of.tr-artslut ent frit on the interior and
exterior skirts, a f~ll9 trahslucent, albeit layered, effet tWas create!! from perpendit ularto the'
facade W h 'ile allct"'1'ing oa~upants from w~hin to havl!. diagonal view corridorsthrough alternatin· y
V ision bands. This effe. et diffus:es aJrd c . ontrols the ear19 momiltg da;i· lilJt\twhile creating .a more
even light quality· within for tesearc hers.
The use of translucency in double-skin facade applications can also be seen globally. One
example is the Laban Centre in London by Herzog and de Meuron that has a 0.6 m (2 ft) cavity
space with a triple layer transparent polycarbonate outer skin and an translucent insulated glass
inner skin (Detail Magazine 2004, 92-95). A second example is the Rafael Moneo's design for the
Kursaal Congress Centre and Auditorium in San Sebastian, Spain that has a 2.5 m (8.2 ft) buffer
facade with a translucent outer pane and sandblasted inner pane of laminated glass (2004, 242-
243).
5.4.7 Sealed, No Access, Pressurized Box-Windows
Classically, mechanized systems within the cavity, such as shading devices, were thought to
require accessibility for maintenance purposes. The lack of a hermetic seal lends the cavity to
possible air and moisture infiltration, resulting in dirt and condensation on the inner glass surfaces
and further heightening maintenance strategy. This thinking is progressing throughout the facade
industry, both in the United States and internationally. Project facades are more frequently
integrating closed cavity solutions that are sealed, not ventilated, and introduce a dry air supply to
mitigate condensation, dust and dirt (De Bleecker, et al. 2012). One such project is the new
Roche Diagnostics AG tower in Rotkreuz, Switzerland (see Section 4.3.8), where Josef Gartner
GmbH introduced a closed cavity facade (CCF) that is completely sealed, has a single-pane
exterior lite, 7.5 in (190 mm) cavity, triple glazed interior lite, perforated aluminum louvers, and
continuously supplies the cavity with clean, dry air to prevent condensation and eliminate the
need to clean the interior glass surfaces (Hell and Kaltenbach 2011). This approach minimizes
the cleaning and maintenance while preserving the advantageous thermal insulation, protected
shading, and acoustical performance qualities of the box-window double-skin facade.
Another project utilizing the pressurized box-window approach is the New Stanford Hospital (see
Appendix C.36) where maintenance and serviceability are, once again, two predominant
concerns for the client and design team. Neither the inner or outer skins are operable. The
resulting 16 cm (6.3 in) deep cavity space creates a thermal buffer and houses 10 cm (4 in) deep
235
horizontal blinds. The solution selected to balance the no-maintenance and no-condensation
requirements was a pressurized double skin facade that introduces a low-level positive airflow or
dry air into each unit's cavity. This is achieved by a source compressor on each floor supplying a
small volume of filtered and processed air through individual feeder loops. This solution is an
effective means of preventing condensation and the deposition of particulate matter within the
cavity. Using modular box-window systems and moving towards a compartmentalization of
controls appears to hold many appealing advantages.
5.4.8 Increased Permeability
A recent trend towards DSFs with increased permeability in the outer skin is also occurring in the
United States - see global trend in Section 4.5.5. Examples of multi-story DSFs with increased
permeability include the Cedars-Sinai Advanced Health Sciences Pavilion (see Appendix C.28)
and the Weill Cornell Medical College - Beller Research Building (see Appendix C.29). Both of
these project's feature facades allow air to enter and exhaust at each floor level; the 0.9 m (3 ft)
cavity of the Cedars-Sinai Advanced Health Sciences Pavilion is more regular and linear than the
Weill Cornell Medical College double-skin which has a varying depth due to a faceted outer skin.
One concern with projects such as these which have fixed openings, cannot be open and closed
via controls, on the outer skin is the ability for birds to enter and nest in the cavity space. In the
case of the Weill Cornell project, there are appropriately spaced tension wires at the opening
locations to minimize this. Increased permeability is also evident in the University of Baltimore
Angelos Law School (see Appendix C.26) where the outer skin is not sealed between the glass
lites, creating baffles in front of an otherwise multi-story cavity space that is 25 cm (10 in) deep.
With regular frequency of openings, there is greater control over the pressure within the cavity
space, potential for outer skin load reduction and exhaust strategies to reduce overheating.
236
5.5 Design Considerations Going Forward
Since the advent of the modern double-skin facade, glazing technology has evolved
tremendously with innovations occurring at a microscopic scale (selective coatings, gas-fillings,
etc.), product scale (triple-glazing, larger glass sizes, curved glass, etc.), system scale (deep
skins, exterior shading, operable controls, etc.), and macro-scale with growing and increasingly
competitive global supply chains. This progress continues to bring increasing attention to the
double-skin facade. As a result, the evaluation of performance attributes must become more
intensive with this facade system type, requiring a critical evaluation of energy performance, life
cycle assessment, embodied energy, recyclability, and cost-benefit. The goal is to create a high
performance enclosure, not an "expensive trouble maker" (Hens, et al. 2008, 1268); yet, the
history of DSFs shows this is easier said than done. DSF technology is not intrinsically high
performance, but careful application of the technology in appropriate applications can produce
high-performance solutions that further the sustainability of the built environment.
In conclusion, the following summary observations are developed as relevant to evaluating the
appropriateness of DSF technology to a given application, and to the development of a DSF
solution for such an application.
5.5.1 Evaluation and Simulation
Design decisions are informed through simulation and modeling throughout the design process.
So why do we end up with over-estimating projections for double-skin facade energy
performance?
While energy modeling is a valuable tool in developing an approach, these systems must be
validated, monitored, and used to inform the next generation of double-skin facade systems. This
process includes thorough commissioning efforts, post-occupancy monitoring and evaluation, and
dissemination of the findings to the architectural/engineering/construction (AEC) community so
the processes of energy modeling and life cycle assessment are informed through a feedback
237
loop; energy models then can be calibrated based on real performance, and design and
construction practices can be modified to eliminate identified shortcomings.
5.5.2 Sensibility and Marketability
The mere use of double-skin facade technology should not be automatically equated with high
performance. Many variables, such as site, climate, program, cost-benefit, energy consumption
and sustainability goals, must be considered in selecting and developing an appropriate facade
strategy. High performance demands the integrated consideration and optimization of many
attributes, among them: energy efficiency, environmental impact, safety, security, durability, cost
benefit, sustainability, functionality, maintenance and other operational considerations, over the
full lifecycle of the building. Though a project may garner attention or visibility for incorporating a
double-skin facade, the AEC community must differentiate between high-performance and
greenwashing, demanding performance metrics that justify and substantiate the application of
any high-performance facade technology including DSF.
Exterior Shading
Evaluation of a DSF system generally includes an energy simulation compared to a baseline
building with a conventional single-glazed facade. This, of course, is not a balanced comparison.
Considering the higher investment associated with DSF's, the baseline facade system should be
a high-performance single-glazed facade that transfers some of the cost differential into other
technological solutions. If a DSF is being considered, the baseline of comparison must include a
well-insulated single-skin facade (perhaps a triple-glazed solution) with exterior solar shading -
though, exposed exterior shading possesses its own set of considerations, including: durability,
maintenance, and aesthetics. When the evaluation between two refined facade alternatives - a
high-performance single-skin facade and a double-skin facade - is genuine, the externally
shaded single-skin is likely to show more reliable performance with respect to energy, thermal
comfort, daylighting, maintenance and fire-safety. Though an in-depth evaluation may debunk
some perceived advantages, the double-skin facade can likely uphold its reputation as a superior
238
acoustic insulator, as well as a strategy for introducing natural ventilation in conditions that would
not permit solely with a single-skin facade.
Responsiveness
The use of a double-skin facade system may incorporate several levels of intelligence:
climatically responsive controls for airflow intakes and exhausts, environmentally responsive
shading, and occupant responsive overrides for comfort Identifying the appropriate levels of
dynamics and responsiveness without inundating the system with controls or operable
mechanisms is important in developing an active facade that will sustain its performance over the
long-term.
Cost-benefit and Lifecycle Costing Analysis (LCCA)
The feasibility of a double-skin facade depends on its performance over the service life of the
building. High-performance features like DSFs will not be justified on a short-term payback basis.
This is why applications are almost exclusively found among institutional building owners or as
commercially owner-occupied buildings. LCCA is the appropriate costing methodology for these
applications, and ultimately for all applications if a sustainable built environment is to be realized.
Operations and Maintenance Access
With twice the number of surfaces to clean, shading devices that require maintenance, and
sensors and control systems to integrate the dynamic components, providing access to the cavity
space is imperative (or have a strategy that doesn't require cavity access and cleaning). This is
one potent justification for using a multi-story approach with a deeper cavity. A prudent DSF
evaluation includes an operation and maintenance plan from the early conceptual design stages,
and considers the increased maintenance costs as part of the LCCA
Retrofit
The potential of double-skin facades in retrofit projects may be an effective and economical
239
alternative to new construction. This approach has rnany potential benefits, including:
preservation of existing building stock, optimal reuse of existing systems and assemblies,
aesthetic upgrade (modernization), improved acoustical insulation, reduced energy consumption,
and provision of natural ventilation. Oivner-occupied buildings, institutional or otherwise, best fit
the profile for facade retrofit candidates. Identifying the appropriate enclosure solution for a
specific building program depends on the condition of the existing building, as well as the
intended use. Arriving at an appropriate facade solution, DSF or not, requires an iterative
process that balances existing conditions with project performance, and economic goals.
5.5.3 Integration and Optimization
Though the balancing and optimization of the rnany considerations characteristic of a high
perforrnance facade is essential, it is the integration of the facade systern in the whole-building
design that produces a successful building over its life span. Where the facade begins and ends
is becoming increasingly ambiguous in the wake of this integration process. High-performance
buildings typically involve complex networks of sensors and controllers coupled to a building
rnanagernent systern that drives dirnrnable lighting systems, dynamic shading systems,
ventilation systems, and mechanical equipment, all responding to changing conditions of interior
and exterior environment, as well as to occupant overrides. A high-performance facade does not
necessarily rnake for a high-performance building, but when properly integrated into a high
perforrnance whole building design strategy, the facade can be a primary contributor to achieving
high-performance buildings, and ultimately, to the sustainability of the built environment.
240
Chapter 5 Endnotes
1
The 23 projects from 201 O or prior are the same profiled and summarized in (Vaglio and
Patterson 2011 ).
2
Kerry Hegedus' presentation, Three Case Studies: Double Skin Facade Design, at the IQPC
Facades Design and Delivery conference, January 24-25, 2011.
3
Ibid.
4
Clifford Gayley's presentation, Sustainability: Tackling Environmental Efficiency and
Performance, at the IQPC Facades Design and Delivery conference, January 25, 2011.
5
Case study is excerpt from Vaglio, Patterson, Hooper and Noble (2010).
6
Norm Shane's presentation, The Cable Stayed Point Supported Chamber Wall's West Coast
7
8
Debut, at the Facade Tectonics #6 conference, July 29, 2011.
Case study is excerpt from Vaglio, Patterson, Hooper and Noble (2010).
Stefan Behnisch's presentation, Technical and Architectural Expectations: The Rapidly
Developing Role of the Building Skin in the Wake of New Technologies, at the
facades+PERFORMANCE conference, October 24, 2013.
9
Ibid.
10
The photos of Columbia University Jerome L. Greene Science Center were taken June 3, 2015
in the middle of construction. There are sheets of plywood in the photos located against the
interior surface of the interior skin that serve as a protective barrier to prevent damage from
work performed by other trades during the remainder of the construction schedule. The
plywood is a temporary measure and will be removed prior to completion.
11
Ibid.
12
Ibid.
13
The Nelson Atkins Museum in Kansas City was not included in the case studies presented in
Section 5.3 and Appendix C. One could argue that it should be included; however, the author
has made a subjective call related to the particulars of the definition outlined in Section 3.1.3.
Since the buffer facade at Nelson Atkins does not fulfill the "introduction of a second
transparent glazed layer", it was excluded. By that measure, one could argue that Loblolly
House (Section C.10) and Museum of Contemporary Art Denver (Section C.12) also do not
belong in this list However, these two were included because the do behave beyond a buffer
mode: Museum of Contemporary Art can function in an air exhaust mode; Loblolly functions as
an outdoor air curtain. So yes, these three projects lie in the grey area or fringe of the definition.
A subjective call was made by the author to include two, but not all three projects.
241
6. Methodology
This section presents the research method as it pertains to this research, including: hypotheses,
research objectives, the research design and procedures for testing and data analysis.
6.1 Hypotheses
Hypothesis 1: Computational fluid dynamics with multiphysics software is an essential
design tool in the determination of pressure coefficients for multi-story double
skin facades.
Hypothesis 2: Airflow openings in multi-story double-skin facades have significant effects on
the design pressure requirements.
6.2 Objectives
The predominant conversations regarding double-skin facades have to do with their perceived
energy benefits. All the while, the unique structural issues presented with double-skin facades are
largely over-shadowed. This research aims to shed light on some of the structural-related
concerns present in double-skin facades, multi-story configurations specifically, by 1) articulating
the existing approaches and challenges to determination of wind pressure coefficients, 2)
identifying special considerations regarding their technical design, and 3) considering how the
design-delivery process may have to evolve.
Existing Problem
In surveying the existing practice of implementing double-skin facades, two questions are of
primary concern in understanding the current context:
242
• What are the most common types of double-skin facades implemented in
the United States?
• What are the current methods of determining wind pressure coefficients
for multi-story double-skin facades?
Special Considerations
In the empirical portion of this research - computational fluid dynamic simulations - several
specific structural questions are initial ambitions for the scope of the research:
• How do various airflow configurations impact the pressure coefficient
profiles of multi-story double-skin facades?
• When a double-skin facade employs a long-spanning exterior skin, at
what point, if any, does fluid-structure interaction (FSI) need to be
considered for a flexible structure supporting the outer layer of a
multi-story double-skin facade?
1
Design Process
Following a review of the existing context of double-skin facade implementation and practice in
the United States, and extensive analysis via CFO simulations, how the design and delivery
process of double-skin facades may evolve will be envisaged, including:
• To what extent may computational fluid dynamics be used to simulate
atmospheric boundary layer wind tunnel tests of double-skin facades?
• What role and at what points in the process of determining wind pressure
coefficients is computational fluid dynamics most appropriate?
243
6.3 Research Design
The research design is comprised of three primary phases: Discovery, Analysis and
Contributions. These primary phases envelop the seven facets of this research, each of which is
described in this section and summarized in Figure 6-1. The location of each facet in this
document is also stated:
6.3.1 Identification of the Problem
• Current Methods and Practices (Section 1.1 to 1.4):
Describe wind pressure distribution characteristics, unique
considerations of them in double-skin facade applications, how current
codes address wind loads on DSF's, what approaches in determining
wind pressure coefficients are used in practice, and new approaches that
may be useful in advancing the methods of wind pressure determination
for double-skin facades.
• Typological Trends (Chapter 7.1)
Gather project information (project name, year, project team, solar
orientation) and evaluate typological trends (ventilation type, ventilation
mode, cavity partitioning: height, depth, width and aspect ratio, airflow
intake and exhaust, and solar control characteristics) of built double-skin
facades in the United States. Create a brief case study of each of these
projects (Appendix C).
• Identify the Most Common Configuration (Section 7.1):
Using a similar evaluation method as Perino (2007), statistically evaluate
the case studies and extract a subset of the most common typological
double-skin facade configuration - based on the project information
gathered for DSFs in the United States.
244
6.3.2 Development of an Analytical Prototype
• Detailed Evaluation of Most Common Configuration (Section 7.2):
Isolate the most common configuration type and statistically evaluate
each project's characteristics: dimensional (height, width, depth and
aspect ratio), airflow type (mechanical and/or natural) and airflow intake
configuration.
• Develop an Analytical Prototype (Section 7.3):
Use the statistical values generated in the evaluation of the most
common configuration to set fixed geometric dimensions for the building
and double-skin facade while allowing the airflow configurations to
remain variable. Create a typological framework of airflow intake and
exhaust configurations to be considered.
6.3.3 Calibration of Analytical Process
• Calibrate CFO to Baseline Wind Tunnel Studies (Section 8.4 to 8.5):
Simple Cube: The first calibration model is a simulation similar to that of
the wind tunnel and full-scale studies in Holscher and Niemann (1998)
that use a 50 m simple cube scaled to ,l = 1/100 with an initial velocity of
U
0
= 7 mis in a 9 m x 3 m x 2 m test section. Vary the power law
exponent (a), turbulence intensity (Ir, mfr,) and roughness length (z
0
) to
gain an understanding of how sensitive the simulation flows - and
resultant pressures - are to these variables.
• Calibrate CFO to DSF Wind Tunnel Studies (Section 8.6 to 8. 7)
Unsheltered and Sheltered Towers: Simulate the 1) unsheltered tower,
2) sheltered tower with a cavity open on all sides (Layout A), and 3)
sheltered tower with full lateral closure (Layout 0) of Marques da Silva
and Gomes' (2005, 2008)
2
wind tunnel experiments using a three-
245
dimensional steady-state flow simulation. Use the research's model
dimensions of a = 32.5 cm wide, b = 20 cm deep and h = 70 cm tall, and
the same tower with a second-skin with the outer skin beginning at h"""=
7.5 cm above ground level for Layout A and Layout 0, at a model scale
of ,l= 1/40 in a wind tunnel section of Lwr= 9 mx Wwr= 3 mxHwr= 2 m
with an inlet velocity of U
0
= 10.5 mis and velocity profile power law
equation exponent of a= 0.18. Compare the results from the CFO
simulations to the results of the wind tunnel experiments.
• Identify Analytical Parameters (Section 8.8):
Use the comparisons of CFO simulations to the power law results of the
wind tunnel experiments to determine the analytical parameters moving
forward, particularly the unknowns of inlet turbulence intensity (fr. min)
and roughness length ( z
0
).
6.3.4 Simulations of Multi-Story Double-Skin Facades
• 20 Transient Simulations of Multi-Story DSFs (Appendix E)
Conduct an expedited investigation into the sensitivity of the analytical
prototype to 1) different dimensional variables (height, depth and
permeability) and 2) various airflow intakes (face, trench, raised) and
exhaust (forward, through and return). These simulations will be quicker,
but their results may be inaccurate. Use the results to prioritize which
variables should be studied further and gain additional understanding to
the transient behavior of the systems.
• 30 Steady-State Simulations of Multi-Story DSFs (Section 9.1to9.7)
Conduct extensive steady-state simulations utilizing a three-dimensional
simulation domain representative of the wind tunnel test section with a
Reynolds Averaged Navier Stokes (RANS) turbulence model type with a
246
k-B turbulence model with incompressible flow. Maintain the parameters
developed from the calibrations and the multi-story analytical prototype
determination. In the simulation environment maintain an open-circuit
wind tunnel with a Lwr = 9 m x Wwr = 3 m x Hwr= 2 m test section, inlet
velocity of U
0
= 10.5 mis and velocity profile power law equation
exponent of a= 0.18; all similar to Marques da Silva and Gomes (2005;
2008). Varying the airflow inlet and exhaust, simulate the twelve possible
combinations of inlet (face, trench, raised and shingled) and exhaust
(forward, through and return) configurations.
6.3.5 Analysis of Results
• 30 Steady-State Results Comparisons (Section 9.8):
First understand each model's wind pressure distribution characteristics
by extracting 1) contour elevations for both the interior and exterior skins,
and 2) extracting characteristic profile curves at key vertical and
horizontal cut locations. Use the local pressure coefficients for each of
the inner and outer skins to determine the net pressure coefficients
occurring across the outer skin. Second, evaluate the results of each
airflow intake/exhaust combination to one another's by comparing the
pressure coefficients on both of the interior and exterior skins, the outer
skin pressure of each double-skin to that of a single-skin, and the net
pressure coefficients that occur across the outer skin. Identify common
and relative behaviors for each variable type.
• 30 Steady-State Compared to 20 Transient (Section 9.9):
Qualitatively compare the findings from the 30 steady-state analysis to
those of the preliminary 20 transient studies to identify any supporting
affirmations or conflicting findings.
247
CALIBRATE
3. Calibration of Analytical Process
3.1 Calibrate CFD analysis procedure
to baseline wind tunnel studies
3.2 Calibrate CFD analysis procedure to
wind tunnel studies of multi-story
double-skin configurations
PROTOTYPE
2. Develop Prototype
2.1 Evaluate, in detail, the sub-typology
of the most common double·skin
facade configuration
2.2 Develop an analytical prototype
3.3 Identify analytical parameters
IDENTIFY
1 . Identification of the problem
1.1 Current methods and practices of
determining pressure coefficients
1.2 Evaluate DSF's in the United States
for typological trends
1.3 Extract a subset of the most common
typological configuration
Figure 6-1: Overview of research methodology.
GUIDELINES
7. Develop Guidelines for:
SIMULATE
4. Simulation of Multi-Story DSF Prototype and Variables
4.1 2D steady-state and transient simulations of multi-story
DSPs with varied airflOIN opening configurations
4.2 3D steady-state simulations of multi-story DSF's with
varied airflow opening configurations
ANALYZE
5. Analysis of Results
5.1 Compare the results of different airflow
configuration combinations' 3D steady-state
simulations to one another's
5.2 Compare the results of different airflow
configuration combinations' 3D steady-state
simulations to that of a single-skin facade
REFLECT
6. Evaluate the Role of CFD in the Design Process
6.1 How does CFD for the determination of wind-pressures
fit into each stage of the design/delivery process?
7.1 Multi-story double-skin facades with various airflow configurations
7.2 A design methodology for the determination of pressure coefficients for multi-story double-skin facades
6.3.6 Reflect on the Role of CFO
• Evaluate the Role of CFO in the Design Process (Section 10.3):
Review the successes and challenges presented by the simulation
approach utilized to replicate wind tunnel studies of double-skin facades.
Develop a set of guidelines for when, why and how computational fluid
dynamics may serve as a useful tool in studying wind pressure behavior.
Additionally, consider how computational fluid dynamics may serve as a
supplemental approach to better inform earlier design decisions for multi
story double-skin facades.
6.3.7 Develop Guidelines
• Multi-Story Double-Skin Facade Guidelines (Section 10.2):
Integrate the outcomes and insights from the analytical models about
wind pressure distribution of double-skin facades with various airflow
configurations into a series of guidelines and tools that address the
advantages and disadvantages of each compared to one another, as
well as a relative understanding of how such configurations perform
compared to a single-skin facade. Also highlight areas that require
special consideration, as well as those that require further study to
understand their potential ramifications.
• Design Methodology of DSF Pressure Coefficients (Section 10.4):
Outline a methodological approach for determining double-skin facade
pressure coefficients within a broader context of the whole building
design and delivery process. Develop a series of decision-making
matrices to aid the user in weighting approach options for paths towards
adequate wind pressure information.
249
Table 6-1: Summary of Research Method Phases and Facets
>
a:
w
>
0
(.)
"' Ci
~
w
"'
(
J:
Q.
!!l
Objective
IOENTIFICA TION OF PROBLEM
1.1 Current methods and practices of
determining pressure coefficients
1.2 Evaluate DSF's in the United States for
typological trends
1.3 Extract subset of the most common
typological configuration
DEVELOP PROTOTYPE
2.1 Evaluate, in detail, the sub-typology of
the most common double-skin facade
configuration
2.2 Develop an analytical prototype
CALIBRATION OF ANALYTICAL PROCESS
3.1 Calibrate CFO analysis procedure to
baseline wind tunnel studies
3-2 Calibrate CFO analysis procedure to
specific wind tunnel studies of multi
story DSF configurations
3.3 Identify analytical parameters
Type of Data
case study
geometric
dimensions
case study
geometric
dimensions
pressure
coefficients
pressure
coefficients
Source of Data
building codes. literature, web-
based databases, conference
proceedings, periodicals,
architects, fayade engineers
literature, web-based databases.
conference proceedings.
periodicals, architects. fai;ade
engineers
project drawings, models (when
available), communication with
architects, literature, periodicals,
field measure, scaling of photos
or diagrams
project drawings, models (when
available), communication with
architects, literature, periodicals,
field measure, scaling of photos
or diagrams
existing wind-tunnel studies
existing wind-tunnel studies of
multi-story double-skin facades
input variables findings from 3.1 and 3.2
~ 4 SIMULATION OF MULTI-STORY DSF PROTOTYPE & VARIABLES
..J
~ 4.1 20 steady-state and transient pressure COMSOL simulations:
ct: simulations of multi-story DSF's with coefficients computational fluid dynamics
«
w
~
Q.
"'
z
0
;:::
:::>
ID
a:
,_
z
0
(.)
.-;
w
~
J:
Q.
4.2
varied airflow opening configurations
30 steady-slate simulations of multi
story DSF's with varied airflow opening
configurations
ANALYSIS OF RESULTS
5.1 Compare the results of different airflow
configuration combinations' 30 steady
state simulations to one another's
5.2 Compare the results of different airflow
configuration combinations' 30 steady
state simulations to that of a single-skin
fao;ade
pressure
coefficients
pressure
coefficients
pressure
coefficients
EVALUATE THE ROLE OF CFO IN THE DESIGN PROCESS
6.1 How does CFO for the determination of
wind pressures fit into each stage of the
designldelivery process
DEVELOP GUIDELINES FOR:
7.1 Multi-story double-skin facades with
various airflow configurations
7.2 A design methodology for the
determination of pressure coefficients
for multi-story double-skin facades
guidelines
guidelines
methodology
COMSOL simulations:
computational fluid dynamics
COMSOL simulations·
computational fluid dynamics
COMSOL simulations:
computational fluid dynamics
Variables
year, location, project team, orientation,
typological characteristics (cavity
partitioning. ventilation type, ventilation
mode), solar shading characteristics
year, location, project team, orientation,
typological characteristics (cavity
partitioning, ventilation type, ventilation
mode), airflow opening configurations
orientation, dimensions (depth, height.
width and aspect ratio), ventilation type,
ventilation mode, airflow opening
configurations
wind flow characteristics: power-law
exponent, turbulence intensity, terrain
roughness
wind flow characteristics: turbulence
intensity, terrain roughness
DSF characteristics: layout, cavity
depth, orientation
wind flow characteristics: turbulence
intensity. terrain roughness
dimensional:
height, depth, opening area rabo
airflow configurations:
intake and exhaust
airflow configurations:
intake and ex~aust
pressure coefficients:
C ,..1 111 and C P.':r'
net pressure coefficients:
C p. 11.·1
outer skin ratios:
C p.DSF /)fifer I C p.Sing!.•
Method
literature review
literature review
statistical
evaluation
observation I
interpretation
statistical
evaluation
quantitative:
simulation,
comparative
quantitative ·
simulation,
comparative
comparative
quantitative:
simulation
quantitative:
simulation
quantitative
comparison
quantitative
comparison
qualitative
qualitative
qualitative
250
6.4 Procedures
All simulations are computational fluid dynamics models of wind tunnel test sections with scaled
building geometries contained within. The calibrations occur at two different model scales, ,l =
1/100, 1/40), and two different reference velocities, U
0
= {7 mis, 10.5 mis}. This is because two
different wind tunnel research papers are referenced for calibration purposes: one of a simple
cube and the other with an unsheltered tower that is also examined with second-skin
configurations. All of the double-skin configurations' simulations that make up this research rely
on the latter parameters, ,l = 1/40 and U
0
= 10.5 mis. The three-dimensional steady-state
simulations, representative of the wind tunnel test section, incorporate a Reynolds Averaged
Navier Stokes (RANS) turbulence flow model type with a k-B turbulence model with
incompressible flow. The simulations are conducted in the CFO module of COMSOL Multiphysics
5.0.0.244. The process for the simulations requires the following steps:
• Modeling the wind tunnel test section and scaled building geometry
• Modeling the airflow within the simulated wind tunnel section
• Assigning the turbulent flow properties
• Assigning boundary conditions, including inlet, outlet, floor and wall parameters
• Assigning regions/edges of forced, user-controlled mesh subdivision
• Meshing the geometry with a fluid dynamic specific mesh
• Assigning the analytical solver properties
• Computing the solution
• Setting up data sets for data extraction and post-processing
• Generating numerical results output
• Generating graphical results output
• Generating comparative numerical results output
• Generating comparative graphical results output
251
6.5 Data Analysis
Once a simulation model successfully converges and its runtime is complete, the model is
prepared for results review, post-processing and data extraction. The data extraction occurs at
the inner and outer skins as surface color maps, or contour maps, of the resultant pressure
coefficient. Additionally, the results are extracted at several vertical and horizontal linear
sectional cuts to generate profiles for comparison. These linear cuts occur at mid-span, near
edges, near the ground and in between, and may provide a clearer view of how the pressure
behaves across the building face. At each of these evaluation sections, the results are compared
in several ways, including:
• the pressure coefficients on both the interior and exterior skins, CP·'"'and cp.n" at mid
span, x/a = 0.50
• outer skin pressure coefficient ratios, cp.DSFcma I cp.smgfr - normalized to the values of a
single-skin configuration - at x/a = {0.50, 0.67, 0.90) in the vertical direction
• outer skin pressure coefficient ratios, Cp.DSFcma I Cp.smgfr - normalized to the values of a
single-skin configuration - at zlh = {0.10, 0.50, 0.67, 0.90) in the horizontal direction'
• net pressure coefficients, CP·"'" at x/a = {0.50, 0.67, 0.90), for each of the simulations.
For consistency purposes, the data from the double-skin facade models are conveyed within the
same ranges for their respective body of information: all contour elevations report the CP values
within a surface color range of -2.00 to 1.00; all CP profile cuts are graphed on an x-axis between
-1.50 and 1.50; all Cp,ner values are graphed on an x-axis between 0 and 2.00; and all Cp,DSFcma I
Cp,smgfr ratios are graphed on an x- or y-axis between 0 and 2.00.
The methods of data analysis are further elaborated on in Section 9.2.5.
252
Chapter 6 Endnotes
Though fluid-structural interaction was an initial ambition of this research, it does not exist
within the results of this research and is acknowledged as future work.
2
The difference between the data presented in Effects of different multi-storey double skin
facade configurations on surface pressures (2005) and Gap inner pressures in multi-storey
double skin facades (2008) was clarified by Fernando Marques da Silva via email
correspondence on October 13, 2011. There is no difference in the data common to the two
papers. The only differences are the extent presented (in 2008, three cavity depths were
studied compared to just one, 0.8 m, in 2005) and the method of presenting the results. The
results for the double-skin configurations in Effects of different multi-storey double skin facade
configurations on surface pressures (2005) were communicated as the net tiCP = cp,osF -
Cp,unsht, where as the same configuration results in Gap inner pressures in multi-storey double
skin facades (2008) were presented as the measured values themselves, cp,osF·
3
The zlh = 0.10 horizontal profile is excluded for the models with Raised inlets since the exterior
skin does not exist in this region.
253
7. Development of a Multi-Story Prototype
Within this dissertation, a search was conducted to identify existing applications of double-skin
facades in the United States to understand the typological distribution used in existing and future
buildings. The search was performed by reviewing 1) existing literature in the form of books,
conference proceedings and journal articles, 2) attending technical conferences where facades
were the topical focus, 3) inquiries with architects, 4) interfacing with United States based facade
contractors familiar with forthcoming projects, and 5) internet articles. Many of the existing
literature review were common to those that informed Chapters two through five. Thirty examples
of built - and documented - buildings utilizing a double-skin facade concept were identified in the
United States. Some projects alluded to but with little to no documentation were excluded. One
such example is the Prudential Life Insurance Company in Princeton, New Jersey that was
designed by Skidmore Owings and Merrill in collaboration with Princeton University in the late
1980s (T. M. Boake 2002, 64). This three story building featured a double exterior glass wall with
a cavity depth of approximately 46 cm (18 in) that was used to harness solar heat gain and
redistribute to building zones that require heating (Horsley 1981). In addition to the thirty built
examples, six forthcoming examples that had substantial information were documented to
demonstrate the fact that double-skin facades in the United States are continuing to be explored
through application in the present and near future. These six forthcoming projects were excluded
from any data analysis presented in this chapter, though they are shown in some tables to
present the double-skin facades' characteristics.
The projects were reviewed and summarized in a manner similar to Perino's evaluation of 215
DSF/AIF projects across the globe (State of the Art Review Vol. 2A: Responsive Building
Elements 2007) with a focus on typological factors: cavity partitioning, ventilation type and
ventilation mode ('flow path' per Perino). The review was used to identify the most common DSF
configurations to inform the development of an analytical prototype.
254
I\.)
(J1
(J1
Table 7-1: Summary of Double-Skin Facades in the United States
Project Name
Warren Petroleum Executive HQs
Occidental Chemical Center
Yazaki North American
Levine Hall - Pennsylvania School of Engineering
Seattle Justice Center
Manulife US Headquarters
Genzyme Headquarters
Foundry Square
UMass Medical School
Labially House
University of Michigan Biomedical Science Research Building
Museum of Contemporary Art
Comcast Tower
C> Riverhouse - One River Terrace
! Loyola Information Commons
"'
)( Walter Cronkite School of Journalism
w
Whatcom Museum
Art Institute of Chicago - Modern Wing
Cleveland Museum of Art - East Wing
Cambridge Public Library
100 Park Avenue
USC Eli and Edythe Broad CIRM Center
New York Presbyterian Hospital
University of Oregon Jaqua Center
New World Center
University of Baltimore Angelos Law School
Harvard Business School Tata Hall
Cedars-Sinai Advanced Health Sciences Pavilion
Weill Cornell Medical College - Seifer Research Building
University of Kansas School of Architecture - The Forum
Northwestern University Recital Hall
w Rodino Federal Office Building
~ A.J. Celebrezze Federal Building
I-
=> PNC Plaza Tower
...
Columbia University Jerome Greene Science Center
New Stanford Hospital
Notes:
Cavity Partitioning
B Barne
BW Box-Window
AF Alternating Faiyade
SB Shaft-Box
SH-C Story-Height Corridor
SH-JM Story-Height Juxtaposed Modules
MS Multi-Story
MS-S Multi-Story Shingled
MS-L Multi-Story Controllable (louvers)
Year
Location
City State
Tulsa
Niagara Falls
Canton
Philadelphia
Seattle
Boston
Boston
San Francisco
Worcester
Taylors Island
Ann Arbor
Denver
Philadelphia
New York
Chicago
Phoenix
Bellingham
Chicago
Cleveland
Cambridge
New York
OK
NY
Ml
PA
WA
MA
MA
CA
MA
MD
Ml
co
PA
NY
IL
AZ
WA
IL
OH
MA
NY
CA
NY
Architect
SOM, Bruce Graham
Cannon Design Inc.
Plantec
Kieran Timberlake Associates
NBBJ, Kerry Hegedus
Skidmore, Owings & Merrill, LLP
Behnisch, Behnisch and Partner
STUDIOS Architecture
Payette Associates
Kieran Timberlake Associates
Polshek Partnership Architects. LLP
David Adjaye w/ Davis Partnership
Robert AM. Stern Architects
Polshek Partnership Architects, LLP
Solomon Cordwell Buenz
Ehrlich Architects
Olson Kundig Architects
Renzo Piano Building Workshop
Rafael Virioly Architects
William Rawn Associates
Moed de Armas & Shannon Architects
Zimmer Gunsul Frasca Architects, LLP
Pei Cobb Freed & Partners
DSF
Orientation
N,S, E,W
N,S, E,W
E.W
E,W
S,W
N,S, E,W
s
N,S, E. W
W,S
w
S,E
NE, SE, SW
S. E, W
W,S,E, N
w
s
w
N,S.E
S, E, W
SW
E
E-SE
S-SW
Typological Characteristics
Cavity Ventilation Ventilation
Partition Type Mode
SH-C
MS
MS
BW
MS
SH-JM
SH-C
MS
MS
B
OAC
OAC
IAC
OAC
IAC
OAC
OAC
OAC
OAC
OAC
AE
AE
OAC
OAC
OAC
OAC
OAC,B
B
OAC
1957
1980
1999
2001
2002
2003
2003
2003
2006
2006
2006
2007
2008
2008
2008
2008
2009
2009
2009
2009
2009
2010
2010
2010
Los Angeles
New York
Eugene OR Zimmer Gunsul Frasca Architects, LLP N, S, E, W
SH-JM
MS
MS
MS
BW
MS
MS
MS
MS
SB
MS
BW
MS
MS
MS
n/a
N
N
M
H
M
N
N
N
N
N
M
M
N
H
N
N
N
M
N
N
N
M
N
M
N
N
N
N
N
B
OAC
AE
OAC
2011 Miami
2013 Baltimore
2013 Boston
2014 Los Angeles
2014 New York
2014 Lawrence
2015 Evanston
2015 Newark
2015 Cleveland
2015 Pittsburgh
2015 New York
2018 Palo Alto
Ventilation Type
N Natural
H Hybrid
M Mechanical
FL Gehry Partners
MD Behnisch Architekten
MA William Rawn Associates
CA HOK
NY Ennead Architects
KS Studio 804
IL Goettsch Partners
NJ Dattner Architects
OH Interactive Design Eight Architects
PA Gensler
NY Renzo Piano Building Workshop
CA Rafael ViMly Architects
Ventilation Mode
OAC Outdoor Air Curtain
IAC Indoor Air Curtain
AS Air Supply
AE Air Exhaust
B Buffer Zone
N MS
S. E. W B
E,W MS
E, W MS
S MS
S, E, W MS
S MS
N, S, E, W MS
N, S, E, W SH-C
N, S, E, W SH-JM, BW
S, E, W MS
N, S, E, W BW
M
N
N
N
M
M
B
OAC
OAC
OAC
OAC
OAC,B
IAC
OAC
B
AS,B
AE
B
Solar Control Strategy
Type Controls
Tinted Glass
Louvers
n/a
Blinds
Blinds
Blinds
Blinds
n/a
Catwalks
Translucent
Blinds
Translucent
Overhang
Blinds
Blinds
Translucent
Translucent
Blinds
Blinds
Louvers
Blinds
Translucent
Fabric Fins
n/a
Auto w/ Override
n/a
Automated
Automated
Manual
Auto w/ Override
n/a
Fixed
n/a
Manual
n/a
Fixed
Manual
Automated
n/a
n/a
Automated
Automated
Automated
Manual
n/a
Automated
Screen; Blinds Fixed; Automated
n/a n/a
Blinds Manual
Louvers Automated
Fins, Screens Fixed
Frits, Screens Fixed
Vertical Louver~ Automated
Blinds Automated
Frit Fixed
Louvers, Fins Fixed
Blinds Auto w/ Override
Blinds Automated
Blinds Automated
7.1 Identifying the Most Common Configuration
This section summarizes the 30 existing projects presented in Table 7-1 and compares the
findings to Perino's findings from a review of 215 DSF/AIF projects across the globe.
Cavity Partitioning: The cavity partitioning configuration that is most common in the United States
is the multi-story (70.0%) followed by the box-window (10.0%). In Perino's evaluation, the multi-
story represented 47.0% of the global projects surveyed (2007, 23).
Ventilation Type: The ventilation type that is most common in the United States is natural (69.0%)
followed by mechanical (24.1%). In Perino's evaluation, natural ventilation represented 58.1% of
the global projects surveyed (2007, 23).
Ventilation Mode: The ventilation mode that is most common in the United States is the outdoor
air curtain (70.0%). In Perino's evaluation, the outdoor air curtain represented 49.1% of the
projects surveyed (2007, 24). It must be noted that many DSFs feature an ability to switch
between modes (e.g. Riverhouse's units can be manually opened to function in an air supply
mode). The primary mode of operation is identified for the purposes of this analysis.
Cavity Partitioning
20
15
tl
a.>
l
0 10
O;
..c
E
:::>
z
5
B BW SB SH S H MS
C JM
Ventilation
Type
N M H B
Ventilation
M od e
AE AS OAC IAC
Figure 7-1: Typological summary of identified buildings with DSFs in the United States.
256
7.2 Dimensional Trends of Multi-Story Double-Skin Facades in the United States
Having identified the multi-story as the predominant cavity partitioning of double-skin facades in
the United States in the previous section, the 21 projects with one are now isolated and evaluated
for trends within this subset The dimensional and airflow characteristics are summarized in Table
7-2. It should be noted that many of the projects include multiple instances of double-skin facades
(e.g. Yazaki North American contains ten unique 20.5 m (67.3 ft) wide by 15 m (50 ft) tall
instances and the Art Institute of Chicago has three instances on the north, three on the south
and others on the east or west elevations). Where a project contains multiple instances, only one
(usually a repeating dimension across occurrences), the largest, was included. Dimensional data
was gathered from projects drawings, technical presentations, journal publications and periodical
articles. Where a dimension could not be determined from the aforementioned resources, it was
scaled from drawing or photograph and noted within the notes of Table 7-2.
Height. The multi-story cavities range in height from 5.5 m (18 ft) to 62.0 m (203.5 ft) with an
average of Ii= 21.1 m (69.1 ft).
Width. The multi-story cavities range in width from 5.5 m (18 ft) to 112.8 m (370 ft) with an
average of a= 46.8 m (153.6 ft).
Cavity Depth. The multi-story cavities range in cavity depth from 0.23 m (8.9 in) to 3.0 m (120 in)
with an average of s = 1.01 m (39.7 in).
Aspect Ratios: The multi-story cavities range in height-to-depth aspect ratio from 6.0 to 101.8 with
an average of h/s = 27.8. The average cavity height-to-width is h/a = 0.672 meaning the cavities
are more frequently wider than they are tall. The average cavity depth-to-width is s/a = 0.031.
257
I\.)
(J1
CXl
w
Cl:
::::>
I
::>
LL
Table 7-2: Dimensional Summary of Multi-Story DSF's in the United States
Dimensional Characteristics
a h s
DSF
Project Name
Orientation
Width Height Cavity Depth
[ft] [ft]
Occidental Chemical Center
1
N,S, E,W 143 120
Yazaki North American E,W 67.3 50
Seattle Justice Center
2
S,W 130.5 130.5
Foundry Square N,S, E,W 116 94
UMass Medical School
3
W,S 180 65
University of Michigan Biomedical Science Research Building S, E 324 66
Museum of Contemporary Art
5
NE, SE, SW 125 46
Comcast Tower S,E,W 130 45
Loyola Information Commons w 150 56
Walter Cronkite School of Journalism
6
s 18 71
Whatcom Museum w 180 36
Art Institute of Chicago - Modern Wing
7
N, S, E 46 43
Cambridge Public Library SW 180 40.5
USC Eli and Edythe Broad CIRM Center E-SE 200 68
New York Presby1erian Hospital S-SW 135 70
University of Oregon Jaqua Center
8
N,S, E,W 125 51
New World Center
9
N 102 18
Harvard Business School Tata Hall E,W 300 37
Cedars-Sinai Advanced Health Sciences Pavilion
10
E,W 370 120
Weill Cornell Medical College - Seifer Research Building
11
s 126 203.5
Universit~ of Kansas School of Architecture - The Forum S,E,W 76.8 20.5
AVERAGE (imperial) 153.6 ft 69.1 ft
AVERAGE (metric) 46.8 m 21.1 m
Northwestern University Recital Hall
12
s 40 50
Rodino Federal Office Building N,S, E,W 265 196
Columbia Universit~ Jerome Greene Science Center S,E,W 125 95
READJUSTED AVERAGE (imperial) 152.3 ft 74.7 ft
READJUSTED AVERAGE (metric) 46.4 m 22.8 m
Notes:
1
Width is based on 29 glazing modules each at 4.92 ft (1.5 m)
' Height is based on a 4.5 ft (1.4 m) overrun at the top and assumes 14 ft (4.3 m) floor-to-floor heights
3
Height is based on 5 floors at 13 ft (4.0 m) spacing; width is based on 6 structural bays each 30 ft (9.1 m) wide
4
Width is based on 9 ft (2.7 m) glazing modules with 36 of 38 encompassing the DSF cavity, Murray (2009, 176-183)
5
Width and height are scaled from 1:500 drawings, Museum in Denver (2008, 119-123)
6
Width is scaled from photo, Ehrlich (2013, 54)
7
Width and height are scaled from drawings available at WNw.fondazionerenzopiano.org
[in]
47.2
120
30
54
36
48
8.9
51.5
36
12
24
32
36
36
30
60
36
36
36
24
41
39.7
1.01
60
57
14
40.2
1.02
6
Width is based on 25 glazing modules each at 5 ft (1.5 m); height scaled from drawings with floor-to-floor height of 13.5 ft (4.1 m)
9
Width is based on 17 glazing modules each at 6 ft (1.8 m)
10
Width is based on 37 bays times two glazing modules each at 5 ft (1.5 m); height is based on 8 floors at 15 ft (4.6 m) each
11
Depth varies from 1.7 ft (0.5 m) to 3 ft (0.9 m) across undulating fayade; depth of 2 ft (0.6 m) averaged
12
Width is based on 5 glazing modules each at 8 ft (2.4 m)
Airflow Intake Configuration
his
Ventilation Inner External
Aspect
Mode Source Source
Ratio
30.5 OAC Trench
5.0 OAC Trench
52.2 OAC Raised
20.9 OAC Raised
21.7 OAC Raised
16.5 OAC Raised
62.0 AE Exhaust
10.5 AE Exhaust
18.7 OAC Trench
71.0 OAC Raised
18.0 OAC Trench
16.1 OAC, B Raised
13.5 OAC Trench
22.7 OAC Trench
28.0 AE Exhaust
10.2 OAC Trench
6.0 B Sealed
12.3 OAC Trench
40.0 OAC Raised
101.8 OAC Raised
6.0 OAC Raised
in 27.8
m
10.0 Sealed
41.3 Raised
81.4 Exhaust
in 29.8
m
7.3 Multi-Story Double-Skin Facade Prototype
As shown in Table 7-2, the outdoor air curtain has been utilized on 17 of the 21 multi-story DSFs
(81.0 %), an increased rate compared to the 30 project sample of all cavity partitioning
configurations. A multi-story double-skin facade with an outdoor air curtain with dimensions of
cavity top height of htop = 21 m (68.9 ft), building height of h = 20 m (65.6 ft), width of a = 45 m
(147.5 ft) and baseline cavity depth of s = 1.0 m (3.28 ft) was deemed the prototype configuration
for analysis. In the following analysis chapters, cavity depth, building dimensions and opening
area ratio are fixed, whereas, airflow configurations are treated as variables.
h
hstart
h
Figure 7-2: Multi-story double-skin facade prototype variations: raised/entry (top) and
trench (bottom).
259
7.4 Airflow and Structural Variables
This section identifies, defines and characterizes the variables considered in the simulation
models specific to this research (implemented in Chapter 9). The scope of these simulations treat
the variables of depth, height, opening area ratio and wind incidence as constants, as outlined in
Section 7.6. The geometric dimensions of height and depth are fixed because 1) the common
configurations outlined in Section 7.2 are driven by practical considerations, such as minimum
dimensions required to access and perform maintenance tasks within the cavity space, and 2)
preliminary two-dimensional transient simulations (see Appendix E) suggested that these
variables had less of an impact on pressure coefficients as compared to opening area ratio and -
even more so - airflow configurations. Although the preliminary two-dimensional transient
simulations (see Appendix E) show the opening area ratio does have a considerable impact on
the peak pressure coefficients of multi-story double-skin facades, it was controlled at a constant
5% opening area ratio throughout the forthcoming chapters.
7.4.1 Airflow Intake Configuration
Airflow into the double-skin facade cavity may occur through the five generalized opening
configurations below. The sealed intake configuration is also shown; this is the mode of many
multi-story double-skin facade systems when the operable mechanisms or louvers close.
EXT INT
Seale d
Sealed
The sealed condition represents a DSF that does not permit
natural ventilation through the cavity space. This is the
equivalent mode of operation for the other airflow inlets when
they are in a closed, sealed position.
260
EXT INT
Face I Flush
EXT INT
Recess I Trench
EXT INT
Raised I Entry
Face/Flush
The face inlet occurs within the vertical plane of the outer skin,
immediately adjacent to the ground or plaza level. Inlet
dampers, louvers or grills are visible. For this reason, this
configuration is not commonly built.
Trench I Recessed I Base
The trench inlet permits airflow into the cavity space through a
recessed volume that is commonly equal depth and inlet
height as the cavity depth. Inlet dampers or louvers are often
concealed beneath the ground level by maintenance grilles,
making it a preferable - and common - configuration.
Raised I Entry
The raised I entry inlet is favorable where a double-skin has an
integral entrance portal, lifting the air inlets at least one story
above ground level, sometimes more. The air inlet occurs at
the horizontal plane, or underside, of the cavity volume. This
configuration has been implemented in the United States at a
similar rate as the trench inlet configuration.
261
EXT 1
J_
Shingled
Louvered
INT
Shingled
The shingled inlet permits airflow into the cavity through
multiple opening bands, often located in alignment with floor
levels, and occurring in the sloping vertical face or the
horizontal offset created by the shingled stacking.
Louvered
The louvered inlet has many smaller openings across the
height of the vertical plane, often possessing higher
permeability than the single elevation regions of the previous
configurations. A louvered outer skin has been applied more
frequently internationally than in the United States.
The last inlet configuration type, louvered, is not considered in the simulations in the remaining
scope of this research for several reasons. First, to maintain a constant opening area ratio, the
gaps between lites on the outer skin would be infinitesimal and result in an extremely large
(number of elements) computational demand. Second, the louvered inlet is difficult to decipher
how exhaust is treated; is it all at the top or is there some exhaust lost through the gaps along the
height of the cavity? Lastly, there were no louvered applications found in the case study review
of multi-story double-skin facades in the United States. Although the louvered configuration is
most aligned with rain-screen principles and may possess key pressure equalization
characteristics and potential load sharing benefits, it is omitted from the analysis in the remaining
portions herein.
262
7.4.2 Airflow Exhaust Configuration
Airflow out of the double-skin facade cavity may occur through the three generalized exhaust, or
outlet, opening configurations below. The sealed exhaust configuration is also shown; this is the
mode of many multi-story double-skin facade systems when the operable mechanisms or louvers
close. The three exhaust opening configurations all occur at the top of the cavity, or on the
adjacent vertical planes, taking advantage of the stack effect. The forward configuration exhausts
air over the building's roof through the inner vertical plane while the return condition thrusts the air
back into the impinging direction via the outer vertical plane.
EXT INT
Sealed
EXT INT
Forward
Sealed
The sealed condition represents a DSF that does not permit
natural ventilation through the cavity space. This is the
equivalent mode of operation for the other airflow outlets when
they are in a closed, sealed position.
Forward
The forward exhaust outlet occurs within the vertical plane of
the inner skin, generally located at the roof level, creating a
parapet condition. Outlet dampers, louvers or grills are visible
on the return side of the parapet. This configuration is the
most common.
263
EXT INT
Return
EXT INT
Through
7.4.3 Permeability
Return
The return outlet exhausts airflow out of the top of the cavity
space along the outermost skin, returning the air to the
direction of impinging flow. Inlet dampers or louvers are visible
to the exterior. For this reason, this configuration is not
commonly built.
Through
The through outlet exhausts airflow through the top of the
cavity space, occurring in the horizontal plane spanning
between the outer and inner skins that would otherwise act as
the cap to the cavity. Exhaust openings in the top horizontal
plane are uncommon likely due to rain or objects falling into
the cavity space.
The permeability of a double-skin facade relates to the exterior skin's ability to allow air to pass
through it. For the purposes of this research, permeability will be characterized by the opening
area ratio. The opening area ratio, OAR = A
0
/(a * hcav), is always equal to 5% for all
simulations with a double-skin facade cavity space. The ratio of inlet opening area to exhaust
opening area always remains 1 :1. Since the cavity depth for the prototype is fixed at s = 1.0 m,
the opening and exhaust heights are also equal to 1.0 m. Across a building with a = 45.0 m and
hcav = 20.0 m, the 1.0 m opening dimensions equates to a 5% opening area ratio.
1
264
7.5 Summary of Analytical Prototype
The following is an outline of the analytical parameters utilized in the primary simulations of this
research, the three-dimensional steady-state analysis that are carried out in Chapter 9. Many
parameters are also common to the transient models presented in Appendix E; exceptions are
noted.
Wind Tunnel Parameters
Lwt = 9 m
Wwr= 3 m
Hwr= 2 m
Simulation Model Parameters
\= 1/40 (per MOS)
U
0
= 10.5 mis (per MOS)
Ir, mfr, = 0.15 (see Chapter 8)
a~0.18 (per MOS)
z
0
= 2.0 m (see Chapter 8)
() = 0° wind incidence
Full-Scale Prototype Geometry
h = 20 m (see Section 7.2 - 7.3)
h"P = 21 m (see Section 7.2 - 7.3)
a= 45 m (see Section 7.2 - 7.3)
b = a I 2 = 22. 5 m
s = 1.0 m
OAR= 5%
hstart = 5 m
(see Section 7.2 - 7.3)
(see Section 743)
2
(only applicable to raised/entry inlet configurations)
265
Chapter 7 Endnotes
For configurations that utilize a Raised inlet, the exterior skin is not the full-height, but instead
equal to h"P - h""" = 21 - 5 = 16 m. This means that for these configurations either the inlet
dimensions need to shrink to maintain a similar opening area ratio as the other configurations
or maintain a constant cavity depth - similar to the other configurations - and allow the opening
area ratio to change. The second option is selected for all Raised configurations simulated in
Chapter 9. The opening area ratio is equal to (A
0
/(a * hrn
0
) = (45m*1 m) I ((45 m) * (16 m)) =
0.0625.
2
Ibid.
266
8. Calibration of CFO Model Using Multi-Story DSF Wind Tunnel Test
Data
8.1 Introduction
The use of simulation for this research requires special attention be given to tuning the modeling
process in an effort to preserve the relevancy of the results. Researchers have identified the
potential of CFO modeling, when used in conjunction with wind tunnel studies, to tackle
challenging design situations. Once standard modeling procedures are established, using wind
tunnel testing to calibrate a CFO model may be used to predict accurate pressure distributions on
facade structures (Zammit, Overend and Hargreaves 2010, 659-660). On the contrary, there are
still many skeptics: CFO is increasingly implemented for qualitative evaluation of wind effects,
especially in the near-ground zone, but its quantitative results are not viewed as sufficiently
accurate for the determination of design wind loads (Irwin, Denoon and Scott 2013, 10). Without a
readily available wind tunnel as an option, and to evaluate a CFO workflow for determining
pressure coefficients for multi-story double-skin facades, this research aimed to simulate
reduced-scaled wind tunnel studies within the simulation domain. The first step in refining an
analytical model was to calibrate the method to existing sources of wind tunnel data.
The initial step in the calibration is replicating simulations of existing wind tunnel data for multi
story single-skin and double-skin facades configurations. Two wind tunnel studies were
conducted to calibrate the CFO modeling. The first calibration (see Section 8.4) exercise was that
of a simplified isolated cube in an atmospheric boundary layer wind tunnel. This experiment was
carried out by 12 separate institutions with wind tunnels and the results compared (Holscher and
Niemann 1998), then carried out separately- and more recently- at full-scale measurements
and reduced-scale wind tunnel experiments (Richards, et al. 2007). Both of these research efforts
were summarized in Overview of Pressure Coefficient Data in Building Energy Simulation and
Airflow Network (Costola, Blacken and Hensen 2009, 2029). The second calibration (see
267
Sections 8.5 - 8.7) exercise was to adopt the most similar physical testing to date: Marques da
Silva and Gomes' tests of Gap inner pressures in multi-storey double skin facades (2008).
8.2 Description of Wind Tunnel Models
The following section outlines the atmospheric boundary layer wind tunnel parameters applied to
the calibration studies of single-skin and double-skin conditions.
8.2.1 Dimensional Configuration
The testing occurred in an open-circuit wind tunnel with a 9 m x 3 m x 2 m test section. The
three-dimensional solution domain includes study models that are centered along the Lwt length
and Wwt width, as well as placed with its base at Hwt = 0, unless noted otherwise.
H =2 m
wt
l:y
x
Figure 8-1: Simulated atmospheric boundary layer wind tunnel configuration.
8.2.2 Boundary Conditions
Inlet: The vertical boundary plane at y = 0, the upstream inlet, has an initial velocity of U
0
, defined
in the next section. The resulting pressure distribution between the inlet and outlet is variable.
The initial turbulence intensity Ur ) and turbulent length scale (LT) are also defined at the inlet.
268
Outlet: The vertical boundary plane at y = 9, or the downstream outlet, is specified as a zero
gauge pressure (p
0
= 0) with no viscous stress.
Floor The floor is defined as a wall function with applied roughness. The roughness is modeled
by the equivalent sand roughness (ks.Aad as described in Section 8.3.3.
8.2.3 Velocity Profile
The velocity profile varies as a function of the reference velocity (Uref ), height (z) and the power
law exponent (a):
Uo = Uref * (- z )"
Zref
u .. ,
Laminar f Transition f
Figure 8-2: Velocity profile at the inlet plane.
8.2.4 Test Configurations
Turbulent
Eq. 8.3.1
Turbulent
Region
Buffer
Layer
Viscous
Sublayer
All calibration simulations use the aforementioned wind tunnel dimensions. What differs are the
model scale ( ,A.), inlet velocity (U
0
) and building configurations dimensions; width (a), depth (b)
height (h) included . The Simple Cube modeled in Section 8.4 uses ). = 1/100, U
0
= 7 mis, a =
{0.18, 0.20, 0.22, 0.24}, and scaled building configurations dimensions; a = 50 cm, b = 50 cm and
h = 50 cm. The Unsheltered and Sheltered Towers modeled in Sections 8.5 and 8.6 uses ). =
1/40, U
0
= 10.5 m/s, a = 0.18, and scaled building configurations dimensions; a = 32.5 cm, b = 20
cm and h = 70 cm, per Marques da Silva and Gomes (2008, 1555).
269
8.3 Considerations for Turbulent Flow
Simulating turbulent flow in an atmospheric boundary layer requires a keen understanding of the
pertinent variables and those most influential on the characteristic flow profile. This section
outlines the importance of turbulence intensity Ur), turbulent length scale (LT), roughness length
(z
0
) and equivalent sand roughness height for the atmospheric boundary layer (ks,Aail·
8.3.1 Turbulence Intensity
A fundamental difference between wind tunnel and computational fluid dynamic simulations is the
role of turbulence. Turbulence level, or relative intensity of turbulence, is a ratio used to
characterize the relationship between the turbulence intensity compared to the mean velocity.
The relative intensity of turbulence increases with surface roughness and decreases with height
Near the earth's surface, turbulence intensity typically ranges from 10-30% (Irwin, Denoon and
Scott 2013, 15). At a height of 10 meters, winds often have a relative turbulence level of 0.20 to
0.30 (Liu 1991, 50). The magnitude of relative intensity of turbulence, Ir, is approximated as,
Eq. 8.3.1
where 15 is the boundary layer thickness, z is the height (must be greater than 5 m) and a is the
power law equation exponent (Liu 1991, 50). The turbulence level varies from one wind tunnel to
another. Turbulence characteristics in atmospheric boundary layer wind tunnels can be
controlled by several common means: vertical spires upstream, screens and developing
roughness along the floor surface. Special tunnels have been constructed to obtain very low
turbulence levels; however, the modeling of turbulence in wind tunnels for structural testing is
generally in the range of 0.10 to 0.30 for Ir (Liu 1991, 156). This is corroborated in the research
of Simiu (2009), who charts the turbulence intensities of wind tunnels participating in the Fritz et
al. (2008) comparison, where open exposure conditions have turbulence intensity between 0.15
and 0.23, and suburban exposure conditions heavily concentrated around 0.25.
270
25
20
10 15 20 25
Turbulence Intensity [ o/o]
e A
!!,. B
DC
ID D
!!,. E
25
20
] 15
"'
10
5
30 20
. [
• D
[
c::
l•o
Gt•
de
1 •
• I jJ
fl D e
• ·1 :.
25
D D
30
Turbulence Intensity [ %]
35
e A
!!,. B
D C
8 D
Figure 8-3: Turbulence intensities of wind tunnels participating in the Fritz et al. (2008)
comparison (Simiu 2009, 12)
1
; open exposure (left) and suburban exposure (right).
To calibrate a CFD simulation to the wind tunnel testing, it is critical to appropriately understand
the turbulence level used in producing the coefficient of pressure data, as well as best integrate
these characteristics into the virtual model. The turbulent flow simulations incorporate a
Reynolds Averaged Navier Stokes (RANS) turbulence model type with a k-e turbulence model
40
with incompressible flow. The parameters used to model turbulence include turbulence intensity,
Ir (same as relative turbulence intensity, Ir) and turbulent length scale, Lr. The turbulent length
scale is a physical quantity related to the size of the large eddies that contain the energy in
turbulent flovvs. For wall-bounded flovvs in which the inlet involves a turbulent boundary layer, the
turbulent length scale, Lr , can be
Lr= 0.4 8 Eq. 8.3.2
where 8 is the boundary layer thickness.
2
In the computational domain, the initial turbulence
intensity (Ir) is defined as a boundary layer condition at the inlet. The definition of turbulence at
271
the inlet is not sufficient on its own to emulate the turbulent nature of the pedestrian zone; terrain
roughness must be introduced along the wind-tunnel floor surface. These two combined influence
the turbulence intensity profile throughout the wind tunnel, defined as,
.~
l=-i::_= -~'-'
u u
Eq. 8.3.4
where u' represents local flow oscillations, U is the time-averaged flow and k is turbulent energy
at a discrete point within the solution domain.
8.3.2 Roughness Classification and Length
In a computational fluid dynamic simulation, the inlet boundary condition includes definitions of
flow profile, the mean wind speed, turbulence intensity and length scale. How the flow develops in
the upstream domain as it approaches the test section is the result of an aerodynamic roughness
length, z
0
, or the appropriate profile for this terrain applied at the inlet's power-law exponent, a
(Davenport 1960), (Wieringa 1992).
Table 8-1: Terrain Type, Roughness and Surface Drag Coefficient (Holmes 2001)
Category Terrain Type
Roughness Length
Surface Drag Coefficient
[m]
Very Flat Terrain 0.001 - 0.005 0.002 - 0.003
2 Open Terrain 0.01 - 0.05 0.003 - 0.006
3 Suburban Terrain 0.1 - 0.5 0.0075 - 0.02
4 Dense Urban 1 - 5 0.03 - 0.3
The surface roughness length, z
0
, represents the physical influence that roughness objects, or
obstacles to wind flow, on the earth's surface have on the shape of the atmospheric boundary
layer wind velocity profile (ASCE/SEI 2010, 541). The classification of surface roughness length
was developed by Davenport (1960). Wieringa (1992) later updated the terrain classification to
272
more specifically develop the roughness lengths. These researchers have informed the range of
surface roughness length (z
0
) by exposure category as expressed in Minimum Design Loads for
Buildings and Other Structures (ASCE/SEI 2010, 540) in Table C26.7-2.
Table 8-2: Davenport Classification of Effective Terrain Roughness (ASCE/SEI 2010, 541)
3
Class Surface Landscape Description
Zo
[m]
Open sea or lake (irrespective of the wave size), tidal flat, snow-
Sea covered flat plain, featureless desert, tarmac and concrete, with a 0.0002
free fetch of several kilometers.
2 Smooth
Featureless land surface without any noticable obstacles and with
0.005
negligible vegetation.
3 Open
Level country with low vegetation and isolated obstacles with
0.03
separations of at least 50 obstacle heights.
Roughly
Cultivated area with regular cover of low crops, or moderately
4 open country with occasional obstacles at relative horizontal 0.10
Open
distances of at least 20 obstacle heights.
Cultivated area with high crops or crops of varying height, and
5 Rough
scattered obstacles at relative distances of 12 to 15 obstacle
0.25
heights for porous objects or 8 to 12 obstacle heights for low solid
objects.
lntensily cultivated landscape with many rather large obstacle
groups separated by open spaces of about 8 obstacle heights.
6 Very Rough Low densly-planted major vegetation like bushland, orchards, 0.50
young forest. Also, moderately covered by low buildings with
interspaces of 3 to 7 building heights.
7 Skimming
Landscape regularly covered with similar-size large obstacles,
1.0
with open spaces comparable to the obstacle heights.
8 Chaotic
City centers with mixture of low-rise and high-rise buildings. Also,
> 2.0
large forests of irregular height with many clearings.
273
Many of the double-skin facades presented in the previous chapters occur in urban environments
surrounded by surface roughness consisting of mid- and high-rise structures. other applications
occur in suburban terrain or rough to very rough terrain classifications. As is evident from the
above Tables 8-1 and 8-2, such urban environments will have a roughness length of 0.5 m z
0

2.0 m, and in some cases even larger.
8.3.3 Equivalent Sand-grain Roughness Height
In three-dimensional turbulence models, smooth walls are assumed by default and require
boundary condition modification by the modeler to introduce wall roughness where appropriate. In
the turbulence model, this requires assigning wall roughness to the wind tunnel floor surface. This
is achieved by defining the equivalent sand roughness height for the atmospheric boundary layer,
ks,AaL, also referred to as kseq in COM SOL. The equivalent sand roughness height, ks,AaL, can be
defined as,
ks,ABL ;::::: 30 Zo Eq. 8.3.4
4
where z
0
represents the aerodynamic roughness length. The equivalent sand height roughness,
or k
5
-type wall function, is what many computational fluid dynamic codes - COMSOL included -
provide as the sole means to modify surface roughness.
For a suburban terrain (per Table 8-1) or a very rough surface (per Table 8-2), a surface
roughness length of z
0
= 0.5 m would yield a ks,ABL "' 15 m. For a dense urban terrain (per Table
8-1) or a chaotic surface (per Table 8-2), a surface roughness length of z
0
= 2.0 m would yield a
ks,ABL "'60 m. The remainder of this research will rely on the basis of ks,ABL = 30*z
0
. When
modeling the built environment, it is clear that ks,ABL will be large.
274
8.4 CFO Model Description - Simple Cube
The first calibration model in this research is a simple cube, similar to that evaluated in the
research of Holscher and Niemann (1998). This study drew comparative wind tunnel test results
from twelve different institutions, seeking the coefficients of pressure calculated on the cube's
various surfaces. This simple model is used to evaluate the accuracy and weaknesses of
simulating built-environment, atmospheric boundary layer wind tunnel studies using three
dimensional computational fluid dynamic simulations.
8.4.1 Introduction
The purpose of Holscher and Niemann (1998) was to conduct and evaluate a comparative study
of scaled wind-tunnel studies modeling the coefficients of pressure on simple building shapes by
twelve institutions. The study sought to identify the variance, or bandwidth, present between the
wind tunnel evaluations. While the full-scale flow field does not suffer from any scaled disparities,
such as Reynolds number, turbulence intensity or turbulence scale, wind tunnel tests may.
However, the primary advantages of wind tunnel tests over full-scale testing is that they are
significantly less expensive, less time consuming and conducted in a controlled environment. For
those reasons, wind tunnel testing holds a prominent role in establishing the structural wind
loading pressures of building designs.
The prescriptive study to the twelve intuitions is summarized below:
1. Simulate boundary-layer flow for a suburban terrain with a= 022 ± 0.02.
2. Determine the mean pressure on a floor-mounted cube equal to 50 m height, full-scale.
3. Measurement of two point sources downstream; one free upstream of building and the
second at the cube center, both points at the 50 m roof height.
The twelve studies were conducted in different size wind tunnels; however, the study suggested
a sufficient boundary layer of approximately three to four times the cube height.
275
- Mean Result
Range of Results
u~
t 0+=-0--------1--'"---------~2=-------------;3
"(J
"' Oi
0
u
~ -0.5 -!---------·
i:l
I'!
Q_
::-L----_ l --'------ _J_ -'---- 1 -'
~
0
a/
Distance Along Trajectory (0-1-2-3]
Figure 84: Range of surface pressures for a wind direction of 0° as modeled by twelve
different wind tunnels. Information for this figure based on Towards Quality Assurance for
Wind Tunnel Tests: A Comparative Testing Program of the Windtechnologische Gesel/schaft
(Holscher and Niemann 1998, 606).
Out of fifteen experiments conducted by the twelve institutions, only two deviated from the
boundary-layer thickness requirement of three to four times the cube height. For roughness, a
suburban terrain was prescribed, z
0
::::: 0.5 to 1.0 m. The twelve experiments ranged in roughness
from z
0
= 0.2 to 1.1 m. The mean velocity profiles of the sample base were more consistent than
the turbulence intensity profiles. At cube height, the turbulence intensity ranged from
approximately 12 to 30%. The range of Reynold's numbers for the tests was 2x10
4
to 3x10
5
. The
coefficients of pressure are shown (Figure 8-4) across the midpoint of the cube's faces with the x-
axis representing distance: 0 to 1 representing the front or impinging face, 1 to 2 the roof, and 2
to 3 the leeward face. The minimum and maximum pressures reported by each experiment are
statistically presented as 1) a complete sample and 2) a reduced sample that throws out two
outliers that do not display similar physics to the others.
276
0.92
Q.
()
c
0.8
Q)
"(3
t:=
Q)
0
()
~
:J
CJ)
0.7
CJ)
Q)
ct
Figure 8-5: Maximum pressure coefficients on the windward face for a wind direction of 0°
as modeled by twelve different wind tunnels. Information for this figure based on Towards
Quality Assurance for Wind Tunnel Tests: A Comparative Testing Program of the
Windtechnologische Gesel/schaft (Holscher and Niemann 1998, 606).
The study then finds the area averaged pressure coefficients in the central portion of the cube
surfaces; (0.25 ~ y/ H ~ 0.75, 0.1 ~ z/ H ~ 0.75) for the windward and leeward surfaces. On the
impinging, front face - for the reduced sample - the study finds an average pressure coefficient
of Cp.front = 0.84 (O"= 0.05) with a minimum of 0.79 and maximum of 0.92. On the roof, the study
finds an average pressure coefficient of Cp,roof = -1.13 (O"= 0.07) with a minimum of -1.25 and
maximum of -1 .02. On the leeward side, the study finds an average of Cp,rear = -0.24 (O"= 0.05)
with a minimum of -0.31 and maximum of -0.17. The study concludes that the atmospheric
boundary-layer wind tunnel can successfully develop representative wind pressures of full-scale
flows if some basic requirements are considered.
277
1.0 ~------------------------~
+
• Range of Results
0.5
- --- Wind Tunnel
- Full-scale
~ -0.5 -i- ---------t ~
Q)
it
-1.0 -i------------H urfJ
0
a/
-1.5 --'-------------------------------'
Distance Along Trajectory [0-1-2-3]
Figure 8-6: Comparison of Silsoe Cube (full-scale and wind tunnel) surface pressures to
the aforementioned range of twelve different wind tunnel results. Information for this figure
based on Wind-tunnel Modeling of the Si/soe Cube (Richards, et al. 2007, 1392).
A later study, influenced by Holscher and Niemann (1998), is that of Richards et al. (2007), where
a 1 :40 scale wind-tunnel model in the University of Auckland's boundary layer wind-tunnel is
compared to full-scale measurements of the 6 m Silsoe Cube in an 'open country' setting located
at the Silsoe Research Institute in the United Kingdom. The wind tunnel experiments were
conducted in a 1.85 m wide by 1.1 m tall with a fetch length of 7.5 m (Richards, et al. 2007,
1390). The following test parameters existed: ,1. = 1 /40, U
0
= 6.4 mis, z
0
= 16.8 mm and scaled
building configuration dimensions of a = 15 cm, b = 15 cm and h = 15 cm. The wind tunnel's
velocity profile was similar to that of the full-scale facility. In Figure 8-6, the results of the two wind
tunnel tests and the full-scale Silsoe Cube of Richards et al. (2007) are overlaid upon the
bandwidth of results from Holscher and Niemann (1998). While the full-scale measurement
displays pressures on the boundary of the results range, both wind tunnel tests of the Silsoe cube
have a greater pressure magnitude with a similar profile curve to the full-scale and others.
278
8.4.2 Test Configuration
The first calibration model simulated in this calibration section was similar to that of Holscher and
Niemann (1998) using a 50 m simple cube scaled to ,l = 1i100 with an initial velocity of U
0
= 7 mis
in a 9 m x 3 m x 2 m test section, simulated in COMSOL.
8.4.3 Comparison of CFO Results to Wind Tunnel Data - Simple Cube
This section reviews the results of calibration models simulated in COMSOL to the
aforementioned wind-tunnel and full-scale studies outlined in Section 8.4.1. The power law
exponent (a), turbulence intensity (Ir, mfr,) and roughness length (z
0
) are varied to gain an
understanding of how sensitive the simulation flows - and resultant pressures - are to these
variables. The two comparative outputs are 1) the profile of the pressure coefficient across the
midpoint of the building faces and 2) the turbulence intensity profile at an approaching position of
y= 4.0 m
Sensitivity to Power Law Exponent
The pressure and turbulence intensity profiles in Figures 8-7 and 8-8 are the results of a model
with U
0
= 7.0 mis, Ir, mfr,= 0.15, a: varies {0.18, 0.20, 0.22, 0.24) and z
0
= 1.0 m All four models
show a magnification of the pressure profile on the windward side, a reasonable match within the
mean bandwidth for the majority of the roof face, and results on the outer bounds of the range for
the leeward face. The general shape of the pressure coefficient curve resembles that of the wind
tunnel studies with a shifted minimum on the roof, indicating the possibility of a more immediate
reattachment of the flows around the vortex shedding region that occurs transitioning around the
front edge to the roof surface. The pressure coefficient on the windward side increases slightly
with an increase in the power law exponent (a). The same is true of the turbulence intensity
profile with the curves fairly parallel. Both the pressure coefficient profile and intensity profile
appear minimally affected by variation of the power law exponent.
279
'E
Cl)
·c::;
:E
Cl)
8
1.0
~ -0.5 -+--------------t
,,,
,,,
~
a..
-1.0 cJID_,
0
a /
-1.5 -+-------------+-------------+-------------;
Distance Along Trajectory [0-1-2-3]
Range of Results
W ind Tunnel
------· Full-scale Test
a=0.18
u=0.20
a= 0.22
a=0.24
Figure 8-7: CP profile comparison; a varies (U
0
= 7 mis, Ir, rnJet = 0.15, a : varies, z
0
= 1.0 m).
4
3
:: 2
N
0
•
-
.....
•
....
...
"'
~
•
•
...
• •
•
-
•
•
...
•
;t ....
•
•
•
• •
•
•
0.10
.....
\
•
•
~
. .....
.....
"!-
•
..
...
._,.:.,. •
.... .
....
•
-
0.20
Turbulence Intensity (I,)
0.30
Figure 8-8: Approaching Ir profiles (Uo= 7 m/s, Ir, inle t = 0.15, a : varies, zo = 1.0 m).
Previous wind
tunnel studies
a =0.18
a =0.20
r =0.22
a= 0.24
280
Sensitivity to Roughness
The pressure and turbulence intensity profiles in Figures 8-9 and 8-10 are the results of a model
with U
0
= 7.0 mis, lr,m1e, = 0.15, a= 0.22 and z
0
: varies {O, 0.1, 0.5, 1.0, 2.0). All five models show
a magnification of the pressure profile on the windward side, a reasonable match within the mean
bandwidth for the majority of the roof face, and results on the outer bounds of the range for the
leeward face. The general shape of the pressure coefficient curve resembles that of the wind
tunnel studies with a shifted minimum on the roof, indicating the possibility of a more immediate
reattachment of the flows around the vortex shedding region that occurs transitioning around the
front edge to the roof surface. On the windward and leeward faces, the zones closest to the
ground surface (Zih 0.50) experience the most variation when the roughness is varied; as would
be expected. Around Zih = 0.20 on the leeward face, the profile curves for simulations with z
0

1.0 m display an inflection point; a clear indicator of laminar-type flow near the floor surface when
significant - or suburban/urban scale - roughness is not applied to the simulated wind-tunnel's
floor surface. The shapes of the pressure profiles for z
0
= 1.0 m and z
0
= 2.0 m show a more
stable, less volatile profile in the near-ground zones on both the windward and leeward sides than
those with z
0
1.0 m. Additionally, the profiles for z
0
= 1.0 m and z
0
= 2.0 m show a reasonably
similar shape to the wind tunnel study range, though they exhibit amplified values on the
windward and leeward surface; more so on the windward face.
The sensitivities to roughness near the floor surface are also evident Figure 8-10 where there is
significant contrast in the Ir profiles for Zih 1.00, convergence between 1.00 Zih 2.50 and
agreement between 2.50 Zih 4.00. Near the ground plane, Zih ~ 0, the turbulence intensity
values range from 0.09 to 0.26. As compared to the near-ground profiles of the previous wind
tunnel studies (ranging from 0.16 to 0.35), the curves for Zih 1.00 show a decrease in intensity
as Z ~ O; opposite of the wind tunnel studies near-ground behavior for turbulence intensity. The
simulations' sensitivity to roughness is noticeably greater than that of the power law exponent.
281
Range of Results
0.5
W ind Tunnel
o~
Full-scale Test
c
0
Q)
0
'(3
2 3
-=
lii
0
0
--- --· = c-= Om
~
·0.5 ~
(/)
(/)
2
~
a..
~
·1 .0
.
0
a /
-1.5
-- - = Olm ~,, .
-- -,- -- .::.~.:.------ ~-~--
I
---· = .= 0.5 m
-- =,,= 1.0 m
--- = .= 2.0 m
Distance Along Trajectory [0-1 -2-3)
Figure 8-9: CP profile comparison; zo varies (U
0
= 7 m/s,lr, inlet = 0.15, a= 0.22, z
0
: varies).
4
3
•
0
•
•
•
•
""'
.....
•
..
... ~
•
•
..
•
•
,,..
•
......
•
•
•
.......
~ • ~
•
-
• •
•
...
0.10
~
'
...
..
....
...
....
\
.
\
\
\
\
\
\
\
\
\
~ \
I
0.20
Turbulence Intensity ( l,)
0.30
Figure 8-10: Approaching Ir profiles (U
0
= 7 mls,lr,znl et = 0.15, a= 0.22, z
0
: varies).
Previous wind
tunnel studies
: ,=0.5 m
:
0
= 1.0 m
: ,,=2.0m
282
Sensitivity to Turbulence Intensity at Inlet
The pressure and turbulence intensity profiles in Figures 8-11 and 8-12 are the results of a model
with U
0
= 7.0 mis, a= 0.22, z
0
= 1.0 mandfr.m1e,: varies{0.10, 0.15, 0.20, 0.25, 0.30). All five
models show a magnification of the pressure profile on the windward side, with curves for Ir. mlee ;::_
0.20 exhibiting broad areas of CP > 1.00 in the upper half (0.50 Zlh 1.00) of the windward face.
There is a reasonable match within the mean bandwidth for the majority of the roof face with the
glaring exception being an over amplification of the pressure in the near-parapet edge area for fr.
mfrl ;::_ 0.20, which drop below CP -1.50. This indicates a heightened sensitivity to flow around
surface edge transitions as the initial turbulence intensity increases. The general shape of the
pressure coefficient curve resembles that of the wind tunnel studies with a shifted minimum on
the roof, reinforcing the notion of a more immediate reattachment of the flows around the vortex
shedding region that occurs transitioning around the front edge to the roof surface than in the
wind tunnel sample. On the leeward face, the profile curves are less volatile than on the
windward and roof surfaces, displaying a constant offset as the initial turbulence intensity
increase. All curves on the leeward side fall outside the mean range of wind tunnel results, with a
peak amplitude nearing CP = -0.50. The pressure profiles suggest that an inlet turbulence intensity
of Ir. mfrl ;::_ 0.20, in combination with these power law exponent and roughness characteristics,
would be more turbulent and not aligned with the conditions outlined in the wind tunnel and full
scale results of Holscher and Niemann (1998) and Richards et al. (2007).
The sensitivities of the approaching flaw's turbulence to varying definitions of inlet turbulence
intensity, Ir, mfr" are exhibited in Figure 8-12. This figure shows a linear increase in the
approaching flaw's Ir as Ir, mfrl is increased. As compared to the previous wind tunnel studies, all
turbulence intensity profiles for Ir, mfrl ;::_ 0.20 fall outside the sample range, reinforcing that an inlet
turbulence intensity of Ir, mfrl ;::_ 0.20 would be more turbulent and not aligned with the conditions
outlined in the previous studies to which this section seeks to calibrate to.
283
c
"'
·5
0
-+-o~~~~~~~~~~--+~~~~~~~~~~~+-2~~~~~~~~~~~3
l _ ____________ _
~
8
~ -0.5
())
())
~
a_
-1.0
-1.5
~--=----------- _,--------
a /
Distance Along Trajectory [0· 1 ·2·3)
Range of Results
Wind Tunnel
Full-scale Test
IT=0.10
1,=0.15
I,= 0.20
I,= 0.30
Figure 8-11: CP profile comparison; Ir, inlet varies (U
0
=7 m/s, I r, inlet: varies, a= 0.22, z
0
= 1.0 m).
4
3
•
0
\
\
...
\
\
....
\
\
.. \
'
....
\
\
..
\
•
.,,.._
\
..
I
....
I
•
• •
I
I
•
.. I
• •
.......
I
•
I
•
..,. ,
\
•
• I
\
• • \
..
\
\
•
\
\
T
\
\
•
\
\
•
\
\
I
I
0.10 0.20
Turbulence Intensity (I, )
'
\
\
\
\
I
I
I
I
\
\
I
_ /
\
I
J
'
I
0.30
Figure 8-12: Approaching Ir profiles (U
0
=7 mis, Ir, mlet: varies, a = 0.22, z
0
= 1 .0 m).
Previous wind
tunnel studies
1],Jttlil =0.15
IT Irr/i i = 0.25
fT.l•I" = 0.30
284
Discussion
Simulating a simple cube based on Holscher and Niemann (1998), the sensitivity of the pressure
coefficient and approach turbulence intensity profiles to variables - power law exponent,
roughness and inlet turbulence intensity-were reviewed in this section. Figures 8-7 through 8-
12 reveal the profiles are most sensitive to variation in the inlet turbulence intensity and least
sensitive to variation in the power law exponent. The sensitivity to variation roughness has a
significant impact in the near-ground flow regions - primarily Zih 1.00 - but has limited impact
on Zih > 2.00. The calibrations exhibit reasonable profiles of pressure on all three surfaces of the
simple cube. There is a general over-amplification of peak c, values on the windward and
leeward surfaces and the initial part of the roof, closest to the front edge. If this were the sole
basis of calibration, future simulations would focus on using an inlet turbulence intensity of 0.10.:::.
Ir, mlee.:::. 0.15 and roughness length 1.0.:::. z
0
.:::. 2.0 m. The calibration models in Sections 8.5, 8.6
and 8.7, based on Marques da Silva and Gomes (2008), will be used to focus these parameters
further for the simulations carried out in future chapters.
8.4.4 Two-Dimensional Steady-State Simulation
The same sets of evaluations presented in Section 8.43 were also carried out using two
dimensional steady-state analysis. These analyses are presented in Appendix Din Figures D-1
through D-3. Though these simulations could be conducted in a fraction of the time and exhibited
reasonable agreement on the leeward face, the two-dimensional simulations' results consistently
amplified the pressure coefficients on the roof and leeward conditions as compared to the existing
wind tunnel (Holscher and Niemann 1998), field-measurement (Richards, et al. 2007) and three
dimensional steady-state simulation data. As compared to the three-dimensional simulations of
Section 8.4.3, the two-dimensional simulations are less-successful in modeling near-ground
turbulent flow and the flow separation and reattachment that occurs at the front roof edge. There
is a consistently extreme and steep pressure as the profile nears Zih = 1.00. This is an indicator
285
that the two-dimensional simulations struggle to model the communication of flows around and
between adjacent skins, especially sharp transitions between these surfaces (e.g. leeward face to
roof). The two-dimensional steady-state simulation may only have merit in conducting relative
studies during the preliminary design phases of multi-story double-skin facades, with only
impinging flow on the leeward face being a reasonable indicator.
Slicing the simple cube of Holscher and Niemann (1998) at mid-span and modeling as a two
dimensional steady-state flow corroborate that the coefficient of pressure and approaching
turbulence intensity profiles are most sensitive to variation in the inlet turbulence intensity and
least sensitive to variation in the power law exponent. The sensitivity to variation roughness has a
significant impact in the near-ground flow regions, primarily Zlh 1.00. These general findings are
consistent with those for the three-dimensional steady-state simulations conducted in Section
8.4.3. For the two-dimensional simulations, when Ir, mlee = 0.15, there is an extensive portion of the
pressure profile on the leeward face (see Figures D-1 and D-2) that falls within the range of
results of Holscher and Niemann (1998). This is most true of the mid-leeward face with deviations
out of the range of results beginning around Zih = 0.85. It is here that the two-dimensional
simulations exhibit a more severe pressure drop nearing the front-roof edge at Zlh = 1.00. For the
approaching flows' Ir profiles, Figures D-1 to D-3 show a tendency for the two-dimensional
steady-state simulations' to be more turbulent - than Holscher and Niemann (1998) and the 30
steady-state simulations - in the region above Zlh > 0. 75, while Zih 0. 75 struggles to achieve a
smooth transition to the turbulent flow nearing the ground. The result is an inflection point in the Ir
profiles around 0.90 Zih 1.10 with a steep decline in turbulence intensity as Zlh approaches O;
a subdued characteristic similar to the near-wall behavior indicative of laminar flow.
Despite the shortcomings of the two-dimensional steady-state simulations, when the results for
the leeward faces are reviewed, there is support for using an inlet turbulence intensity of inlet
turbulence intensity of 0.15.::: Ir, mlee.::: 0.20 and roughness length 1.0.::: z
0
.::: 2.0 m.
286
8.5 CFO Model Description - Unsheltered Tower
8.5.1 Introduction
This section covers the first of three calibration models conducted to understand how accurately
the wind tunnel experiments of Marques da Silva and Gomes (2005, 2008)
5
can be replicated in
the multi physics, computational fluid dynamic environment of COM SOL as a three-dimensional
steady-state flow simulation. The calibrations herein the next three sections include simulating
and comparing an 1) unsheltered tower, 2) sheltered tower with a cavity open on all sides (Layout
A, see Section 8.6), and 3) sheltered tower with full lateral closure (Layout 0, Section 8.6). The
research of Marques da Silva and Gomes was modeled using an unsheltered tower with model
dimensions of a= 32.5 cm wide, b = 20 cm deep and h = 70 cm tall, and the same tower with a
second-skin with the outer skin beginning at h"""= 7.5 cm above ground level for Layout A and
Layout 0. Following the calibration studies of these configurations, a set of testing parameters
and simulation criteria are outlined (see Section 8.8) for the simulations performed in Chapter 9.
8.5.2 Test Configuration
All calibration models maintain the study parameters, including: model scale of ,l = 1/40; wind
tunnel section of Lwr = 9 m x Wwr= 3 m x Hwr = 2 m; inlet velocity of U
0
= 10.5 mis; and velocity
profile power law equation exponent of a= 0.18. The tested specimen was a rectangular
geometry with the aforementioned dimensions, which equate to a relative building height, h/a =
2.15 and relative aspect ratio b /a= 0.615, where a is the building width. A turbulence length
scale of Lr = 0.415 was applied and only flow normal (fJ = 0°) to the leeward surface was
compared. The wind tunnel implemented a terrain roughness simulating a suburban exposure
(Marques da Silva and Gomes 2008, 1555). This section will explore what inlet turbulence
intensity (Ir, mfr,) and characteristic roughness length (z
0
) achieve pressure coefficient profiles for
the leeward face most similar to the existing wind tunnel research's findings.
287
8.5.3 Comparison of CFO Results to Wind Tunnel Data - Unsheltered Tower
The simulations of the unsheltered tower are compared to the results of Marques da Silva and
Gomes at the mid-span, x/a = 0.50, as shown as the section cut on the CP distribution elevation
presented in Figure 8-15. This CP distribution is shown as it was presented in Perino's State-of
the-Art Review (2007, 30).
6
The interpreted intersection of CP contours and their respective
heights are presented in Appendix D's Table D-1. These points make up the through points of the
spline that is shown as the Unsheltered Tower wind tunnel study profile, dashed in Figures 8-13
and 8-14. The results at the mid-span show a stronger agreement than those of the previous
calibrations in Section 8.4. This is also true of the side-by-side CP distribution shown in Figure 8-
15. These contours are similar in shape and magnitude, both fundamentally symmetrical, with the
key exceptions being 1) a greater maximum in the simulation at a slightly higher location and 2) a
more significant drop off in the near-edge regions than the wind tunnel results. Part of the reason
for the smoother contour shapes and the significant drop-off near the edges may be attributed to
the fact that the simulation can access thousands of distinct points from the surface's mesh as
compared to the limited pressure tap locations in the wind tunnel model, as shown in Figure 2-5.
This is characterized as the simulation possessing a higher fidelity than the wind tunnel tests.
Sensitivity to Roughness
The sensitivity to variation of z
0
is primarily limited to Zlh 0.60. The profile for z
0
= 1.0 mis
closest in shape' to the wind tunnel study and is slightly greater along the full-height while the
curve for z
0
= 0.5 m underestimates CP for Zlh 0.50 and overestimate for 0.50 Zlh 1.00.
Sensitivity to Turbulence Intensity at Inlet
The sensitivity to variation of I
7
• mlee is linear and significant with the approaching I ,profile
magnifying as the Ir, mlee increases. Ir, mlee = 0.15 is closest' to the wind tunnel study results.
288
\
'
'
::: 0.5
"
0
0 0.50
Pressure Coefficient (l)
\
\
I
I
1.00
Unsheltered Tower
wind tunnel study
= ,,= 0.5 m
= u= 1.0 m
:, = 2.0 m
w
e/ o
Figure 8-13: CP profile comparison; zo varies (U = 10.5 mis, Ir, inlet = 0.15, a = 0.18, z
0
: varies) .
::: 0.5
"'
0
0 0.50
Pressure Coefficient ((~)
....
'
" \
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
1.00
Unsheltered Tower
wind tunnel study
l, =0.10
!, =0.15
I , = 0.20
LO
e/o
Figure 8-14: CP comparison; Ir,iniet varies (U =10.5 mis, Ir, inlet: varies, a = 0.18, z
0
= 2.0 m).
289
N
CD
0
0
Marques da Silva and Gomes (2005) • Wind Tunnel Studies
Gap Inner Pressures in Multi-Storey Double Skin Facades
0.1 0.2 0.3 0.4 0.5
x/a
0.6 0.7 0.8 0.9
0.9
0.8
0.7
0.6
0.5
-t:
'
0.4
0.3
0.2
0.1
o
1 -
0.9
-!
1 1
0.8
-
I
I
I
0.7
0.6
-t:
0.5
'
0.4
-
0.3
0.2
D
Front
0.1
i
0' 0
0
Figure 8-15: Comparative elevations of CP distribution for Unsheltered Tower.
0.1
Vaglio (2015) • 30 Simulation COMSOL Multiphysics
(U
0
= 10.5 mis, /
1
;,.
1
,, =0.1 5, a =0.18, = ,,= 2.0 m)
.8
0.2 0 .3 0.4 0.5
x l a
0.8
0.6 0.7 0 .8
The CP distribution for Marques da Silva and Gomes (2005) is shown as it was presented in Perino's State-of-the-Art Review5.
0.6
0.9
I
I
11
8.6 CFO Model Description - Sheltered Tower
8.6.1 Introduction
This section simulates the multi-story DSF configuration, Layout A, of Marques da Silva and
Gomes (2008, 1554). This is best described as a sheltered tower on the front and one lateral (left)
side where both second skins have a raised inlet, a through outlet and are unsealed laterally.
8.6.2 Test Configuration
The tested specimen is a rectangular geometry similar to that of the Unsheltered Tower
presented in the previous section. The primary difference is the addition of a second skin on two
sides of the model - front and left - starting at approximately 0.107 h above the ground level and
continuing the remaining height of the building. The cavity depth between skins is varied in
Marques da Silva and Gomes (2008), but the configuration considered in this comparative
calibration is the equivalent of a full-scale cavity depth of s = 0.8 m.
LAYOUT A:
Sheltered Tower
Model Scale
;., = 1 /40
Model Dimensions
a = 32.5 cm
b = 20 cm
h = 70 cm
hstan= 7 .5 cm
hcav = 62.5 cm
s = 2.0 cm
Velocity Profile
U
0
= 10.5 m/s
a = 0.18
CP Section Locations
Midspan
x/L: 0.35, 0 .67
Figure 8-16: Sheltered Tower (Layout A) configuration and section cut locations.
291
8.6.3 Comparison of CFO Results to Wind Tunnel Data - Sheltered Tower A
The simulations of the sheltered tower are compared to the results of Marques da Silva and
Gomes at the mid-span, x/a = 0.50, as shown as the section cut on the CP distribution elevation
presented in Figure 8-19. This CP distribution is shown as it was presented in Marques da Silva
and Gomes (2008, 1555). The CP·"" (or 6Cp) profiles for x!L (equivalent to x/a) = 0.35 and 0.67,
presented in Figures 8-17 and 8-18, are extracted for the graphs presented in Marques da Silva
and Gomes (2005).
9
The simulations' results at the mid-span show a strong agreement with the
referenced wind tunnel studies, both in profile shape and CP·"" magnitude. The side-by-side Cp.ml
distribution shown in Figure 8-19 for both the wind tunnel model and simulation's building face
(inside wall of the cavity space; not the exterior second skin) have similarities, but do vary in
magnitude. Both possess a negative peak pressure zone just above the raised inlet around 0.10
Zlh 0.20 with the wind tunnel and simulation reaching Cp.ml ~ -1.40 and -1.10, respectively.
The wind tunnel study exhibits greater pressure magnitudes at both the peak locations and near
the edges. The near-ground edge (x/a ~ {0.10, 0.90) and Zlh ~ 0) magnitudes shown in Figure 8-
19 are Cp.ml ~ 1.20 for the wind-tunnel and CP·'"' ~ 0.60 for the simulations.
Sensitivity to Roughness
The sensitivity to variation of z
0
is primarily limited to Zlh 0.60, as seen in Figure 8-17. The
profile for z
0
= 2.0 mis closest in shape
10
to the wind-tunnel study and is most similar in shape to
the inflection point that occurs between 0.10 Zlh 0.20.
Sensitivity to Turbulence Intensity at Inlet
Figure 8-18 shows the sensitivity to variation of Ir. mlee is linear and significant with the approaching
fr profile increasing as the Ir, mfrl increase. It appears that an inlet turbulence intensity of Ir, mfrl =
0.15 correlates well to the wind-tunnel study results.
292
1.0
0.9 - Layout A:
0.8
wind tunnel study
0.7
- -- =
0
= 0.1 m
0.6 -- :
0
=0.5m
""'
'
0.5
--- - :,,= 1.0 m
"
A 0 61
0.4
+
0.3
0.2
0,1
OJ
0 a/ o
·3 ·2 ·1 0
Net Pressure Coefficient (AC~)
Figure 8-17: Cp,net profiles at mid (simulation) and 1/3 (wind tunnel) points with varied
roughness length (U = 10.5 mls,lr = 0.15, a= 0.18, z
0
: varies).
1.0
0.9
0.8
0.7
0.6
""'
...._
0.5
.•
0.4
0.3
0.2
0.1
0
A x/L=
-3
I
./
/
/
!
I
I
I
·,
j
i
I
i
I
I
I
-2
=-- -~-~ -~~_ """:! _ "'5"'~-...:::~~-
-
r
/ --""
I
I
I
I
I
I
I
I
l
I
I
I
I
I
I
~ I
I
I
I
I
.-!'
/
/
...._,,_
t
- 1
Net Pressure Coefficient (AC~)
0
- Layout A:
wind tunnel study
l, = 0.10
l, = 0.1 5
I, = 0.20
t, = 0.25
l,=0.30
[]
a / o
Figure 8-18: C p,net profiles at mid (simulation) and 1/3 (wind tunnel) points with varied inlet
turbulence intensity (U = 10.5 m/s,lr : varies, a= 0.1 8, z
0
= 2.0 m).
293
0
Marques da Silva and Gomes (2008) - Wind Tunnel Studies
Gap Inner Pressures in Multi-Storey Double Skin Facades
- 1
I
- 1
0.9
0.8
0.7
0.6
0.5
.:;:
......
. .
0.4
0.3
0.9
0.8
0.7 -
0.6 -
:::: 0.5 -
"
0.4 -
0.3 -
Vaglio (2015) - 30 Simulation COMSOL Multiphysics
(U
0
= 10.5 mis, IT:'"'" = 0.15, a = 0.1 8, = n = 2.0 m)
_/
;
0
---- )
----------- ·-- - -------. ._/ 0.2 0.2
-
0.1
ID
0.1
Front
0.9 m
0
i
0
01 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0
o•
0.1 0.2 0.3
x! a
Figure 8-19: Comparative building-face elevations of Cp,a.t distribution for Sheltered Tower.
The CP distribution is shown as reported in Marques da Silva and Gomes (2008, 1555).
0.4 0.5
x ! a
0.6 0.7 0.8 0.9
8.7 CFO Model Description - Laterally Sealed Sheltered Tower
8.7.1 Introduction
This section simulates the multi-story DSF configuration, Layout D, of Marques da Silva and
Gomes (2008, 1554). This is best described as a laterally sealed sheltered tower on the front and
one lateral (left) side where both second skins have a raised/portal inlet and a through outlet.
8.7.2 Test Configuration
The tested specimen is a rectangular geometry similar to that of the previous Sheltered Tower-
Layout A- presented in the previous section. The primary difference is the sealing of each
cavity's sides, starting at approximately 0.107 h above the ground level and continuing the
remaining height of the building. The cavity depth between skins is varied in Marques da Silva
and Gomes (2008); the configurations considered in this comparative calibration include the
equivalent of full-scale cavity depths of s = 0.8 and 1.6 m.
LAYOUT D:
Laterally Sealed
Sheltered Tower
Model Scale
"A= 1 /40
Model Dimensions
a = 32.5 cm
b = 20 cm
h = 70 cm
hstart = 7.5 cm
h = 62.5 cm
°''
s = 2.0 cm
Velocity Profile
U
0
= 10.5 mis
(1 =0.18
Cp Section Locations
Midspan
x/L: 0.35, 0.67
Figure 8-20: Laterally Sealed Sheltered Tower (Layout D) configuration and section cuts.
295
8.7 .3 Comparison of CFO Results to Wind Tunnel Data - Sheltered Tower D
The simulations of the sheltered tower are compared to the results of Marques da Silva and
Gomes at the mid-span, xl a = 0.5, as shown as the section cut on the Cp distribution elevation
presented in Figure 8-22. This CP distribution is shown as it was presented in Perino's State-of
the-Art Review (2007, 30)
5
. The C p,net (or ~Cp) profiles for x/L (equivalent to xla) = 0.35 and 0.67
presented in Figure 8-21 are extracted for the graphs presented in Marques da Silva and Gomes
(2005)
6
. The simulations' results at the mid-span show a strong agreement with the referenced
wind tunnel studies, both in profile shape and Cp,net magnitude. Varying the cavity depth between
s = 0.8 and 1.6 m has a minimal overall impact with slight differences occurring in the zone just
above the raised inlet (0.10 Zlh 0.20). Figure 8-22 displays similarities in the two models' C p,mt
magnitudes. Once again, the simulation shows a greater magnitude at the near-ground condition,
but the overall elevation is closer in magnitude than Layout A. This suggests that the wind tunnel
studies have a greater sensitivity to the communication with other skins presented by the
unsealed condition (Layout A) as opposed to the sealed condition (Layout D). An inlet turbulence
intensity of Ir mlet = 0.10 exhibits the profile closest to the wind tunnel's c p,net profile.
1.0
0.9
r_
- Layout D:
wind tunnel study
0.8
0.7
0.6
..:;:
'- 0.5
"
0.4
0.3
0.2
0.1
0
[!]
e / o
*
+
D: &O
\
,.,,
r
- /
1 11
.:1
111
=0.10
- / r.11. 1 ,., = 0.15
- I =0.20
r. .i.• ...
J, 16fll = 0.20
--- -- ---- ---- --- -~ ----~ --- " -~- · -·~~ ---~- -~ ----L ----~ ----~ ---~ ----~ ---------------
·3 ·2 · 1
Net Pressure Coefficient (AC~ )
Figure 8-21: C p,nt profiles at mid (simulation) and 1/3 (wind tunnel) points with varied inlet
turbulence intensity (U = 10.5 m/s, fr : varies, a = 0.18, z
0
= 2.0 m, s = {0.8 m, 1.6 m}).
296
0
0
Marques da Silva and Gomes (2005) - Wind Tunnel Studies
Gap Inner Pressures in Multi-Storey Double Skin Facades
0.1 0.2 0.3
0
0.4 0.5
xla
\
-0.4
0.6 0.7 0.8 0.9
0.9
O.B
0.7
0.6
0.5
-:::
'
"
0.4
0.3
0.2
0.1
0
1
0.9
O.B
0.7
0.6
-:::
0.5
'
"
0.4
0.3
0.2
D 0.1
0 .8m
Fr011 I
i
0
()"
-
0 0.1
Vaglio (2015) - 30 Simulation COMSOL Multiphysics
(U
0
= 10.5 mis,/~'"'"= 0.15, u = 0.1 B , z
0
= 2.0 m)
0.2 0.3 0.4 0.5
xla
-0.6
0.6 0.7 0.8
Figure 8-22: Comparative building-face elevations of Cp,int distribution for Laterally Sealed Sheltered Tower.
0.9
The CP distribution for Marques da Silva and Gomes (2005) is shown as it \Nels presented in Perino's State-of-the-Art Review (2007, 30)
4
.
8.8 Evaluation Criteria Moving Forward
The use of computational fluid dynamic simulations holds distinct advantages over
commissioning wind tunnel studies, including cost-saving, time, flexibility to vary parameters and
a higher fidelity of data points to extract results from. A simulation mesh can be refined to contain
many times more data points compared to the number of pressure taps that may be integrated in
a physical wind tunnel model. The disadvantages of simulations compared to wind tunnel studies
are quality assurance of modelling techniques, their consistency, and thus, questions regarding
their accuracy. What the calibrations herein have highlighted is an approach that yields results
with reasonable agreement to existing wind tunnel data with a general tendency to over-magnify
pressures on some portions of the leeward face.
The two calibration exercises compared simulation data to that of simple, unsheltered cube wind
tunnel studies (Holscher and Niemann 1998) (Richards, et al. 2007) and multi-story double-skin
wind tunnel studies (Marques da Silva and Gomes 2005, 2008). The simulations show the
importance - and sensitivity - of modeling of turbulence intensity and roughness in the near
ground boundary condition. The disagreement between the wind tunnel studies and simulations is
generally focused on regions of flow separation and reattachment, such as the parapet or
transition from leeward face to the roof surface. This is in alignment with Overend and Zammit's
(2006, 8) observation that CFO displays less accurate results in flow separation and reattachment
zones. Pressures on windward and leeward faces are predicted with much greater accuracy and
consistency. The overarching parameters that consistently yielded agreeable, or near-agreeable,
results for the multiple layouts considered were an inlet turbulence intensity of Ir, mlee = 0.15 and a
roughness length of z
0
= 2.0 m. The CP·"" profiles with these parameters for Layout A and Layout
D of Marques da Silva and Gomes (2005; 2008) are shown in Figure 8-23. Since the previous
simulations delineated the simulations' CP·"" profiles at x/a = 0.5, this graph looks to match the
points of the wind-tunnel studies with profiles for x/a = 0.35 and 0.67.
298
1.0
Wind tunnel study:
0.9
- Layout A
--- Layout D
O.B
0.7
- A., =0.35
I
0.6
..::
...__
0.5
"
0.4
0.3
0.2
0.1
0
•-----+ I
\1 \
1 \. I
I I
I I
l I
I )
I \
I i
I I
// .//
/ ~ /
,,,,.-' "' ...... '
------------------"~~~~~ --= -- -~ ---- ;;_.:.:; ---.:.:..:. --- -~ -- :.::..:.: ---.:..:.:. ----= --------- --
~ '
r ,\
---- A., = 0.67
D., = 0.35
A: x/L= & 0.6
D,, = D.67
D
0
e/ LayoutA
-3 0
8/ LayoulD
·2 · 1
Net Pressure Coefficient (AC~)
0
Figure 8-23: Cp,net profiles at x/a = 0.35 and 0.67 for simulations compared to wind tunnel
studies (U = 10.5 rn/s,lr = 0.15, a = 0.18, z
0
= 2.0 m, s = 0.8 m).
The graph shows that for both Layout A and D, the selected set of input parameters yield Cp.net
profiles that are near-matches for the peak pressures just above the air inlet, 0.10 zl h 0.20.
For the remaining height, zlh > 0.20, the simulations follow a similar profile shape but have a
greater magnitude. This overestimate is viewed as acceptable for the simulations carried out in
the following chapters since it is encompasses the pressure profile and has a consistent ability to
match the peak magnitudes in the inlet and outlet regions.
The findings of the calibration exercises simulating existing wind tunnel studies of unsheltered
and multi-story double-skin facade building configurations conducted in this chapter have led to a
selection of the following input parameters for use in the simulations of other configurations in the
following chapters: U = 10.5 rn/s, Ir = 0.15, a = 0.18, z
0
= 2.0 m.
299
Chapter 8 Endnotes
Information for this figure based on Wind Loading Codification in the Americas: Fundamentals
for a Renewal as presented in (Simiu 2009, 12).
2
Within the atmosphere, the region closet to the ground, within approximately one kilometer,
contains the atmospheric boundary layer (ABL). It is the layer where the atmosphere is
influenced by the earth's surface: ground, land or sea. Above this layer is the gradient wind
which is does not vary with height (Liu 1991, 41 ). Within the atmospheric boundary layer, the
velocity profile generally consists of turbulent flow. The velocity is zero at the surface, but with
increasing height, the velocity increases to a maximum known as the free stream velocity. It is
this height that delineates the edge of the boundary layer. The velocity profile approaches the
free stream velocity height asymptotically; therefore, the boundary layer thickness is
customarily defined as the height where the velocity is 99% of the free stream velocity. The
gradient height, zg, is also presented for each terrain roughness classification in Table C26.7-2
of ASCE/SEI 7-10 (2010, 541 ); commentary is provided in the footnotes for this table.
3
The landscape descriptions are judiciously modified to shorten their length.
4
The definition of equivalent sand roughness height, ks,ABL - specifically the coefficient - varies
from one CFO program to the next. For instance, for Ansys CFX the coefficient is 29.6. Per
Blacken, Stathopoulos and Carmeliet (2007), ks,ABL "' 30 z
0
.
5
The difference between the data presented in Effects of different multi-storey double skin
facade configurations on surface pressures (2005) and Gap inner pressures in multi-storey
double skin facades (2008) was clarified by Fernando Marques da Silva via email
correspondence on October 13, 2011. There is no difference in the data common to the two
papers. The only differences are the extent presented (in 2008, three cavity depths were
studied compared to just one, 0.8 m, in 2005) and the method of presenting the results. The
results for the double-skin configurations in Effects of different multi-storey double skin facade
configurations on surface pressures (2005) were communicated as the net tiCP = cp,osF -
Cp,unsht, where as the same configuration results in Gap inner pressures in multi-storey double
skin facades (2008) were presented as the measured values themselves, cp,osF·
6
These diagrams are lifted directly from Figure 2.19 in Perino's State-of-the-Art Review (2007,
30). In that document, these CP distributions are presented as for an "unsheltered" envelope
and a "MS-DSF", for a perpendicular wind incidence where the multi-story configuration is open
at the top and bottom, closed laterally, has an inlet starting 3 m above the ground plane (full
scale) and a full-scale cavity depth of 0.8 m.
7
This refers to the profile that is closest in shape while not underestimating any significant
portion across the profile height. The areas of overestimate are viewed as acceptable for the
simulations carried out in the following chapters since it is encompasses the pressure profile
and has a consistent ability to match the peak magnitudes in the inlet and outlet airflow opening
regions.
8
Ibid.
300
9
These CP,"" profiles for Layout A and D at x!L = 0,35; 0,67 are extracted directly from Figure 3a
in Marques da Silva and Gomes (2005), These curves are derived as the "difference between
the local CP values for each layout and the correspondent value for the unsheltered envelope,
for 0° incidence," This figure also presents similar CP,""profile curves for other layouts and wind
incidences,
10
This refers to the profile that is closest in shape while not underestimating any significant
portion across the profile height The areas of overestimate are viewed as acceptable for the
simulations carried out in the following chapters since it is encompasses the pressure profile
and has a consistent ability to match the peak magnitudes in the inlet and outlet regions,
301
9. Steady-State Simulation to Determine Wind Pressure Distribution
on Multi-Story DSF
This section presents the three-dimensional steady-state simulation models and results for one
single-skin and twelve double-skin facades with various multi-story cavity configurations.
9.1 Introduction to Steady-State Simulations
As opposed to three-dimensional transient simulations, steady-state simulations are preferred for
many engineering applications because 1) often, only time-averaged values are of concern, 2)
they require less simulation time and 3) they have simplified post-processing. Like most
engineering flow problems, the multi-story double-skin facade prototypes in an atmospheric
boundary layer wind tunnel test section - or in reality - are turbulent in nature. Since the flow of
eddies in a turbulent environment are innately unsteady and three-dimensional, it is vital to
appropriately model turbulence and in three dimensions. For the simulations herein, a Reynolds
Averaged Navier-Stokes (RANS) simulation is preferred as it is time-averaged, can simulate
turbulent flows and is a widely used approach.
9.2 Description of Experimental Model
The steady-state simulations utilized a three-dimensional simulation domain representative of the
wind tunnel test section, as shown in Figure 9-1. The turbulent flow simulations incorporate a
Reynolds Averaged Navier Stokes (RANS) turbulence model type with a k-B turbulence model
with incompressible flow. The simulations maintain the parameters developed in the preceding
calibration Chapter 8 and multi-story prototype configuration established in Chapter 7. The
simulations were conducted in the CFO module of COM SOL Multiphysics 5.0.0.244. Further
information regarding geometry, boundary conditions, loading, turbulence model and data
reporting locations follows.
302
9.2.1 Geometry
The simulations occur in an open-circuit wind tunnel with a 9 m x 3 m x 2 m test section, similar to
Marques da Silva and Gomes (2005; 2008). The three-dimensional solution domain includes
study models that are centered along the Lwt length and W w t width, as well as placed with its
base at Hwt = 0, unless noted otherwise.
H =2 m
wt
Figure 9-1 : Simulated wind tunnel configuration with single-skin building configuration.
Base Building
The building geometry is the same as the multi-story prototype outlined in Section 7.5: height of
h= 20 m, width of a = 45 m and depth of b = a I 2 = 22.5 m, at a model scale of ,1 = 1/40.
Second Skin
The second skin is placed upstream of the leeward building face, offset a distance of s = 1.0 m.
Each inlet condition (face, trench, raised and shingled) has a unique location of the inlet (see
Figure 9-2), but for all instances, the top of the second skin and height of the cavity's top is at h
10
P
= 21 m, a distance of s = 1.0 m above h= 20 m.
303
Openings
The inlet and exhaust openings are of equal dimension, each maintaining a total area of s *a.
The primary exception, the Shingled inlet, distributes the equivalent opening area across four inlet
bands equally spaced across the outer skin height: (4 locations)* (s/4) * a = s *a. There are
two of these smaller inlets visible in Figure 9-2's diagram for the Shingled inlet. Each model's
opening geometry is introduced at the beginning of their respective results in Sections 9.3 to 9.7,
but here, Figure 9-2 gives a closer view of each opening within the simulation mesh domain.
Forward Through Return
Face Trench
Exhaust Opening Area
- Inlet Opening Area
Raised Shingled
Figure 9-2: Opening configurations for each inlet and exhaust type.
304
9.2.2 Boundary Conditions
Inlet: A vertical boundary plane at y= 0, the upstream inlet, has an initial velocity of U
0
. The
resulting pressure distribution between the inlet and outlet is variable. The initial turbulence
intensity Ur) and turbulent length scale (Lr} are also defined at the inlet.
Outlet: The vertical boundary plane at y = 9, or the downstream outlet, is specified as a zero
gauge pressure (p
0
= 0) with no viscous stress. The simulation suppresses backflow at the outlet.
Floor. The floor is defined as a wall function with applied roughness. The roughness is modeled
by the equivalent sand roughness, ks,ABL ""' 30 z
0
, as described in Section 8.3.3.
Fluid Properties: The fluid flow in the test section is air with a density of p = 1.22521 kg/m
3
.
Mesh: A user-controlled mesh calibrated for fluid dynamics is used. The mesh parameters include
a minimum element size of 0.002 m, maximum of 0.30 m, maximum growth rate of 1.20, a
curvature factor of 0.80, and a resolution of narrow regions equal to 0.50. The minimum element
size is determined by the thickness given to the outer skin solid. To ensure greater density of
nodes in the mesh in areas of interest, several edge locations receive a user-defined subdivision,
including the inlet, vertical edges and horizontal edges of both the interior and exterior skins.
Horizontal Edges:
150 Elements
Figure 9-3: Key edge locations of user-controlled, forced mesh subdivision.
1
305
9.2.3 Loading
The wind loading introduced at the inlet is characterized by Equation 9.2.1 where Ur,f = 10.5 mis
and a= 0.18.
Uo = Uref * (-z )a
Zref
Eq. 9.2.1
Additionally, for these simulations, the wind load approaches the test subject from a wind
incidence of e = 0° for all models.
9.2.4 Turbulence Model Selection
The turbulent flow simulations incorporate a Reynolds Averaged Navier Stokes (RANS)
turbulence model type with a k-B turbulence model with incompressible flow. The standard k-B
turbulence model has the following free constant values:
Cle= 1.44
C2c = 1.92
Cµ = 0.09
(Jk = 1.00
u, = 1.30
The parameters defined at the inlet that characterize the turbulent flow include turbulence
intensity (IT,inler), turbulent length scale (LT) and roughness length (z
0
).
lr,mlet = 0.15 (see Chapter 8)
LT = 0.4* 0 ~ 0.4* Zoe[ (see Chapter 8)
zo = 2.0 m (see Chapter 8)
Many of these variables have been reviewed and described at length in previous chapters.
306
9.2.5 Data Reporting Locations
Having the ability to define the mesh density allows the number of nodes from which data is
extracted to be controlled. In these models the vertical profile cuts have a minimum of 70 data
points across the vertical elevation, and in some cases, like the Shingled geometries, over 300
data points. Highlighted in Figure 9-4, the evaluation surfaces included the exterior skin (blue)
and the interior skin (green) that occurs behind the exterior. The space between is the air cavity.
Exterior ~
1 -
,
Figure 9-4: Mesh layers of a Trench-Fotward configuration with the exterior skin (blue),
mesh subdivision (grey) and interior skin (green).
The primary data extracted from the simulation models are the local pressures at each mesh
node to determine the local CP value using Equation 1.1.3. This data is presented in two formats
in Sections 9.3 to 9.7: 1) colored contour elevations and 2) profile graphs extracted from
horizontal and vertical locations. In all colored contour elevations the CP values are normalized to
a range of -2.00 to 1.00. Due to the symmetrical condition at the middle of the building, the result
elevations only present 0.50 x/a 1.00 for many of the figures in Section 9.4 through 9.7.
307
To compare the results from one model to another, several vertical slices are extracted from the
building geometries to evaluate the pressure dynamics across the height. These vertical reporting
locations - for both the interior and exterior skins - are diagrammed in Figure 9-5a.
~----_
xla == 0.90
xla == 0.67
xla == 0.50
I
I
I
I
I
I I
I I
L--_J_, __
- - I I
I
Figure 9-5a: Vertical reporting locations: x/a = {0.50, 0.67, 0.90}.
h
Additionally, several horizontal slices are also extracted from the building geometries to evaluate
and compare the pressure dynamics across the width of one model to another. These horizontal
reporting locations - for both the interior and exterior skins - are diagrammed in Figure 9-5b.
~
~
~
~----
:- -
:---
--- I
t ----~
Figure 9-5b: Horizontal reporting locations: zl h = {0.10, 0.50, 0.67, 0.90}.
h
308
9.3 Single-Skin Condition
This section presents the 30 steady-state simulation results for a simple configuration of a Single-
Skin condition with the base building geometry. There is no second skin, outer skin, or difference
between interior and exterior skin; just one building elevation. With no second skin, there are no
inlet or outlet conditions. The results from this Single-Skin model are used in Section 9.8 to
compare and normalize evaluations of the double-skin configurations of Sections 9.4 to 9.7.
9.3.1 Singles-Skin Condition
Single-Skin: The Single-Skin model is a simplified, rectangular building geometry of the following
dimensions: height of h= 20 m, width of a= 45 m and depth of b =a 12 = 22.5 m, at a model
scale of A= 1/40. There is no secondary skin and therefore no inlet or exhaust openings. The
single facade is sealed.
h
z
Figure 9-6: Single-skin configuration.
The Cp,single profiles for all three vertical profile locations, x/a = {0.50, 0.67, 0.90}, are positive
between 0.46 C p,Smgle 0.93 for zlh 0.95. Near the roof edge, zlh > 0.95, they decrease
towards and below zero to -1 .13 for zlh > 0.98 as the profile approaches the edge. Both central
profiles, x/a = {0.50, 0.67}, exhibit similar Cp.S ingle profiles, both hovering around Cp,Single z 0.83 for
0.10 zl h 0.85.
2
The C p,smgle ro.9o J profile displays a smaller negative magnitude but similar profile
in the 0.10 zlh 0.85 region, ranging from 76% to 99% the magnitude of Cp,Szngle ro.
50
;.
309
""
' ~
0.90
0.67
0.50
0.10
0
0
Single-Skin Exterior Skin Pressure Coefficient (Cp.sin,)
0.50
xla
r
0.67
r
1. 0
0 .5
-1.0
-1.5
-2.0
0 c,
0.90 1.0
1.0 -
Pressure Coefficient for Single-Skin (C
1
, ..
0
,.,,) - Vertical
1.0
Pressure Coefficient for Single-Skin (C
1
, ""•') - Horizontal
-cpS"'Qle(O~l
-- c p . s.notcio.(171
-- -c
p,Stnglio(090/
DTI!l
1
'
0
e~
0.5
0
~
u'
·0.5
· 1.0
[
-Cp~(o10)
-cp,S.ngle(O~)
--cp,S.ngle(Ol!l7)
- - - C p,S.nglf'(OOO )
o~-------------+-----.--'----'..._-~
· 1.5-+-----.--..----.----.-----.--..---...----.
·1 .5 ·1.0 -0.5 0 0.5 1.0 1.5 0 0.25 0.5
xla
0.75 1.0
Figure 9-7: Cp,sa.g 1 , elevation (top), vertical Cp ,singl• profiles (bottom left) and horizontal Cp,sa.g 1 ,
profiles (bottom right) for a single-skin sealed configuration.
All four horizontal Cp,Smgle profiles locations, zlh = {0.10, 0.50, 0.67, 0.90}, are positive between
0.01 x/a 0.99. In the horizontal's center zone, 0.10 x/a 0.90, the C p,s ingle values for the zlh =
{0.10, 0.50, 0.67} profiles fall between 0.67 C p.single 0.86 with the following mean values:
Cp,Smgle(a JO) = 0.83, Cp,Smgle(aso) = 0.82 and C p,Smgle (0.67) = 0.82. In the same center zone, C p,:Jngle(0.90) is
smaller in magnitude between 0.62 Cp,S ingle(o. 9 o) 0.67 with a mean value of Cp,smgle ro. 9 o) = 0.64.
310
9.4 Face Inlet
This section presents the 30 steady-state simulation results for three configurations with a face
(a.k.a. flush) inlet condition: Face-Forward, Face-Through and Face-Return.
9.4.1 Face - Forward
Face-Forward (FF): The face inlet occurs within the vertical plane of the outer skin at its lowest
extents (between the simulated test section's floor at 0 z s), immediately adjacent to the
ground or plaza level. The forward outlet exhausts the cavity airflow at roof level, in between h
and h
1 0
P , in the same direction as the impinging load.
I
I
_____ .J __ _
Figure 9-8: Face-Forward configuration.
h
The C p,mt profiles for all three locations, xla = {0.50, 0.67, 0.90}, are negative except for the near-
ground region, zlh 0.025, where there is exposure to the direct impinging load via the exterior
skin's inlet condition. Both central profiles, x/a = {0.50, 0.67}, exhibit similar Cp,mt profiles, both
hovering around Cp,int -;:; -1.00 for 0.10 zlh 0.95. The profile at x/a = 0.90, C p,znt (o.
9
oJ . displays a
smaller negative magnitude but similar profile in the 0.10 zlh 0.95 region, ranging from 79% to
98% the magnitude of C p,znt(o.so ) · The exterior skin profiles also have commonality between the
central profiles, xla = {0.50, 0.67}, that range from 0.97 > C p,ext > 0.75 for the 0.10 zlh 0.90
region. Cp,ext(o.
9
o J is smaller in magnitude than the central profiles' with the sole exception occurring
near the exhaust at roof level, zlh > 0.95.
311
Interior Skin Pressure Coefficient (CP·'"' ) Exterior Skin Pressure Coefficient (C )
P - "xl
...... - - -
1.0
1.0 1.0
0.5
•0.5
-1.0
-1.5
-0.9-000
o .
0
_..-~ ~. 2000 .;....~-------1..>.00{..._ ...... ~---.. ............ --..--i
·2.0
0.5 x ! a 1.0 0.5 x l a
1.0 - 1.0
- c
p.••l050J -CP""!O~l
--- cp. ""1 (0.~) -- c p.....t(0.8 7)
-- - c p, nt'l [0.90)
mr
.
_.,
0 ...._
,~ ...
- - c p. ftt (OJ117)
- c
p, ••• (OQO]
--· C
p ... t (OgQ)
o -,-~~-.-~~...-~___;~.__~.,....~--.~~.....,
-1.5 -1 .0 -0.5 0 0.5 1.0 1 .5 0 0.5 1.0 1.5
Pressure Coefficient for Interior (C~ . ,) and Exterior (C, .• ) Net Pressure Coefficient (~C,)
Figure 9-9: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) for the Face-Fotward configuration.
The net pressure coefficient for both central locations, x/a = {0.50, 0.67}, range from 1.74 C p,net
2.01 for the 0.90 > zlh > 0.10 region. Throughout the central height zone Cp,net(o.
9
o) is smaller in
magnitude. In both the near-roof edge (z/h > 0.95) and near-base condition (z/h 0.10) the C p,11Rt
values at the central locations, x/a = {0.50, 0.67}, reduce significantly from their peak values. The
same is true for x /a = {0.90} except for a peak c p,net(0.90) value occurring at zlh ~ 1.00.
312
9.4.2 Face - Through
Face-Through (FT): The face inlet occurs within the vertical plane of the outer skin at its lowest
extents (between the simulated test section's floor at 0 z s), immediately adjacent to the
ground or plaza level. The through outlet exhausts the cavity airflow directly upward; the cavity is
not capped at z = h top• but instead open to above.
h
I
___ ..! __ _
Figure 9-10: Face-Through configuration.
Similar to the Face-Forward configuration, the Cp,mt profiles for all three locations, x/a = {0.50,
0.67, 0.90}, are negative except for the near-ground region, zlh 0.025, where there is exposure
to the direct impinging load via the exterior skin's inlet condition. Both central profiles, xla = {0.50,
0.67}, exhibit similar Cp,mt profiles, both hovering around Cp,int -:::; -1.00 for 0.10 zlh 1.00. The
profile at xla = 0.90, Cp,int (o .
9
oJ. displays a smaller negative magnitude but similar profile in the 0.10
zlh 1.00 region, ranging from 79% to 97% the magnitude of Cp,mt(o.sor
The exterior skin profiles also have commonality between the central reporting locations, x/a =
{0.50, 0.67}, that possess a pressure coefficient range of 0.96 > Cp,ext > 0.76 for the 0.10 zlh
0.90 region. At the outer reporting location, x/a = 0.90, the exterior skin pressure coefficient profile
Cp,ext(o.
9
oJ is smaller in magnitude than the central profiles' with the sole exception occurring near
the exhaust at roof level, zlh > 0.95, where C p,ext(o .
9
oJ mildly exceeds the others' magnitudes.
313
Interior Skin Pressure Coefficient (CP· '"' )
~··
Exterior Skin Pressure Coefficient (C )
p . .. xt
"I
1.0
1.0 -
0.5
"""
'
•0.5
...
-1.0
·1.5
-0.8000
·2.0
M oo 1 ·0,1000 ll..i
0
0.5 x ! a 1.0 0.5 x l a
-- ---) ,
1.0 - 1.0 t
1 :
,,
••
"
- c
- c p..,..,050)
!'
p,.-1{050)
i: --- cp. irlot (O.~) -- c p.....t(0.87)
1 :
- c p . 11n•to s11 -- - c p,M'l[0.90)
II
' • - - c p. N (OJ117)
"
[]]'
••
- c
'
-:
, , p, .. 1(0QOl ..,
0
'
--I I
--· C ' ,~ ...
I I
p.nt(090)
"'
I I
i
I
~
I
I I
:1
+
• I
' I
I
;1
'
.,_
T
..
,~ I
I
: 1 (
'-".:::::.~
-----
0 0
·1.5 ·1 .0 ·0.5 0 0.5 1.0 1 .5 0 0.5 1.0 1.5
Pressure Coeffic ient for Interior (C~ . ,) and Exterior (CP .. ) Net Pressure Coefficient (~C,)
Figure 9-11: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) for the Face-Through configuration.
The net pressure coefficient for both central locations, x/a = {0.50, 0.67}, range from 1.80 C p.net
2.03 for the 0.90 > zlh > 0.10 region. Throughout the central height zone C p,net(o.
9
o) is smaller in
magnitude. In both the near-roof (z/h > 0.95) and near-base (zlh 0.10) conditions the C p,net values
at x/a = {0.50, 0.67} reduce significantly from their peak values. For xla = 0.90, a peak C p,net(o 90 J
value of 1.75 occurs at z/h :::: 0.81. Cp,net(o.
9
o) is slightly greater than the others for zlh > 1.0.
314
9.4.3 Face - Return
Face-Return (FR): The face inlet occurs within the vertical plane of the outer skin at its lowest
extents (between the simulated test section's floor at z = 0 and z = s), immediately adjacent to the
ground or plaza level. The return outlet channels the cavity airflow back into the impinging
direction via the outer vertical plane between h z htor
Figure 9-12: Face-Return configuration.
I
I
___ .J __ _
h
-- -
Unlike the Face-Forward and Face-Through configurations, the Cp,znt profiles for all three
locations, x/a = {0.50, 0.67, 0.90}, are positive, ranging from 0.40 C p,int 0.59. Both the
minimum and maximum C p,mt occur within the inlet region, zlh 0.10. Both central profiles, x/a =
{0.50, 0.67}, exhibit similar average values of Cp,int :::: 0.42 while the profile at x/a = 0.90, displays
c p,int(O.W)= 0.45. The profile at x/a = 0.90, cp.mt(0.90) • displays a slightly larger but similar positive
magnitude in the 0.10 zlh 1.00 region, ranging from 103% to 111 % the magnitude of Cp,mt (osoJ ·
The exterior skin profiles for the central reporting locations, x/a = {0.50, 0.67}, exhibit similarities,
possessing a pressure coefficient range of 0.94 > Cp,ext > 0.75 for the 0.10 zlh 0.90 region. At
the outer reporting location, xl a = 0.90, the exterior skin pressure coefficient profile Cp,ext(o.9oJ is
smaller in magnitude than the central profiles', 75% of Cp,ext(o.soJ at zlh = 0.25; the sole exception
occurs near the return exhaust, 0.95 zlh 1.00, where Cp,ext(o.?oJ mildly exceeds the others.
315
Interior Skin Pressure Coefficient (Cp. ,.,) Exterior Skin Pressure Coefficient (C )
p . .. xt
1.0
1.0
0.5
·1.0
·1.5
·2.0
x! a 1.0 0.5 x l a
1.0 -
/I
1.0
f1
,,
f!
- cp( ... 1(0.!>0)
1 !
-Cp.~osoi
f l
--- cP." '°!!i°'
, ; -- c
p, net (D.117)
- cp . .... (0 .67) ti
I /
---c
~
p, n11 (0.90l
- -cp.irt!()67)
DIJI]'
- c p_...,,o.90)
"
'
-:
..,
0
' - -· C p n 0.901 h ' • ..x
t-
"' "'
I '
I
I
-+--
1 !
fl
,;
I'
,!
1 !
'I
0 0
· 1.5 ·1 .0 ·0.5 0 0.5 1.0 1 .5 0 0.5 1.0 1.5
Pressure Coefficient for Interior {C~ . ,,) and Exterior (CP .. ) Net Pressure Coefficient (M')
Figure 9-13: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) for the Face-Return configuration.
Due to the positive interior pressures, Cp.rnl• the net pressure magnitudes are significantly less
than the Face-Forward and Face-Through configurations. The net pressure coefficient for both
central locations, x/a = {0.50, 0.67}, range from 0.34 Cp,net 0.54 for 0.90 > zlh > 0. 10.
Throughout the central height zone Cp,ne t(o.
9
o) is smaller in magnitude. In both the near-roof (z/h >
0.95) and near-base (z/h 0.10) conditions the Cp,net values at x/a = {0.50, 0.67} approach 0.
316
9.5 Trench Inlet
This section presents the 30 steady-state simulation results for three configurations with a trench
(a.k.a. recess or base) inlet condition: Trench-Forward, Trench-Through and Trench-Return.
9.5.1 Trench - Forward
Trench-Forward (TF): The trench inlet permits airflow into the cavity space through a recessed
(beneath the simulated test section's floor, z 0) volume that maintains equal depth and inlet
height as the cavity depth, s. The forward outlet exhausts the cavity airflow at roof level, in
between h and htop• in the same direction as the impinging load.
h
I
___ J __ _
Figure 9-14: Trench-Fotward configuration.
The Cp,mt profiles for all three locations, x/a = {0.50, 0.67, 0.90}, are negative. Both central profiles,
xla = {0.50, 0.67}, exhibit similar C p,mt profiles, both hovering around Cp,int -::::. -0.97 for 0.10 zlh
0.95. The profile at x/a = 0.90, Cp,tnt (o.
9
oJ . displays a smaller but similar negative magnitude in the
0.10 zl h 0.95 region, ranging from 78% to 98% the magnitude of Cp,mt (O .JOJ· The exterior skin
profiles also have commonality between the central profiles, x/a = {0.50, 0.67}, that range from
0.94 > c p,ext > 0.78 for the 0.10 zlh 0.90 region. cp,ext(0. 90) is smaller in magnitude than the
central profiles' with the sole exception occurring near the exhaust at roof level, zlh > 0.95.
317
Interior Skin Pressure Coefficient (Cp. ,.,) Exterior Skin Pressure Coefficient (C )
p . o •x/
- ---
-- ~ -
1.0
1.0
0.5
·0.5
·1.0
·1.5
·2.0
0 'V. !IOOO
0.2000
0
0.5 x l a 1.0 0.5 x l a 1.0
-- + " 1.0 - '- h
""i
I '
1 1
1 ;
- c
I'
p, .. 1(050)
1 ! --- cp. int(O.~)
l! - C p, e.H0 6?)
!1
- - c p. n (oJn)
I '
• I
-: ~ '
- c
' I
p, .. 1 (0QO)
' ~ '
--· C ...
~ I
p, .. 1(090)
I o
~ I
• I I
I I
• I
I
1.0
+
+
- c P ·""'o50) ~
-- c ... ....t(0.87 )
--- c
p,nt'l[0.90)
CIT]]'
.
0
,~
+ t r
'
-1-:--
• I
'
.~ + ~ r r
I{
I
,
tJ
,,
I
'~-
...
0
- ---...,
· 1.5 ·1 .0 ·0.5 0 0.5 1.0 1 .5 0 0.5 1.0 1.5
Pressure Coefficient for Interior (C~ . ,) and Exterior (Cp .• ) Net Pressure Coefficient (~C,)
Figure 9-15: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) for the Trench-Fotward configuration.
The net pressure coefficient for both central locations, x/a = {0.50, 0.67}, range from 1.76 C p.net
1.95 for the 0.90 > zlh > 0.10 region. Throughout the central height zone C p,net(o.
9
o) is smaller in
magnitude. In the near-roof edge (z/h > 0.95) condition the C p,net values at the central locations, x/a
= {0.50, 0.67}, reduce significantly from their peak values. The same is true for x/a = 0.90 except
for a peak c p,net(Q90)value of 1.77 occurring at zlh :::: 1.0.
318
9.5.2 Trench - Through
Trench-Forward (TF): The trench inlet permits airflow into the cavity space through a recessed
(beneath the simulated test section's floor, z 0) volume that maintains equal depth and inlet
height as the cavity depth, s. The through outlet exhausts the cavity airflow directly upward; the
cavity is not capped at z = h1op. but instead open to above.
h
I
___ J __ _
Figure 9-16: Trench-Through configuration.
Similar to the Trench-Forward configuration, the Cp,mt profiles for all three locations, x/a = {0.50,
0.67, 0.90}, are negative. Both central profiles, x/a = {0.50, 0.67}, exhibit similar C p,znt profiles, both
hovering around C p,int -:::; -1.00 for 0.10 zlh 1.00. The profile at xla = 0.90, C p,mt(o
9
o J. displays a
smaller negative magnitude but similar profile in the 0.10 zlh 1.00 region, ranging from 78% to
96% the magnitude of cp.znl {0.50) ·
The exterior skin profiles also have commonality between the central reporting locations, xla =
{0.50, 0.67}, that possess a pressure coefficient range of 0.94 > Cp,ext > 0.80 for the 0.10 zlh
0.90 region. At the outer reporting location, x/ a = 0.90, the exterior skin pressure coefficient profile
Cp,ext(o.9oJ is smaller in magnitude than the central profiles' with the sole exception occurring near
the exhaust at roof level, zlh > 0.95, where Cp.ext(o.
9
oJ mildly exceeds the others' magnitudes.
319
Interior Skin Pressure Coefficient (CP·'"')
1.0
1.0 -
0.5
•0.5
-1.0
·1.5
·2.0
0 -
- n.2000
-0.7000
- 1.S.S
0.5 x l a 1.0 0.5
1.0 - 1.0
-c,.~"'"'
--- cp. int (O.~)
- c p . 1JJ>•tos11
- - c p.nttoJn)
- Cp ... ,(O~)
--· C
p ,n! (O~)
-+---
t
Exterior Skin Pressure Coefficient (C )
p . o'.d
- C P""!Oe-Ol
-- c ... ....t(0.87 )
-- - cp, nt'I 10.90)
[[TIJ
1
.
0
,~
+
x l a
,
,
,
I
'
' I
I
+ : - 1
'
' I
,
I
\
\
\
\
I
\
I
I
I
I
I
I
I
\
I
I
I
I
/,
1.0
0 -i.~ --~ -~ -~ ;~ --~ -~ -~ -~ --~ -~ -~ -~ --~ -= -= -- = -~ -= -= - -=-~ -~ -= -==:;::==...,..~--.
·1.5 ·1 .0 ·0.5 0.5 1.0 1 .5 0 Q5 1~ 1~
Pressure Coefficient for Interior (C~ . ,) and Exterior (Cp .• ) Net Pressure Coefficient (~C,)
Figure 9-17: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) for the Trench-Through configuration.
The net pressure coefficient for both central locations, x/a = {0.50, 0.67}, range from 1.80 C p.net
1.97 for the 0.90 > zlh > 0.10 region. Throughout the central height zone, C p,net(o.
90
; is smaller
magnitude. In both the near-roof (z/h > 0.95) and near-base (z/h 0.10) conditions, the C p,net
values at xla = {0.50, 0.67} reduce greatly from their peak values. For xla = 0.90, a peak C p,net(o.
9
oJ
value of 1.72 occurs at z/h :::: 0.86. Cp,net(o.
9
o) is slightly greater than the others for z/h > 1.00.
320
9.5.3 Trench - Return
Trench-Return (TR): The trench inlet permits airflow into the cavity space through a recessed
(beneath the simulated test section's floor, z 0) volume that maintains equal depth and inlet
height as the cavity depth, s. The return outlet channels the cavity airflow back into the impinging
direction via the outer vertical plane between h z htop·
I
I
___ J __ _
Figure 9-18: Trench-Return configuration.
h
-- -
Unlike the Trench-Forward and Trench- Through configurations, the C p,mt profiles for all three
locations, x/a = {0.50, 0.67, 0.90}, are positive, ranging from 0.43 C p,mt 0.56. Both the
minimum and maximum C p,mt occur within the inlet region, zlh 0.10. Both central profiles, xla =
{0.50, 0.67}, exhibit similar average values of Cp,int :::: 0.46 while the profile at x/a = 0.90 displays
c p,int (O.W) = 0.49. The profile at x/a = 0.90, c p.mt (0.90) • displays a slightly larger but similar positive
magnitude in the 0.10 zlh 1.00 region, ranging from 104% to 110% the magnitude of C p,mt (osoJ·
The exterior skin profiles for the central reporting locations, x/a = {0.50, 0.67}, exhibit similarities,
possessing a pressure coefficient range of 0.89 > C p,ext > 0.76 for the 0.10 zlh 0.90 region. At
the outer reporting location, xla = 0.90, the exterior skin pressure coefficient profile Cp,ext(o.9oJ is
smaller in magnitude than the central profiles', 76% of C p,ext(o.so J at zlh = 0.25; the sole exception
occurs near the return exhaust, 0.95 zlh 1.00, where Cp,ext(o.?oJ mildly exceeds the others.
321
Interior Skin Pressure Coefficient (C P· '"' ) Exterior Skin Pressure Coefficient (C )
p.o>.:t
1.0
1.0
0.5
·1.0
-1.5
-2.0
0 0
0.5 x l a 1.0 0.5 x l a 1.0
1.0 -
, ,
ii
1.0
I'
1 !
-cp~•1(050)
1 !
- Cp.n.t!MO)
1 1
--- c
p."11!0 .!IO) i
-- c
--·
p.no!I0,6 7)
- C p.-•1 1os1i
~
' I - - - cp. ""110.90,
'
- -c
I l
p. .,llO.fl?)
_J
I
[]]'
'
- cp. .. 110 .aoi
"
I
I
-:
..,
I 0
' 1 1 '
'
I . .,.
... --· C ...
p ...... (090)
I '
I
1 !
I
+
I
1 ! I
1 1
' I
I
1 :
. I
' I
' I'
'
I
I
1 !
' I
'
1 !
'
I
'
I
I!
I
-
,
0 0
·1.5 ·1 .0 ·0.5 0 1.0 1.5 0 0.5 1.0 1.5
Pressure Coeffic ient for Interior {C~ . ,) and Exterior (CP .. ) Net Pressure Coefficient (~C,)
Figure 9-19: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) for the Trench-Return configuration.
Due to the positive interior pressures, C p.inl• the net pressure magnitudes are significantly less
than the Trench-Forward and Trench-Through configurations. The net pressure coefficient for
both central locations, x/a = {0.50, 0.67}, range from 0.30 C p.oct 0.44 for 0.90 > zlh > 0.10.
Throughout the entire building height Cp,net(o.
9
o) is smaller in magnitude. In both the near-roof (z/h >
0.95) and near-base (z/h 0.10) conditions, the C p,net values at x/a = {0.50, 0.67, 0.90} approach 0.
322
9.6 Raised Inlet
This section presents the 30 steady-state simulation results for three configurations with a raised
(a.k.a. entry) inlet condition: Raised-Forward, Raised-Through and Raised-Return.
9.6.1 Raised - Forward
Raised-Forward (RF): The raised air inlet is lifted one story beginning at zlh = 0.25 and occurs in
the horizontal plane, or underside, of the cavity volume and has a continuous cavity depth, s. The
forward outlet exhausts the cavity airflow at roof level, in between h and h
0
P, in the same direction
as the impinging load.
hcav
h
Figure 9-20: Raised-Fotward configuration.
All three Cp,int profiles, x/ a = {0.50, 0.67, 0.90}, are negative above zlh > 0.25 where there is a
second skin and positive beneath zlh 0.25 where the interior skin is exposed. Both central
profiles, x/a = {0.50, 0.67}, exhibit similar Cp,int profiles, hovering around Cp,int ::= -0.64 for 0.30 zl h
0.95. The profile at x/a = 0.90, Cp,rnt(o .
9
oJ. displays a smaller but similar negative magnitude in the
0.30 zlh 0.95 region, ranging from 58% to 97% the magnitude of Cp,znJ(o.so) · The exterior skin
profiles also have commonality between the central profiles, x/a = {0.50, 0.67}, that range from
0.96 > Cp,ext > 0.76 for the 0.30 zlh 0.90 region . Cp,ext(o.9o) is smaller in magnitude than the
central profiles' with the sole exception occurring near the exhaust at roof level, zlh > 0.95.
323
1.0
0
0.5
1.0-
Interior Skin Pressure Coefficient (Cp.;m)
-0.7000
'"'°.1000
-- - ----- '
-~
-...
~' ,,
,.
x!a
'
\
\ - C P."•(o5oJ
~ I c
:1 \ - -- p,in.t(0.50)
~I \ - C ri. •nH067l
:i \ - -cp. ini(o.en
:•
:1
: •
:1
,'I
'/
- - · c p. .. 1 (090)
' I
'- -..:..:::::..~
-0~
'
1.0
·1.5 ·1.0 ·0.5 0 0.5 1.0 1.5
Pressure Coeffic ient for Interior (C~ . ,.,) and Exterior (CP·"')
1.0
0.5
-1.0
-1.5
·2.0
0.5
1.0
0
Exterior Skin Pressure Coefficient (C, .• )
- c p.ne((0.50)
-- cp, -(CUl7)
--- CP,~(o.oo)
[]JIJ
1
.
0
,--"
xla
Q5 1~ 1~
Net Pressure Coefficient (l.r)
0
1.0
Figure 9-21: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) for the Raised-Fotward configuration.
The Cp,net values for central locations, x/a = {0.50, 0.67}, range from 1.34 Cp,net 1.69 for 0.90 >
zl h > 0.30. In the near-roof edge (z/h > 0.95) condition C p,net(o.
5
oJ and C p.net(o.
67
J experience a positive
increase before reducing considerably. The same is true for C p,net(o.
9
o) with a peak value of 1.64
occurring at zlh :::: 1.00. Throughout the central height zone Cp,net(o.
9
o) is smaller in magnitude. It is
only at zlh > 0.95 that the C p,net(o. 9 o) profile has values that exceed Cp.net(o. 5 o) and C p,net(o.67 ) .
324
9.6.2 Raised - Through
Raised-Through (RT): The raised air inlet is lifted one story beginning at z/h = 0.25 and occurs in
the horizontal plane, or underside, of the cavity volume and has a continuous cavity depth, s. The
through outlet exhausts the cavity airflow directly upward; the cavity is not capped at z = htop• but
instead open to above.
hcav
h
I
hstart
- - ..J - -
Figure 9-22: Raised-Through configuration.
Similar to the Raised-Forward configuration, all three Cp,mt profiles, x/a = {0.50, 0.67, 0.90}, are
negative above zl h > 0.25 where there is a second skin and positive beneath zlh 0.25 where the
interior skin is exposed. Both central profiles, x/a = {0.50, 0.67}, exhibit similar Cp,znt profiles,
hovering around Cp,in t -::; -0.74 for 0.30 zlh 0.95. The profile at xla = 0.90, Cp,int(o .9 o J . displays a
smaller negative magnitude but similar profile in the 0.30 zlh 0.95 region, ranging from 60% to
97% the magnitude of C p,inJ (o.so J·
The exterior skin profiles also have commonality between the central reporting locations, x/a =
{0.50, 0.67}, that possess a pressure coefficient range of 0.97 > Cp,ext > 0.82 for the 0.30 zlh
0.90 region. At the outer reporting location, x/a = 0.90, the exterior skin pressure coefficient profile
Cp,ext(o.
9
oJ is smaller in magnitude than the central profiles' with the sole exception occurring near
the exhaust at roof level, zlh > 0.95, where Cp, ext(o .? oJ mildly exceeds the others' magnitudes.
325
1.0
0
0.5
1.0 -
Interior Skin Pressure Coefficient (Cp . ;m)
-0.600D
-0,2000
- - - -----
--- --..,_
ti
! I
• '
i\
:1 \
:1 \
I '
x!a
- c I
l" .,.;• (0.!50)
--- cp, in.t(O.M)
- Cp.•c• ioen
•I I
:
1
\ - -cp. ini (o.en
I I
: ' ' - c p.•105'0)
•I
I
,,
- - · C ,,..,11 090J
,
•I
/1 I
, I
'--~-
0 10000
1.0
-1.5 -1.0 -0.5 0 0.5 1.0 1.5
Pressure Coefficient for Interior ((~ ·'"') and Exterior (CP·"')
1.0
0.5
·0.5
-1.0
-1.5
-2.0
0.5
1.0
0
Exterior Skin Pressure Coefficient (C,. ")
- c p.ne1(0.so)
-- c p,-(0.87)
--- c
p,nt:(0.90)
[]]
1
.
0
,--"
x l a
Q5 1~ 1~
Net Pressure Coefficient ~,)
0
1 .0
Figure 9-23: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) for the Raised-Through configuration.
The Cp,net values for central locations, x/a = {0.50, 0.67}, range from 1.52 Cp,net 1.79 for 0.90 >
zl h > 0.30. Throughout the central height zone Cp,net(o .
9
oJ is smaller in magnitude. In both the near-
roof (zlh > 0.95) and near-base (zl h 0.10) conditions, the C p,net values at xla = {0.50, 0.67} reduce
significantly from their peak values. For xla = 0.90, a peak Cp,net(o.
9
oJ value of 1.44 occurs at zlh :::::
0.89. cp,net(0.90) is slightly greater than the others for zlh > 1.00.
326
9.6.3 Raised - Return
Raised-Return (RR): The raised air inlet is lifted one story beginning at z/h = 0.25 and occurs in
the horizontal plane, or underside, of the cavity volume and has a continuous cavity depth, s. The
return outlet channels the cavity airflow back into the impinging direction via the outer vertical
plane between h z h1op·
h
Figure 9-24: Raised-Return configuration.
Unlike the Raised-Forward and Raised-Through configurations, the C p,mt profiles for all three
locations, x/a = {0.50, 0.67, 0.90}, are positive, ranging from 0.53 C p,mt 0.95 with the greatest
values occurring beneath the inlet, zlh 0.25, where the inner skin is exposed. Both central
profiles, x/a = {0.50, 0.67}, exhibit similar Cp,int profiles, hovering around Cp,int -:::: 0.55 for 0.30 zlh
0.95. The profile at x/a = 0.90, Cp,int( o.
9
oJ. displays a larger positive magnitude but similar profile in
the 0.30 zlh 0.95 region, ranging from 104% to 110% the magnitude of C p,mt (o.so;.
The exterior skin profiles also have commonality between the central reporting locations, x/a =
{0.50, 0.67}, that possess a pressure coefficient range of 0.92 > Cp,ext > 0.78 for the 0.30 zlh
0.90 region. At the outer reporting location, x/a = 0.90, the exterior skin pressure coefficient profile
Cp,ext(o.
9
oJ is smaller in magnitude than the central profiles' with the sole exception occurring near
the exhaust at roof level, zlh > 0.95, where Cp,ext(o.?oJ mildly exceeds the others' magnitudes.
327
Interior Skin Pressure Coefficient (Cr.'",)
1.0
x/a 1.0
1.0 -
- c
f>, iecl(050J
--- cp.ini(0.50)
+- +-
·1.5 ·1.0 ·0.5 0 0.5 1.0 1.5
Pressure Coeffic ient for Interior {C~ . ,) and Exterior (Cp .• )
1.0
0.5
0.5
-1.0
-1.5
·2.0
0.5
1.0
Exterior Skin Pressure Coefficient (C , . ,,)
x l a
+-
-cp,~wooi
-- cp, net (o. 111)
--- c
p, nlll (0.90)
[]]]
' .
0
.~
+- --+--
1.0
0 -+----...-----.--..--~--.---~---...-----.
0 0.5 1.0 1.5
Net Pressure Coefficient (6.CP)
0
Figure 9-25: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) for the Raised-Return configuration.
The Cp,net values for central locations, x/a = {0.50, 0.67}, range from 0.22 Cp,net 0.38 for 0.90 >
zlh > 0.30. Throughout the central height zone Cp,ext( o.
9
oJ is smaller in magnitude. In both the near-
roof (zlh > 0.95) and near-inlet (0.25 zlh 0.30) conditions, the Cp,net values at xla = {0.50, 0.67},
reduce significantly from their peak values towards 0. For x/a = 0.90, a peak C p,net(o.
9
o J value of 0.18
occurs at zlh == 0.84.
328
9.7 Shingled Inlet
This section presents the 30 steady-state simulation results for three configurations with a
shingled inlet condition: Shingled-Forward, Shingled-Through and Shingled-Return. For the
following subsections, exterior Level 1 refers to 0.0125 zlh 0.2625, Level 2 refers to 0.275
zlh 0.525, Level 3 refers to 0.5375 zlh 0.7875, and Level 4 refers to 0.80 zlh 1.0.
9.7.1 Shingled - Forward
Shingled-Forward (SF): The shingled inlet permits airflow into the cavity through four opening
bands, each 0.25*s tall, that occur at ground level and hypothetical floor levels at quarter points
up the facade. The forward outlet exhausts the cavity airflow at roof level, in between h and h
1
op, in
the same direction as the impinging load.
s/44== i
s/4 +=== i --===--
~=~
h
s/4
Figure 9-26: Shingled Forward configuration.
All three C p,mt profiles are negative for the entire height. Both central profiles, x/a = {0.50, 0.67},
exhibit similar profiles, incrementally stepping from Cp,mt ~ -0.11 for Level 1, Cp,mt ~ -0.23 for Level
2, Cp ,int~ -0.44 for Level 3, and C p,int~ -0.77 for Level 4. The C p,int(o.
9
oJ profile has a smaller
magnitude but similar profile for zlh 0.95. The exterior skin profiles also have commonality
between x/a = {0.50, 0.67}, both incrementally stepping from Cp,ext~ 0.82 for Level 1, Cp,ext~ 0.85
for Level 2, Cp,ext~ 0.81 for Level 3 and Cp,ext~ 0.77 for Level 4 . C p,ext(o. 9 o J is smaller in magnitude.
329
1.0-
0.5
1.0 -
Interior Skin Pressure Coefficient (Cp . ;m)
~
1'
,.
,,
xla
\ '
~-
~ ... , "
I '
I I
- c
p. e•t(050)
--- c
p,nf {0.501
- C p..,,1(061)
- -c
p,nt {0.67)
- c pM(090)
I '
' I
I ,
~ I,
,,
,,
' I
. '
11
I '
,f I
\ \
"' i:
,,
,;
,,
t!
0.1
1.0
·1.5 ·1.0 ·0.5 0 0.5 1.0 1.5
Pressure Coeffic ient for Interior ((~ ·'"' ) and Exterior (CP·"')
1.0
0.5
·0.5
-1 0
-1.5
·2.0
0.5
1.0
1
..
..
0
Exterior Skin Pressure Coefficient (C, . "")
..
I -';
' r
'
"
xla
.... ':.~~~=-=-= '- ==-
- c p.rttll !0.5-0)
-- c p.ne•(0.6 ?J
- - - cp,Ml (090~
:ITJIJ
1
.
0
. ,,.
Q5 1~ 1~
Net Pressure Coefficient ~r)
0
1 .0
Figure 9-27: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) ) for the Shingled-Forward configuration.
The Cp,net values also display an incremental increase in magnitude with height, up to zl h 0.85. In
the central locations, x/a = {0.50, 0.67}, Cp,n et ranges incrementally step: 0.86 Cp,n et 1.04 for
Level 1 ; 0 .91 Cp.n et 1 .16 for Level 2; 1 .09 Cp.net 1 .38 for Level 3; and 1 .40 Cp.net 1 .67 for
Level 4. Throughout the central height zone Cp,ne t(o .9o) is smaller in magnitude before encountering
a peak of Cp,n e t(o .
9
o) = 1.69 occurring at zlh ~ 1.02.
330
9.7 .2 Shingled - Through
Shingled-Forward (SF): The shingled inlet permits airflow into the cavity through four opening
bands, each 0.25*s tall, that occur at ground level and hypothetical floor levels at quarter points
up the facade. The through outlet exhausts the cavity airflow directly upward; the cavity is not
capped at z = h1op. but instead open to above.
s/4f== :
s/4f== :
I
- ..J - -
Figure 9-28: Shingled-Through configuration.
h
Similar to the Shingled-Forward configuration, all three C p,int profiles are negative for the entire
height. Both central profiles, x/a = {0.50, 0.67}, exhibit similar profiles, incrementally stepping from
Cp,int ~ -0.17 for Level 1, C p,int ~ -0.29 for Level 2, Cp, rnt ~ -0.51 for Level 3 and Cp,int ~ -0.87 for Level
4. The Cp,tnt( o.9o) profile has a smaller magnitude but similar profile for zlh 0.95, ranging from
66% to 97% the magnitude of c p,in/ (050) ·
The exterior skin profiles also have commonality between xla = {0.50, 0.67}, both incrementally
stepping from 0.70 Cp,ext 0.92 for Level 1, 0.65 Cp,ext 0.91 for Level 2, 0.54 C p,ext 0.92 for
Level 3 and 0.64 C p,ext 0.90 for Level 4. At the outer reporting location, xla = 0.90, the exterior
skin pressure coefficient profile Cp,ext(o.
9
o) is smaller in magnitude than the central profiles' with the
sole exception occurring near the exhaust at roof level, zlh > 0.95, where C p,ext(o.9o) mildly exceeds
the others' magnitudes. Cp,ext(o.
9
o) achieves a peak value of 0.83 at zlh ~ 0.84.
331
Interior Skin Pressure Coefficient (Cp . ;m) Ex1erior Skin Pressure Coefficien1 (C, .• )
1.0
1.0-
-0.9000
0.5
·0_5
-1.0
-0.2000
-1.4 4
-1.5
·2.0
0
0.5 x ! a 1.0 0.5 x l a 1.0
1.0 - 1.0
- c
p. ert(050) - c.,_rtt1110.50)
--- c
p,nf {0.501
- C p..,.l(067)
- -c
p,nt {0.87)
-cp~l{090)
-- · C" _ ..... 1090)
·1.5 ·1.0 ·0.5 0 0.5 1.0 1.5 0 Q 5 1~ 1~
Pressure Coeffic ient for lnterior (C~ .,J and Exterior (CP·"') Net Pressure Coefficient ~r)
Figure 9-29: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) for the Shingled-Through configuration.
The Cp,net values also display an incremental increase in magnitude with height, up to zl h 0.85. In
the central locations, x/a = {0.50, 0.67}, Cp,netranges incrementally step: 0.87 Cp,net 1.09 for
Level 1; 0.94 Cp.net 1.23 for Level 2; 1.16 Cp.net 1.47 for Level 3; and 1.40 Cp.net 1.81 for
Level 4. Throughout the central height zone C p,n et(o.9o) is smaller in magnitude before encountering
a peak of Cp,n e t(o .
9
o ) = 1.68 occurring at zlh ~ 0.87.
332
9.7 .3 Shingled - Return
Shingled-Return (SR): The shingled inlet permits airflow into the cavity through four opening
bands, each 0.25*s tall, that occur at ground level and hypothetical floor levels at quarter points
up the facade. The return outlet channels the cavity airflow back into the impinging direction via
the outer vertical plane between h z h
1
w
sl4=F=" i
h
I
__ .J __ _
Figure 9-30: Shingled-Return configuration.
Unlike the Shingled-Forward and Shingled-Through configurations, the Cp,znt profiles for all three
locations, x!a = {0.50, 0.67, 0.90}, are positive, ranging from 0.50 Cp,tnt 0.59 The incremental
stepping is not present on the interior skin. The Cp,znt(o.
9
oJ profile primarily has a slightly larger
magnitude and similar profile ranging from 97% to 108% the magnitude of Cp,mt (o so J·
The exterior skin profiles have commonality between x/a = {0.50, 0.67}, both incrementally
stepping from 0.81 Cp,ext 0.89 for Level 1, 0.82 Cp,ext 0.88 for Level 2, 0.80 Cp,ext 0.89 for
Level 3 and 0.56 C p,tn1 0.84 for Level 4. At the outer reporting location, xla = 0.90, the exterior
skin pressure coefficient profile Cp,ex1(09o J is smaller in magnitude than the central profiles' with the
sole exception occurring near the exhaust at roof level, zlh > 0.95, where Cp,ext(o?oJ mildly exceeds
the others' magnitudes. Cp,ext(o.
9
oJ achieves a peak value of 0.78 at zlh:::: 0.76.
333
Interior Skin Pressure Coefficient (Cp . ;m)
1.0
""'
' ...
0
0.5 xla 1.0
1.0 -
- c,. .. '"'°'
- - - cp ....... 1o.5a1
- C .,,..:1(0611
- -c
p. n to.e1)
- c jl, M(090)
- - · C
p. mlt090)
·1.5 ·1.0 ·0.5 0 0.5 1.0 1.5
Pressure Coefficient for ln1erior {C~ . , J and Exterior (Cp.•)
1.0
0.5
·0.5
-1.0
-1.5
·2.0
0.5
1.0
0
Ex1erior Skin Pressure Coefficien1 (C )
p . •!XI
xla
l
- c, .... ,"'''
-- c1 ,,.110.s7)
--- c
p , 11"1 C0.90)
:JI]J
1
I
0
. ,,.
Q 5 1~ 1~
Net Pressure Coefficient (ti.r,J
0
1.0
Figure 9-31: CP elevations interior (top left) and exterior (top right), CP profiles (bottom left)
and Cp ,net profiles (bottom right) for the Shingled-Return configuration.
The Cp,net values display a subtle incremental increase in magnitude with height, up to zlh 0.60
for x/a = {0.50, 0.67}, and to zlh 0.85 for xla = 0.90. In the central locations, xla = {0.50, 0.67},
C p,net ranges step: 0.25 C p,net 0.32 for Level 1; 0.26 Cp,net 0.33 for Level 2; 0.27 Cp,net
0.36 for Level 3; and 0.13 Cp,net 0.34 for Level 4. Throughout the central height zone C p,net(o. 90 ; is
smaller in magnitude before encountering a peak of Cp,net(o.
9
o) = 0.24 occurring at zlh ~ 0.84.
334
9.8 Discussion of Results
This section overlays and reviews the simulation data for models with 1) a common inlet condition
and varied outlet (see Section 9.8.1), and 2) a varied inlet condition and common exhaust (see
Section 9.8.2). The results are evaluated in several ways within these two sections. The values
presented are:
• the pressure coefficients on both the interior and exterior skins, CP·'"'and Cp.n" at mid
span, x/a = 0.50
• outer skin pressure coefficient ratios, Cp.DSFcma I Cp.smgfr - normalized to the values of a
single-skin configuration under equivalent loading conditions - at x/a = {0.50, 0.67, 0.90)
in the vertical direction
• outer skin pressure coefficient ratios, cp.DSFcma I cp.smgfr - normalized to the values of a
single-skin configuration under equivalent loading conditions - at zih = {0.10, 0.50, 0.67,
0.90) in the horizontal direction'
• net pressure coefficients, CP·""' at x/a = {0.50, 0.67, 0.90), for each of the simulation
models.
9.8.1 Comparison of Inlet Configurations
The first comparisons, shown in Figures 9-32 to 9-47, present the results from three different
simulation models each, holding the inlet condition constant (either Face, Trench, Raised or
Shingled), while varying the exhaust (Forward, Through and Return). These profiles are found in
the previous sections in their respective model, but here they are presented alongside multiple
configurations.
In the first section, CP Values, only the profiles at x/a = 0.50 are shown for the interior and exterior
pressure coefficients (Cp.m,and Cp.n,). The other sections present results of several locations from
each model.
335
CP Values
The CP profiles shown in Figures 9-32 and 9-33 exhibit similarities amongst Face and Trench
configurations, given a similar exhaust. Both the profile shapes and the CP magnitudes coincide.
One of the few variations between the Face and Trench configurations is in the near inlet zone,
zlh 0.15, where the Trench-Forward and Trench-Through exhibit a greater local peak magnitude
(Cp.m,(TF) ~ Cp.m,(TT) ~ -1.18) as compared to the Face-Forward (Cp.m,(FF) = -1.05) and Face-Through
(Cp.m,(FT! = -1.08) models. Above that in the central portion of the height, 0.20 zlh 0.90, both
models hover right around -1.00.
For the Raised inlet conditions, shown in Figure 9-34, the cp.m,(RF) and cp.m,(RT) profiles transition
from a negative to a positive magnitude around zlh ~ 0.25. This occurs where the exterior skin is
not present - beneath its inlet - and also where there is no value for Cp.n,·
The pressure coefficient profiles for the Shingled conditions are shown in Figure 9-35. For these
models, Cp.m,(SFJ and Cp.m,(STJ are negative and follow a stepping profile due to the four inlet
locations across the facade height. The stepping profile is barely pronounced in the Cp.m,(SRJ
profile that is positive in magnitude, hovering around 0.50 to 0.60. The Cp.n, profiles are positive in
magnitude but experience a transition to negative at the top in the near-exhaust region. They also
display a series of four spikes that bring Cp.n, towards zero; these locations correspond to the
location of the four inlets. In reality, the Cp.n, value presented is the pressure measured at the
elevation within the inlets' air gaps.
All four inlet configurations (Figures 9-32 to 9-35) display a distinct positive profile for all Return
exhaust configurations' cp.m, while the Forward and Through exhausts possess negative cp.m,
magnitudes. All of the cp.n, profiles are of similar magnitude with the Forward and Through
exhausts having slightly greater magnitudes than the Return configurations.
336
1.1
1.0
0.9
0.8
0.7
0.6
""
' N
0.5
0.4
0.3
0,2
0.1
0
-1.5
---;- =:;t
~
I:
I:
I:
I:
I ,
1 :
1 :
1 :
u __
1 :
I:
r :
I:
1 . :
1 :
IL
1 , 1
I ,' I
l ~ ... t
- ~-~-~-=.-=.-=.-~-=--~-~-----~-
-1.0 -0.5 0 0.5
Pressure Coefficient for Interior (C,. ,J and Exterior( ~ .• ) Skins
Face Forward
-- Cp.er1 ~FF)
---- c
p. int (FA
Face Through
- - c
p,.lfl(FT}
-- c
•~IFTI
Face Return
---cp.lnl (F~)
1.0 1.5
Figure 9-32: Comparison of CP values at mid-span (x/a = 0.5) for Face inlet configurations
with varied exhaust configurations.
1 .1
-- ::--~ .. -
\ 1.0
~-
-
' \,
Trench Forward
0.9
~ ~
I'
,:
0.8
I
1 :
1 :
---- c
p.int (l'A
Trench Through
- - c
p,.!Cl(ll)
0.7 1 :
1 :
0.6
I'
'
1 :
""'
I '
'
0.5
I
"
1 :
I '
I
Trench Return
-- c p.•d C IHJ
--- cp.llll(TR)
0.4
·':
1 :
0.3
1 :
1 :
I '
0,2
I
'T
I,'
0.1
_ J- 'r
,,,,,,"
0
--.:::..- ----
- -- -=--~-=.-=.- ...... _~
-1.5 -1.0 -0.5 0 0.5 1.0 1.5
Pressure Coefficient for Interior (C, .. J and Exterior (C;_ .) Skins
Figure 9-33: Comparison of CP values at mid-span (x/a = 0.5) for Trench inlet configurations
with varied exhaust configurations.
337
1.1
1.0
0.9
0.8
0.7
0.6
""
' N
0.5
0.4
0.3
0,2
0.1
0
-1.5
---~--: '!. :;-=:- ,--::- _ ~1-=--=t-===:::::::-=--1_
---- "
-1.0
~--
1 '
'
'
'
' T
'
'
'
!--/-
/ ,'
I I
'
....... - '- :... - -::.:-::._- ":...="~~~-=-=-~-
-0.5 0 0.5
Pressure Coefficient for Interior (C,. ,J and Exterior (I ~ .) Skins
Raised Forward
-- Cp.e r1 CRF)
---- c
p.int~F)
Raised Through
- - c
p,.lfl(RT)
- - c p. nl(RT)
Raised Return
--- cp.lnl(RR)
1.0 1.5
Figure 9-34: Comparison of CP values at mid-span (x/a = 0.5) for Raised inlet configurations
with varied exhaust configurations.
1.1
1.0
0.9
0.8
0.7
0.6
""'
'
"
0.5
0.4
0.3
0,2
0.1
0
-1.5
- --:::.·r :-:: _ -: :-- \.- _ --1~---i-~===1....
-1.0
I
I
I -,
'
' I
'
---:.. ---...
- -,
" \
I !
I
I
I ,'
( ',
...... _ ... _ ...
---
-0.5
~ \
I :
I ,'
, ,,_
I '.
..... ..::. .........
I \
_I_!_
I :
I :
I [
\ ',
--,
0 0.5
Pressure Coefficient for Interior (C, .. J and Exterior (C;_ .) Skins
1.0 1.5
Figure 9-35: Comparison of CP values at mid-span (x/a = 0.5) for Shingled inlet
configurations with varied exhaust configurations.
Shingled Forward
---- c
p, int(SF)
Shingled Through
- - c
p ... 111t'S TI
- - C P,W (ST)
Shingled Return
--- C p.m (SR)
338
Outer Skin Ratios - Vertical
Each outer skin's Cp,ext (or Cp,DsFouter) is compared to the equivalent for a single-skin facade,
Cp,Szngle· The ratio is calculated for each of the twelve models at vertical profile locations x/a =
{0.50, 0.67, 0.90}, charted in Figures 9-36 to 9-39 and summarized in Table 9-1. The intent is to
evaluate the relative pressure of the exterior skin by normalizing it to a single-skin.
Table 9-1. Vertical profiles of exterior skin pressure ratios, Cp,DsFortter I Cp,s1ng1 ••
4
G.I
~
Ill
LL.
~
~
c:
G.I
...
I-
N
...:
"C
G.I
"' "iii
0::
"'
"C
ell
"Ei'i
c:
:c
Cl)
Model
Face Forward
Face Through
Face Return
Trench Forward
Trench Through
Trench Return
Raised Forward
Raised Through
Raised Return
Shingled Forward
Shingled Through
Shingled Return
Notes:
Profile
xla
0.50
0.67
0.90
0.50
0.67
0.90
0.50
0.67
0.90
0.50
0.67
0.90
0.50
0.67
0.90
0.50
0.67
0.90
0.50
0.67
0.90
0.50
0.67
0.90
0.50
0.67
0.90
0.50
0.67
0 .90
0 .50
0.67
0 .90
0.50
0.67
0.90
Near Inlet
1
·
3
zlh 0.10
min. mean max.
0.24 0.68 1.07
0.24 0.73 1.09
0.31 0.52 1.22
-0.12 0.49 1.07
-0.10 0.52 1.09
-0.31 0.46 1.22
0.61 0.65 1.06
0.58 0.80 1.06
0.70 0.90 1.06
-0.90 0.49 0.96
-0.63 0.44 0.98
-0.92 0.53 1.03
-0.89 0.52 0.96
-0.88 0.38 0.99
-0.90 0.55 1.04
0.58 0.85 1.01
0.54 0.84 1.00
0.83 0.93 1.00
0.62 0.77 1.07
0.52 0.67 1.11
0.56 0.78 1.18
0.53 0.93 1.05
0.81 0.96 1.10
0.96 1.13 1.17
0.95 1.04 1.06
1.01 1.04 1.06
1.07 1.08 1.09
0.27 0.80 1.05
0.29 0.80 1.06
0.27 0.78 1.07
0.26 0.77 1.05
0.26 0.80 1.06
0.22 0.78 1.06
0.74 0.94 1.03
0.75 0.95 1.02
0.88 0.97 1.02
1
Near inlet zone for a Raised inlet is 0.25 z/h 0.325.
2
Center zone for a Raised inlet is 0.325 z/h 0.85.
Center
2
0.10 zlh 0.85
min. mean max.
1.05 1.08 1.12
1.04 1.08 1.11
1.01 1.05 1.21
1.05 1.08 1.14
1.05 1.08 1.11
1.00 1.05 1.21
1.02 1.04 1.11
1.02 1.04 1.08
0.98 1.00 1.05
1.02 1.07 1.12
1.04 1.07 1.10
1.01 1.04 1.10
1.00 1.07 1.14
1.01 1.08 1.14
1.01 1.04 1.10
1.01 1.04 1.11
1.01 1.03 1.09
0.99 1.00 1.04
1.08 1.10 1.14
1.08 1.10 1.14
1.04 1.07 1.16
1.09 1.11 1.17
1.08 1.12 1.18
1.04 1.07 1.16
1.05 1.07 1.12
1.05 1 06 1.10
1.01 1 03 1.07
0.77 1.04 1.21
0.71 1.02 1.15
0.51 1.02 1.14
0.77 1.04 1.23
0.76 1.02 1.17
0.62 1 .01 1.15
0.76 1 .00 1.14
0.76 0.99 1.10
0.83 1.00 1.08
3
Inlet for Shingled is distributed across four locations (see Section 9. 7).
Near Exhaust
0.85 zlh
min. mean max.
-2.41 1.89 5.46
-4.20 0.56 3.50
-9.47 -0.20 2.18
-2.60 1.95 5.68
-4.55 0.55 3.57
-10.33 -0.17 2.30
-2.51 1.93 5.57
-4.47 0.54 3.56
-10.09 -0.21 2.29
-2.37 1.82 5.11
-4.15 0.58 3.49
-9.34 -0.13 2.24
-2.58 1.99 5.85
-4.53 0.45 3.66
-10.31 -0.10 2.34
-2.54 1.97 5.64
-4.56 0 .55 3.64
-10.04 -0.20 2.31
-2.44 1.92 5.51
-4.16 0.48 3.57
-9.50 -0.14 2.24
-2.75 2.10 6.15
-4.97 0.44 3.93
-10.60 -0.22 2.41
-2.73 2.02 5.92
-4.93 0.53 3.84
-10.62 -0.25 2.37
-2.48 0.84 5.62
-1.16 0.75 3.55
-3.11 0.33 2.30
-2.77 0.77 6.16
-3.06 0.66 3.88
-3.37 0.22 2.44
-2.75 0.38 5.71
-4.91 0.13 3.75
-10.57 -0.51 2.35
339
For the center portion of the building, 0.10 zlh 0.85, the outer skins receive an increase of up
to 23% (Shingled-Through) when compared to the single-skin facade. The Face and Trench
configurations have very similar ratios with the primary exception occurring in the near inlet zone.
The mean value of the center zones' ratios range from 1.00 -1.08 for the Face and Trench
models, 1.03 -1.12 for the Raised, and 0.99 -1.04 for the Shingled configurations. For all
models it is the x/a = 0.90 profile that is smallest and closest to 1.00. This is supported in Table 9-
1 where all models except the Shingled-Return have the lowest mean value at the x/a = 0.90
profile.
5
The three Shingled models all exhibit ratios below 1.00 within the center zone. This is
largely due to the rapid reduction in ratio around the four inlets that are distributed across the
height of the exterior skin. The values within the gap are excluded from the profiles that define
the means to avoid misrepresenting the overall mean across the zones' heights as a reduction
compared to a single-skin, where Figure 9-39 shows profiles that have ratios that straddle 1.00.
As seen in Figures 9-36 to 9-39, there are areas that exhibit a decrease in pressure, generally
associated with the vicinity of inlet opening features. In the lower half, several spikes occur for the
Face (at zlh ~ 0.10) and Raised (at zlh ~ 0.30) conditions for x/a = 0.90 profiles, indicating an
increase in outer-skin loads near the building's lateral edges as compared to a single-skin. The
dynamics near the lateral edges will be more evident in the next section where horizontal data
profiles are extracted and charted.
The vertical profiles show that in the near exhaust upper region (zlh > 0.85) and near-roof edge,
the ratio dramatically increases, peaking anywhere from 2.2 to 6.2 times (beyond the extents of
the graphs' x-axis) that of a single-skin before dramatically reverting back across the y-axis to
negative magnitudes that exceed -10 for most configurations' x/a = 0.90 profile curve. These
severe and volatile zones above zlh > 0.95 are an indicator of the sensitivity at the near edge
condition as well as variation in mesh resolutions of the double-skin models as compared to the
single-skin.
6
340
1.0
0.9
OB
0.7
0.6
-:
'
0.5
"
0.4
0.3
0.2
0.1
0
0
' ' ,,
I \~
'
'
'
'\~~ ~'
,,
,,
- "
----------------· ---------- ::::.r-
0.5
Outer Skin Pressure Ratio((;, ,,_w "'"ff I ( · ,,. '"'""· )
1 .5 2
Face Forward
x /a = 0.50
-- x la= 0 .67
--- x la= 0.90
Face Through
x/a = 0.50
x la= 0.67
la= 0.90
Face Return
x/a = 0.50
x la= 0.67
xla = 0.90
Figure 9-36: Comparison of outer skins' pressure ratio vertical profiles for Face inlet
configurations with varied exhaust configurations.
1.0
_ .
0.9
Trench Forward
x /a= 0.50
OB -- la= 0.67
--- x /a = 0.90
0.7
Trench Through
x/a = 0.50
0.6
-:
'
0.5
"
0.4
=t=
xla = 0.67
la = 0.90
Trench Relurn
x /a= 0.50
I
K ia= 0 .67
~
'
xla = 0.90
0.3
'
0.2
0.1
0
0 0.5 1.5 2
Outer Skin Pressure Ratio((;,
1
,_w .,,,,ff I·,,.,, ,,,,, . )
Figure 9-37: Comparison of outer skins' pressure ratio vertical profiles for Trench inlet
configurations with varied exhaust configurations.
341
1.0
0.9
OB
0.7
0.6
-:
'
0.5
"
0.4
0.3
0.2
0.1
0
0 0.5
+
1 .5
Outer Skin Pressure Ratio((',,
1
,_w
0
,,,",.1 ( · ,.. '"'""· )
2
Raised Forward
x /a = 0.50
-- x /a= 0 .67
--- x /a= 0.90
Raised Through
x/a = 0.50
x la= 0.67
x la= 0.90
Raised Return
x/a = 0.50
x la= 0.67
xla = 0.90
Figure 9-38: Comparison of outer skins' pressure ratio vertical profiles for Raised inlet
configurations with varied exhaust configurations.
1.0
0.9
OB
0.7
0.6
-:
'
0.5
"
0.4
0.3
0.2
0.1
0
0
+
0.5
; ,
'
' •
' •
Outer Skin Pressure Ratio((',, ,,_.w "'"ff I', .. ,,.,,, , . )
1.5 2
Shingled Forward
x /a= 0.50
-- x /a =0.67
--- K /a =0.90
Shingled Through
x/a = 0.50
x la = 0.67
x /a = 0.90
Shingled Return
x /a= 0.50
•la= 0.67
xla = 0.90
Figure 9-39: Comparison of outer skins' pressure ratio vertical profiles for Shingled inlet
configurations with varied exhaust configurations.
342
Outer Skin Ratios - Horizontal
Once again, each outer skin's Cp,n, (or Cp,DSFOute,) is compared to the equivalent for a single-skin
facade, Cp,smgk However, in this section the ratio is calculated for each of the twelve models at
horizontal profile locations zih = {0,10, 0,50, 0,67, 0,90), charted in Figures 9-40 to 9-43 and
summarized in Table 9-2, The intent is to evaluate the relative pressure of the exterior skin by
normalizing it to a single-skin model's coefficient of pressure, Cp,smgk Using horizontal profiles
provides a more detailed insight into the outer skins' Cp,DSFcmml Cp,smgfr ratios at the lateral edges,
For the Raised inlet models the zih = 0, 10 profile's values are excluded since they occur beneath
the elevated start of the outer skin, For all the other horizontal profiles the elevation is divided into
a center zone between 0, 10 x/a 0,90 and the edge zones that occur at x/a 0, 10 and 0,90
x/a, Since there is geometric symmetry to the models, the two edge zones lumped together are
similar,
For the center portion of the building, 0, 10 x/a 0,90, the outer skins of the double-skin facade
models evaluated receive an increase of up to 36% (Shingled-Through) when compared to the
single-skin facade, All four inlet types exhibit similar ratios in Figures 9-40 to 9-43 with the most
noticeable difference being the varying maximum magnitude of each model's zih = 0,90 profile,
The Face and Trench configurations have very similar ratios with the primary exception occurring
in the lower elevation, zih = 0, 10 profile, similar to the near inlet zone differences shown in the
previous section, The mean value of the center zones' ratios range from 1,01 - 122 for the Face
and Trench models, 1,04 -127 for the Raised and 1,00 - 1,31 for the Shingled configurations,
The largest mean ratio value for each model occurs at the zih = 0,90 elevation,
7
This is in the
midst of the severe and volatile zone above zih > 0,95 that was identified in the previous section,
When only the mid-height, center elevations (zih = 0,50 and 0,67) are considered, the mean ratio
value ranges are 1,01 -1,05 for Face, 1,02 -1,05 for Trench, 1,04 -1,09 for Raised, and 1,00 -
1,09 for Shingled configurations,
343
Table 9-2. Horizontal profiles of exterior skin pressure ratios, Cp,DSF Outer I Cp,Single.
8
GI
CJ
Ill
LL
..c:
CJ
c:
GI
....
.....
N
"C
GI
IJI
'iij
0:::
"C
..!!:!
en
c:
..c:
(/)
Model
Face Forward
Face Through
Face Return
Trench Forward
Trench Through
Trench Return
Raised Forward
Raised Through
Raised Return
Shingled Forward
Shingled Through
Shingled Return
Notes:
Profile
z/h
0.10
0.50
0.67
0.90
0.10
0.50
0.67
0.90
0.10
0.50
0.67
0.90
0.10
0.50
0.67
0.90
0.10
0.50
0.67
0.90
0.10
0.50
0.67
0.90
0.50
0.67
0.90
0.50
0.67
0.90
0.50
0.67
0.90
0.10
0.50
0.67
0.90
0.10
0.50
0.67
0.90
0.10
0.50
0.67
0.90
Edge Zones
1
x/a 0.1 and 0.9 x/a
min. mean max.
0.80 1.53 7.98
0.88 1.04 1.72
0.61 1.44 12.4
0.81 1.06 1.38
-1.59 1.39 6.02
0.61 1.03 1.90
0.45 1.42 12.2
0.62 1.17 2.55
0.86 1.06 1.51
0.15 1.01 1.74
-7.83 0.71 3.91
0.53 1.18 2.99
0.35 1.14 2.08
0.29 1.08 2.34
0.48 1.24 7.26
-1.80 0.88 1.45
0.33 1.20 3.29
0.02 1.00 2.24
0.52 1.42 12.7
0.12 0.95 1.14
-2.56 0.95 2.07
0.40 0.92 1.93
0.11 1.22 8.48
0.17 1.02 2.19
0.56 0.99 1.30
0.67 1.08 3.73
0.72 1.24 3.10
0.62 1.08 2.00
0.43 1.37 10.9
0.36 1.10 1.78
0.79 1.01 1.26
-0.02 1.07 4.64
0.88 1.24 3.48
-1 .16 1.15 5.41
-0.73 1.01 3.11
-0.71 1.81 19.8
-3.38 0.81 1.33
-0.69 1.11 4.52
-0.60 0.98 2.90
-0.60 1.61 18.6
-2.68 0.83 1.29
-0.33 1.04 3.69
-0.33 0.97 2.42
-0.54 1.09 8.95
-1 .36 1.59 13.2
1
Edge zone includes both sides, 0.10 > xla and x/a 0.90.
Center
0.1 x/a 0.9
min. mean max.
1.08 1.11 1.22
1.02 1.05 1.06
1.01 1.04 1.06
1.10 1.19 1.23
1.08 1.11 1.22
1.01 1.05 1.06
1.01 1.05 1.06
1.13 1.22 1.26
1.05 1.06 1.07
0.98 1.01 1.03
0.98 1.02 1.04
1.11 1.19 1.23
0.98 1.00 1.07
1.01 1.04 1.06
1.01 1.04 1.06
1.11 1.19 1.23
0.98 1.00 106
1.01 1.05 1.06
1.01 1.05 1.07
1.14 1.22 1.27
0.99 1.00 1.00
0.99 1.02 1.03
0.99 1.03 1.04
1.12 1.20 1.24
1.06 1.09 1.10
1.04 1.07 1.09
1.14 1.22 1.25
1.05 1.09 1.11
1.03 1.08 1.10
1.16 1.27 1.32
1 02 1.04 1.05
1.01 1.04 1.06
1.12 1.22 1.27
1.05 1.06 1.07
0.99 1.00 103
1.06 1.08 1.09
1.19 1.27 1.32
1.05 1.05 1.06
0.99 1.00 1.02
1.05 1.09 1.10
1.20 1.31 1.36
1.00 1.01 102
1.01 1.02 1.04
1.02 1.05 1.06
1.13 1.22 1.26
2
For Raised, z/h = 0.10 falls within the zone that the exterior skin does not extend to.
344
2.0
1.8 l
"
t
1.6
u'
~
:§
1.4
':::l
1.2
0
:;
1.0
ll'.
--
e
:>
0.8 (/)
(/)
e
Q_
0.6
c
+
~
en
~
0.4
.Sl
:>
0
+
0.2
+
0
0 0.5
x ! a
+
I
.
.
•
•
T
.
•
• , ,
"
, ,
,/
~' ,,.
,,-
Face Forward
z/h= 0 .1 0
z/h = 0.50
z/h = 0.67
z/h = 0.90
Face Through
z/h = 0.10
d h = 0.50
z/h = 0.67
z/h = 0.90
Face Return
lfh= 0 .1 0
z/h = 0.50
z/h = 0.67
z/h = 0.90
Figure 9-40: Comparison of outer skins' pressure ratio horizontal profiles for Face inlet
configurations with varied exhaust configurations.
2.0
1.8
"
t
1.6
u'
'
0
1.4
'::::!..'
1 .2
0
ia
1.0
ll'.
e
:>
0.8 (/)
(/)
e
Q_
c
0.6
~
en
.Sl
0.4
:>
0
0 .2
0
0
---~ --
+
0.5
x ! a
I
~
'
'
"
JI
,,
,,
Trench Forward
lfh= 0.10
z/h = 0.50
z/h = 0.67
z/h = 0.90
Trench Through
z/h = 0.10
dh = 0.50
z/h = 0.67
z/h = 0.90
Trench Rel.tlrn
lfh= 0.10
z/h = 0.50
z/h = 0.67
z/h = 0.90
,,,~ ·
0.67
'~
'"~1
e
Figure 9-41: Comparison of outer skins' pressure ratio horizontal profiles for Trench inlet
configurations with varied exhaust configurations.
345
2.0
1.8
"
t
1.6
u'
~
:§
1.4
':::l
1.2
0
/.
:;
1.0
11'.
e
:J
0.8
"'
"' e
Cl.
0.6
c
~
en
~
.Sl
:J
0
0.4
0.2
0
0
-+-+
+
t
0.5
x!a
I
I
ii
:1
I
I
Raised Forward
- lih=0.50
z/h = 0.67
- · - z/h = 0.90
Raised Through
z/h = 0.50
lih = 0.67
z/h = 0.90
Raised Return
dh = 0.50
z/h = 0.67
z/h = 0.90
'~~ ·
~.~)
~Vi-. --
0/ 1
e
Figure 9-42: Comparison of outer skins' pressure ratio horizontal profiles for Raised inlet
configurations with varied exhaust configurations.
"
t
u'
'
0
'::::!..'
0
iC
11'.
e
:J
"'
"' e
Cl.
c
~
en
.Sl
:J
0
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
0 0.5
x ! a
Shingled Forward
-- - z/h=0.10
- z/h = 0.50
-- z/h =0.67
- · - z/h = 0.90
Shingled Through
z/h = 0.10
dh = 0.50
z/h = 0.67
z/h = 0.90
Shingled Relurn
--- z/h=0.10
- z/h= 0.50
z/h = 0.67
- · - z/h=0.90
'~~ · 0.67
050
"~1
e
Figure 9-43: Comparison of outer skins' pressure ratio horizontal profiles for Shingled
inlet configurations with varied exhaust configurations.
346
Cp,net Values
This section presents the net pressure coefficient, Cp,mee (or !1CP = Cp,n, - Cp,m,), values at x/a =
{0,50, 0,67, 0,90) for each of the twelve models, grouped by inlet type (Face, Trench, Raised or
Shingled) resulting in Figures 9-44 to 9-47,
The figures show that Cp,mee is always greatest at x/a = 0,50 and noticeably reduced,
comparatively, at x/a =0,90, The outer most profiles at x/a = 0,90 for all twelve configurations has
the smallest magnitude (as compared to x/a = 0,50 and 0,67) for zlh 0,95, Above zlh > 0,95 the
x/a = 0,90 profiles increase to match or become the greatest magnitude, indicative of increased
net pressure loads in the upper corners, The overall greatest Cp,mee values occur just above ground
level around zlh = 0, 125 for the Trench-Through (nearing 2,00) and Face-Through configurations
Uust over 2,00), The Forward and Through conditions also display similarities in profiles for the
Raised and Shingled inlets, In both the Raised-Through and Shingled-Through results, the Cp,mee
magnitudes are slightly greater than Raised-Forward and Shingled-Forward, respectively, for zlh
0,95, Similar to the Face and Trench models, peak values for the Raised configurations also
occur just above the inlet; for Raised models, that occurs around zlh = 0,30,
In the center region of the elevation, 0, 10 zlh 0,85 and x/a = {0,50, 0,67),
9
the net pressure
coefficients for Face-Forward, Face-Through, Trench-Forward and Trench-Through all primarily
occupy the range between 1,80 Cp,mee 2,05 as compared to 1 A4 Cp,mee 1,79 for Raised
Forward and Raised-Through The Shingled-Forward and Shingled-Through also peak around
1,80 in the center zone, but have a smaller magnitude in the lower portions, primarily falling
between 1,00 Cp,mee 1,80, The Return exhaust models are the smallest magnitude for each
inlet group, All Return models have net pressure coefficients primarily in the range of 0,26 Cp,mee
0,53 for the center zone, The results suggest a favorable Cp,mee profile for configurations with a
Return outlet condition as compared to Forward and Through
347
1.1
1.0 + +
' '
0.9
t
'
'
' I
I I
o I
0.8
' ' ' I
I I
I '
I '
0.7
t :----:-
' '
' '
' ' 0.6
t ' '
' ' I I
""
I I
I I
'
0.5 + -r '
M
1,'
'
,,
I ,,
0.4
, ,
1
I I
r
' I
,,
'
,,
( (
,,
0.3
, ,
,,
,,
I 0
+
' •
0,2 _j \
I I
, ,
0.1
t
\~'.
,,-
0
0 0.5 1.5
Net Pressure Coefficient (AC)
Figure 9-44: Comparison of Cp,net values (at xla = {0.50, 0.67, 0.90}) for Face inlet
configurations with varied exhaust configurations.
1.1
1.0
0.9
0.8
0.7 '
'
'
'
'
0.6
'
' I
""
'
'
'
0.5 I
" '
' I
0.4 ' I
I
' I
0.3
I
I
0,2
'
'
'
0.1 ' I
------
0
0 0.5
Net Pressure Coefficient (AC)
'
' '
' '
' '
I '
I '
' '
' '
I ' , ,
J ,
I '
' I
' I
' •
' • .,
,,
' • ,,
' • ,,
' • ,,
' • ~.
,,
,,
"
,,
"
,,
' •
; .
, I
' I
"
,,
,!J
- -- _ .:r-t"'
1.5
Face Forward
-- x/a= 0.50
x/a= 0.67
--- x/a=0.90
Face Through
-- x/a=0.50
x/a= 0.67
--- x/a=0.90
Face Return
-- x/a = 0.50
xla = 0.67
--- xla=0.90
Trench Forward
-- x/a=0.50
xla= 0.67
- - - xia= 0.90
Trench Through
-- x/a=0.50
x/a = 0.6 7
- -- x/a=0.90
Trench Return
-- x/a=0.50
xla=0.67
--- x/a=0.90
Figure 9-45: Comparison of Cp,ner values (at xla = {0.50, 0.67, 0.90}) for Trench inlet
configurations with varied exhaust configurations.
348
1.1
1.0
0.9
0.8
0.7
0.6
""
' N
0.5
0.4
0.3
0,2
0.1
0
0 0.5
+
t
- - ----==----
,
'
'
'
'
7
,I
'
'
Net Pressure Coefficient (AC)
1.5
Raised Forward
- x/a = 0.50
x/a= 0.67
--- x/a=0.90
Raised Through
- x/a=0.50
x/a = 0.67
--- x/a=0.90
Raised Return
- xia=0.50
x/a = 0.67
--- x/a=0.90
Figure 9-46: Comparison of Cp,net values (at xla = {0.50, 0.67, 0.90}) for Raised inlet
configurations with varied exhaust configurations.
1.1
1.0
0.9
0.8
0.7
0.6
""'
' 0.5
"
0.4
0.3
0,2
0.1
0
0 0.5
t
'
'
'
+ _J
------=----~--~ -:..::
- -~ ....
' '
' '
' ,
' ,
' ' , '
~ ~
_ ... -~ ... _ ...... '
' I
'--•
' '
' '
I '
' '
' I
I \
' '
I
'
'
'
'
'
'
Net Pressure Coefficient (AC)
1.5
Shingled Forward
- x/a=0.50
-- x/a=0.67
--- xla = 0.90
Shingled Through
- x/a=0.50
x/a = 0.6 7
- -- x/a= 0.90
Shingled Return
- xia=0.50
x/a =0.67
--- x/a=0.90
Figure 9-47: Comparison of Cp,ner values (at xla = {0.50, 0.67, 0.90}) for Shingled inlet
configurations with varied exhaust configurations.
349
9.8.2 Comparison of Exhaust Configurations
The second set of comparisons (Figures 9-48 to 9-59) show the results of four different simulation
models each, holding the exhaust condition constant (either Forward, Through or Return), while
varying the inlet (Face, Raised, Shingled and Trench).
CP Values
The CP profiles shown in Figures 9-48 and 9-49 exhibit similarities amongst Forward and Through
configurations given a similar inlet. One variation is near the exhaust zone, zlh > 0.95, where the
Through configurations exhibit a smoother interior profile decay towards C p,;nt :::: -1.20 as
compared to the Forward configurations that approach C p,int :::: -1.30. All Return exhaust
configurations (Figure 9-50) display a positive profile for all inlet configurations' C p.znt and Cp.ext
while the Forward and Through exhausts possess predominantly negative C p,mt magnitudes.
1.1
Face Forward
1.0
- Cp.ol~Ff)
0.9 - ---- c
p.intf'A
O.B -
Raised Forward
0.7 -
- - c~lnl(Rfl
0.6 - Shingled Forward
""'
-....
0.5
"'
-+--
- cp, • rt(SF)
--- Cp.int(SFl
0.4 -
Trench Forward
0.3 - -Cp.etfll')
0.2 - ----r--_
-·-- c
r-1111(TA
0.1
0
·1.5 ·1.0 ·0.5 0 0.5 1.0 1.5
Pressure Coefficient for Interior (C,. ,J and Exterior( ~ . ,) Skins
Figure 9-48: Comparison of CP values at mid-span (x/a = 0.50) for Forward exhaust
configurations with varied inlet configurations.
350
1.1
-~-;(:::,- - \
Face Through
1.0
~
,., \
-- C p. er1FT>
\\ ; '
0.9 ' ---- c
I p.int (rn
'
'- I
Raised Through
0.8
:. --, ___
,. '
- - c
d '
,.
p,.lfl (RT)
0.7
: ·i '
I
- -
c p. nl(RT) ,, .
,, . ,
0.6
ii I
r •
'
Shingled Through
,.
' ii
-- c p. • d (ST)
""'
,.
-
'
0.5
:ti N
l
I ' --- C p.int (STJ
:i
0.4 I'--
,.
I
Trench Through
11
'
"
I
l
0.3
,.
' --Cp ,&i:1tm
i'I
I ' \
__Jj__
- - - - - L.. -
- --- c
0,2
I . ... (111
,.
l
, !
•:
'
0.1
}' I
D~ 0
( _-~c ----- - - --------------~
-
-- ~--
/
-1.5 -1.0 ·0,5
·o
0,5 1.0 1.5 8
Pressure Coefficient for Interior (C,. ,J and Exterior( ~ ... ) Skins
Figure 9-49: Comparison of CP values at mid-span (x/a = 0.50) for Through exhaust
configurations with varied inlet configurations.
1.1
' I J J Face Return
1.0
I ' ,
I
' I ' -- CP.•" l'Q) I
I : I
0.9
l _j I
---- C p.int (f'R)
: : j
I I '
0.8
: ~ \
\1
Raised Return
I I
-- cp , ""1tR~ : :
;1
0.7
I j
' ·
11
- -
c p.W (AR)
I
!1 I I
' .
0.6
I • 11
Shingled Return
' I
I
: :
\I
--cp.•d (Si.1~
""
I I
\
'
0.5
: : ..
1 1
--- c p.llll (SR) , I
1 ! ' .
0,4
' ' 1 1
-:-:
1 1
Trench Return
' '
0.3
' I
I ' --Cp,&'!lt~l
: : \
I I
"'
: : I ' -
---- C r-...i (TR)
0,2
+-f
' ,._
: : ! \
I I I
\
0.1
: :
o:
' L
: '"'-
0
\
/
-1.5 ·1.0 ·0,5 0 0,5 1.0 1.5 8
Pressure Coefficient for Interior (C, .. J and Exterior (C;_ .) Skins
Figure 9-50: Comparison of CP values at mid-span (x/a = 0.50) for Return exhaust
configurations with varied inlet configurations.
351
Outer Skin Ratios - Vertical
Each outer skin's C p,ext (or Cp,DsFouter) is compared to the equivalent for a single-skin facade,
Cp,Szngle· The ratio is calculated for each of the models at vertical locations x/a = {0.50, 0.67, 0.90},
charted in Figures 9-51 to 9-53. These are the same profiles presented in Section 9.8.1 and also
summarized in Table 9-1 but are instead grouped by exhaust type. The intent is to evaluate the
relative pressure of the exterior skin by normalizing it to a single-skin.
The results shown in Figures 9-51 and 9-52 possess strong similarities in the profile shapes and
outer skin pressure ratio magnitudes for Forward and Trench exhaust configurations.
Comparatively, Figure 9-53 exhibits smaller ratios, closer to 1.00, and has milder peak
magnitudes for zlh 0.85. It can be deduced from these charts and Table 9-1 that the outer skins'
pressure ratio is lowest for a configuration with a Return exhaust as compared to Forward and
Through with all other variables remaining constant.
0.5 1.5 2
Outer Skin Pressure Ratio((~.""'"""" I ( ; • . s ;,,_ ,,.)
Face Forward
- x /a=0.50
-- x /a =0.67
--- x la= 0 .90
Raised Forward
- x la=0.50
-- x /a =0.67
--- x /a= 0.90
Shingled Forward
- x/a= 0.50
-- li:la = 0.67
--- x /a= 0.90
Trench Forward
- x/a= 0.50
-- x /a = 0.67
- - - x /a = 0.90
Figure 9-51: Comparison of outer skins' pressure ratio vertical profiles for Fo1Ward
exhaust configurations with varied inlet configurations.
352
1.0
0.9
0.8
0.7
0.6
..:;::
...._
N
0.5
0.4
0.3
02
0.1
0
0 0.5
~
'
'
'
---
+
1.5
Outer Skin Pressure Ratio((~.°°'"""" I ; .. s;,,.,,. )
Face Through
x/a = 0.50
-- x /a= 0.67
- --
xla= 0.90
Raised Through
x/a = 0.50
x /a = 0.67
x /a = 0.90
Shingled Through
x /a= 0.50
xla = 0.67
x /a= 0.90
Trench Through
x/a= 0.50
x/a= 0.67
x /a= 0.90
~ 1
0
2
/
8
Figure 9-52: Comparison of outer skins' pressure ratio vertical profiles for Through
exhaust configurations with varied inlet configurations.
1.0
0.9
0.8
0.7
0.6
..:;::
...._
N
0.5
0.4
0.3
02
0.1
0
0 0.5
r:
,,
•"
,,
,,
'"
'"
"
,,
-L;
' I
,,
,,
+' ' , ,
•
l
Outer Skin Pressure Ratio((~. °°'"""" I; .. s;,,.,,. )
Face Return
x/a= 0.50
-- x /a = 0.67
--- x la= 0.90
Raised Return
x/a = 0.50
x /a = 0.67
x /a = 0.90
Shingled Return
x/a= 0.50
xla = 0.67
x /a = 0.90
Trench Return
x/a= 0.50
x/a= 0.67
x /a= 0.90
~ 1
o
1.5 2
/
8
Figure 9-53: Comparison of outer skins' pressure ratio vertical profiles for Return exhaust
configurations with varied inlet configurations.
353
Outer Skin Ratios - Horizontal
The outer skin to single-skin ratio Cp,DSF outer I Cp,stngle is calculated for each model at horizontal
locations z/h = {0.10, 0.50, 0.67, 0.90} and charted in Figures 9-54 to 9-56. These are the same
profiles presented in Section 9.8.1 and Table 9-2 but are instead grouped by exhaust type.
Overall, the Forward and Through configurations have very similar ratios, particularly in the center
zone - between 0.10 x/a 0.90 for zlh = {0.50, 0.67} - with the Forward configurations falling
within 0.99 to 1.10, Through ranging from 0.99 to 1.11, and Return between 0.98 to 1.06. In
Figure 9-54, the center portion between 0.10 x/a 0.90 reaches a peak ratio for Forward
configurations of 132% at the zlh = 0.90 profile for Shingled-Forward. In Figure 9-55, a similar
peak ratio of 136% occurs for Through outlets at the zlh = 0.90 profile of Shingled-Through. In
Figure 9-56, a similar peak ratio of 126% occurs at the zlh = 0.90 profile of Shingled-Return. All
peak ratio values in the center portion occur towards the upper edges in a Shingled condition.
2.0
1.8
,
l
1.6
u'
ii
~
1.4
S:!.,~ 1.2
0
:;
1.0
a:
~
:::>
0.8 Ul
Ul
~
Cl.
c
0.6
:i'
Cf)
2
0.4
:::>
0
0.2
0
0
r
0.5
x ! a
Face Forward
--- z/h-0.1 0
- z/h = 0.50
-- z/h = 0.67
- · - z/h = 0.90
Raised Forward
- z/h= 0.50
-- z/h = 0.67
--- z/h = 0.90
Shingled Forward
--- zlh = 0 .1 0
- z/h= 0.50
' ( -- z/h = 0.67
I - · - lih =0.90
Trench Forward
- - - zlh = 0.10
- z/h= 0.50
-- z/h = 0.67
- · - z/h = 0.90
' . . ··.· ·----ri
J.~~_u
'"0~
e
Figure 9-54: Comparison of outer skins' pressure ratio horizontal profiles for FoJWard
exhaust configurations with varied inlet configurations.
354
2.0
1.8
"
t
1.6
u'
~
:§
1.4
':::l
1.2
0
:;
1.0
11'.
e
:>
0.8
"'
"' e
a..
0.6
c
~
en
~
.Sl
:>
0
0.4
0.2
0
0
+-
t
t
0.5
x!a
Face Through
z/h=0.10
z/h = 0.50
z/h = 0.67
z/h = 0.90
Raised Through
z/h = 0.50
-- zih=0.67
- ·- z/h = 0.90
Shingled Through
zlh= 0.10
z/h = 0.50
z/h = 0.6 7
d h = 0.90
Trench Through
zlh=0.1 0
z/h = 0.50
z/h = 0.67
z/h = 0.90
"'~
0.61
o.~
,,1\1 ___ ·
0/ 1
e
Figure 9-55: Comparison of outer skins' pressure ratio horizontal profiles for Through
exhaust configurations with varied inlet configurations.
2.0
1.6
"
t
1.6
u'
'
0
1.4
'::::!..'
1.2
0
~
1.0
11'.
e
:>
0.8
"'
"' e
a..
c
0.6
~
en
.Sl
0.4
:>
0
0.2
0
0
+ +- +-
I +
.. ~~::~~~~~- .- - .-- .- ~.- - "' ·"' - ;-, ·'= = "° ·- -- -- =-- ·- '°'·-= - ·"'" -- 1~ ~===~~~=:~-~-~~~~
_...,. ----~----,----- -----r------"""------i--~---- -----
t
I
0.5
x ! a
+-
' I
I
-,-
'
'
'
Face Return
d h=0.10
z/h = 0.50
z/h = 0.67
z/h = 0.90
Raised Return
z/h = 0.50
-- z/h = 0.67
- · - z/h = 0.90
Shingled Return
zlh= 0.10
z/h = 0.50
z/h = 0.67
d h = 0.90
Trench Return
zlh =0.10
z/h = 0.50
z/h = 0.67
z/h = 0.90
'~~
o.e1
rn
~.h-, __ _
0/ 1
e
Figure 9-56: Comparison of outer skins' pressure ratio horizontal profiles for Return
exhaust configurations with varied inlet configurations.
355
Cp,net Values
This section presents the net pressure coefficient, C p,net (or AC P = C p,ext - C p,int) . values at x/a =
{0.50, 0.67, 0.90} for each of the twelve models, grouped by exhaust type (Forward, Through or
Return) resulting in Figures 9-57 to 9-59. The figures show that Cp,net is always greatest at x/a =
0.50 and noticeably reduced at xla =0.90. The outer most profiles at xla = 0.90 for all twelve
configurations has the smallest magnitude (as compared to x/a = 0.50 and 0.67) for zlh 0.95. As
it relates to Figures 9-44 to 9-47 where the Return exhaust configurations are always significant
outliers compared to Forward and Through, Figures 9-57 to 9-59 show more commonality in C p,net
due to exhaust type than inlet type.
10
This is evident by the inclusive range of values in the center
zone - between 0.10 zlh 0.85 for xl a = {0.50, 0.67} - for all Forward (0.98 C p,net 2.01) and
Through (1.04 C p,net 2.05) configurations. The range of values in the same center region for
Return exhaust configurations is much narrower at (0.26 Cp,net 0.53).
1 .1
1.0
0.9
0.8
0 .7
0.6
-'
......
0.5
"
0.4
0.3
0 .2
0 .1 l
0
0 0.5
'
'
'
'
'
I
I
I
j
I
'
'
t
'
'
'
'
Net Pressure Coefficient (M~)
"
i-0-
, '
' \
' '
,! ~
... _ .. _ -- .,
1.5 2
Face Forward
- •la = 0.50
-- x/a -0.6 7
- - - x/a = 0.90
Raised Forward
- x/a =0.50
-- x/a= 0.67
--- xla = 0.90
Shingled Forward
- x/a=0.50
-- x/a = 0.67
--- x/a = 0.90
Trench Forward
- x/a = 0.50
xla = 0.67
--- x/a = 0 .90
Figure 9-57: Comparison of Cp,net values (at x/a = {0.50, 0.67, 0.90}) for Forward exhaust
configurations with varied inlet configurations.
356
1.1
1.0
0.9
0.8
0.7
0.6
""
' ...
0.5
0.4
0.3
0,2
0.1
0
t
.. _ -_ - _ -_ -_-~---- .......... =-- - - - -
+
---------------------------------~---
0 0.5
'
,I
I
I
'
'
'
'
'
Net Pressure Coefficient (AC)
·'
1.5
Face Through
-- x/a=0.50
x/a= 0.67
-- - x/a = 0.90
Raised Through
-- x/a= 0.50
x/a = 0.67
--- x/a=0.90
Shingled Through
-- x/a = 0.50
x/a = 0.67
-- - x/a = 0 .90
Trench Through
-- x/a= 0.50
x/a = 0.6 7
-- - x/a = 0.90
Figure 9-58: Comparison of Cp,net values (at xla = {0.50, 0.67, 0.90}) for Through exhaust
configurations with varied inlet configurations.
1.1
1.0
0.9
0.8
f
0.7
0.6
""
'
0.5
"
0.4
0.3
++
0,2
0.1
0
0 0.5
Net Pressure Coefficient (AC)
..
1.5
Face Return
-- x/a=0.50
x/a = 0.67
--- xia=0.90
Raised Return
-- x/a=0.50
x/a = 0.6 7
- - - xla=0.90
Shingled Return
-- xia= 0.50
xla = 0.67
-- - x/a = 0 .90
Trench Return
-- x/a= 0.50
x/a = 0.67
-- - x/a = 0.90
Figure 9-59: Comparison of Cp,ner values (at xla = {0.50, 0.67, 0.90}) for Return exhaust
configurations with varied inlet configurations.
357
9.9 Conclusions Regarding Steady-State Analysis
The results of the comparative analysis for the single-skin and twelve multi-story double-skin
facade steady-state simulation models are summarized in this section. Additionally, the findings
from the three-dimensional steady-state simulations are briefly compared to the observations
made in the preliminary two-dimensional transient simulations with similar airflow opening
configurations.
11
9.9.1 Summary of Findings from 30 Steady-State Analysis
The results of the three-dimensional steady-state simulations with turbulent flow indicate the
potential for considerable disparity in the CP·"" values for multi-story double-skin facades with
various combinations of airflow inlets and exhausts. This is most evident in Figures 9-44 to 9-47
and 9-57 to 9-59 which overlay multiple models' CP·"" values simultaneously. Further investigation
into the cause of the variation in the resulting values reveals that the predominant reason is the
Cp.m, distribution (see Figures 9-32 to 9-35 and 9-48 to 9-50) of the interior skin are impacted by
the airflow openings. The outer skins, however, remain relatively consistent in their cp.n, values.
This becomes most clear when the outer skin is isolated and studied comparatively to cp.smgfr
values in the vertical and horizontal outer skin ratios (see Figures 9-36 to 9-43 and 9-51 to 9-56).
Of course there are exceptions and peak values
12
- primarily near edges - but the outer skin
ratios show the multi-story exterior skin's average center zone pressure coefficient may see an
increase of up to approximately 10% compared to a single skin (see Tables 9-1 and 9-2).
Meanwhile, the interior skin has negative CP·'"' values for Forward and Through exhaust
configurations, but positive values for Return. The range of values is considerably wider than the
range of exterior values. In short, the simulations presented in Sections 9.3 to 9.7 indicate the
airflow inlet and exhaust configuration have a noticeable impact on the pressure coefficients of
multi-story double-skin facades, particularly the negative pressure on the inner skin within the
cavity space.
358
The following observations summarize the results of the three-dimensional steady-state
simulations:
• The greatest variation is in the Cp,m, profiles of multi-story DSF configurations
• Face and Trench inlet conditions exhibit similar CP and the Cp,•ee profiles
• Forward and Through exhaust conditions exhibit similar CP and the cp,•ee profiles with
slight variation near-opening regions
• Inlet conditions that elevate (Raised) or distribute the opening across the vertical height
(Shingled) exhibit a reduced magnitude of negative pressure in the cavity space through
the lower and center portions of the elevation
• Return exhaust configurations exhibit a positive Cp,m, profile, contrary to the other exhaust
configurations that are negative
• The positive Cp,,., of the Return configurations reduces the corresponding Cp,•ee profiles,
9.9.2 Comparison of 30 Steady-State to 20 Transient Analysis
Figure E-3 indicates a noticeably different behavior for the Return exhaust configurations. As was
the case in the three-dimensional steady-state, these two-dimensional transient simulations
exhibit a positive Cp,m, profile, contrary to the other exhaust configurations that are negative.
Unlike the three-dimensional steady-state simulations, the Cp,,., and Cp,•ee profiles for the Forward
and Through exhaust configurations do not show as much similarity, instead having a bigger
disparity in magnitudes. The positive Cp,m, profile for the Return exhaust models impacts the Cp,•ee
profiles, reducing the magnitude greatly compared to the Forward and Through configurations,
although the sign convention flips. This is an indicator of these pressure coefficients
instantaneous qualities during a transient state, not like the steady-state simulations.
Nevertheless, these preliminary simulations do not contradict the findings of the three
dimensional steady-state simulations, particularly the Return configurations' distinctive behavior.
359
Chapter 9 Endnotes
2
3
4
5
6
The Raised configurations' exterior skin's vertical edges are not full height; nevertheless, they
still have an imposed subdivision of 75 elements. The Shingled configurations have four
exterior surfaces that makeup the exterior skin; each of these four surfaces has an imposed
subdivision of 75 elements, resulting in an aggregate number of nodes along the vertical edges
in excess of 300.
The Single-skin simulation's average pressure coefficient through the central portion - as
defined in Holscher and Niemann (1998) and presented in Section 84.1 - (0.25.:::. y/H.:::. 0.75,
0.1 .:::. z/ H.:::. 0.75, is cp,Single = 0.84, a strong agreement to the study's average pressure
coefficient of Cp,front = 0.84 (er= 0.05) with a minimum of 0.79 and maximum of 0.92.
The zlh = 0.10 horizontal profile is excluded for the models with Raised inlets since the exterior
skin does not exist in this region.
The graphs present the Shingled ratios across the openings. The Table has excluded the
areas on the Shingled exterior elevation where there is no facade due to the inlets across the
exterior's height
Visually it is the smallest in the figures, but due to the dips around the inlets, the mean value is
reduced.
This is also an indication of turbulent flow CFO simulations' challenges in modeling near-wall
flow behaviors.
7
This comment is in reference to the largest mean ratio value of the four horizontal extraction
locations evaluated in the simulations' models. It is possible that the largest ratio exists
elsewhere, but it is likely that it occurs in the vicinity of the zlh = 0.90 profile.
8
The zlh = 0.10 horizontal profile is excluded for the models with Raised inlets since the exterior
skin does not exist in this region.
9
The center region for a model with a Raised inlet configuration is different since the outer skin
does not exist in the zlh 0.125 region. The Raised models' center zones are considered to be
0.325 zlh 0.85.
10
That said, this is still a wide range. For the exhausts, Return configurations' results differ
significantly from Forward and Through. For the inlets, Shingled configurations' results vary
significantly, as do Raised (but less so), compared to the similar Face and Trench.
11
The two-dimensional transient simulation did not include a set of models with a Shingled inlet
type.
12
Up to 36% per Table 9-2.
360
10. Discussion and Conclusions
The overall aim of the research was 1) to understand the effects that airflow openings in multi
story double-skin facades have on the design pressure requirements, and 2) to evaluate the role
of computational fluid dynamics as an essential or supplemental design tool in the determination
of pressure coefficients for multi-story double-skin facades. While the pressure effects have
traditionally been investigated in wind tunnel tests, the alternative of computation fluid dynamics
is more and more attractive for consultants and design engineers because of recent increases in
computational power and advancements of commercial software. It is possible that computational
fluid dynamic analysis may be used to inform facade design decision making earlier in the
process, prior to wind tunnel testing, or be used for projects that are not large enough to warrant
commissioning a wind tunnel study. Ideally, a project would have computational fluid dynamics as
a supplemental tool both before and after wind tunnel testing - prior for the purposes of
optioneering to identify the facade solution that does get tested in the wind tunnel and following to
address the ramifications of any post-wind tunnel study design decisions such as details of solar
control elements in the cavity or as-built conditions not considered. This may be even more
appropriate for double-skin facade systems: "CFO analysis and wind tunnel testing will be used
together to enable the accurate design of double skin facades" (Zammit, Overend and
Hargreaves 2010, 666). The research herein begins to investigate how and when each of the
two, CFO analysis and wind tunnel testing, may work together in the design process for multi
story double-skin facades.
A summary of the work and presentation of the contributions of this research to the body of
knowledge in the field of facade engineering are presented in the following sections, including a
set of recommendations for 1) multi-story OSF design guidelines, 2) the role of computational
fluid dynamics in the design process, and 3) pressure coefficient determination methodology.
361
10.1 Experimental Summary
To evaluate the structural response of multi-story double-skin facades, this research investigates
how various airflow configurations affect the pressure coefficient distribution by using
computational fluid dynamics to simulate a wind tunnel testing environment with scaled building
geometries. The simulation method is first calibrated and reviewed against existing wind tunnel
studies. A prototypical geometry of a multi-story double-skin facade is derived from an in-depth
review of thirty existing double-skin facades in the United States - of which twenty-one are multi
story configurations - and three-dimensional steady-state computational fluid dynamic
simulations were run for twelve configurations of airflow inlet and outlet combinations. The interior
skin's coefficient of pressure profile varied considerably depending on the different inlet and outlet
configurations. Computational fluid dynamics with multi physics software is evaluated to determine
its potential effectiveness as a design tool for the determination of coefficients of pressure for
multi-story double-skin facades.
10.2 Research Limitations
10.2.1 Geometric
The geometric configurations considered in the simulation models are simple, rectilinear in
nature, and relatively low-rise.
1
In reality, projects may exhibit unique geometric conditions that
require special consideration regarding wind flow. Features such as corners, fly-bys, parapets,
canopies and portals impact the local flow behavior. In some instances, there may be building
forms that are more advantageous at mitigating flow separation around corners than the hard,
orthogonal transitions in rectilinear forms.
2
10.2.2 Approaching Flow
The simulation models are limited to considering wind normal to the building face, or wind
aligned, at a wind incidence of 0°. Cross-wind and torsional wind approaching flows should be
understood in the design of a building skin. Though they may not govern the structural design,
362
anomalies may arise that require localized mitigation strategies. Additionally, each building site
has a unique surrounding context. Effects of wind flow around adjacent structures may need to
be considered. This issue is of heightened importance for projects sites located in dense urban
environments. Similar to how the surrounding urban environment is modeled on the turntable in
wind tunnel studies (see Figure 1-3), this environment may be modeled in the CFO analysis
model.
10.2.3 Airflow Configurations and Openings
Though Section 1.7 presents a Louvered inlet configuration, this was not modeled in the
simulations. Their modeling challenge is more complex and requires many smaller elements exist
in the simulation domain. This results in meshes of equivalent quality requiring many more nodes
and elements, thus increasing the computational demands and runtime. If the results of the
Shingled inlet configuration are any indication of pressure distribution behavior, there may be
advantages in the pressure distribution and magnitudes of a Louvered approach worth studying
further. Also, only an opening air ratio of 5% was simulated in the 30 steady-state models. As
seen in Appendix E, varying this ratio would impact the results. In practice the opening area ratio
is a function of the design geometry, performance requirements and product selections. Project
specific simulations would want to accurately reflect these in the determination of a project
specific opening air ratio.
Additionally, there are limitations to the level of detail that is present in the simulation models at
the opening inlets and exhausts. In these simulations, flow altering, obstructing or diffusing
elements, such as airflow dampers, louvers or grating were not consider. The openings were
considered to be completely free air in the region where they were assigned. Though most
projects have something in these openings - controllable or fixed - the breadth of possible
configurations was too wide to consider as a variable herein. Practically speaking, elements like
louvers may aid in minimizing water penetration into the cavity space. However, they also create
363
resistance to airflow that creates some pressure loss over the louver free area. Effects on air flow
velocity into the cavity and the resulting pressures adjacent to openings and within the cavity
would be impacted if considered.
10.2.4 Cavity Contents
Often, the cavity volume is used to house solar control systems so that they are not fully exposed
to the exterior environment and may be accessed for maintenance and cleaning. The solar
control strategies vary tremendously from one project to the next - as is evident in Appendices A
and C. For this reason, they are omitted from consideration in the simulations. Neither shading
devices nor grating - often located at each floor slab level to provide access for cleaning - are
modeled in the simulations. These elements would impact the fluid flow dynamics through the
cavity space.
10.2.5 Structure
The supporting structure of the exterior skin is not modeled within the cavity space. The
simulations treat the exterior skin as a single surface of consistent depth, similar to how Plexiglas
is used in front of the test section of Marques da Silva and Gomes' (2005, 2008) wind tunnel
studies. Additionally, though fluid-structure interaction was expressed as an initial motivation into
this research topic, the simulations do not consider the influence of structural deformation of
various structural systems with different levels of rigidity on the fluid dynamics.
10.2.6 Analytical Method
The primary simulations in this research are focused on turbulent flow with a Reynolds Averaged
Navier Stokes (RANS) turbulence model type with a k-B turbulence model with incompressible
flow. Some researchers have acknowledged that other simulations, such as Detached Eddy
Simulations (DES) and Large Eddy Simulations (LES) may be preferred and see increased use in
the future in lieu of RANS (see Section 1.3.3). Additionally, since the primary simulations are
steady-state, they do not reveal the dynamic loading that a transient simulation would.
364
10.3 Recommended Design Guidelines for Multi-Story Double-Skin Facades
This section outlines a series of essential questions to be considered during each phase of the
design and delivery process for a project that utilizes a multi-story double-skin facade. Many of
these questions would be similar to those asked in planning a high-performance single-skin
facade. This section elaborates on each question to provide a view of how each may be
responded to with consideration for the unique design hurdles that should be confronted in the
development of a multi-story double-skin facade. Following, a theoretical framework for a
responsibility matrix and project timeline are outlined in Table 10-1 and Figure 10-3 to give a view
of these considerations in the macro context of a project's delivery.
10.3.1 Preliminary Planning Phase
The earliest stages of a project begin with concept design. During this phase the primary
considerations are often defining the program, researching potential solutions, and developing a
preliminary budget Addressing the following questions as early as possible in a project
considering a multi-story double-skin facade begins to reveal which airflow mode and
configuration strategies may be most appropriate:
1. VI/hat is the building program I space use?
2. VI/hat are the top priorities for the building enclosure's performance?
3. VI/hat is the ventilation mode? Is natural ventilation a priority? Does the ventilation mode
permit mechanical control of cavity pressure?
Identifying Programmatic Requirements
The reflection of a building program's performance requirements in a double-skin facade can
range tremendously. A client with a commercial office space has to balance the trade-off between
enhanced performance and the ramifications on leasable area. Meanwhile, an education facility
365
may be owned by the same institution for as long as that building stands, whereas a developer
may be looking to sell a commercial space in the first decade after its completion. That
educational institution may also place a greater emphasis on the facility management component
of a DSF's operations than speculative commercial developers. Another example would be a
high-end residential project that seeks access to natural ventilation while also mitigating acoustic
distractions to a heightened degree compared to a work space. Articulating the desired
functionality of the space early may point towards a high-performance double-skin facade or steer
the project team in the direction of a more conventional exterior solution. Addressing these
questions early advances an understanding of a client's willingness and project budget's ability to
support a probable upfront premium for a double-skin in lieu of a single-skin facade system.
Identifying Motivations
The desire for improved energy performance while obtaining a high level of transparency may
lead a project team in the direction of a double-skin facade. Other motivations may include
creating an improved acoustic buffer compared to a single-skin facade, providing a buffer cavity
to permit access to natural ventilation where it would otherwise not be possible, or for protection
of solar control elements that would otherwise be exposed to the exterior. Understanding these
priorities will aid the selection of appropriate ventilation modes and airflow configurations.
Less
Acoustic
Buffer
Air
Supply
Outdoor
Air Curtain
Indoor
Air Curtain
Air
Exhaust
Buffer
(Sealed)
More
Acoustic
Buffer
Figure 10-1: Relative comparison of acoustic buffering by ventilation mode.
366
Ventilation Mode
If natural ventilation is not a priority for the air supply, consider using an Air Exhaust instead of an
Outdoor Air Curtain ventilation mode and positively pressurize the cavity space for a controlled,
predictable understanding of the cavity pressure's influence on both skins' net pressure for
design. This controlled approached will mitigate some of the volatility associated with a naturally
ventilated cavity and may permit greater control of potential load reduction during design. The
caveat is that this is not the most energy efficient approach. Figure 10-2 shows a relative scale of
how the different ventilation modes' controllability ranges to one another's. The ventilation modes
that provide natural ventilation are the most susceptible to volatility in the cavity pressure profile.
More
Controlled
Indoor
Air Curtain
Buffer
(Sealed)
Air
Exhaust
Outdoor
Air Curtain
Air
Supply
More
Volatile
Figure 10-2: Relative comparison of cavity pressure volatility by ventilation mode.
10.3.2 Schematic Design
As a project advances from concept design it enters into the schematic design phase. During this
phase site and floor plans are generated, the building mass begins to take shapes, structural and
energy approaches may be selected, and potential systems solutions are explored. This phase
generally sees an increase in the number of collaborative participants on the project team, often
expanding the design team to include key consultant parties and potential contractors may have
input into the project's development. During schematic design the following are important
concerns to address:
367
1. If a double-skin facade is being considered, what is a single-skin alternative with
comparable performance?
2. VI/hat is the relationship between the facade and the 30 massing's site plan and solar
orientation?
3. VI/hat is the whole building energy approach? How does the double-skin facade integrate
with mechanical systems as part of a whole building energy approach?
4. VI/hat is the relationship between the double-skin facade and the primary structure?
Single-Skin Baseline
Too often the attraction to a double-skin facade is for its perceived sustainable qualities and
energy performance. However, jumping to this conclusion should not happen hastily and should
instead be compared against a high-performance single skin solution. Single-skin facades,
particularly those with high-performance glazing, perhaps triple glazing, and exterior shading
elements can achieve many of the desired performances that a double-skin facade can. A system
of this caliber should be the baseline for a cost evaluation that includes the operational costs
associated with all elements - specifically the maintenance of exterior shading elements on a
single-skin facade versus that of the solar control elements within the double-skin facade's cavity
space and the added cleaning costs associated with a double-skin's many surfaces.
Facade's Relationship to Form and Context
A double-skin facade should be located based on a number of possible environmental responses,
including solar-orientation, adjacencies to acoustical concerns (vehicles, trains, etc.) and
prevailing winds if natural ventilation harnessing is integral. In the northern hemisphere, the most
common orientation of double-skin facades is towards the southern exposure. This is deliberate
when the system seeks to harness solar heat gain during summer months, particularly for
climates that experience colder winter conditions at greater latitudes.
368
VI/hole Building Energy Approach
A multi-story double-skin facade may exist as a stand-alone buffer, independent of the
mechanical systems, or it could be an integral component used to provide natural ventilation,
supply air to mechanical systems, exhaust air and be used as part of a heat-recovery scheme.
There are many possibilities and the earlier the role of the double-skin facade within the whole
building energy approach can be defined, the more informed and aligned the decisions regarding
ventilation mode and airflow configuration selection will be. Additionally, the double-skin facade
will likely be required to support several different airflow modes through the various seasons.
Planning for measures of adaptability to these seasonal modes should be considered during the
schematic design phase.
Relationship to Structure
When decisions are made regarding the building's primary structural system approach, then there
is a context for a relationship between it and the multi-story double-skin facade's structure. One
consideration is how frequently the facade structure will attach back to the primary structure; does
this happen at every floor level or does the system need to span from base to roof? Is there a
different relationship for the inner skin compared to the outer skin? Since a multi-story cavity is
generally designed with a cavity of 0.4 m (16 in) to 0.9 m (36 in) or more - deep enough to permit
access for cleaning and maintenance purposes -then providing structure to the outer skin and
supporting grating in the cavity space should be considered. These elements can be developed
to serve multiple purposes (e.g. a grating at each slab level can provide solar shading for the
inner skin). There are many possible solutions for the structural relationship of the multi-story
double-skin facade to the primary structure. The two layers can be part of prefabricated unitized
curtain wall (see Section 5.3.9), they can be attached to a shared structural frame that spans from
ground to roof independent of intermediate floor levels (see Section 5.3.4), or the two skins can
have separate structural approaches with a simplified curtain wall inner and tensioned cable wall
369
exterior (see Sections 5.3.3 and 5.3.5) - just to describe a few possible solutions. Each of these
approaches has a different transfer of reaction forces back to the primary structure and early
attention should consider these relationships, including coordination between the primary
structural engineer and facade engineer consultants.
10.3.3 Design Development
The design development phase is the most critical in a double-skin facade's evolution. By this
stage the architect and owner have identified their priorities and mechanical engineers and
facade consultants are advising on best practices and the most feasible design solutions.
Discussions with and input from general and facade contractors may occur in this stage. The
architect is compiling all the analytical insights from consultants into a balanced envelope design
that requires special attention with respect to ensuring proper building code compliance. Prior to
advancing to the construction documentation phase, refined cost estimates converge towards
finalizing pricing with potential support from independent cost estimators, contractors' pricing
and/or facade contractors' pricing input. Several vital questions to be discussed during this pivotal
stage include:
1. VI/hat criteria make up the performance specifications for the double-skin facade
envelope?
2. VI/hat is the glass make-up of each skin layer?
3. VI/hat unique wind conditions - unusual geometric or spatial response characteristics -
exist that require special consideration in wind tunnel testing?
4. VI/hat operational requirements does the owner and facility management team have that
inform the double-skin, its operation, or the components within the cavity?
5. VI/hat visual and performance mock-up scopes are to be developed with the facade
contractor?
370
Performance Specifications
Articulating performance specifications is essential to define the expected level of quality and
performance that a building's systems will embody. During the bidding phase, the importance of
specifications is elevated as it is a tool to ensure all potential prices from facade contractors
reflect comparable or acceptable levels of performance.
Glazing Make-up
Identifying the glass make-up for a double-skin facade's multiple layers is a balancing of
structural requirements, energy requirements as required by the whole building energy model,
building codes, and aesthetics. Most often, insulated glass units are used on the layer that
represents the primary air and water barrier while laminated lites are used for the other skin. In an
outdoor air curtain mode, this would equate to laminated exterior lites and IGUs for the inner skin.
Unique Wind Conditions
It is the opinion of the author that just having a multi-story double-skin facade is an unusual
geometric characteristic as it relates to the wind flow. other unusual characteristics may relate to
the 3D massing of the form - it may not be rectilinear but a more organic, curvilinear shape (see
Section 4.3.6) - or be surrounded by an unpredictable wind environment in an urban context with
surrounding tall buildings.
Layers of Operational Performance
Defining the operational mode is a significant step towards selecting the layers of controls
required in a multi-story double-skin facade. Are all solar control elements automatically
controlled, perhaps responding to sensor data, or is there a layer of manual control? Which is
371
primary? Is there a manual override in some systems or is the primary operation manual with an
automatic override to reset a system due to some period of inactivity? These questions help
define the level of controls as well as complexity. Additionally, how these controls tie into the
overall building management system should be discussed during design development.
Mock-ups
There are two primary types of mock-ups used by the project team to ensure that quality and
performance are aligned with expectations prior to construction commencing: visual and
performance. The visual mock-ups are used to evaluate the qualities of products - glass often
being a primary concern - as well as proportions of shapes such as mullion profiles. Performance
mock-ups occur later than visual mock-ups when design decisions have been solidified and
performance of the air and water barrier as it relates to craftsmanship are the key concern. This is
an opportunity for the facade contractor to validate the performance of the multi-story DSF's
detailing prior to widespread fabrication, assembly and installation on site.
10.3.4 Construction Documents
The construction document phase is the point in the design process where the facade contractor
has input in the design team's conversations. In general, the more complex or less common a
facade system, the earlier the design team will seek participation from facade contractors (usually
multiple) to gain several perspectives and maintain a competitive field. Regardless of when the
formal milestone of commencing construction documents begins, the following questions are
important to consider when the design has advanced to a point when the facade contractor is
developing system design details and calculations for review by the design team.
1. How does the double-skin facade system perform thermally, structurally and operationally
to meet the performance specifications? Do the chosen facade contractor's design details
and calculations adequately reflect this?
372
2. How do the intended airflow openings get detailed and what is their impact on the
required design loads from the wind tunnel studies?
3. VI/hat components require physical samples for material approval by the design team?
4. VI/hat was learned from the mock-ups that can improve project execution?
Facade Contractors Integration of Performance Requirements
Each facade contractor will have different design approaches to achieve the architectural intent of
a double-skin facade. It is the job of the architect and pertinent consultants to review the facade
contractor's system detail development, provide feedback and ensure it is achieving desired
aesthetic and performance specification requirements.
Airflow Opening Details
Air inlet and exhaust opening details can vary widely and are not always fully designed during the
design development phase. With the facade contractor developing a project specific approach,
the integration of either standard or custom opening components occurs with collaboration
amongst the facade contractor, architect, facade consultant and mechanical engineers.
Material Samples
It is not practical to be able to review all materials or finish options at the full scale of a visual
mock-up, so material samples are used as early as possible to define acceptable directions.
Mock-up Lessons Learned
The mock-up process may reveal opportunities for improvement in the system design that require
revisions to the system design details and possibly the structural calculations.
373
10.3.5 On the Construction Site
There is a significant transition in responsibilities and activities when a project advances passed
construction documents and are in the field, on site with multiple trades, and numerous material
preparations occurring at many shops at different locations near the job site, in the region or even
across the globe. The construction administration and project execution phase is a coordination
challenge that is primarily focused on schedule, coordination of trades and ensuring the levels of
quality that were defined and accepted in the preceding phases are upheld in the execution of the
final installed systems. Several necessary considerations related to this phase include:
1. How are the two layers of the double-skin facade assembled?
2. How is the double-skin facade installed?
3. VI/hat special considerations should be considered for coordination of trades involved with
executing the work of, or immediately adjacent to, the double-skin facade?
4. VI/hat are the expectations for site coordination meetings amongst the owner, contractor,
architect and facade contractor?
5. VI/hat precautions may be taken to assure an acceptable level of quality control in the
shop, on the site during installation and post-installation while other trades are still
working near?
6. VI/hat is the process of commissioning required to provide confidence in system
performance prior to close-out and hand-off to the owner?
Assembly of Double-Skin Facade Systems
Enclosures can use varying degrees of off-site assembly to take advantage of shop-controlled
environmental conditions. The primary drivers for decisions regarding an assembly approach are
schedule, cost and quality. Generally speaking, maximizing off-site assembly will equate to a
more efficient installation, thus resulting in a more economical solution.
374
Installation of Double-Skin Facade
With the double-skin facade it is possible to consider delivery of both layers as a single pre
assembled unit if detailed appropriately. This approach reduces multiple passes of installation for
the exterior, often complicated by the first installed skin obstructing access to one of the sides of
the second skin.
Coordination of Trades
Depending on the level of integration of mechanical systems, solar control systems, and any
electrical requirements for operable components and sensors, the double-skin facade represents
a reasonably shallow depth where a lot of performance is desired and access for many parties
during installation is required. Coordinating the timing of each trade's activities sequentially with
the others is paramount to an efficient construction operation.
Quality Control I Quality Assurance
Measures must be taken to ensure that the installed product meets both the visual and
performance requirements agreed to and tested in mock-ups during earlier phases. Additionally,
the quality control and approval process should consider measures to protect the installed DSF
enclosure while other trades may be working to complete their adjacent scopes of work.
Commissioning
Defining a commissioning plan takes place in earlier phases, but it is at the transition from
construction to occupancy and operation when the implemented plan is executed. The steps for
commissioning should include optimizing the building systems relative to the double-skin facade
as well as education for continued optimal operation by the facility management team.
375
10.3.6 Post-Occupancy
While design and construction of a project with a double-skin facade may span a handful of
years, its life-span of intended use is likely on the order of a handful of decades. It is in the
efficiency of its use, not its construction, that the payback period is most significantly impacted.
The design team should not wait until the project is complete to first ask these questions, but
should work with the owner and facility management team throughout the design process to
continually address them. Following completion of construction, the owner and facility
management team should periodically revisit these questions, working to sustain longevity of the
double-skin facade and its role in the larger building energy approach.
1. VI/hat monitoring periods following commissioning may be required?
2. VI/hat is the operational maintenance plan?
3. VI/hat is the life-expectancy of the system and what incremental periods for updates are
considered as part of the facade design? Is the system structured for ease of
recyclability?
Post-occupancy Monitoring
Though the commissioning phase may occur at the end of construction, just prior to occupancy
and operation, there may be milestone check-ins and performance tracking to ensure the
systems related to the double-skin facade and whole building energy approach are operating as
intended. A plan for follow-up check-ins should be articulated amongst the owner, designer and
facility management teams during earlier phases of the project. The more that can be learned
from the actual built performance of the mechanical and control systems integral to a double-skin
facade the more informed they may be designed in future applications. Additionally, any building
systems interaction with building's occupants, such as manual overrides in solar control systems,
may require an educational process with the occupants to further improve performance.
376
Operational Maintenance Plan
When the design and construction teams have completed their tasks and vacated the jobsite, this
is when the building is just emerging from its infancy. It is during the remainder of the building life
that the merits of the upfront design decisions either prove true or unnecessary. An essential
element of a double-skin facade's performance success or failure is dependent on a simple and
effective operational maintenance plan. The design decisions that inform the functionality of the
systems over time should be informed by input of the facility management team. Involving the
facility management parties responsible for overseeing a double-skin facade early in the design
phases will streamline the learning curve and understanding during the implementation and start
up phases near the end of construction, as well as in operation.
Life-expectancy and Sustaining Performance
A comprehensive life-cycle plan for the double-skin facade enclosure will include a timeline and
plan for performance check-ins, potential system updates, and at some point, replacement.
10.3.7 Responsibility Matrix
The key considerations from the preceding sections are summarized in Table 10-1 where each of
the possible responsible parties for the DSF delivery have key tasks assigned for the six primary
phases of the project process. Figure 10-3 strives to map out these responsibilities in a
theoretical timeline with overlapping responsible parties for various tasks. The items highlighted in
a heavier black are directly relevant to the development of a double-skin facade within a larger
project. It is evident in this map that a successful implementation of a double-skin facade project
requires support and accountability from many parties, including the client, architect, multiple
consultants, general contractor and facade contractor. The early phases place tremendous
responsibility on the architect and supporting consultants, while the latter phases command the
facade contractor to deliver on meeting the designed and engineered intent of the enclosure.
377
Table 10-1: Responsibility Matrix of a Sample Construction Project.
Concept Schematic Design Development Construction Documents
Construction Admin I
Operation
Project Execution
I-
•Primary Structural
• Building Energy Model
z
• Envelope System Development • Ensure Contractors Adhere to
( Approach
• Consultants' Drawings Project Specifications or • Commissioning I-
•Whole Building Energy ...I
• Begin Performance Specifications Acceptable Alternates in • Response to Site Issues not
::>
Approach: Relationship
Cl)
·Wind Tunnel Studies Development of Shop Drawings Adhering to Specifications
z of Envelope and
0
Mechanical
(if applicable) and Calculations
u
• Code Compliance
• Identify Program •Building Code Compliance • Review Shop Drawings
I-
• Review Zoning •Preliminary Site Plan • Detailed Floor Plans • Answer Contractor Questions
u
Requirements • Develop Floor Plans • Integrate Consultants' Design • Material Sample Approval • Observation Services
w
I- • Research Design & • 3D Massing Details into Architectural Drawings • Construction Details •Additional Drawings and Design
:c
Materials • Conceptual Geometry • Code Compliance • Submissions to Building Feedback as Required
u
0:: • Feasibility Studies •Visualizations • 3D Model of Building Exterior Department for Approval • Site Meetings w/ Contractor & Owner
(
• Preliminary Budget •Propose Materials & • Compile Project Specifications • Assemble Punch List
Analysis Systems • Final Review
•Close-out
•Articulate Space
•Commissioning
0::
•Occupancy
w
Requirements • Approve Floor Plans • Review Contracto~s Cost
• Contractor's Contract •Sile Meetings with Contractor and • Space Management
z •Client • Select Contractor Estimate
~
Questionnaire • Review Updated Budget •Approve Design of Major Systems
Execution Architect • Systems/Equipment
0 Maintenance & Mgmt.
• Approve Program
• Energy Management
• Renovation Evaluation
0::
• Key Subcontractor Selection
• Construction Management
0 • Execution of Work
I- • Site Logistics Plan • Demolition, Site Preparation
u •Coordination of Trades
i2
• Contractor's Cost Estimate and Grading Plans
• Detailed Scheduling
• Repair of Work within
I-
• Division of Scope • Construction Review
• Response to Site Errors
Warranty Period
z • Preparation of Bid Packages •Schedule
0
• Final Pricing
• Close-out and Hand-over
u •Quality Control/Assurance
0::
• Develop Shop Drawings and
0 Structural Calculations • Project Management
w I-
cu • Material Vendors and Pricing • Procure Materials
• Repair of Work within
~'2
• System Design Details • Visual and Performance • Assemble Systems
Warranty Period
( I- Mock-ups (if applicable) • Installation of Enclosure
u. z
0
• Develop Means and Methods •Quality Control/Assurance
(,)
u
• Schedule Coordination
-....i
())
(,,)
Wind Tunnel Studies
~ Performance Specification Development j
::1 Envelope System Development!
~ Building Energy Modeling j
Response to Site Issues not
Adhering to Specifications
Ensure Contractors Adhere to
Performance Specification
Commissioning
Support
~ Whole Building Energy Approach I Consultanls'I
8 Primary Structural Approacl o.-ings
-t ld~:gp~:=- ~~~ft:? I Bu~il~~~~ -I~~: - - ~on~i:Deta~~- ~.:;;-m~s:n~o- - - j~v~w~;,-o~wi~~~~~L - ~i i~~~i~ - - - - - - - - - - - - - - -
w s· Pl Consultants' lls Building Dept. Drawmgs as Needed and Contractor
1- [Feasibility ite an n-i..AI D · for Approval
:I: ]Research Floor Plans ~ rawings Material Sample] Observation Services !Assemble Punch List
~ l Budget . j [ 3D Model Approval Answer Contractor l j Final
[ Analys 3D Massing of Exterior Questions Review
CONCEPT SCHEMATIC DESIGN CONSTRUCTION CONSTRUCTION ADMIN. I OPERATION
DEVELOPMENT DOCUMENTS PROJECT EXECUTION
a::
w
z
Approve)
Piogram
Ap~rovel
]
Budget
Review
Commissioning l
Systems Energy
Management Management
~
Floor Plans
Approve Designf
Space
Requirements
__ _ c~n~ O~e~o~a.'.'.. e _ _ _ _ 8~~gioi:_ __ I
Review
Cost E stimate
Contractor's Contract j
Execution
Space Management]
,..__':.c~u ,pancy l I
.,._...,.., Renovation
Evaluation
- --- - ------ - - -
a:: [ ] l [ Close-out &
8 Contractor's Cost Estimate j Final Pricing Hand-over
~ Site Log1st1cs Plan j I ] Demoh t1on, Site Preparation Construction [ ~
1
~~ ~:'c9~~
1 1
] Oualily . Control/ j Repair of .
,_ Preparation of Bid Packages and Grading Plans Management and Owner 1 j Ouahty Assurance Work within
z Warranty
o Key Subcontractor Selection J Cooldlnatlon Response to Site Errors Period
_ ~ _________________ ~ _ _ _ _ S_0~u le J Construction Review J Execution of Work of Trades Detailed Scheduling __ _ _ _______ _
w~ [ Schedule I l{t'g~r:,:hin
go Coordination W Pe a r
1
. r
0
ra dnty
_ System Design Details I
~I= Quality Control/
u.S Material Vendors and Pricing I Quality Assurance
Project Installation of Enclosure
Develop Shop Drawings and Management
Structural Calculations Assemble Systems
Performance Mock-ups
Develop Means and Methods Visual Mock-ups Procure Materials
(!l Figure 10-3: Conceptual design construction process related to developing the building skin.
10.3.8 Summary of Guidelines
The following is a set of guidelines related to the process and actual design of multi-story double
skin facades extracted from the previous sections. Some of the guidelines are practical in nature,
while others focus more specifically on the various airflow inlet and exhaust configurations
considered in the simulation portions of this research.
Practical Guidelines for DSF Implementation
• The driving motivations for a double-skin facade should be multi-faceted and
performance driven.
• The baseline single-skin facade for cost comparisons (DSF vs. single-skin) should be of
equivalent performance metrics (e.g. U-value, SHGC, acoustics STC and OITC, etc.) and
include the cost of external solar shading if required to meet performance values.
• For practicality and maintenance considerations, design for a minimum multi-story cavity
depth of0.4 m (16 in), preferably closer to 0.9 m (36 in).
• Insulated glass units should be present in the layer that represents the primary air and
water barrier between interior and exterior environmental conditions.
• To avoid direct entry of unwanted moisture or objects entering the cavity space, avoid
using a through exhaust if possible.
Guidelines Relating to Airflow
• Avoid communication between airflow cavities on adjacent building surfaces.
• Increased distribution of the airflow openings across the outer skin height, such as the
louvered or shingled inlets, may reduce concerns regarding cavity pressure.
• Where supply of natural ventilation to the interior is not an objective, a mechanically
pressurized cavity space may aid in mitigating volatility of cavity pressure.
380
10.4 Recommendations for the Role of CFO in the Design Process
The selection of a building's skin is one of the most critical features of building design. This
selection dictates how comfortable occupants will be, the energy demand of the building to create
comfort within the external climatic conditions, and the appearance of the building. The range of
possibilities for enclosure design is overwhelming. The earlier the multifaceted expectations of
performance are identified the sooner design development can mature. During this process
simulation tools become useful for a number of purposes and at different levels of detail that are
appropriate for each stage of design development. Amongst those simulation tools is
computational fluid dynamics. CFO can be used to study a multitude of configurations more
readily than a wind tunnel model may be able to. Some of the concern regarding computational
fluid dynamics is with respect to the accuracy of the results, often driven by the ability and
understanding of the modeler. These concerns regarding variation of data analysis, accuracy of
results, and the role of human error in the process also exist in wind tunnel tests, as noted by
Holscher and Niemann. Their research also advised that the atmospheric boundary layer wind
tunnel with simulated full-scale flows can produce reliable results if some fundamental
requirements are considered (Holscher and Niemann 1998, 607). The same level of some
fundamental requirements should be expected of computational fluid dynamic modeling, and
potentially the confidence level related to CFO as a tool for determination of wind pressure
coefficients will improve. This process will require a feedback loop for validation. In this research,
existing wind tunnel studies were simulated, with the wind tunnel study's results providing a
metric of confidence for the calibrations. If the same experimental procedure was carried out in
another wind tunnel facility, the results would vary indefinably. The objective in future iterations
would be to use simulation to model another building and facade configuration in the same
simulated wind tunnel configuration and then compare the results to the outcome of the actual
wind tunnel tests. The first step, simulation of another DSF prototype, has been taken in this
research with wind tunnel validation as potential future work. Ultimately, the standard deviation of
results, for both simulated and actual wind tunnel tests, would both decrease.
381
For the time being, CFO cannot be relied upon solely to determine the wind pressure coefficients
of a double-skin facade. CFO simulations may be used more frequently, though, to provide earlier
insight into the qualitative differences between possible multi-story double-skin facade airflow
configurations. Below is an outline of some opportunities for computational fluid dynamics to
supplement wind tunnel testing in the design process of multi-story double-skin facades.
Opportunities for CFO in the Design Process of Multi-story Double-Skin Facades
• CFO may be used prior to wind tunnel testing to study the implications of various three
dimensional building mass form geometries.
• CFO may be used prior to wind tunnel testing to understand the impact of surrounding
structures on the approaching wind flow traits, particularly in a dense urban setting. If
possible, the simulation is calibrated to the known wind tunnel to be used later in design.
• CFO may be used for preliminary studies of various multi-story double-skin facade
geometric and airflow configurations.
• CFO may be used to conduct extensive transient simulations, perhaps using wind
conditions recorded from the project site, to identify dynamic loading concerns that do not
reveal themselves in steady-state wind tunnel tests. Transient simulations require
extensive pre- and post-processing comparatively, as well as significantly longer
computational time, and should be used selectively.
• CFO may be used to study the areas of high detail concentration - such as at airflow
openings, dampers, louvers, grating, and around shading elements in the cavity space -
that is beyond a reasonable scale to be studied in the wind tunnel model.
• If producing consistent results to a wind tunnel, CFO may be used to evaluate alternate
design solutions that arise after a wind tunnel study has been completed, including a) as
part of the project design development, b) after its completion when a renovation is being
considered, or c) a new adjacent structure is altering the local wind flow environment.
382
10.5 Recommended Pressure Coefficient Determination Methodology
This section describes seven steps, some essential while others are suggested, in the process of
determining pressure coefficients for the design and engineering of multi-story double-skin
facades. Figure 10-4 outlines these seven steps.
2
Identify Code
Requirements
Determine if
Wind Tunnel Studies
are Required
Select Wind Tunnel
Testing Facility
6
Perform Project's
Wind Tunnel Tests
,___ _ ___,;. Design Pressure
Coefficients
4
Perform
Simple
Geometry
Calibration
Calibrate
CFO to
Wind
Tunnel
Select OSF
Configuration
for Wind Tunnel
Testing
Preliminary
1-----1•1 OSF Evaluations
Using CFO
-
Recalibrate
Simulation Model
to Wind Tunnel
• I
I I
1
FEEDBACK
1
~------- ------ ----------------J
LOOP
Figure 104: Suggested pressure coefficient determination methodology.
383
Identify Code Requirements
The first question the engineer must ask is what are the code requirements for this building and
building skin configuration? Figure 10-5 diagrams what this process looks like when using
ASCE/SEI 7 (2010) for the determination of wind loading. The first two questions relate to
determining if the building has any unusual geometric irregularities or response characteristics.
Several examples of irregular building geometries include "buildings with multiple setbacks,
curved facades, or irregular plans resulting from significant indentations or projections, openings
through the building, or multi-tower buildings connected by bridges" (ASCE/SEI 2010, 506).
ASCE/SEI 7 also lists a number of response characteristics that would require special
consideration, including "site locations that have channeling effects or wakes from upwind
obstructions" such as topographic features or buildings, and "buildings with response
characteristics that result in substantial vortex-induced and/or torsional dynamic effects, or
dynamic effects resulting from aeroelastic instabilities such as flutter or galloping," (2010, 506). If
a building's characteristics result in a "no" response for both Questions 1 and 2 in Figure 10-5,
then the appropriate procedural method is a function of its height and whether it is partially
enclosed or enclosed. Regardless of which procedure a building may qualify for, the wind tunnel
approach is always an acceptable option for the determination of wind pressures. In reality, many
of the buildings that possess a monumental multi-story double-skin facade are surrounded by an
urban building landscape, have an irregular geometry, or have significant indentions and
projections, thus pointing towards a wind tunnel approach for the determination of wind
pressures.
The research herein points towards the potential for a unique response characteristic any time a
multi-story double-skin facade exists. Therefore, the recommendation is that the wind tunnel
procedure be the baseline approach for determining wind pressures for all projects that possess a
multi-story double-skin facade.
384
Site Plan
(end of SD)
Building Mass
(end of SD)
Building Skin Model
(be9i1nning of DD)
\·Vr'll'"l'dl 1tir-·1q
1~;1 ~ J~ rat1 '.·:-1 ·1·:
ASCE/SE I 7 Ch. 31
Wind Tunnel Ptocedute
. - _,
~ Figure 10-5: Process of determining procedure for wind load determination.
CJ1
ASCE/SEI 7 Ch. 30
Envelope Procedure
r 1,,rt 1: Low-Rise Buildings
ASCE/SEI 7 Ch. 30
Envelope Procedure
p,,rt 2: Low-Rise Buildings
(Simplified)
ASCE/SEI 7 Ch. 30
,.. Direction11I Procedure
Part 3: Buildings w/
h > 18.3 m (60 ft)
ASCE/SEI 7 Ch. 30
Direction11I Procedure
Part 4: Buildings w/
h ~ 48.8 m (160 ft)
(Simplified)
Decide When in Design to Conduct Wind Tunnel Studies
The primary benefit of conducting a wind tunnel study for a building is that the results in design
loads are more accurate than those derived from codes. In an academic sense, or to the
structural engineer, this would be the preferred path for determination of wind loads. However, an
owner or developer has to weigh the benefits compared to the associated costs of such a study.
There is an up-front cost and the time required to conduct these studies, however, the wind loads
produced from wind tunnel testing are often lower than those from code, resulting in a more
efficient structural framing and glazing assembly. Occasionally, the wind tunnel results can
produce loads greater than the building codes, requiring greater safety factors be accounted for in
the engineered solution. In both instances, the owner is benefiting from either 1) reduced cost of
materials in the enclosure assembly, or 2) mitigating risk and potential damage (as well as
replacement) in the long-term performance of the enclosure. For these reasons, as well as the
unique geometric and the potential for unusual response characteristics, it is suggested that
buildings utilizing a multi-story OSF use wind tunnel testing for the determination of wind loading.
Select Wind Tunnel Testing Facility
Selecting a credible wind and climate consultant with boundary layer wind tunnel facilities and an
established reputation is important. There are not many of these consultancies, so in practice, the
field of wind engineering and environmental engineering is dominated by very few.
Perform Simple Geometry Calibration
Performing a baseline test in the wind tunnel of a regular geometry, such as the Simple Cube in
Section 8.4, can be used as an initial data set for a computational fluid dynamic simulation to be
calibrated to. If there is no correlation between wind tunnel and CFO for a simple geometry then
there is no chance to develop confidence in the accuracy of the CFO analysis of the OSFs.
386
Preliminary Evaluations Using CFO
Following a sufficient calibration ofCFD modeling to the simple wind tunnel tests to be used in
the final project-specific wind tunnel tests, simulations may be used to compare different
configurations, whether they vary in geometry, airflow mode, or another manner. These
evaluations are, at a minimum, qualitative comparisons, and in time could prove to have some
level of quantitative accuracy. Below are two figures qualitatively comparing results from the
twelve simulations conducted in Chapter 9 for multi-story DSFs with varied airflow configurations:
Raised Raised
Through Forward
Face Face Shingled Face S hingled
Through Forward Through Return Return
Smaller
0
Greater
C. c
"""
p.rnl
Trench Trench Shingled Trench Raised
Through Forward Forward Return R eturn
Figure 10-6: Relative comparison of inner skin pressures of twelve multi-story double-skin
facade configurations.
Trench Return
Raised Face Shingled
Return Return Forward
Shingled
Return
Shingled
Through
Raised
Forward
Raised
Through
Face Forward
Face
Through
Trench
F orwa rd
Tre nch
Th rough
Greater
c
p: xei
Figure 10-7: Relative comparison of net pressure across outer skin of twelve multi-story
double-skin facade configurations.
387
Wind Tunnel Testing
During design development there comes a point where the team has identified a building skin
approach that is articulated substantially enough to be integrated into atmospheric boundary layer
wind tunnel studies. This milestone generally occurs around the time that the performance
specifications are being compiled, and may be integrated or amended to them as part of the bid
documentation required to solicit proposals from qualified facade contractors. These results are
utilized for the building enclosure design as well as the primary structure. The wind tunnel
pressure results may reveal reduced loads compared to the code derived values, heightened
anomaly zones that require special treatment (e.g. reinforcing curtain wall mullions), or
occasionally, greater loads that may be accounted for - reducing risk and resulting in a safer,
more accurately engineered building. The wind tunnel report provided by the wind engineer will
include positive and negative net pressures that are required for the facade engineer and
contractor to design the building enclosure. Special considerations, which may require revision to
the traditional wind tunnel data reporting approach, are necessary for evaluating the net pressure
related to multi-story double-skin facades, especially when openings permit them to vary airflow
modes during different seasons, daily operational hours, or under comfortable climatic conditions.
Recalibration of Simulation Model to Wind Tunnel Testing
The best way to close the gap between the findings of preliminary computational fluid dynamics
simulations and the wind tunnel studies of the building configuration with a double-skin facade is
to compare the two following the wind tunnel studies, refine the simulation model accordingly, and
determine what modeling approaches require altering in future simulations to better emulate the
actual wind tunnel's atmospheric boundary layer. In practice, this step is best suited for either the
wind engineer or the facade consultant who intends to utilize the same atmospheric boundary
layer wind tunnel for future project applications.
388
10.6 Conclusions
This section compiles the conclusions from the respective chapters of the primary investigations
of this research: 1) how are double-skin facade applications being implemented in the United
States, 2) can computational fluid dynamics be used to simulate wind tunnel tests of multi-story
double skin facades, and 3) what are the unique wind pressure coefficient dynamics that the
simulations reveal for multi-story double-skin facades with varied airflow configurations?
DSF in the USA
• Double-skin facade applications in the United States are using multi-story configurations
at a greater rate than those summarized globally by Perino (2007).
• Double-skin facade applications in the United States are using the outdoor air curtain
ventilation mode at a greater rate than those summarized globally by Perino (2007).
• Double-skin facade applications in the United States are diverging in scale, gravitating
either towards the multi-story or box-window solutions.
CFO Calibration of Wind Tunnel Studies
• Computational fluid dynamic models can be tuned to replicate wind tunnel conditions for
steady-state analysis of impinging loads on multi-story double-skin facades.
• Replication of a specific wind tunnel's turbulence characteristics in a computational fluid
dynamics model requires special attention and iterative refinement, as these
characteristics are unique to each wind tunnel and have a noticeable impact on the
results produced from the simulation models.
• Computational fluid dynamics models should be used as a supplemental tool to wind
tunnel testing to provide pre- and post-experimental insights into multi-story double-skin
facade's pressure response characteristics.
389
Pressure Coefficient Determination for Multi-Story Double-Skin Facades
• Building codes do not adequately address the determination of wind pressure coefficients
for double-skin facades, particularly deeper cavity multi-story double-skin facades and
their possible airflow configurations.
• A multi-story double-skin facade's outer skin sees greater loads as compared to a single
skin, particularly near the building's edges and airflow openings.
• The airflow inlet configuration has a noticeable impact on the pressure coefficients of
multi-story double-skin facades, particularly the negative pressure on the inner skin within
the cavity space.
• The greatest variation of multi-story double-skin facade configurations is the Cp.m, profiles.
• Face and Trench inlet conditions exhibit similar CP and the CP·"" profiles.
• Forward and Through exhaust conditions exhibit similar CP and the CP·"" profiles with
slight variation near opening regions.
• Inlet conditions that elevate (Raised) or distribute the opening across the vertical height
(Shingled) exhibit a reduced magnitude of negative pressure in the cavity space through
the lower and center portions of the elevation.
• Return exhaust configurations exhibit a positive Cp.m, profile, contrary to the other exhaust
configurations that are negative.
• The positive CP·'"' of the Return configurations reduces the corresponding CP·"" profiles.
With the double-skin facade applications presented in the case studies of Chapters 4 and 5, as
well as the Appendices, it is evident that the range of conceivable structural and airflow solutions
for multi-story DSFs is diverse. It is impossible to address all of these possible solutions in a
building code, but an increased industry dialogue, advisement on best practices, suggested
methodological frameworks, and design guidelines related to the determination of wind pressure
coefficients for multi-story double-skin facades are all vital to advancing the industry as a whole.
390
10.7 Suggestions for Future Research
This section identifies several topic areas that would further the research conducted herein and
the understanding of how multi-story double-skin facade structures respond to airflow conditions:
• Validation via Wind Tunnel Study
Some level of assurance was provided by first simulating the known configuration of the
calibration studies. The simulations of the twelve different airflow configurations are for a
prototype of different dimensions. Ultimately, wind tunnel studies of this research's
simulated configuration would provide some understanding of accuracy. That said, the
simulations must be tailored to reflect the unique characteristics of any specific
atmospheric boundary layer wind tunnel.
• Analysis with Different Flow Directions
Only a wind incidence of 0° was considered in this research. Further simulations are
required to understand the effects of cross-wind and torsional wind loading.
• Analysis with Consideration to Flow Redirection at the Openings
The variable that is most urgent to study further is arguably the detailing of flow
redirection at airflow opening features. The possibilities are many, but starting with some
geometrically simple louvered or grating configurations seems to be a good starting point
An incremental step that may be less computationally difficult may be to introduce a
permeable surface at the airflow opening and specify the free flow area as a percentage.
This approach may keep the number of elements required in the simulation mesh, and
the simulation time, within reason.
• Analysis with Fluid-Structure Interaction
Fluid-structure interaction (FSI) problems are complex because they consist of structural
nonlinear boundary conditions imposed on fluid moving boundaries where the position is
part of the solution. This iterative computation requires substantial computational power
on the magnitude of a non-linear structural finite-element analysis. Pressure coefficients
391
in a less rigid facade are a heightened concern for systems that utilize a more flexible
skin, such as exterior skins designed as a cable wall (see Appendices C15, C22 and
C31). The magnitude of the structural deformation's effect on the fluid flow through the
cavity space is unknown and requires further investigation under a series of loading
conditions. In order to manage model sizes and computation time, the modeler should
first consider methods of simplifying the problem size, in particular by using symmetry
boundary conditions where applicable.
• In-Situ Measurements on Multi-Story DSFs
Full-scale tests are incredibly costly during the development of a multi-story double-skin
facade, as was mentioned in Section 1.3.1. Preparing a full-scale mock-up is not as easy
for a multi-story cavity as it is for a box-window DSF (see Sections 2.3.3 and 2.3.4). This
future research path would investigate existing built applications of multi-story double
skin facades. At a minimum, this research would capture data for upstream wind flow
from the surrounding environment and pressure response for the exterior and interior
skins. Multiple data locations captured simultaneously- both on the exterior and within
the cavity on the inner skin - would be preferable, including: typical central, near-ground,
near-roof, near-edge, near-inlet and near-exhaust zone locations.
The future research identified above can be classified into two categories: 1) analysis of variables
not considered in this research effort, and 2) steps towards empirical validation, or an
understanding of the inaccuracies, of the proposed methods outlined in this research's
simulations. The validation efforts would be of greater benefit, first, than the explication of other
simulation variables, in order to understand the merits of the research herein prior to generating
more simulation results. The validation processes are most appropriately conducted in an
academic research environment outside of project-specific applications. Additional research into
the existing practices of wind pressure determination for multi-story double-skin facades would
also be a beneficial contribution.
392
Chapter 10 Endnotes
2
The building prototypes simulated have a building height of h = 20 m (66 ft) and facade cavity
height of h"P = 21 m (69 ft). Low-rise buildings are defined as having a mean roof height h.:::.
18.3 m (60 ft) in ASCE/SEI 7 (2010, 241).
For example, the KIW Westarkade project in Frankfort, Germany (see Appendix A.33).
393
Bibliography
(GANA), Glass Association of North America. "GANA Effort in ASH RAE Appeal Successful"
GANA Advocacy October 2010.
http/twww.glasswebsite.com/advocacy/pdf/ASHRAE%20Appealpdf (accessed July 30, 2013).
Addis, B. Building: 3000 Years of Design Engineering and Construction New York: Phaidon,
2007.
Adjaye, D, P Allison, and Whitechapel Art Gallery. David Adjaye. making public buildings.
specificity, customization, imbrication. London: Thames & Hudson, 2006.
American Institute of Architects. Architectural Graphic Standards. Edited by A Pressman.
Hoboken: John Wiley & Sons, 2007.
American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE). Standard
901-2010 Energy Standard for Buildings Except Low-Rise Residential Buildings. ASHRAE,
2010.
Andreotti, G. "Past and Future of Double-Skin Facades." GPO Conference Proceedings.
Tampere, Finland: Glass Processing Days, 2005. 1-5.
Architects, William Rawn Associates. William Rawn Associates Architects. Images Publishing,
2014.
Architectural Record. "Building Types Studies, Museums: Lightcatcher Building at the Whatcom
Museum" December 2010.
394
Arons, D. Properties and Applications of Double-Skin Building Facades. MSc thesis in Building
Technology, USA Massachusetts Institute of Technology (MIT), 2000.
AS/NZS. ASINZS 1170.2: Structural design actions - Part 2: Wind actions. Standards Australia,
Standards New Zealand, 2002.
ASCE/SEI 7. Minimum Design Loads for Buildings and Other Structures. Reston, Virginia:
American Society of Civil Engineers, 2010.
Banham, R. The Architecture of the Well-tempered Environment. 2"' ed. Chicago: The University
of Chicago Press, 1984.
Baskaran, A "A Numerical Model to Evaluate the Performance of Pressure Equalized Rainscreen
Walls." Journal of Building and Environment 29, no. 2 (1994) 159-171.
Bauer, M, P Measle, and M Schwarz. Green Building: Guidebook for Sustainable Architecture.
New York: Springer, 2009.
Bernard, M. "The Tower at PNC Plaza." eco-structure, September - October 2012: 80-83.
BESTFACADE. Best Practice for Double Skin Facades. WP 5 Report Best Practice Guidelines,
Intelligent Energy Europe, 2007.
BESTFACADE. Best Practice for Double Skin Facades. WP 1 Report State of the Art, Intelligent
Energy Europe, 2005.
Blacken, B, T Stathopoulos, and J Carmeliet. "CFO simulation of the atmospheric boundary layer:
wall function approach." Atmospheric Environment41, no. 2 (2007): 238-252.
395
Boake, T M. "Structure and Its Enclosure." ACSA Technology Conference. Portland, Oregon:
ACSA, 2002. 64-73.
Boake, T, K Harrison, and A Chatham. The Tectonics of the Double Skin: North American Case
Studies. University of Waterloo, 2001, 1-13.
Bodart, M, and E Gratia. Bibliography study of control strategies in buildings equipped with
Ventilated Double-skin Facades. Universite Catholique de Louvain, 2003, (15).
Bodart, M, and E Gratia. Ventilated Double Facades Project- Maintenance and Durability of
Ventilated Double Facades. Architecture et Climat, Belgium: Universte Catholique de Louvain,
2002.
Booth, D. The Double Shell Solar House. Canterbury, New Hampshire: Community Builders,
1980.
Braham, W Active Glass Walls: A Typological and Historical Account. University of Pennsylvania
Department of Architecture, 2005, 1-4.
Brock, L. Designing the Exterior Wall: an architectural guide to the vertical envelope. Hoboken,
New Jersey: Wiley, 2005.
BSI. BS 6399-2.1997 Loading for buildings. Code of practice for wind loads. London BSI, 1997.
BSI. BS EN 1991-1-4. Draft 4.10a, 2005 UK. National Annex to Eurocode 1 - Actions on
structures - Part 1-4: General actions - wind actions. London: BSI, 2005.
Bucci, F. "Building the Shard. Icon of Time." Casabella, October 2012: 66-99.
396
Bungartz, HJ, and M Schafer. Fluid-Structure Interaction: Modeling, Simulation, Optimisation The
Netherlands: Springer-Verlag Berlin Heidelberg, 2006.
"Cedars-Sinai Advanced Health Sciences Pavilion Receives LEED-NC Gold Certification."
Medical Construction and Design. Los Angeles, February 11, 2014.
CEN!TC250. "Eurocode 1 Part 1-4 Actions on structures: General actions: Wind Actions." In EN
1991-1-4.2005 (E) 2005.
Choi, ECC, and Z Wang. "Study on pressure-equalization of curtain wall systems." Journal of
Wind Engineering and Industrial Aerodynamics 73, no. 3 (1998) 256-266.
Clarke, J A. Energy simulation in building design Oxford: Butterworth-Heinemann, 2001.
Compagno, A. "Innovative Tower Block Facades: Swiss Re London Headquarters and Westhafen
Tower, Frankfurt." Detail: Facades, July-August 2003: 820-826.
Compagno, A. "From Double Windows to Double Building Envelopes." Proceedings of Glass
Processing Days Conference. Tampere, Finland, 2003. 244-248.
Compagno, A. Intelligent Glass Facades - Material, Practice, Design Basel: Birkhauser, 2002.
Costola, D, B Blacken, and JLM Hensen. "Overview of pressure coefficient data in building
energy simulation and airflow network programs." Building and Environment 44, no. 10 (October
2009) 2027-2036.
CTBUH. CTBUH Height Criteria. 2013. http//www.ctbuh.org!TallBuildings/HeightStatistics/
Criteria/tabid/446/language/en-GB/Default.aspx (accessed September 16, 2013).
397
Cu no, J. The Modern Wing: Renzo Piano and the Art Institute of Chicago. Chicago: Art Institute of
Chicago, 2009.
Davenport, AG. "Rationale for determining design wind velocities." Journal of the Structural
Division, Proceedings of the America/ Society of Civil Engineers 86 (1960): 39-68.
De Bleecker, H, M Berckmoes, P Standaert, and L Aye. "MFREE-S Closed Cavity Facade: Cost
Effective, Clean, Environmental." Proceedings of the CTBUH 9th World Congress - Asia
Ascending: Age of Sustainable Skyscraper City: A collection of state-of-the-art, multi-disciplinary
papers on tall buildings and sustainable cities. Shanghai, China CTBUH, 2012. 310-318.
de Wit, S. Uncertainty in predictions of thermal comfort in buildings. PhD Thesis, Delft University
of Technology, 2001.
Detail Magazine. Architectural Details 2003. Munich: Routledge, 2004.
Detail. "Museum in Denver." October 2008: 1119-1123.
Detail: Facades. "Media Centre in Sendai." October- November 2001: 1263-1277.
Detail: Glass Construction "Administration Building in Hanover." April - May 2000: 397-405.
Detail: Refurbishment. "Kraanspoor" Office Building in Amsterdam." November 2009: 1206-1210.
Detail: Solar Architecture. "Administration Building in Worzburg." April - May 1997 343-348.
Detail: Solar Architecture. "Bank Tower in Fankfurt-on-Main." April - May 1997: 349-354.
Detail: Solar Architecture. "Company Headquarters Tower in Essen." April - May 1997: 355-363.
398
Dickinson, EE. "Global Hospitals: The Cleveland Clinic Abu Dhabi." Architect, February 2010: 56-
61.
Ehrlich, S. Ehrlich Architects - Learning, Working, Living. Guilin: Guangxi Normal University
Press, 2013.
Elkadi, H. Cultures of glass architecture. Hampshire, England: Ashgate, 2006.
ESDU International. ESDU Data Item 82026, Revision E: Strong winds in the atmospheric
boundary layer. London: ESDU International, 2002.
ESDU International. ESDU Data Item 83045, Revision C: Strong winds in the atmospheric
boundary layer. London: ESDU International, 2002.
Etheridge, D W, and M Sandberg. "Building Ventilation: theory & measurement. John Wiley &
Sons, 1996.
Fortmeyer, R. "Transparency: Literal and Sustainable." Architectural Record, July 2009: 100-106.
Franke, J, A Hellsten, H Schlunzen, and B Carissimo. Cost Action 732 - Best practice guideline
for the CFO simulation of flows in the urban environment. Brussels: COST, 2007.
Frechette, R, and R Gilchrist. "Towards Zero Energy: A Case Study of the Pearl River Tower,
Guangzhou, China " Proceedings of CTBUH 8th World Congress. Dubai CTBUH, 2008. 13.
Fritz, WP, et al. "International Comparison of Wind Tunnel Estimates of Wind Effects on Low
Rise Buildings Test-Related" Journal of Structural Engineering 134 (2008) 1887-1890.
399
Gardner, K. "Owning an Icon: What We've Learned frorn the Gherkin." CTBUH Jounal: no. II
(2013) 60-62.
Gensler. "Shanghai Tower." Design Update. 12 22, 2010.
http://www.gensler.corn/uploads/docurnents/Shanghai_ Tower_ 12_22_2010. pdf (accessed 11 9,
2012).
Gerhardt, HJ, and F Janser. "Wind loads on wind permeable facades." Journal of Wind
Engineering and Industrial Aerodynamics 53 (1994) 37-48.
Gerhardt, HJ, and 0 Kruger. "Double skin glass facades - Investigations into the load sharing
possibilities." International Conference on Building Envelope Systems and Technologies
Conference. Bath, UK ICBEST, 1997.
Geurts, C.P.W, P.C. van Staalduinen, and M.S. de Wit 'Towards a reliable design of facade and
roof elements against wind loading." HERON 49, no. 2 (2004) 171-187.
Gonchar, J. "Legible and Luminous: William Rawn's pristine glass box for the Cambridge Public
Library quadruples the size of the original historic building." Architectural Record, October 2010:
76-81.
Gonchar, J. "Let the Fresh Air In: A tower offers an alternative to the hermetically sealed glass
box." Architectural Record, October 2012: 100-101.
Gratia, E, and A Herde. "The rnost efficient position of shading devices in a double-skin facade."
Energy and Buildings 39 (2007) 364-373.
400
"H. R. 6 -- 11 o'" ongress: Energy Independence and Security Act of 2007." www Gov Track.us.
2007. http://www.govtrack.us/congress/bills/11 O/hr6> (accessed September 12, 2013).
Haase, M. "Development of a double-skin facade system that combines airflow windows with
solar chimneys." Proceedings of The 2005 World Sustainable Building Conference. Tokyo, 2005.
763-770.
Hagberg, E. "The House that Phil Knight Built" Metropolis, August 2010 51-57, 85, 87.
Hagentoft, CE. The Impact of Heat, Air and Moisture Transport on Energy Demand and
Durability ECBCS Annex Publications - Annex 24 Heat, Air and Moisture Transfer in Insulated
Envelope Parts, Leuven: International Energy Agency, 1998.
Hanjalic, K. "Will RANS survive LES? A view of perspectives." ASME Journal of Fluids
Engineering 127 (2005) 831-839.
Harrison, K, and T Meyer-Boake. The Tectonics of the Environmental Skin. Ontario, Canada:
University of Waterloo, School of Architecture, 2003.
Hart, S. "Seeking innovative alternatives, Kieran Timberlake Associates pumps research into
architecture." Architectural Record, October 2003: 42.
Hausladen, G, MD Saldanha, and P Liedl. ClimateSkin: Building-skin Concepts that Can Do More
with Less Energy 1 '
1
ed. Basel, Switzerland: Birkhauser, 2008.
Hegedus, K. "Convincing the Client: A Double-Envelope Facade at the Seattle Justice Center."
High-Performance Commercial Building Facades. April 30, 2001.
http://gaia.lbl.gov/hpbf/design_g4.htm (accessed December 21, 2008).
401
Hell, A, and F Kaltenbach. "Low-Energy Office Tower- Innovative Double-skin Facade with
Decentralized Facade Ventilation" Detail, April 2011 • 404-412.
Hendriksen, OJ, H Sorensen, A Svenson, and P Aaqvist. "Double Skin Facades -- Fashion or a
step towards sustainable buildings." Proceedings of /SES. Eurosun, 2000. (8).
Hens, H, D Saelens, V De Meulenaer, and P Elsen. "Multiple-skin facades• high tech blessing or
not?" Proceedings of the 8th Symposium on Building Physics in the Nordic Countries.
Copenhagen, Denmark, 2008. 1261-1268.
Hensen, J, M Barlak, and D Frantisek. "Modeling and simulation of a double-skin facade system."
ASHRAE Transactions 108, no. 2 (2002) 1251-1259.
Herzog, T, and et al. Facade Construction Manual. 1st ed. Vol. Construction Manuals. Basel•
Birkhauser, 2004.
Herzog, T, R Krippner, and W Lang. Facade Construction Manual. 1 '
1
ed. Vol• Construction
Manuals. Basel• Birhauser, 2004.
Hill, D. Architectural Record. October 26, 2007. httpJ/archrecord.construction.com/news/
daily/archives/071026mca.asp (accessed April 21, 2008).
Hindrichs, D, and W Heusler. Facades - Building Envelopes for the 21st Century 3'' ed. Basel•
Birkhauser, 2010.
Holmes, J D. Wind Load of Structures. Span Press, 2001.
402
Holscher, N, and HJ Niemann. "Towards quality assurance for wind tunnel tests: A comparative
testing program of the Windtechnologische Gesellschaft." Journal of Wind Engineering and
Industrial Aerodynamics 7 4 (1998) 599-608.
Horsley, CB. "Prudential Project Includes 'Ice Pond'." The New York Times, May 17, 1981.
lnculet, DR, and AG Davenport "Pressure-equalized rainscreens: A study in the frequency
domain." Journal of Wind Engineering and Industrial Aerodynamics 53 (1994): 63-87.
lrminger, JOV, and CHR N0kkentved. "Wind-pressure on buildings." lngeni0r-Videnskabelig
Skrifter (K0benhavn: Danmarks naturvidenskabelige samfund) A.42 (1936): 88.
Irwin, P. "Wind engineering challenges of the new generation of super-tall buildings." Journal of
Wind Engineering and Industrial Aerodynamics 97, no. 7-8 (September-October 2009) 328-334.
Irwin, P, R Denoon, and D Scott Wind Tunnel Testing of High-Rise Buildings (CTBUH Technical
Guide) New York Routledge, 2013.
Ishida, H. "Double-Skin Glass Facade - Design Method of Wind Pressure." Glass Processing
Days. Tampere, Finland GPO, 2003. 253-255.
Jensen, M, and N Franck. Model-scale tests in turbulent wind - Part II. Copenhagen: The Danish
Technical Press, 1965.
Jiru, TE, and F Haghighat. "Modeling ventilated double skin facade - A zonal approach." Energy
and Buildings 40 (2008) 1567-1576.
Jodidio, P. Jean-Paul Viguier: Architecture 1992-2002. Basel: Birkhauser, 2002.
403
Kawai, H. "Pressure in a cavity of building-high double facade." Journal of Wind Engineering 108
(2006) 741-744.
Kawai, H. 'Wind force on double skin facade with building high glazing." Proceedings of the 12th
International Conference on Wind Engineering. 2007. 311-318.
Kawai, H. 'Wind Loads on Double Skin Facades with Opening Windows." Proceedings of The Six
Asia-Pacific Conference on Wind Engineering. 2005. 231-247.
Kawai, H, H Nishimura, M Suzuki, and Y Oura. "Field Measurement of Wind Pressure on a
Double Skin with a Ventilator" Edited by C Barri, G Augusti and G Bartoli. Proceedings of the 5th
European and African Conference on Wind Engineering EACWE 5. Florence, Italy: Firenze
University Press, 2009. 229-232.
Kieran, S, and J Timberlake. Loblolly House. New York: Princeton Architectural Press, 2008.
Kirchhoff, J, J Vaglio, M Patterson, and D Noble. "Double-skin cable-net facade case study:
Loyola Information Commons." BESS 2010: High Performance Building Enclosures - Practical
Sustainability Symposium. Pomona: California State Polytechnic University, 2010. 267-277.
Knaack, U, K Tillmann, M Bi low, and T Auer. Facades: Principle of Construction. Basel:
Birkhauser, 2007.
Kragh, M. "Monitoring of Advanced Facades and Environmental Systems." Whole-Life
Performance of Facades. Bath, UK: University of Bath I CWCT, 2001. 71-80.
Kwok, A, and W Grondzik. The Green Studio Handbook: Environmental Strategies for Schematic
Design. New York: Routledge, 2011.
404
Lang, W, and T Herzog. "Using multiple glass skins to clad buildings." Architectural Record, July
2000 181.
Lee, E, S Selkowitz, V Bazjanac, V lnkarojrit, and C Kohler. High-Performance Commercial
Building Facades. Building Technologies Program, Environmental Energy Technologies Division,
Ernest Orlando Lawrence Berkeley National Laboratory (LBNL), Berkley, California: University of
California, 2002.
Liddament, MW Air Infiltration Calculation Techniques - An Application Guide. AIVC Document
AIC-AG-1-86, AIVC, 1986.
Liu, H. Wind Engineering: A Handbook for Structural Engineers. Prentice Hall, 1991.
Lancour, X, A Deneyer, M Blasco, G Flamant, and P Wouters. Ventilated double facades
classification and illustration of facade concepts. Division of Building Physics, Belgium: Belgian
Building Research Institute (BBRI), 2004.
Lubell, S. "Walter Cronkite School of Journalism." Architect, July 2009.
Manz, H. "Total solar energy transmittance of glass double facades with free convection." Energy
and Building 36 (2004) 127-136.
Manz, H, and T Frank. "Thermal simulation of buildings with double-skin facades." Energy and
Buildings 37 (2005) 1114-1121.
Marques da Silva, F, and MG Gomes. "Effects of different multi-storey double skin facade
configurations on surface pressures." CISBAT 2005 Renewables in a changing climate!
405
Innovation in building envelopes and environmental systems. Lausanne, Switzerland, 2005. 353-
358.
Marques da Silva, F, and MG Gomes. "Gap inner pressures in multi-storey double skin facades."
Energy and Buildings 40 (2008) 1553-1559.
Masera, G. "The Shard at London Bridge, UK - Renzo Piano Building Workshop." II Sole 24 ORE
Arketipo, November 2012 46-59.
Mehta, RD, and P Bradshaw. "Design rules for small low speed wind tunnels." The Aeronautical
Journal of the Royal Aeronautical Society, November 1979: 443-449.
Metals in Construction. "Riverhouse: Double Vision." Fall 2008: 28-33.
Moazami, K, and R Slade. "Engineering Tall in Historic Cities: The Shard." CTBUH Journal, no. II
(2013) 44-49.
Murray, S. Contemporary Curtain Wall Architect. New York: Princeton Architectural Press, 2009.
National Oceanic and Atmospheric Administration (NOAA). National Coastal Population Report:
Population Trends from 1970 to 2020. NOAA's State of the Coast Report, Silver Spring, MD:
NOAA, 2013.
Nawab, M. "Full Scale Testing and Computer Simulation of a Double Skin Facade Building."
Building Simulation. Montreal: International Building Performance Simulation Association, 2005.
831-838.
Newhouse, V. "New World Center." Architectural Record, May 2011: 114-121.
406
Oesterle, E, Lieb, RD, Lutz, M & Heusler, W Double-Skin Facades: Integrated Planning. Munich:
Prestel, 2001.
Olmstead, D, and D Neely. Genzyme Center Shedding New Light on Sustainable Building
Strategies. A Case Study of Sustainable Building Strategies, Tufts Climate Initiative & Tufts
University Department of Urban and Environmental Policy and Planning, 2005.
Overend, M, and K Zammit. 'Wind loading on cladding and glazed facades." Proceedings of the
2nd International Symposium on the Application of Architectural Glass. Munich, 2006.
Patterson, M. Structural Glass Facades and Enclosures. Hoboken: John Wiley & Sons, Inc.,
2011.
Perino, M (ed.). State of the Art Review Vol. 2A: Responsive Building Elements. Annex 44:
Integrating Environmentally Responsive Elements in Buildings, Aalborg University: International
Energy Agency, 2007.
Perino, M, and V Serra. "Advanced Integrated Facades: An Overview." 4th Annex 44 Forum.
Torino IEA-ECBCS-ANNEX 44, 2006.
Poirazis, H. Double skin facades for office buildings: a literature review Report EBD-R--04/3,
Department of Construction and Architecture, Division of Energy and Building, Lund, Sweden:
Lund University, 2004.
Poirazis, H. Double skin facades: a literature review IEA SHC Task 34 ECBCS Annex 43,
Sweden: Department of Construction and Architecture, Lund University, 2006.
407
Potangaroa, R, and R Aynsley. "Pressure distributions behind double facades." Proceedings of
the 11th International Conference on Wind Engineering. Lubbock, Texas, 2003. 713-720.
Prendergast, J. "A Passion for Putting Things Together." Pennsylvania Gazette, November 2003.
Quirouette, RL, and J Rousseau. "A review of pressure equalization and compartmentalization
studies of exterior walls for rain penetration control." Water Leakage Through Building Facades
(ASTM STP, 1314), 1998 47-56.
Reinhold, TA. Wind tunnel modeling for civil engineering applications. Cambridge: Cambridge
University Press, 1982.
Richards, PJ, RP Hoxey, BO Connell, and DP Lander. 'Wind-tunnel modelling of the Silsoe
Cube" Journal of Wind Engineering and Industrial Aerodynamics 95 (2007) 1384-1399.
Rio Declaration on Environment and Development. June 3-14, 1992.
http//habitat.igc.org/agenda21/rio-dec. htm (accessed August 14, 2010).
Saelens, D. Energy Performance Assessment of Multiple-Skin Facades. PhD dissertation,
Leuven: KU Leuven, Laboratory for Building Physics, 2002.
Saelens, D, J Carmeliet, and H Hens. "Energy Performance Assessment of Multiple Skin
Facades." lnternationalJournalofHVAC&R Research9, no. 2(2003):167-186.
Schittich, C (ed.). In Detail. Building Skins. Basel Birkhauser, 2001.
Schittich, C, ed. In Detail. Building Skins. Enlarged Edition. Basel Birkhauser, 2006.
408
Schittich, C, G Staib, D Balkow, M Schuler, and W Sobek. Glass Construction Manual. 2"' ed.
Basel: Birkhauser, 2007.
Schneider, J. "Cronkite Communication School Speaks to Phoenix Redevelopment." Building
Design and Construction, May 2009: S1O-S12.
Selkowitz, SE. "Integrating Advanced Facades into High-performance Buildings." Proceedings of
the 7th International Glass Processing Days Conference. Tampere, Finland: GPO, 2001. 52-59.
Sharma, RN. "Internal and Net Envelope Pressures in a Building having Quasi-static Flexibility
and a Dominant Opening." Journal of Wind Engineering and Industrial Aerodynamics 90, no. 6-7
(2007) 107 4-1083.
Shi, X, and E Burnett "Mechanics and test study of flexible membranes ballooning in three
dimensions" Building and Environment 43, no. 11 (2008) 1871-1881.
Shurcliff, WA Superinsulated Houses and Double-Envelope Houses: A Preliminary Survey of
Principles and Practice. 2"' ed. Cambridge, Massachusetts: William Shurcliff, 1980.
Shurcliff, WA Superinsulated Houses and Double-Envelope Houses: A Survey of Principles and
Practice. Andover, Massachusetts: Brick House Pub, 1981.
Simiu, E. 'Wind Loading Codification in the Americas: Fundamentals for a Renewal." 11th
Americas Conference on Wind Engineering. San Juan, PR, USA, 2009.
Simiu, E, and R H Scanlan. Wind effects on structures: fundamentals and applications to design,
Volume 1. Wiley, 1996.
409
Sinclair, R, D Phillips, and V Mezhibovski. "Ventilating Facades" ASHRAE Journal, April 2009
16-27.
Smith, A The architecture of Adrian Smith, SOM: toward a sustainable future. Edited by Robyn
Beaver. Victoria: The Images Publishing Group, 2007.
Smith, T, and LP Butler. The Energy Producing House: Handbook/Case Study of the Passive
Solar House. San Francisco: Ekose'a, 1978.
Sobek, W "The Transparent High Rise Building of Deutsche Post in Bonn." Proceedings of Glass
Processing Days Conference. Tampere, Finland: GPO, 2003. 127-130.
Sokol, D. "First Look: John and Frances Angelos Law Center at the University of Baltimore."
Architectural Record, April 16, 2013.
Staib, G. "From the origins to classical modernism." In Glass Construction Manual, by C Schittich,
G Staib, D Balkow, M Schuler and W Sobek, 10-29. Basel Birkhauser, 2007.
Stephens, S. "Robert AM. Stern Architects raises the bar with Philadelphia's Comcast Center."
Architectural Record, May 1, 2009 169.
Stough, B. "2015 Top Glass Fabricators: Feats of Fabrication." Glass Magazine,
January/February 2015: 50.
Strabala, M. Shanghai Tower, Shanghai. 2011. http://www.flickr.com/photos/architectural
design/5625187336/ (accessed August 17, 2013).
410
Streicher, Wet al. On the Typology, Costs, Energy Performance, Environmental Quality and
Operational Characteristics of Double Skin Facades in European Buildings. Vol. 1, in Advances in
Building Energy Research, edited by M Santamouris. London: Earthscan, 2007.
Suresh Kumar, K, and JA Wisse. "Pressure equalization of rainscreen facades: Analysis of the
field data in the frequency domain." Wind and Structures 40, no. 2 (2001) 101-118.
Tanaka, H, M Okumiya, H Tanaka, GY Yoon, and K Watanabe. "Thermal characteristics of a
double-glazed external wall system with roll screen in cooling season." Building and Environment
44 (2009) 1509-1516.
The Plan 'TiFS Engineering Headquarters - Padua, Italy." April 2005 84-92.
Tigerman, S. Bruce Graham, SOM. New York: Rizzoli, 1989.
Timberlake, J, R Maiman, and Y Gauthier. "On the performance characteristics of an active,
pressure-equalized, double-skin curtain wall (active wall)." Kieran Timberlake. 2000.
www.kierantimberlake.com (accessed May 22, 2009).
TNO Building Reseach. Cp Generator: Cp Info n.d. http://cpgen.bouw.tno.nl/cp/info.asp
(accessed August 9, 2013).
United Nations. United Nations Framework Convention on Climate Change. Rio de Janeiro:
United Nations, 1992, 1-33.
US Department of Energy. Building Envelope: Technology Roadmap. Developed by
representatives of the commercial buildings industry, facilitated by the Office of Building
411
Technology, State and Community Programs, Energy Efficiency and Renewable Energy, U.S.
Department of Energy, no date>.
US Department of Energy. High-Performance Commercial Buildings: A Technology Roadmap.
Developed by representatives of the commercial buildings industry, facilitated by the Office of
Building Technology, State and Community Programs, Energy Efficiency and Renewable Energy,
U.S. Department of Energy, June 2001.
USGBC. lnfographics. LEED in the World. May 3, 2013.
http://www.usgbc.org/articles/infographic-leed-world (accessed June 3, 2013).
USGBC. LEED in New and Existing Markets. June 5, 2013. http://www.usgbc.org/articles/usgbc
offer-free-leed-certification-groundbreaking-projects-new-markets (accessed August 14, 2013).
Uuttu, S. Study of Current Structures in Double-Skin Facades. MSc thesis in Structural
Engineering and Building Physics, Finland: Department of Civil and Environmental Engineering,
Helsinki University of Technology (HUT), 2001.
Vaglio, J, and M Patterson. "Seeing double: applications of double-skin facade (DSF) systems on
the rise" USG/ass Metal and Glazing 46, no. 2 (2011) 34-36, 62.
Vaglio, J, M Patterson, S Hooper, and D Noble. "Emerging Trends and Applications of Double
Skin Facades in the United States." International Conference on Building Envelope Systems and
Technologies (ICBEST). Vancouver: National Research Council Canada, 2010. 191-199.
van Meurs, J, and PG Luscuere. "Double Skin Facade and Hybrid Ventilation of the ING House
Amsterdam (ING Bank HQ)." Proceedings of the 9th International Conference on Indoor Air
412
Quality and Climate. Monterey, California: International Academy of Indoor Air Sciences, 2002.
1044-1049.
van Schijndel, AWM, and SFC Schols. "Modeling pressure equalization in cavities." Journal of
Wind Engineering and Industrial Aerodynamics 7 4-76 (1998) 641-649.
Vazquez, JGV. Fluid-Structure Interaction Analysis: Development with Finite Elements. Germany:
VDM Verlad Dr. Mueller, 2008.
Weller, B, K Harth, S Tasche, and S Unnewehr. Glass in Building (Detail Practice). Basel:
Birkhauser, 2009.
Wellserhoff, F, and M Hortrnanns. "Wind loads on double facade systems with gaps greater than
15 crn" Proceeding of the Wind Engineering into the 21st Century Copenhagen, Denmark:
Balkerna Publishers, 1999.
Wieringa, J. "Updating the Davenport roughness classification." Journal of Wind Engineering and
Industrial 41 (1992) 357-368.
Wigginton, M. "Glass architecture and the interactive building envelope principles and precedent."
Proceedings of EU COST C13 Glass and Interactive Building Envelopes. Delft University Press,
2007. 27-31.
Wigginton, M, and J Harris. Intelligent Skins. Boston: Architectural Press, 2002.
Wong, PC. Natural Ventilation in Double-Skin Facade Design for Office Buildings in Hot and
Humid Climate. PhD Thesis, Australia: The University of New South Wales, 2008.
413
Wood, A, and R Salib. Natural Ventilation in High-Rise Office Buildings (CTBUH Technical
Guide) New York Routledge, 2012.
Wright, G. "Enhanced enclosures: dual-wall facades, long common in Europe, are finding
applications in a few U.S. projects, but the concept still seems foreign to most owners and
developers." Building Design & Construction, 2005 64 (6).
Xie, J, GD Schuyler, and HR Resar. "Prediction of net pressure on pressure equalized cavities."
Journal of Wind Engineering and Industrial Aerodynamics 41-44 (1992) 2449-2460.
Young, C. "Building a Double Wall." Building Design and Construction, October 12, 2011.
Zammit, K, M Overend, and D Hargreaves. "Improved computational methods for determining
wind pressures and glass thickness in facades." Challenging Glass 2. Delft University of
Technology, 2010.
Zeiger, M. "University of Oregon John E. Jaqua Academic Center for Student Athletes." Architect,
November 2010: 95.
Zeljic, AS. "Shanghai 632 Meter Tower Facade Design: The Complexities and Decision-Making
Process Involved in Developing the Final Design." Proceedings of the International Conference
on Building Envelope Systems and Technologies (ICBEST 2010). Vancouver, Canada: National
Research Council Canada Institute for Research in Construction, 2010. 169-187.
Zerofootprint. "Zerofootprint Re-Skinning Awards 2010." Zerofootprint Software. 2010.
http://zerofootprintsoftware.com//wp-
content/uploads/2012/04/ZFP _ReSkinning_Winners_2010.pdf (accessed November 11, 2011).
414
Appendix A: International Double-Skin Facade Case Studies
This appendix presents 40 case studies, chronologically, of double-skin facades from across the
globe. With the exception of the first case study, United States based double-skin facade
examples are outlined in Appendix C.
A.1 Occidental Chemical Center
Completed: 1980
Location: Niagara Falls, NY, USA
Architect: Canon Design
Cavity Partition: Multi-Story
Ventilation Type: Natural
Ventilation Mode: Outdoor air curtain
Cavity Depth: 1.2 m (3.9 ft)
Shading: Louvers, automated with override
Sources: Boake, Harrison & Chatham (2001, 1-5);
Wigginton & Harris (2002, 163-168); Compagno (2003,
245); Poirazis(2006, 205-206); Wright (2005, (4))
Figure A-1: Occidental Chemical Center - Overall view.
Source: Wigginton & Harris (2002, 165); ©Barbara Elliott Martin I Cannon Design.
This building is widely recognized as the first modern double-skin facade. The 18,580 m
2
(200,000 sf) building consists of a nine-story square plan around a central core resulting in
column free office space for approximately 500 employees at its time of completion. The nine
story double-skin facade was adopted to protect solar control devices at the highly wind-exposed
415
site, which would have made use of exterior solar control devices with a single-skin facade
challenging (Compagno 2003, 245). Additionally, during operations the envelope remains
transparent giving views on out towards Niagara Falls. When the building is completely vacant,
the louvered shading system is shut to provide insulation, retaining internal heat that was
generated during the day to reduce heating cost. The double-skin reduces the impact of severe
external temperatures by minimizing infiltration from the cavity to the interior conditioned space.
During the winter, the cavity acts as a thermal buffer. A mockup done at Arizona State University
found the double-skin to have a U-value of 0.27 Btui°F-ft
2
-h (Wigginton and Harris 2002, 164).
The 1.2 m (3.9 ft) cavity space is encompassed by a blue-green tinted IGU exterior with a visible
transmittance up to 80% and single clear glazing for the interior. Within the cavity depth are
operable white-painted louvers spaced 200 mm (8 in) apart. The louver position varies between
horizontal to 45° angle based on sunlight hitting a single sensor placed at the center of each
elevation. The white louvered surface reflects daylight reaching over half the usable floor space.
The original system relied on the BMS to control the louver position and airflow dampers at the
cavity inlet and exhaust. Occupant override was only provided for the louver position at corner
offices.
The original design was projected to perform at 39,000 Btu/sf, but has actually performed as low
as 32,000 Btu/sf (Wright 2005, (4)). One complaint about the operations is that the louvers are
not utilized as originally intended; they typically remain open at night except on occasional
weekends. The building has been only partially occupied since 2000 and has fallen into dire
states including extended periods of time with failed, inoperable controls.
416
A.2 Super Energy Conservation Building
Completed: 1982
Location: Tokyo, Japan
Architect: Ohbayashi-Gumi
Cavity Partition: Multi-story
Ventilation Type: Hybrid
Ventilation Mode: Air supply,
outdoor air curtain
Cavity Depth: n/a
Shading: Blinds, automated
Sources: Wigginton & Harris (2002,
159-162); Wong (2008, 87-89)
Figure A-2: Super Energy Conservation Building- Overall view.
Source: Wigginton & Harris (2002, 160); © Ohbayashi Corporation.
This is the first double-skin facade application in Japan. This four-story building was originally
designed to be an exhibit building shCM'casing innovative technologies that can be used in
generating a low-consumption office building for one of the country's largest general contractors,
Ohbayashi Corporation. The building uses an inclined double-skin facade on the south exposure
with automated controls to vary between two seasonal operation modes. During the winter
months, the cavity preheats incoming air before directing it to the air handling units for distribution
within the building. In the summer, air intakes located ICM' in the wall and air-extracts at the top of
the cavity are opened to create natural ventilation to help reduce peak cooling loads. The
building uses night cooling to reduce cooling loads during the follawing day. This approach is
improved by lowering insulating blinds outside the inner facade (within the cavity) during night
cooling to minimize losses.
417
A3 Briarcliff House
Completed: 1984
Location: Farnborough, England
Architect: Arup Associates
Cavity Partition: Multi-Story
Ventilation Type: Natural
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 1.2 m (3.9 ft)
Shading: Blinds, automated
Sources: Compagno (2002, 119);
Poirazis (2006, 191-192)
Figure A-3: Briarcliff House - Overall view.
Source: Compagno (2002, 119).
This double-skin facade was selected for it acoustic protection from traffic and aircraft and its
ability to give the interior offices solar control. The intermediate cavity space is 1.2 m (3.9 ft)
deep encompassed by an exterior skin with 10 mm (3/8 in) heat-absorbing single glazing and an
inner skin of insulated glass. The cavity space contains sensor-controlled louver blinds at the
interior skin and exposed ventilation ductwork. During the winter months, outdoor air enters at
ground level and is lifted up the cavity space before being recovered by a heat exchanger.
During the summer months, the heat exchanger is bypassed and the air exits the cavity at
louvered extract openings atop the roof return. Poirazis describes the primary facade function as
an active wall or climate wall (2006, 191 ). The cavity space provides access for cleaning and
contains the ventilation ducts.
418
A.4 Caixa Geral de Dep6sitos, Av. Da Republica
Completed: 1983-1987
Location: Lisbon, Portugal
Sources: BESTFACADE (2005, 54 and A 11);
Streicher (2007, 1-28)
Figure A-4: Caixa Geral de Dep6sitos - Overall Str eet View.
Source: maps.google.com.
This is the first known double-skin facade constructed in Portugal. Most double-skin facades in
Portugal are located in Lisbon. These designs are usually taller than five stories tall. Common
typologies are corridor facades and multistory facades. Primary reasons for double-skin facades
in this region are aesthetics and energy conservation (BESTFACADE 2005, 54). Further
information regarding this facade was not available.
A.5 EAL, School of Architecture of Lyon
Completed: 1987
Location: Vaulx en Velin, France
Sources: BESTFACADE (2005, 55 and A 7)
Cavity Partion: Story-Height Corridor
Width: 119 m (390 ft)
Height: 12 m (39 ft)
Figure A-5: EAL, School of Architecture of Lyon - Overall Street View.
Source: BESTFACADE (2005, A 7).
419
A.6 Business Promotion Centre
Completed: 1993
Location: Duisburg, Germany
Architect: Foster and Partners w/ Kaiser
Bautechnik
Cavity Partition: Multi-Story
Ventilation Type: Mechanical
Ventilation Mode: Indoor air curtain
Cavity Depth: 0.2 m (8 in)
Shading: Blinds, automated with override
Sources: Wigginton & Harris (2002, 125-128);
Compagno (2002, 120)
Figure A-6: Business Promotion Center - Overall view.
Source: Wigginton & Harris (2002, 125); ©Dennis Gilbert I VIEW.
This eight-story, 27 m (86 ft), lens-shaped building includes rentable office space and galleries.
The entire perimeter of the building is surrounded with a curved, full-height, multi-story double
skin facade. The exterior layer of glazing follows a 1.5 x 3.3 m (4.9 x 10.8 ft) module and is a
faceted representation of a 46 m ( 151 ft) radius curve comprised of 12 mm (1 /2 in) tempered
single glazing 200 mm (8 in) in front of the inner skin. The outer layer is point-fixed using a
Pilkington Planar™ glass bolt system tied back to aluminum mullions that are hung at roof level
and transfer lateral loads back at each floor slab. The vertical mullions are hung above from a
steel ring beam along the curvature of the building profile. The interior skin is a thermally broken
aluminum glazed IGU comprised of 6 mm (1/4 in) outer lite, 12 mm (1/2 in) argon-filled air space
and an 8 mm (5/16 in) laminated inner lite with a low-e coating. The inner glazing consists of
side-hung windows that are operable only for maintenance (Wigginton and Harris 2002, 127).
420
- I
--! '
_,,
-'----.e=~iif-~ -
,_
li~ 1 1
Figure A-7: Business Promotion Centre - Cavity drawing (left) and exterior photo (right).
Source (left): Wigginton & Harris (2002, 126); ©Foster+ Partners. Source (right): Compagno
(2002, 120).
Since the building is situated next to a busy roadway on its east elevation, natural ventilation was
deemed not feasible. Instead, air is supplied to the cavity by a displacement ventilation system
and rises as the result of the stack effect. Issues of overheating in the higher floors have been
reported (Compagno 2002, 121). Wthin the cavity space, 7% perforated aluminum louver blinds
are located closer to the inner skin. The computer-controlled blinds are integrated into a central
BMS that adjusts tilt angle based on heat and light sensors located in each room. Natural
daylighting is experienced throughout the building due to the high levels of transparency, shallow
depth floor plan and integration of the motorized blind controls with the BMS. The occupant has
manual control to override the blind position, but the BMS frequently checks and repositions for
optimal performance.
421
A.7 Bibliotheque nationale de France
Completed: 1995
Location: Paris, France
Architect: Dominique Perrault
Architecture
Cavity Partition: Box-Window
Ventilation Type: Mechanical
Ventilation Mode: Indoor air curtain
Cavity Depth: 90 mm (3.5 in)
Shading: Fins (interior wood panels)
Sources: Compagno (2002, 106-108)
Figure A-8: Bibliotheque nationale de France - Tower view.
Source: www.archdaily.com; ©Yuri Pal min.
Four 20-story, L-shaped towers facing inwards towards one another create a central courtyard on
the site. Each tower primarily consists of floors of book storage but also has support levels for
building services. The enclosure contains story-height box-window double-skin facade modules
that are pressurized with mechanical air. The pressurized air can be pre-heated or pre-cooled to
respond to the exterior temperatures accordingly and mitigates against condensation on the
interior side. Additionally, the humidity can be controlled in a manner that is best for conserving
the books stored within the towers. Both layers follow a 1.8 x 3.6 m (5.9 x 11.8 ft) facade module.
The outer skin is comprised of laminated lites, either of two 6 mm (1 /4 in) or 8 mm (5/16 in)
panes, structurally glazed with silicone to aluminum frames. The interior skin also uses a
laminated lite of two 10 mm (3/8 in) lites - additionally, this layer is fire-resistant. Placed 0.9 m
(3.0 ft) behind the inner glass lites are pivoting wooden panels to mitigate solar radiation.
422
A.8 Gotz Office Building
Completed: 1996
Location: Wi.irzburg, Germany
Architect: We bier & Geissler
Cavity Partition: Multi-Story
Ventilation Type: Natural
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 0.6 m (2 ft)
Shading: Blinds, automated with override
Sources: Hendriksen et al. (2000, 6); Administration
Building in Wurzburg (1997, 343-348); Herzog et al. (2004,
244-245); Wigginton & Harris (2002, 93-98)
Figure A-9: Gotz Office Building - Corner view with fans in the cavity.
Source: Administration Building in Wurzburg (1997, 347); ©Roland Halbe I Contur.
The two-story, 8.2 m (27 ft) tall building is located in the industrial outskirts of Wi.irzburg. A multi
story DSF is applied to all four building elevations. The outer skin is comprised of an 8 mm (5/16
in) tempered glass, 22 mm (7/8 in) gas-filled cavity, and 6 mm (1/4 in) low-e coated internal pane.
The internal glazing is thermally broken with an IGU of 6 mm (1/4 in) tempered glass, 16 mm (5/8
in) gas-filled airspace, and 6 mm (1/4 in) low-e coated internal pane. The inner skin also includes
sliding doors that permit natural ventilation, access for maintenance and escape routes. 10%
perforated louver blinds within the cavity space occur equidistant from the inner and outer skins
and are reversible with one surface reflective and the other absorptive. Air enters the cavity
through vent flaps at the DSF base. At the corner conditions, fans equilibrate solar gain between
the four faces. The BMS has an ability to learn and adapt to energy performance over time, but
occupants can prescribe the blinds' tilt and whether they are lowered or retracted.
423
,
n
'~
-~
T
l
¢
¢11
~
·~·
~
1 I
q-;
i
~
..
d
'
.
lJ
Figure A-10: Giitz Office Building - DSF section and detail.
Source: Adminisration Building in WUrzburg (1997, 346).
Figure A-11: Giitz Office Building - Entrance elevation.
Source: Ad min is ration Building in WUrzburg (1 997, 34 3); @Roland Hal be I c ontur.
424
A.9 Commerzbank
Completed: 1997
Location: Frankfurt, Germany
Architect: Foster+ Partners
Cavity Partition: Box-Window
Ventilation Type: Hybrid
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 20 cm (8 in)
Shading: Blinds, automated with override
Sources: Wigginton &Harris (2002, 59-64);
Compagno (2002, 142); Wood & Salib (2012, 32-41);
Bank Tower in Fankfurt-on-Main (1997, 349-354)
Figure A-12: Commerzbank - Overall view.
Source: Wood & Salib (2012, 32).
Commonly considered one of the most ecological tall buildings ever built (Wood and Sa lib 2012,
33), Commerzbank is a 56-story office building that is arranged with a soft-cornered triangular
plan. In cross-section, the building is arranged into 12-story villages, each environmentally and
socially centering around four-story sky gardens that act as a quasi-external space. The sky
garden contains natural ventilation and daylight that interior facing offices can access. The
double-skin facade consists of an inner facade with double-glazed, low-e coated units which are
side/bottom hung with motor-driven sashes (Wigginton and Harris 2002, 59). These operable
windows tilt inwards to a maximum of 15 degrees (see Figure 4.30) and have an insulated glass
makeup of 6 mm (1/4 in) inner lite, 14 mm (9/16 in) air space and 8 mm (5/16 in) outer lite. The
exterior skin is single 6 mm (1/4 in) toughened safety glass pane creating a 200 mm (8 in)
ventilated cavity space (Bank Tower in Fankfurt-on-Main 1997, 354) .
425
Figure A-13: Commerzbank - DSF airflow diagram (left) and interior photo (right). Source:
Wood & Salib (2012, 37-38), ©Foster+ Partners (left) and Ralph Richter/archenova (right).
The box-window cavity is one unit wide, one-story tall and houses grey-painted, motorized
aluminum blinds. The cavity is ventilated at the top and bottom through continuous slots 125 mm
(5 in) wide. The building has two operational modes, fully mechanical or naturally ventilated and
is controlled by a central BMS. In natural ventilation mode, the air conditioning is deactivated,
and motorized controls open the windows. The natural ventilation is single-sided. Occupants
have control of their interior and environment and can override the BMS system, except when
external wind loads are too high. When this occurs, windows oriented towards the prevailing
winds are automatically closed by the BMS. Airflow patterns around the building, through the
building, and through facade openings were simulated using CFO early in the design stage.
Thanks to Frankfurt's warm temperate climate, the building can operate in a natural ventilation
mode approximately 80% of the year, exceeding original design estimates.
426
A.10 RWE Headquarters
Completed: 1997
Location: Essen, Germany
Architect: lngenhoven Overdiek Kahlen and Partner
Cavity Partition: Box-Window
Ventilation Type: Hybrid
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 500 mm (20 in)
Shading: Blinds, automated with override
Sources: Herzog et al. (2004, 256-257); Compagno (2002, 138-140);
Wood & Salib (2012, 24-31); Schittich et al. (2007, 290-295);
Company Headquarters Tower in Essen (1997, 355-363)
Figure A-14: RWE Headquarters - Overall view.
Source: Schittich et al. (2007, 290); © Holger Knauf.
This project is marketed as the first contemporary high-rise tower to be naturally ventilated.
Every office has access to at least one operable window. The approximate percentage of the
year that natural ventilation can be used is 75% (Wood and Salib 2012, 24). Essen's mild climate
can largely be attributed to proximity to the coastline. The 31-story cylindrical tower is wrapped in
a double-skin facade made from prefabricated box-window modules. The exterior facade
consists of tempered 'extra clear' glass supported by drilled glass point supports occurring at four
locations along the each window's 3.5 m (11 .5 ft) height. The interior consists of floor-to-ceiling
sliding glass doors made with thermally insulated glazing. The 500 mm (20 in) cavity houses
centrally controlled perforated aluminum venetian blinds, and the interior space has translucent
textile anti-glare roller blinds along the inner-skin. Alternating sliding doors are operable but are
limited to 150 cm (6 in) opening for safety reasons, except for maintenance and cleaning.
427
Figure A-15: RWE Headquarters- Diagram and photo of box-window with Fish Mouth air
openings. Source: Wood & Salib (2012, 27-28). © lngenhoven Architects .
The 2.0 m (6.6 ft) wide box-window units follow an alternating pattern that staggers air supply
modules and air exhaust modules. These modules are separated from one another by a vertical
glass fin between units and at each floor level. The alternating pattern prevents recontamination
of exhaust air into an adjacent units supply air. The key feature that permits the type and rate of
airflow in each unit while providing acoustic insulation is the "Fish Mouth". There are air emit and
exhaust variations as well as two different opening sizes; one for higher wind zones in the upper
half of the building and one for the bottom. The design of the Fish Mouth and its openings was
determined using CFD and large scale wind-tunnel testing. The hybrid ventilation mode can
switch between mechanical and natural ventilation as needed, on a daily or seasonal basis.
Post-occupancy evaluation was conducted between 1998-2000 measuring room temperatures,
facade temperatures and air exchange rates.
428
A.11 Stadttor Dusseldorf (City Gate)
Completed: 1997
Location: Di.isseldorf, Germany
Architect: Petzinka Pink und Partner
Cavity Partition: Story-Height Juxtaposed Modules
Ventilation Type: Natural
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 0.9 m (3 ft) or 1 .4 m (4.6 ft)
Shading: Blinds, automated with override
Sources: Oesterle et al. (2001, 21-22); Lee et al. (2002, 99);
Wigginton & Harris (2002, 65-70); Bodart & Gratia (2003, 7);
Compagno (2002, 140-142); Herzog et al. (2004, 252-253)
Figure A-16: Stadttor Dusseldorf - Overall view.
Source: Compagno (2002, 140); © Gundelfingen.
This 19-story, 70 m (230 ft) rhomboidal building consists of two-office towers offset and bridged at
the top three stories. The entire building exterior is wrapped in glass. The use of a double-skin
envelope on three sides was implemented to reduce traffic noise and wind pressure on the
exposed site. The outer facade consists of 12 mm (1 /2 in) or 15 mm thick tempered single
glazing with a cavity behind it that is either 0.9 m (3 ft) or 1.4 m (4.6 ft). The outer glazing is
continuously captured along top and bottom edges and has an additional glass bolt point-fixing at
the railing. It is also low-iron to permit maximum transparency. 200 mm (8 in) behind the exterior
skin are venetian blinds that are lowered and raised automatically according to light, insolation
levels and the need for nighttime insulation (Wigginton & Harris 2002, 68). The interior skin is
made of 1.5 x 2.85 m (4.9 x 9.4 ft) modules of IGU's with a low-e coating in a wood frame.
Alternate bays pivot about their center to provide access to the cavity and for natural ventilation .
429
Figure A-17: Stadttor Dusseldorf - DSF elevation (left) and cavity (right).
Source: Compagno (2002, 141); © Gundelfingen (right).
The story-height DSF cavities are divided into 20 m (66 ft) lengths by an escape staircase, the
atrium or corners. The double-skin facade, in conjunction with other design elements, allows for
natural ventilation 60% of the year (Compagno 2002, 142). Natural ventilation into the cavity
space is through staggered air inlet and outlet vents that occur at the outer skin along the floor
levels. Alternating between the inlet and outlet vents is done to avoid short-circuiting effects. The
ventilation flaps can be 0%, 10% or 100% opened (Wigginton & Harris 2002, 67) and are
automatically closed when wind speed exceeds 9 m/s (20 mph). Within the cavity space,
aluminum louver blinds directly behind the outer skin (and in-line with the railing) provide solar
shading. Both the air inlet and exhaust flaps , as well as blinds within the cavity, are controlled by
the BMS. However, occupants do have the ability to override whether the blinds are up or down
and choose between one of three tilted positions (0°, 45° and 90°).
430
A.12 debis Headquarters
Completed: 1997
Location: Berlin, Germany
Architect: Renzo Piano Building Workshop
Cavity Partition: Multi-Story Louvered I Corridor
Ventilation Type: Natural
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 0.7 m (2.3 ft)
Shading: Blinds, automated with override
Sources: Lee et al. (2002, 97); Oesterle et al. (2001,
170-172); Wigginton & Harris (2002, 55-58); Compagno
(2002, 144-145)
Figure A-18: debis Headquarters - Overall View.
Source: Compagno (2002, 128); ©Michel Denance.
The 21-story high-rise is located in the center of Berlin, an area that experiences warm and
sometimes humid summers, as well as cold winters. The east, west and south elevations
incorporate a double-skin facade strategy, increasing the portion of the year that natural
ventilation could be leveraged while also improving the acoustical insulation. Though the outer
skin is always partially permeable, it does provide protection to shading devices located in the
cavity and helps to reduce wind pressures. This project was the first application of a frameless,
louvered, exterior skin on a double-skin facade to create a thermal buffer during the winter, but
pivot into a neutral position effectively adapting to a single skin with external shading protection.
The outer skin is composed of two 6 mm (1/4 in) glass panes laminated together to create louvers
attached to pivoting cast aluminum brackets. A 0.7 m (2.3 ft) cavity separates the outer layer
from the interior skin of side/bottom hung windows and hopper windows at the transom height.
431
Figure A-19: debis Headquarters - Exterior louvers photo (left) and diagrams (right).
Source: Compagno (2002, 144-145); ©Renzo Piano Building Workshop (diagrams).
Each story of the exterior skin is subdivided into eight louvers, seven of which can tilt to an open
position of 70°. A BMS controls the position of the exterior louvers to adjust the level of
ventilation into the cavity air space. When the louvers are closed a 10 mm (3/8 in) air gap
remains between, permitting small amounts of air to penetrate the outer skin and enter the airflow
cavity. The design permits natural ventilation for 40-55% of the year (Wigginton & Harris 2002,
57) and is achieved through manual operables and automated hopper windows on the inner skin.
To prevent smoke and fire spread, 10 mm (3/8 in) tempered glass covers gantries at each floor
level, also easing maintenance access. Located external to the inner facade are venetian blinds,
used for solar control. The louvers are powder coated a brick color. The BMS controls the blinds
position and tilt, but can be overridden by occupants. The building's inner-windows and external
louvers are automatically opened at night for cooling of thermal mass.
432
A.13 UCB Centre
Completed: 1998
Location: Brussels, Belgium
Architect: Assar
Cavity Partition: Box-Window
Ventilation Type: Mechanical
Ventilation Mode: Indoor air curtain
Cavity Depth: 143 mm (5.6 in)
Shading: Blinds, automated
Sources: Kragh (2001, 72-74 );
Loncour, et al. (2004, 37-40);
Poirazis (2006, 188-189)
Figure A-20: UCB Centre - Overall view.
Source: Loncour, et al. (2004, 38).
During the design phases of the project, traditional mechanical strategies were not permitting a
fully transparent facade, a desire of the ownership. To achieve maximum transparency and a
smooth glass exterior, a combination of a chilled ceiling and double-skin facade presented
opportunities. Additionally, the extra glazed layer provided improved acoustical insulation and
protection of solar controls located in the cavity space. The box-window double-skin acts as an
Active Wall; a term used by Permasteelisa to describe a system that is ventilated with room air.
The system is comprised of an external IGU, a 143 mm (5.6 in) deep cavity air space with
motorized blinds and an internal layer of single glazing. The cavity space is contained within the
1.5 m (4.9 ft) module width and contained between spandrel panels. The cavity airflow rate is 40
m
3
/h (23.5 CFM) (Kragh 2001, 73). Heating of the cavity is provided by the supply air while
cooling is provided by chilled ceilings. Air is dehumidified to combat potential condensation.
433
A.14 GSW Headquarters Tower
Completed: 1999
Location: Berlin, Germany
Architect: Sauerbruch Hutton
Cavity Partition: Multi-Story, Alternating Facade
Ventilation Type: Hybrid
Ventilation Mode: Air exhaust, air supply, outdoor air curtain
Cavity Depth: East- 1 m (3.3 ft); West - 200 mm (8 in)
Shading: Blinds, automated with override
Sources: Compagno (2002, 150-152);
Wood & Salib (2012, 64-73)
Figure A-21: GSW Headquarters Tower - Overall view.
Source: Wood & Salib (2012, 64).
In part to the large seasonal swings in Berlin, multiple DSF strategies were instituted for the 23-
story office tower to address the hot summer and potential for severely cold winters. The building
is a gentle curving arc in plan with a width ranging from 7 .2 to 11 m (23.6 to 36.1 ft). The building
relies on natural ventilation for approximately 70% of the year (Wood and Salib 2012, 64), utilizing
a cross-ventilation strategy within each floor, where the west facade acts as a thermal flue that
leverages the stack effect and exhausts air through the full-height, 67 m (220 ft), cavity space.
The west facade is comprised of a single-glazed outer lite and an IGU inner lite, resulting in a 1 m
(3.3 ft) enclosed cavity depth. Alternating operable and fixed windows, the inner skin provided
access to cavity airflow and natural ventilation. The east facade utilizes 200 mm (8 in) cavity box windows alternating with fixed louver screens that allow air to penetrate. The box-windows have
single glazing on the outer skin and IGU's on the inner, operable only for cleaning purposes.
434
Figure A-22: GSW Headquarters - West facade multi-story DSF configuration.
Source: http://www.flickr.com/photos/an_untrained_eye/1185704829/> (accessed July 6, 2012).
When outside temperatures fall outside the permissible range of comfort for natural ventilation, a
BMS initiates mechanical ventilation. The BMS is able to adjust the air inlet and exhaust
dampers at the top (an articulated Wing feature) and bottom of the west facade's exhaust flue.
Additionally, occupants have access to wall-mounted controls to adjust or override the ventilation
mode. The BMS also mitigates solar heat gain by controlling the vibrant (to the outside only)
blinds located in the west facades exhaust flue and venetian blinds on the east facade. The west
facade's shading devices are hues of red and pink, differentiating it in appearance from its
surrounding context. Once again, the BMS can be overridden by occupants to modify the blinds
and lighting in the office space. In addition to provide natural ventilation via cross-ventilation, the
double-skin facade strategies protect against heat loss, high wind loads and provide improved
acoustic insulation.
435
A.15 Deutsche Messe AG (Trade-Fair Tower)
Completed: 1999
Location: Hanover, Germany
Architect: Herzog + Partner
Cavity Partition: Story-Height Corridor
Ventilation Type: Hybrid
Ventilation Mode: Air supply
Cavity Depth: 1.4 m (4.6 ft)
Shading: Blinds
Sources: Wood & Salib (2012, 58-63); Schittich (2006, 158-163);
Administration Building in Hanover (2000, 397-405); Herzog et al.
(2004, 254-255);
Figure A-23: Trade-Fair Tower- Overall view.
Source: Administration Building in Hanover (2000, 397).
The 20-story (third tallest in Hanover), square building vvas designed for EXPO 2000 and serves
as the administrative offices for a venue that hosts numerous events annually along the river
Leine. Hanover is a temperate climate with cold winters and vvarm summers. Part of the design
strategy vvas to employ a DSF to control the positive and negative pressures around the building
to extract excess heat gain. This vvas achieved with two large story-height corridor facades; one
along the north and west, and one along the south and east. Where the facades are interrupted
by circulation cores, air ducts located in the ceilings link the elevations to create a horizontally
continuous 1.4 m (4.6 ft) corridor cavity encompassed by two double glazed skins. The outer skin
is an insulated glass unit with 8 mm (5/16 in) outboard lite, 16 mm (5/8 in) air, and 8 mm (5/16 in)
inner lite. The inner skin is wood and glass with a tempered inner lite, part of an IGU with 4 mm
(3/16 in) outboard lite, 16 mm (5/8 in) air, and 6 mm (1 /4 in) inner lite.
436
Figure A-24: Deutsche Messe AG (Trade-Fair Tower) - DSF story-height corridor cavity
(left) and interior office (right). Source: Wood & Salib (2012, 61-63). © MoritzKorn.
Air is exhausted from the corridor cavity through horizontal louvers at eight locations around each
level. The glass blade louvers are made up of a similar insulated makeup as the outer skin, and
are controlled by a central BMS that relies on real-time data from six temperature measurement
locations. The louver bays around the building were located based on wind-tunnel testing during
design. Air enters each interior office through sliding casement windows, made up of wood, on
the inner skin. When functioning in mechanical ventilation mode, air is supplied via air ducts
located beneath the inner facade's sill (see Figure 4.41). These provide ventilation when the
sliding operables are closed but are automatically deactivated when the windows are open. Solar
control is provided by adjustable sunshades installed behind the outer layer of glass as well as
behind the office interior skin. Firebreaks between floors are provided by extending the floor slab
beyond the interior wall and out to the exterior facade.
437
A.16 ARAG
Completed: 2000
Location: Di.isseldorf, Germany
Architect: RKW Architektur + Stadtebau,
in collaboration with Norman Foster
Cavity Partition: Shaft-Box
Ventilation Type: Natural
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 0.7 m (2.3 ft)
Shading: Blinds
Sources: Oesterle et al. (2001, 17-19); Compagno
(2002, 155-157)
Figure A-25: ARAG - Overall view.
Source: Compagno (2002, 156).
The 120 m (394 ft), 32-story tower is located in the temperate climate of Di.isseldorf in the lower
Rhineland, an area that experiences moderate winters and mild to warm summers. Sited near a
major traffic interchange with high noise levels, the tower implemented a double-skin facade
scheme to provide improved acoustical insulation and natural ventilation. The tower is subdivided
into four tiers, each eight-stories tall with a shaft-box configuration spanning seven floors before
terminating in front of a mechanical story. Air is drawn into the facade through 37 cm (14.6 in)
ventilation inlets that are located in front of each floor slab. The 0.7 m (2.3 ft) deep shaft axes
occur between box-windows that have bypass airflow openings that vary in size depending where
they occur along the shaft height. The exterior skin is an aluminum curtain wall construction with
12 mm (1/2 in) laminated glass and the interior skin consists of sliding operable IGU windows with
low-e glazing occurring between finished floor and dropped ceiling levels.
438
Figure A-26: ARAG -Air inlet (left) and elevation (right).
Source: Compagno (2002, 157); © RKVV Architek1ur + Stadtebau and Foster+ Partners.
The shaft-box DSF configuration is not as common as some other configurations. Its application
on ARAG tower required extensive testing in the design phase to address cavity airflow
dynamics. Between 1992 and 1994, larg~scale models (scale 1 :7) were tested to evaluate the
shaft-box DSF configuration and refine the aerodynamic design elements (Oesterle et al. 2001,
17). Early design simulations found the shaft-box capable of targeting natural ventilation for 50-
60% of the year. The ventilation rate within the cavity has means of control to avoid disturbing
rates of airflow to occupants as well as the aluminum venetian blinds that are located behind the
outer skin. Building controls are able to adjust facade flaps on the exterior skin to vary wind
speeds and control cavity airflow velocity. When wind speeds reach 8 m/s (18 mph) or greater,
the facade openings are sealed on the impinging face as a safety precaution (Oesterle et al.
2001'17)'
439
A.17 Administration Building in Kronberg
Completed: 2000
Location: Frankfurt/Main, Germany
Architect: Schneider + Schumacher
Cavity Partition: Box-Window
Ventilation Type: Natural
Ventilation Mode: Air supply, outdoor air curtain
Cavity Depth: 175 mm (6.9 in)
Shading: Blinds, automated with override
Sources: Schittich (2006, 150-157); Herzog et al. (2004,
250-251)
Figure A-27: Building in Kronberg - Overall view.
Source: Schittich (2006, 152). © Jorg Hempel.
The three-story administrative office building is laid out in a U-shaped plan. The facade uses a
box-window DSF configuration spanning story heights. The outer skin is 12 mm (1 /2 in)
tempered glass. A unique feature of the design is the outer lite opens vertically, pivoting about
one vertical edge, to provide natural ventilation and prevent overheating. This allows the facade
to transform from a smooth glazed surface to a sawtoothed scalar surface. The inner skin is an
IGU with a narrow magnetic ventilation flap made of aluminum sheet that is manually operated to
provide natural ventilation. The IGU is comprised of 6 mm (1/4 in) outer lite, 14 mm (9/16 in) air
space and 8 mm (5/16 in) laminated inner lite and is only opened for maintenance and cleaning
purposes. Solar shading is provided by perforated louver blinds located midway in the cavity
depth and are controlled by the BMS - as are the operable outer skin lites. Occupants can
override the outer skin operable and the blinds located within the cavity.
440
Figure A-28: Building in Kronberg - DSF closed (left) and open (right).
Source: Schittich (2006, 153). © Jorg Hempel.
! I
/ 1
""'
I
I
,
;;,_ rL .f .1......-k
a..... ~ i.. J.). l' ik.,,. ..+
Figure A-29: Kronberg - Naturally ventilated DSF concept diagram.
Source: Schittich (2006, 153). © Schneider+ Schumacher.
441
A.1'8 Aurora Piace
~ompleted: 2000
Loe3ti'on; Sydney. Australia
Architect Renzo Pianu Building· Workshop
Cavity Partitian :story-Height Corridor with Juxtaposed Modules
and Louvered Outer Skin
Vent1l:1tian Typ