Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Daylight prediction: an evaluation of daylighting simulation software for four cases
(USC Thesis Other)
Daylight prediction: an evaluation of daylighting simulation software for four cases
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
1
DAYLIGHT PREDICTION
AN EVALUATION OF DAYLIGHTING SIMULATION SOFTWARE FOR FOUR CASES
by
Sebanti Banerjee
B.Arch., Jadavpur University, Kolkata, India
A Thesis presented to
THE SCHOOL OF ARCHITECTURE
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
Master of Building Science
“THE CHASE L. LEAVITT GRADUATE BUILDING SCIENCE PROGRAM”
May 2015
Copyright 2015 Sebanti Banerjee
2
Sebanti Banerjee
sebantib@usc.edu
sebantib1989@gmail.com
Thesis Committee
Chair
Marc Schiler
marcs@usc.edu
School of Architecture
Member 2
Karen Kensek
kensek@usc.edu
School of Architecture
Member 3
Douglas Noble
dnoble@usc.edu
School of Architecture
3
Dedication
" Perhaps this could have stayed unstated.
Had our words turned to other things
In the grey park, the rain abated,
Life would have quickened other strings.
I list your gifts in this creation:
Pen, paper, ink and inspiration, "
~ Vikram Seth, An Equal Music
This work is first and foremost dedicated to my mother, father, sister, brother-in-law and my partner for their
support and guidance and secondly to everyone in the USC Master of Building Science Class of 2015 for being some
of the most inspiring and good people I have met and for creating a very positive environment to work in.
4
Acknowledgements
This work would not have been possible without the contributions of my Chair, Professor Marc Schiler, who was
always willing to share his extensive knowledge on the subject of the physics of lighting and offer valuable guidance;
Professor Karen Kensek, whose detailed scrutiny significantly improved the quality of the project; Professor Douglas
Noble for his guidance throughout the process of doing a thesis; and Professor Joon Ho Choi, who was extremely
generous with his camera and fish eye lens that was used for testing physical models.
I would also like to thank Doug Ross from the technical support team at Lighting Analysts, Inc. for his prompt and
comprehensive responses to all my doubts about AGi32; Kera Lagios, the developer and founder of DIVA for Rhino
for her help with understanding the technical aspects of the software and my fellow classmates for answering any
question I had, work or otherwise.
5
List of Figures
Chapter 1
Figure 1-1: Waldram diagram for CIE overcast sky and vertically glazed apertures, including corrections for glazing losses. As an
example a large window and an obstructing tower are indicated. Each square indicated in fine lines corresponds with a daylight
factor of 0.1% (Kota and Haberl 2007)………………………………………………………………………………………………………………………………………..14
Figure 1-2: a) Radiosity solution in Lightscape b) Radiosity+Raytracing solution from Lightscape. Notice the glare on the tables in
b (Inanici 2001)……………………………………………………………………………………………………………………………………………………………………………15
Figure 1-3: a) Radiosity solution in AGi32 b) Raytracing solution in AGi32. Notice the scallops on the wall due to the luminaires
being placed too close to the wall……………………………………………………………………………………………………………………………………………….15
Figure 1-4: Photometric summary of a ceiling recessed luminaire (Bega website 2014)……………………………………………………………..16
Figure 1-5: The ElectroMagnetic spectrum (Lillesand and Kiefer, 1994)………………………………………………………………………………………17
Figure 1-6: Absorption of light…………………………………………………………………………………………………………………………………………………….17
Figure 1-7: Specular and diffuse reflection………………………………………………………………………………………………………………………………….18
Figure 1-8: Refraction of light……………………………………………………………………………………………………………………………………………………..18
Figure 1-9: CRI (http://www.ledvista.ie/colour-rendering-index) ………………………………………………………………………………………….……20
Figure 1-10: Anatomy of the human eye (www.sciencesuperschool.com) ……………………………………………………………….………………..21
Figure 1-11: The range that the human eye can perceive at a given time (Adapted from Suk 2014) ……….………………………………...22
Figure 1-12: The resulting effect of visual adaptation (Baker and Steemers 2002) ………………………………….………………………………....22
Figure 1-13: Azimuth and Altitude………………………………….…………………………………………………………………………………………………………..23
Figure 1-14: Side Lighting (left) and Side Lighting enhanced with external or internal light shelves (right) …………………………………23
Figure 1-15: Toplighting strategies (Hastings 1994) ……….………………………………………………………………………………………………………...23
Figure 1-16: Aspect Ratio - a) calculation of top-lit zones b) calculation of side-lit zones (Lighting Control and Design) …………….24
Figure 1-17: Scene rendered in AGi32………………………………….……………………………………………………………………………………………………..25
Figure 1-18: Rendering and false color in Ecotect/Radiance….………………………………………………………………………………………………….…25
Figure 1-19: Rendering and false color in Rhino/DIVA….…………………………………………………………………………………………………………..…26
Figure 1-20: Connection of Rhino to Honeybee….………………………………………………………………………………………………………………..…..…26
Chapter 2
Figure 2-1: Raytracing Process (Wikipedia Images).….…………………………………………………………………………………………………………………29
Figure 2-2: Radiosity Process ((Ochoa, Aries and Hensen 2010) ….……………………………………………………………………………………………..29
Figure 2-3: Examples of Quantitative (left) and Qualitative output (right) ….…………………………………………………………………….………..32
Figure 2-4: Summary Chart.….………………………………………………………………………………………………………………………………………………….…33
Figure 2-5: A brief history of LEED simulation-based daylight credit compliance (Jakubiec 2014) ….…………………………………………..35
Figure 2-6: sDA Calculation Logic (Jakubiec 2014) ….…………………………………………………………………….………………………………………..…..37
Figure 2-7: Graphical Representation of sDA (Jakubiec 2014) ….…………………………………………………………………….……………….……..….38
Figure 2-8: Calculation Logic of ASE (Jakubiec 2014) ….………………………………………………………………………………….……………….……..…..39
Figure 2-9: Comparative charts used in study (Panitz, Garcia-Hansen 2013) …………………………………………….……………….…………..…..40
Chapter 3
Figure 3-1: Workflow………………………………………………………………………………………………………………………………………………………………….41
Figure 3-2: Specular reflection test model layout plan and 3D view……………………………………………………………………………………………42
Figure 3-3: Solar Gnomon for Southern California (Brown 1985)………………………………………………………………………………………………..42
Figure 3-4: Test 1: A open……………………………………………………………………………………………………………………………………………………………43
Figure 3-5: Test 2: B open……………………………………………………………………………………………………………………………………………………………43
Figure 3-6: Test 3: C open……………………………………………………………………………………………………………………………………………………………43
Figure 3-7: Physical model for specularity test…………………………………………………………………………………………………………………………….44
Figure 3-8: Gardco Statistical Novogloss…………………………………..…………………………………………………………………………………………………44
Figure 3-9: Images captured at 4 different exposures using a fish eye lens…………………………………………………………………………………45
Figure 3-10: Example of an image generated by photography from physical model using a fish-eye lens and Photolux software.45
Figure 3-11:Reflection test model layout plan and 3D view………………………………..………………………………………………………………………46
6
Figure 3-12: Physical model for 100% reflection test………………………………………………………………………………………………………………….47
Figure 3-13:Luminous flux conservation test model layout plan and 3D view…………………………………………………………………………….48
Chapter 4
Figure 4-1: Fish eye images taken at 1/2, 1/15, 1/125 and 1/1000 (from left to right) exposure settings for Test1…………………….52
Figure 4-2: False color image created in Photolux software showing luminance for Test1………………………………………………………...52
Figure 4-3:Fish eye images taken at 1/2, 1/15, 1/125 and 1/1000 (from left to right) exposure settings for Test2………………………53
Figure 4-4: False color created in Photolux software showing luminances for Test2……………………………………………………………………53
Figure 4-5:Fish eye images taken at 1/2, 1/15, 1/125 and 1/1000 (from left to right) exposure settings for Test3……………………..53
Figure 4-6:False Color created in Photolux software showing luminances for Test3……………………………………………………………………54
Figure 4-7: Renderings in AGi32 using radiosity (left) and raytracing (right) for Test1…………………………………………………………………55
Figure 4-8: Illuminance false color (left) and luminance false color (right) for Test1…………………………………………………………………...55
Figure 4-9: Calculation points on north wall (left) and floor (right) for Test1…………………………………………………………………………….…55
Figure 4-10: Renderings in AGi32 using radiosity (left) and raytracing (right) for Test2……………………………………………………………….56
Figure 4-11: Illuminance false color (left) and luminance false color (right)for Test2…………………………………………………………………..56
Figure 4-12: Calculation points on north wall (left) and floor (right) for Test2…………………………………………………………………………....56
Figure 4-13: Renderings in AGi32 using radiosity (left) and raytracing (right) for Test3………………………………………………………………57
Figure 4-14: Illuminance false color (left) and luminance false color (right) for Test3………………………………………………………………...57
Figure 4-15: Calculation points on north wall (left) and floor (right) for Test3…………………………………………………………………………...57
Figure 4-16: Test 1 renderings in Ecotect/Radiance - illuminance (left) and luminance (right)…………………………………………………..58
Figure 4-17: Test 1 false color diagrams in Ecotect/Radiance - illuminance (left) and luminance (right)…………………………………….59
Figure 4-18: Test 2 renderings in Ecotect/Radiance - illuminance (left) and luminance (right)…………………………………………………..59
Figure 4-19: Test 2 false color diagrams in Ecotect/Radiance - illuminance (left) and luminance (right)…………………………………….60
Figure 4-20: Test 3 renderings in Ecotect/Radiance - illuminance (left) and luminance (right)……………………………………………………61
Figure 4-21: Test 3 false color diagrams in Ecotect/Radiance - illuminance (left) and luminance (right)…………………………………….61
Figure 4-22: Test 1 luminance rendering in Rhino/DIVA……………………………………………………………………………………………………………..62
Figure 4-23: Test 1 luminance false color rendering in Rhino/DIVA…………………………………………………………………………………………….62
Figure 4-24: Test 2 luminance rendering in Rhino/DIVA……………………………………………………………………………………………………………..63
Figure 4-25: Test 2 luminance false color rendering in Rhino/DIVA…………………………………………………………………………………………….63
Figure 4-26: Test 3 luminance rendering in Rhino/DIVA……………………………………………………………………………………………………………..64
Figure 4-27: Test 3 luminance false color rendering in Rhino/DIVA…………………………………………………………………………………………….64
Figure 4-28: Test 1 renderings in Honeybee - illuminance (left) and luminance (right)……………………………………………………………….65
Figure 4-29: Test 1 false color renderings in Honeybee - illuminance (left) and luminance (right)………………………………………………65
Figure 4-30: Test 2 renderings in Honeybee - illuminance (left) and luminance (right)……………………………………………………………….66
Figure 4-31: Test 2 false color renderings in Honeybee - illuminance (left) and luminance (right)………………………………………………66
Figure 4-32: Test 3 renderings in Honeybee - illuminance (left) and luminance (right)……………………………………………………………….67
Figure 4-33: Test 2 false color renderings in Honeybee - illuminance (left) and luminance (right)……………………………………………..67
Figure 4-34: Illuminance values in footcandles recorded for Test1 in the four software packages………………………………………………68
Figure 4-35: Illuminance values in footcandles recorded for Test2 in the four software packages……………………………………………..68
Figure 4-36: Illuminance values in footcandles recorded for Test 3 in the four software packages……………………………………………..69
Figure 4-37: Luminance values in cd/m
2
recorded for Test 1 in the physical model and four software packages………………………..69
Figure 4-38: Luminance values in cd/m
2
recorded for Test 2 in the physical model and four software packages………………………..70
Figure 4-39: Luminance values in cd/m
2
recorded for Test 3 in the physical model and four software packages………………………..70
Figure 4-40: Fish eye images taken at 1/2, 1/15, 1/125 and 1/1000 (from left to right) exposure settings………………………………….71
Figure 4-41: False Color created in Photolux software showing luminances……………………………………………………………………………….71
Figure 4-42: Fish eye images taken at 1/2, 1/15, 1/125 and 1/1000 (from left to right) exposure settings………………………………….72
Figure 4-43: False Color created in Photolux software showing luminances……………………………………………………………………………….72
Figure 4-44: Illuminance levels on the north wall (left) and exitance levels on the north wall (right)…………………………..……………..73
Figure 4-45: Radiosity rendering and luminance false color in AGi32 for a 100% reflective wall test…………………………………………..73
Figure 4-46: Illuminance levels on the north wall (left) and exitance levels on the north wall (right)…………………………………..……..74
Figure 4-47: Radiosity rendering and luminance false color in AGi32 for a 50% reflective wall test…………………………………………….74
Figure 4-48: Comparison between Illuminance and Luminance distributions for a 100% reflective surface in Ecotect/Radiance..76
Figure 4-49: Comparison between Illuminance and Luminance distributions for a 50% reflective surface in Ecotect/Radiance….77
7
Figure 4-50: Illuminance Plot and Luminance rendering in Rhino/DIVA for 100% reflective surface…………………………………………..78
Figure 4-51: Illuminance Plot and Luminance render in Rhino/DIVA for 50% reflective surface……………………………………………….…78
Figure 4-52: Comparison between Illuminance and Luminance distributions for a 100% reflective surface in
Rhino/Grasshopper/Honeybee…………………………………………………………………………………………………………………………………………………..79
Figure 4-53: Comparison between Illuminance and Luminance distributions for a 50% reflective surface in
Rhino/Grasshopper/Honeybee…………………………………………………………………………………………………………………………………………………..80
Figure 4-54: Percentage of deviation for all the software packages for the reflectivity tests……………………………………………………….81
Figure 4-55: Graph showing percentage of deviation in AGi32 for all the tests……………………………………………………………………………85
Figure 4-56: Graph showing percentage of deviation in Ecotect/Radiance for all the tests………………………………………………………...89
Figure 4-57: Graph showing percentage of deviation in Rhino/DIVA for all the tests…………………………………………………………………..93
Figure 4-58: Graph showing percentage of deviation in Rhino/Grasshopper/Honeybee for all the tests…………………………………….97
Figure 4-64: Percentage of deviation shown by the software packages with increasing size of opening……………………………………..97
Chapter 5
Figure 5-1: From left to right - a)White foam core (4.6), b)Black foam core (1.0), c)Grey paper (2.7), d) Stainless Steel plate
(301.1), e) Double sided gold plastic (546.0), f) Silver plastic sheet (722.1), g) Aluminum sheet (544.6)……………………………………..98
Figure 5-2: Physical model for the specular reflection test………………………………………………………………………………………………………….99
Figure 5-3: Test model setup in Agi32……………………………………………………………………………………………………………………………………….100
Figure 5-4: Test model setup with calculation points………………………………………………………………………………………………………………..100
Figure 5-5: Defining the planar object in the surface edit window…………………………………………………………………………………………….101
Figure 5-6: Daylight option in the toolbar………………………………………………………………………………………………………………………………….101
Figure 5-7: The daylighting parameters window……………………………………………………………………………………………………………………….102
Figure 5-8: Run 1 - No light in the interior…………………………………………………………………………………………………………………………………102
Figure 5-9: Run 2 - Hazy spot at the center of north wall……………………………………………………………………………………………………….….103
Figure 5-10: Hotspot at the center of the north wall (left) and floor (right)…………………………………………………………………………….…103
Figure 5-11: Flipping the surface normal of the planar object to direct light towards the interior………………………………………….…104
Figure 5-12: Run 3 - Spot missing on the floor………………………………………………………………………………………………………………………..…104
Figure 5-13: Run 3 - North wall exitance values…………………………………………………………………………………………………………………………105
Figure 5-14: Run 3 - Floor illuminance values……………………………………………………………………………………………………………………………105
Figure 5-15: Raytrace with high reflectivity and specularity (left) and raytrace with low reflectivity and high specularity
(right)………………………………………………………………………………………………………………………………………………………………………………………..106
Figure 5-16: Rendering with no adaptive subdivision and default mesh levels (left) and rendering with high adaptive subdivision
and increased mesh levels (right) (Lighting Analysts, Inc. 2015)………………………………………………………………………………..………………106
Figure 5-17: Model setup for Test1 in Ecotect/Radiance………………………………………………………………………………………………………..…107
Figure 5-18:Radiance options selection panel………………………………………………………………………………………………………………………..…109
Figure 5-19:Radiance options selection panel………………………………………………………………………………………………………………………..…109
Figure 5-20: Run 1 for Test 1,2,3 (from left to right) in Ecotect/Radiance……………………………………………………………………………….…110
Figure 5-21: Run2 for Test 1,2,3 (from left to right) in Ecotect/Radiance, illuminance maps…………………………………………………..…111
Figure 5-22:Run2 for Test 1,2,3 (from left to right) in Ecotect/Radiance, luminance maps……………………………………………………..…111
Figure 5-23:Radiance options selection panel……………………………………………………………………………………………………………………………112
Figure 5-24: Radiance Control Panel……………………………………………………………………………………………………………………………………….…112
Figure 5-25: Model setup for Test1 in Rhino…………………………………………………………………………………………………………………………..…113
Figure 5-26: Workflow in DIVA………………………………………………………………………………………………………………………………………………….114
Figure 5-27: The materials tab in Rhino/DIVA……………………………………………………………………………………………………………………………114
Figure 5-28: Simulation options tab in DIVA………………………………………………………………………………………………………………………………115
Figure 5-29: Test1 Run1 in Rhino/DIVA……………………………………………………………………………………………………………………………..………115
Figure 5-30: Sky type parameters in AGi32 (top) and DIVA (bottom) ……………………………………………………………………………………….116
Figure 5-31: Test1 Luminance distributions in Ecotect/Radiance (left) and DIVA (right) ……………………………………………………………117
Figure 5-32: Illuminance plots for Test1 on floor (left) and north wall (right) ………………………………………………………………….…….…117
Figure 5-33: Model setup for Test1 in Rhino…………………………………………………………………………………………………………..…………………118
Figure 5-34: Workflow in Honeybee…………………………………………………………………………………………..…………………………..…………………119
Figure 5-35: Defining surface-wise geometry and materials in Honeybee…………………………………………………………………………………119
Figure 5-36: Adding sky…………………………………………………………………………………………..…………………………..……………………………………120
Figure 5-37: Image based analysis recipe in Honeybee…………………………………………………………………………………………..…………………120
8
Figure 5-38: Running the analysis…………………………………………………………………………………………..…………………………………………………121
Figure 5-39: EmbryoViz module for visualizing images in the Grasshopper canvas……………………………………………………………………122
Figure 5-40: Module for creating a false color image in Honeybee……………………………………………………………………………………………122
Figure 5-41: Test1 run1 in Honeybee…………………………………………………………………………………………..……………………………………………123
Figure 5-42: Increasing the ambient bounce (top) and the resultant visualization of Test1 run2 in Honeybee………………………….124
Figure 5-43: With ab set to 1, the path of light from the point on the table would have terminated at the first contact with the
horizontal blind. Raising it ensures that it bounces and reached the source of light (Reinhart 2010)……………………………………..…124
Figure 5-44: Physical model for the 100% reflection test…………………………………………………………………………………………………………..125
Figure 5-45: Model setup in AGi32……………………………………………………………………………………………………………………………………………126
Figure 5-46: Model setup in Ecotect/Radiance………………………………………………………………………………………………………………………….128
Figure 5-47: Model setup in Rhino…………………………………………………………………………………………………………………………………………….129
Figure 5-48: WXFalseColor window…………………………………………………………………………………………………………………………………….…….129
Figure 5-49: Model setup in Rhino…………………………………………………………………………………………………………………………………………….130
Figure 5-50: Grid based simulation module in Honeybee………………………………………………………………………………………………………….131
Figure 5-51: Setting up the visualization of a grid based analysis in Honeybee (left) and some ways of visualization (right)…..…131
Figure 5-52: Model setup for Test1 in AGi32…………………………………………………………………………………………………………………………….132
Figure 5-53: Setting calculation points on all surfaces (left) and on the panel (right)………………………………………………………………..132
Figure 5-54: Illuminance levels on panel facing inside (left) and facing outside (right)………………………………………………………………132
Figure 5-55:Model setup in Ecotect/Radiance for Test 1…………………………………………………………………………………………………………..133
Figure 5-56: Imported sensor point values on a wall for Test 1…………………………………………………………………………………………………133
Figure 5-57:Data Chart created for each test in Excel………………………………………………………………………………………………………………..134
Figure 5-58: Test1 setup in Rhino………………………………………………………………………………………………………………………………………………134
Figure 5-59: Illuminance values on some surfaces for Test1 as calculated in DIVA…………………………………………………………….………134
Figure 5-60: Test1 setup in Rhino……………………………………………………………………………………………………………………………………………..135
Figure 5-61: Illuminance value text file for the outer surface of the aperture……………………………………………………………………………135
9
List of Tables
Chapter 1
Table 1-1:Radiometry unit chart (Palmer 1999)………………………………………………………………………………………………………………………….19
Table 1-2:Photometry unit chart (Schiler 1992) …………………………………………………………………………………………………………………………19
Chapter 2
Table 2-1:Non-exhaustive summary of analysis tools and their properties……………………………………………………………….…………………30
Table 2-2: Output in lighting simulation (Ochoa, Aries and Hensen 2012)..…………………………………………………………………………………32
Chapter 3
Table 3-1: Surface properties (Namburi 2006); Notes: n/a - Not Applicable, n/r - Not required………………………………………………….44
Table 3-2: Surface properties for a 100% reflective north wall (Namburi 2006); Notes: n/a - Not Applicable, n/r - Not required..47
Table 3-3: Surface properties for a 50% reflective north wall (Namburi 2006); Notes: n/a - Not Applicable, n/r - Not required….47
Table 3-4: Analysis Software Interoperability Observations………………………………………………………………………………………………………..49
Table 3-5: Analysis Software Modeling Observations………………………………………………………………………………………………………………….49
Table 3-6: Analysis Software Metrics Observations…………………………………………………………………………………………………………………….49
Table 3-7: Comparative chart for specular reflection test………………………………………………………….………………………………………………..50
Table 3-8: Comparative chart for reflection test………………………………………………………………………………………………………………………….50
Table 3-9: Comparative chart for luminous flux transmission test………………………………………………………………………………………………50
Chapter 5
Table 5-1: Surface properties Notes: n/a - Not Applicable, n/r - Not required…………………………………………………………………………….99
Table 5-2: Surface properties in AGi32 Notes: n/a - Not Applicable, n/r - Not required…………………………………………………………….100
Table 5-3: Surface Properties in Ecotect/Radiance Notes: n/a - Not Applicable, n/r - Not required……………………………………..……107
Table 5-4: Surface Properties in Rhino/DIVA Notes: n/a - Not Applicable, n/r - Not required……………………………………………………113
Table 5-5: Surface Properties in Rhino/Honeybee Notes: n/a - Not Applicable, n/r - Not required……………………………………………118
Table 5-6: Surface Properties for 100% reflection test; Notes: n/a - Not Applicable, n/r - Not required……………………………………125
Table 5-7: Surface Properties for 50% reflection test; Notes: n/a - Not Applicable, n/r - Not required……………………………………..125
Table 5-8: Surface Properties for 100% reflection test in AGi32; Notes: n/a - Not Applicable, n/r - Not required………………………126
Table 5-9: Surface Properties for 50% reflection test in AGi32; Notes: n/a - Not Applicable, n/r - Not required………………………..126
Table 5-10: Surface Properties in Ecotect/Radiance for 100% reflection test; Notes: n/a - Not Applicable, n/r - Not required….127
Table 5-11: Surface Properties in Ecotect/Radiance for 50% reflection test; Notes: n/a - Not Applicable, n/r - Not required……127
Table 5-12: Surface Properties in Rhino/DIVA for 100% reflection test; Notes: n/a - Not Applicable, n/r - Not required…………..128
Table 5-12: Surface Properties in Rhino/DIVA for 100% reflection test; Notes: n/a - Not Applicable, n/r - Not required…………..129
Table 5-14: Surface Properties in Rhino/Honeybee for 100% reflection test; Notes: n/a - Not Applicable, n/r - Not required…..130
Table 5-15: Surface Properties in Rhino/Honeybee for 50% reflection test; Notes: n/a - Not Applicable, n/r - Not required…….130
Chapter 6
Table 6-1: Comparative chart for specular reflection test………………………………………………………………………………………………………….136
Table 6-2: Comparative chart for reflection test……………………………………………………………………………………………………………………….136
Table 6-3: Comparative chart for luminous flux transmission test…………………………………………………………………………………………….137
Table 6-4: Analysis Software Interoperability Observations………………………………………………………………………………………………………137
Table 6-5: Analysis Software Modeling Observations………………………………………………………………………………………………………………..138
Table 6-6: Analysis Software Metrics Observations…………………………………………………………………………………………………………………..139
10
Abstract
The use of daylighting to supplement and offset artificial lighting requirements in interior spaces is an important
step in achieving California’s current and future energy goals. The analysis of daylighting is not only difficult for
lighting designers, but also the results are not always accurate. Four popular lighting analysis software programs
(AGi32, Ecotect/Radiance, Rhino/Diva, and Rhino/Grasshopper/Honeybee) failed to provide results that matched
the data gathered in physical models and from calculations based on the laws of physics for the Specular Reflection
Test, Reflectance Test (Namburi 2006) and Luminous Flux Conservation Test (CIE 171:2006).The results recorded for
most of the simulations have shown more than 20% variation (the threshold for validation tests according to CIE),
between the software packages and real data.
Additional parameters, such as the clarity of the Graphical User Interface (GUI), time and training required for full
use, quality of scene renderings, ability to calculate climate-based metrics and code compliance, and interoperability
with other software have been compared across the same software programs and the results tabulated. The
tabulation also lists the critical nuances for daylighting simulation in each of the four software programs. This
together with a narrative describing the change of metrics used in digital daylight simulation will help designers
analyze their buildings more accurately.
Hypothesis: AGi32, Ecotect/Radiance, Rhino/Diva, and Rhino/Grasshopper/Honeybee cannot consistently and
accurately simulate the physical properties of daylight for Specular Reflection, Reflection, and Luminous Flux
Transmission Test.
Note: All references in this work are included at the end in alphabetical order by author last name and year.
11
Table of Contents
SERIAL
NO.
NAME PAGE NO.
Dedication 3
Acknowledgements 4
List of figures 5
List of tables 9
Abstract 10
1 An introduction to daylight design and analysis 13
1.0 Shortcomings in daylight analysis by current software 13
1.1 Importance of finding the right tool for daylight analysis 14
1.1.1 Issues and concerns faced in daylight analysis 14
1.1.2 Evidence of issues in daylight analysis in the market 15
1.1.3 Previous studies conducted on daylight analysis 15
1.1.4 Who benefits from a comprehensive identification of the issues? 16
1.2 Important terms and concepts in the analysis of daylighting 16
1.3 Definition of the scope of this project: What is not included? 26
1.4 Conclusion: A comparative evaluation chart of different analysis software 27
2 Daylight simulation in software 28
2.0 A state-of-the-art narrative on daylight simulation and analysis in software 28
2.1 Pathological cases for evaluation of lighting simulation tools 33
2.2 Discussion of test cases and parameters while evaluating software 34
2.3 A narrative on development of daylight analysis metrics 34
2.4 Software testing 39
2.5 Conclusion 40
3 The methodology of software validation and comparison 41
3.0 Inference from background research 41
3.1 Selection of pathological cases 41
3.1.0 Specular reflection test 42
3.1.1 Reflection test 46
3.1.2 Luminous flux conservation test 48
3.2 Selection of additional parameters and comparative charts 49
3.3 Conclusion 50
4 Simulation results of daylight analysis tests 51
4.0 Executive summary of methodology 51
4.1 Specular reflection test 51
4.1.1 Results of physical model test 51
4.1.2 Results of simulation in AGi32 54
4.1.3 Results of simulation in Ecotect/Radiance 58
4.1.4 Results of simulation in Rhino/Diva 61
4.1.5 Results of simulation in Rhino/Grasshopper/Honeybee 65
4.1.6 Conclusion of specular reflection test 67
4.2 Reflection test 71
4.2.1 Results of physical model test 71
4.2.2 Results of simulation in AGi32 72
12
4.2.3 Results of simulation in Ecotect/Radiance 74
4.2.4 Results of simulation in Rhino/Diva 77
4.2.5 Results of simulation in Rhino/Grasshopper/Honeybee 78
4.2.6 Conclusion of reflection test 80
4.3 Luminous flux conversion test 81
4.3.1 Results of simulation in AGi32 81
4.3.2 Results of simulation in Ecotect/Radiance 85
4.3.3 Results of simulation in Rhino/Diva 89
4.3.4 Results of simulation in Rhino/Grasshopper/Honeybee 93
4.3.5 Conclusion of luminous flux conservation test 97
5 Evaluation and critique of simulation results 98
5.0 Executive summary of the tests 98
5.1 Specular reflection test 98
5.1.1 Discussion of physical model test 98
5.1.2 Discussion of simulation in AGi32 99
5.1.3 Discussion of simulation in Ecotect/Radiance 106
5.1.4 Discussion of simulation in Rhino/Diva 113
5.1.5 Discussion of simulation in Rhino/Grasshopper/Honeybee 117
5.2 Reflection test 124
5.2.1 Discussion of physical model test 124
5.2.2 Discussion of simulation in AGi32 125
5.2.3 Discussion of simulation in Ecotect/Radiance 127
5.2.4 Discussion of simulation in Rhino/Diva 128
5.2.5 Discussion of simulation in Rhino/Grasshopper/Honeybee 130
5.3 Luminous flux conversion test 131
5.3.1 Discussion of simulation in AGi32 131
5.3.2 Discussion of simulation in Ecotect/Radiance 132
5.3.3 Discussion of simulation in Rhino/Diva 134
5.3.4 Discussion of simulation in Rhino/Grasshopper/Honeybee 135
6 Conclusion: Comparative charts 136
6.0 Conclusions 136
6.1 Comparative charts for tests 136
6.2 Comparative charts for additional parameters 137
6.3 Future work in the field of analysis of daylighting in software 139
6.4 Conclusion 140
References 141
Bibliography 145
13
Chapter 1
An introduction to daylight design and analysis
1.0 Shortcomings in daylight analysis by current software
The incorporation of daylighting in architectural design was always considered by designers. For example, Frank
Lloyd Wright, often sized the wings of his buildings based on the double loaded corridor width to admit daylight
(Wright and Lipman 1986). With the advent of fluorescent lights, designers found a way to substitute natural
lighting with artificial lighting for tasks. This gradually led to an unnecessary use of electricity and hence, energy.
Recently, designers have realized that daylighting strategies should have greater priority when designing
buildings. As a result, several software programs which traditionally analyzed artificial lighting in spaces have
started developing daylight analysis modules.
Traditionally, design methods like the sunpath diagram, the CIE method, the IESNA lumen method, developed
in 1928 (Kota and Haberl 2007), the Waldram diagram, dating back to 1923 (Geebelen 2015) (Fig. 1-1), and BRE
protractors were used to calculate daylight harvesting. Physical models were built and measured. Nowadays,
due to the increasing complexity of design, computer simulation software is used. However, lighting simulation
is challenging as it is difficult to accurately represent the physical laws of light (Ochoa, Aries and Hensen 2010).
The prediction of light levels in a space during the schematic design phase is very important. The lighting
designer, however, is faced with questions such as which software to use, what are the limitations, how much
time is required, what inputs are required and how accurate are the outputs. Several daylighting analysis
software programs exist in the market today, all used for different purposes, provided that a set of conditions
are met. As such, knowledge of the conditions required, applications served and shortcomings in these software
programs will prove handy. This information has been laid out to serve as a detailed guide for students and
professionals of lighting design.
14
Figure 2-1: Waldram diagram for CIE overcast sky and vertically glazed apertures, including corrections for glazing losses. As an example a
large window and an obstructing tower are indicated. Each square indicated in fine lines corresponds with a daylight factor of 0.1% (Kota
and Haberl 2007)
1.1 Importance of finding the right tool for daylight analysis
The necessity of finding the right tool for daylight analysis is important to lighting designers, energy analysts, and
software developers who regularly face problems with the existing daylight analysis software. The aim is to not
only find the bugs in each software, but to find out what each software does well. It may very well be that a test
(for specular reflections, for instance) which is not done well in one software is done very well in another.
Knowing the features and weakness of each tool will help people choose the correct software applicable for
their particular use.
1.1.1 Issues and concerns faced in daylight analysis
The usual best response to questions faced during daylight analysis in a software is for the designer to
approximate and make assumptions (for surface reflectances, for instance) and reach a conclusion. Various
studies like the Post occupancy evaluation of daylight in buildings (Hygge, Staffan, Löfberg 1999), have shown
that design decisions reached as a result of computer simulation often fail in reality since many issues like glare,
overheating, etc. are not considered or checked. There is a gap between simulation results and real data and a
need exists to raise awareness about this gap.
15
1.1.2 Evidence of issues in daylight analysis in the market
An example of such discrepancies in software can be seen when using Ecotect along with the Radiance plugin for
daylighting analysis. Typically, in daylighting analysis of an architectural space, three days of the year are chosen:
March 21, September 22, and December 21 at three different times of day, often 9:00 AM, 12:00 PM, and 4:00
PM. Ecotect has a built-in option of incorporating Daylight Savings Time (DST) when simulating the condition for
the month of March, but this information is not translated to the Radiance plugin. As such, the results of the
simulation are slightly different from real life. This discrepancy is not noticed by many people. Another example
is using raytracing versus the radiosity method of calculation in certain software programs. Raytracing can
calculate specular reflection while radiosity cannot. If a scene is rendered using only radiosity, glare is not
apparent (Inanici 2001) (Fig. 1-2 and 1-3).
Figure 1-2: a) Radiosity solution in Lightscape b) Radiosity+Raytracing solution from Lightscape. Notice the glare on the tables in b (Inanici
2001)
Figure 1-3: a) Radiosity solution in AGi32 b) Raytracing solution in AGi32. Notice the scallops on the wall due to the luminaires being
placed too close to the wall.
1.1.3 Previous studies conducted on daylight analysis
Gradually the term daylighting (letting light into the building) came to be replaced by daylight harvesting (using
daylighting to the extent that electric lights can be turned off) (Steffy 2008) (Schiler 1995). As awareness
16
increased and greater stress was given on sustainable design, daylight harvesting and artificial lighting analysis
became more important in the schematic design phase. As far as artificial lighting was concerned, lighting fixture
manufacturers started developing detailed photometric files having different properties like lumen output,
beam spread, wattage, lamp lumen depreciation, inclination and tilt, which made the measurement of
illuminance levels fairly straightforward (Fig. 1-4).
Figure 1-4: Photometric summary of a ceiling recessed luminaire (Bega website 2014)
The analysis of daylighting in computer software programs, however, proved to have many complications. The
paper Comparative Evaluation of Four Daylighting Software Programs demonstrated that none of the software
existing at the time was capable of predicting the simplest of daylighting designs (Ubbelohde et al. 1988). The
software programs since then have increased in number as well as in sophistication.
1.1.4 Who benefits from a comprehensive identification of the issues?
An identification of the issues plaguing the analysis of daylighting in various software programs will have
significant impacts on the lighting design community. Most projects are undertaken in offices using one software
that the people in the office are familiar with. Many times, the fact that the results are not accurate go
unnoticed until a problem crops up in a future project when the same method is used or if commissioning is
done (Hygge, Staffan, Löfberg 1999). The chances of either of these two conditions happening are minimal and
as such, a gap exists that people may not even notice but which can have significant impacts on project
outcomes nonetheless.
1.2 Important terms and concepts in the analysis of daylighting
Introduction
Different individuals will have different responses when asked to explain the term “light” or, more to the point,
"daylight." One definition is that light is an electromagnetic radiation that is emitted by the sun that reaches the
earth after travelling through space, is reflected off objects, hits the eye, and cause it to stimulate the sense of
vision (Reinhart 2014) (Fig. 1-5). Architects, interior designers and lighting designers tend to focus more on the
17
qualitative aspects of light and how it can used to render a space. An engineer, on the other hand, may define it
as the minimum number of lux or footcandles required by a broad spectrum of people for a particular task in a
habitable space. All of the aspects of light have to be considered to effectively understand the use of daylight in
a space.
Figure 1-5:The ElectroMagnetic spectrum (Lillesand and Kiefer, 1994)
Light has a dual character, that of lateral waves (composed of cross axis waves at 90 degree angles to each other)
and particles. The wave characteristics of light are responsible for the processes of interference, diffraction, and
polarization whereas the particle characteristics of light are responsible for the Compton effect and Photo-electric
emission (Namburi 2006). This complex nature of light is often difficult to simulate in software programs.
Some of the properties of light are as follows (Namburi 2006):
A material medium is not required for the propagation of light.
Energy is carried in the form of waves.
The velocity of light is 3 x 10
8
ms
-1
.
Light incident on a surface is either absorbed, reflected or transmitted, refracted, or a combination of
these.
Absorption of light
Light incident on an object can be absorbed and converted to heat. If 100% of the light incident on a object is
absorbed by the object, it will not be visible or visible only in silhouette (Fig. 1-6).
Figure 1-6: Absorption of light
18
Reflection of light
When light incident on an object at a particular angle is reflected back either at the same or a different angle,
reflection occurs. There are two types of reflections:
1) Specular - When light incident on a smooth (or polished) surface is reflected at the same incident angle to
produce direct reflection (Fig. 1-7).
2) Diffuse - When light incident on a rough (or matte)surface is reflected back at many different angles to
produce diffuse reflection (Fig. 1-7). An ideally isotropic (non-directional) diffusing surface is called a Lambertian
surface. An anisotropic (directional) surface is an ideal specular and directional reflecting surface. A bi-
directional reflectance distribution function (BRDF) has the diffuse, directional diffuse and specular components
of reflected light (Bass 1995).
Figure 1-7:Specular and diffuse reflection
Refraction of light
Refraction is the bending of a light wave when it passes from one medium to another. The speed of light
changes in the new medium, giving the impression that it is bending while its frequency (wavelength) remains
the same (Fig. 1-8).
Figure 1-8:Refraction of light
Units
Light can be measured in two alternate sets of units.
1) Radiometry - Measurement of light power at all wavelengths (Table 1-1).
19
Quantity Unit Notes
Radiant Energy Joule Energy
Radiant Flux Watt Radiant Energy per unit time/ Radiant Power
Spectral Power Watt per meter Radiant Power per Wavelength
Radiant Intensity Watt per Steradian Power per unit solid angle
Spectral Intensity Watt per Steradian per meter Radiant Intensity per Wavelength
Radiance Watt per Steradian per square
meter
Power per unit solid angle per unit projected
source are
Spectral Radiance Watt per Steradian per meter
cube
Quantity of radiation that passes through or is
emitted from a surface and falls within a given
solid angle in a specified direction
Irradiance Watt per square meter Power Incident on a surface/
Radiant Flux Density
Spectral Irradiance Watt per meter cube Solar Flux Unit
Radiant Exitance/
Radiance Emittance
Watt per square meter Power Emitted from a surface
Spectral Radiant
Exitance/
Spectral Radiant
Emittance
Watt per meter cube Power Emitted from a surface per unit wavelength
or frequency
Radiosity Watt per square meter Emitted plus reflected power leaving a surface
Spectral Radiosity Watt per meter cube Emitted plus reflected power leaving a surface
per unit wavelength
Radiant Exposure Joule per square meter The accumulated physical quantity of visible light
energy (weighted by the luminosity function) per
area applied to a surface during a given exposure time
Radiant Energy
Density
Joule per meter cube The measure of the amount of radiant energy per
Unit volume at a given location and time
Table 3-1:Radiometry unit chart (Palmer 1999)
2) Photometry - Measurement of light with wavelengths weighted to a standardized model of human brightness
perception (Table 1-2).
Quantity Units Notes
Luminous Energy (Q) Lumen-Seconds The amount of energy transmitted in the visual spectrum.
Luminous Flux ( ɸ =
dQ/dt)
Lumen The amount of light energy transmitted through a
surface. It is the light energy flow rate.
Luminous Intensity
(I = d ɸ/dw)
Candela The amount of light emitted by a source in a particular
direction.
Luminance (L =
d
2
∏/dwdA)
Candela per square foot The light energy leaving a real surface in a particular
direction.
Illuminance (E =
d ɸ/dA arriving)
Lux The light energy arriving at a real surface.
Exitance (M = d ɸ/dA
leaving)/ Emittance
Lux The light energy leaving a real surface without regard to
direction.
Table 1-4:Photometry unit chart (Schiler 1992)
Advantages of daylighting
According to the book Daylighting Handbook 1, "Daylighting is the controlled use of natural light in and around
buildings." (Reinhart 2014).
20
Daylight has several advantages. It is utilized by almost all life processes on the earth. The two most important
benefits from the point of view of architects and lighting designers are the psychological benefits and energy
savings. Good daylighting is related to attitude, satisfaction and well-being of building occupants. Adequate
exposure to daylighting can help improve health and circadian rhythm of human beings. A number of research
studies have shown the direct relationship between daylighting and improved worker productivity, reduced
absenteeism, increased student performance and improved patient recovery times (Mahone Group 2003). A
series of studies between 1999 and 2003 established a direct relation between the availability of daylight and
improved performance in schools and improved retail sales in commercial establishments (Mahone Group
2003).
The importance of daylighting was established at the National Institute of Mental Health where a controlled
study using light therapy was used to conclude that too little light in winter led to depression (Seasonal affective
disorder or SAD) (Rosenthal et al. 1980). The first study on this behavior was uncontrolled, and the single person
tested had positive reaction to bright light (Lewy et al. 1982). In the second study, bright light produced 52%
greater reduction in symptoms as compared to dim light (Rosenthal et al. 1984). Subsequent studies have shown
considerably lower advantages of bright light over dim light (Terman et al. 1989).
An additional benefit of daylight is color rendering. Color Rendering Index(CRI) is a measure of how well light
shows the true colors of an object. It shows the ability of a light source to render the color of objects truly
compared to a black body source (Steffy 2008) (Fig. 1-9). The CRI of lamps are rated on a scale which goes to
100, which is considered to be the CRI of daylight (Steffy 2008). It is important to balance these advantages with
glare and overheating.
Figure 1-9:CRI (http://www.ledvista.ie/colour-rendering-index)
21
Historically, daylight was ubiquitous in almost all architecture. In the 1940s, with the advent of mass produced
cheap fluorescent lights, daylight was sidelined(Reinhart 2014). Its importance was felt again in the 1970s
oil/energy crisis (Reinhart 2014).
The human eye
The human eye is responsible for sight (Fig. 1-10). Reflected rays from the objects enter the eye and get focused
on the retina. This information is then sent to the brain via the optic nerve. The retina has two classic
photoreceptive cells: rods and cones. Rods are sensitive to motion but not to color in objects. They aid in
scotopic vision and require around 1 hour for adaptation. Cones see color in objects and aid in photopic vision
and require 2-7 minutes for adaptation. Between scotopic and photopic vision lies the mesopic vision that is the
transition phase between photopic and scotopic vision and is done by both rods and cones. The photosensitive
ganglion cells were discovered as an additional class of photoreceptors in the 1990s. They do not aid in sight
directly but are believed to contribute to the body's circadian rhythm and pupillary reflex (Foster et al. 1991).
Two important parameters for seeing an object are the distance between the object and the eye and the angle
subtended by the object to the eye called the Visual Angle. The greater the distance, the smaller the angle and
the smaller the apparent size of the object. The focal distance of an object for a person with perfect eyesight is
called Least Distance of Distinct Vision. The point closest to the eye at which an object can be seen clearly
without strain to the eye is called the Near Point and the farthest point for the same is called the Far Point. The
distance between these two points is called the Range of Vision (Foster et al. 1991).
Figure 1-10: Anatomy of the human eye (www.sciencesuperschool.com)
The eye is capable of adapting itself to different levels of brightness. It takes approximately 20 - 30 minutes to
adapt from bright sunlight to complete darkness and around 5 minutes to adapt to bright sunlight from
darkness. The brightest and the darkest light signal that the human eye can see at a given time are a factor of
22
roughly 1,00,000apart (Fig. 1-11) (Suk 2014), but at any given time the eye can perceive a maximum contrast
ratio of 1000:1 (Encyclopedia Britannica 1987) (Fig. 1-12).
Figure 1-11: The range that the human eye can perceive at a given time (Adapted from Suk 2014)
Figure 1-12: The resulting effect of visual adaptation (Baker and Steemers 2002)
Designing with daylighting
Compared with artificial lighting, designing with daylighting is complex because daylighting is variable. It
depends on the location of the sun (azimuth and altitude), the location of the site/building (latitude and
longitude), cloud cover, terrain, etc. Azimuth is the angle along the horizon with zero degrees corresponding to
north and increasing in clockwise fashion. Altitude is the vertical angle of the sun above the horizon (Fig. 1-13).
23
Figure 1-13: Azimuth and Altitude
Daylighting can be incorporated into buildings by side lighting (single storey/multiple stories) (Fig. 1-14) and top
lighting (single storey) (Fig. 1-15). Skylights and clearstories are effective ways of toplighting spaces and can be
configured in various ways to increase their efficiency.
Figure 1-14:Side Lighting (left) and Side Lighting enhanced with external or internal light shelves (right)
Figure 1-15: Toplighting strategies (Hastings 1994)
While sidelighting, it is important to maintain the aspect ratio. The relative dimension between the height of
window and the depth of the space is very important in determining the amount of daylight in a space. The
primary daylit zone extends upto a depth of 15' perpendicular from the window with the width being the width
of glazing plus 2' on either side. The secondary daylit zone is either 25' or 2.5 x height of glazing, whichever is
lesser (Title 24). This can be used for furniture layouts so that the workstations in an office are placed in the first
and second zones while the circulation area that requires less light can be placed in the farthest zone (Fig. 1-16).
24
Figure 1-16: Aspect Ratio - a) calculation of top-lit zones b) calculation of side-lit zones (Lighting Control and Design)
A key point to keep in mind while designing with daylighting is to avoid glare. Glare is of two kinds: absolute
glare and relative glare. Absolute glare occurs due to the presence of an excessive bright source, the sun or a
luminaire of very high luminance that will cause discomfort regardless of the environment since the absolute
luminance is too high for the photoreceptors to process. Relative glare occurs due to the contrast between the
glare source and background (Suk 2014).
Software
AGi32
AGi32 is a computer program which performs calculations that show the behavior of light, artificial and natural
in a space via point-by-point calculations. Complex models can be built and it supports the import of 3D
geometry via the DWG, DXF and LaiDex formats. It can employ both raytracing and radiosity algorithms and can
produce good quality photorealistic renders and pseudocolor images. Recent versions of the software (Version
15.3) have incorporated Daylight Study as well (http://www.agi32.com/) (Fig. 1-17).
25
Figure 1-17: Scene rendered in AGi32
Ecotect/Radiance
Radiance is a lighting analysis tool that can be used by itself or as a plugin with Ecotect and can help calculate
daylight factor, illuminance, luminance and daylight autonomy. The models have to be built in Ecotect and
cannot be imported from other formats, which creates some limitation in the modeling of complex space types
(http://sustainabilityworkshop.autodesk.com/buildings/radiance-accurate-day lighting) (Fig. 1-18).
Figure 1-18: Rendering and false color in Ecotect/Radiance
Rhino/Diva
Diva is a plugin for Rhino that does optimized and parametric daylighting and energy modeling developed by the
Graduate School of Design at Harvard University. It allows users to conduct whole building simulations along
26
with daylighting integrated with electric lighting. It can do certain code compliance tests and can measure the
new LEED climate based metrics (http://diva4rhino.com/) (Fig. 1-19).
Figure 1-19: Rendering and false color in Rhino/DIVA
Rhino/Grasshopper/Honeybee
Honeybee connects Rhino3D to validated simulation engines such as EnergyPlus, Radiance, Daysim and
OpenStudio for building energy and daylighting simulation (http://www.food4rhino.com/project/ladybug-
honeybee?ufh) (Fig. 1-20). At present, the Radiance and DaySim connections work while the connections to
OpenStudio and EnergyPlus are being developed.
Figure 1-20: Connection of Rhino to Honeybee
1.3 Definition of the scope of this project: What is not included?
Many large architectural design firms have developed their own in-house software for the prediction of
daylighting in spaces (Namburi 2006). These software programs have not been taken into consideration. Two
27
software programs that are in widespread use in offices, AGi32 and Ecotect/Radiance are compared with two
relatively new tools, Rhino/Diva and Rhino/Grasshopper/Honeybee.
The tests are divided into two parts:
1) Testing the accuracy of simulation of the physical properties of light, like specularity, reflectivity and luminous
flux conservation.
2) Comparing additional parameters like import and export options, analysis plugins used, modeling complexity,
required familiarity, availability/cost, types of sky recognized, types of render algorithms used, types of metrics
calculated, types of code compliancy tests possible, speed of render, types of output visualizations possible,
clarity of Graphical User Interface(GUI).
Three pathological tests have been defined for testing the simulation of the physical properties of light.
Pathological test cases are those that use the least number of variables to set up an extreme case where the
user can predict what is going to happen based on knowledge of physics. For example, the 100% reflection test
has only one parameter - the reflectance value of the north wall which is extreme at 100% (Refer to Section
3.1.1) Extraneous parameters like glazing type, surface properties have been either eliminated or kept constant
in all the tests. Additionally, the result can be anticipated by the exitance-illuminance equation. Therefore, it is
easy to recognize the success or failure of the software. Physical models have been built for two of the three
cases for the collection of real data with simulation data. Based on these tests and additional parameters, a
number of comparative charts have been created.
1.4 Conclusion: A comparative evaluation chart of different analysis software
The aim of the current project has been organized into a group of objectives listed as follow:
1) To create a comprehensive easy-to-refer chart listing the tests and properties that work well on one software
and do not on another to help users in making better design analysis decisions. For instance, if the objective is to
create a photorealistic rendering, the chart would suggest a particular software, or if the objective is to get
accurate light levels considering the different material properties of various surfaces, the chart would suggest
another software. Again, if the objective is to get a rough qualitative idea of a daylit interior space in a very short
time, the chart would suggest a software.
2) To analyze the accuracy level of each software based on the pathological tests.
3) To study the ability of each of the software to calculate code compliancy and its ability to calculate dynamic
climate based metrics.
28
Chapter 2
Daylight simulation in software
2.0 A State-of-the-Art narrative on daylight simulation and analysis in software
The field of lighting simulation in software gradually came into being and started developing from 1968 (Kota
and Haberl 2007). Different calculation algorithms like radiosity in AGi32 and raytracing in Radiance and Relux
were explored in different software (Ochoa, Aries and Hensen 2010).
The analysis of lighting in buildings in software programs started with repetitive electric lighting installations in
closed rectangular rooms (Ochoa, Aries and Hensen 2010).At the same time, integration of daylight and artificial
lighting in fixed conditions using simplified formulas was attempted (Plant and Archer 1973). Besides numerical
results, the ability of the software to produce physically based renderings was developed (Knisset et al. 2003).
The limitations faced by early simulation programs included the inability to model complex geometries, minimal
daylighting studies, and inaccuracy of output (Svendenius and Pertola 1995).
Lighting simulation typically falls into two main categories - photorealistic rendering and predictive rendering -
(Ochoa, Aries and Hensen 2010):
Photorealistic rendering deals with the production of good quality rendered images that might not be
completely accurate.
Predictive rendering deals with the accurate representation of a scene based on the physical laws of light.
Some of the different algorithms used in lighting simulation can be summarized as follows (Table 2-1):
1) Direct Calculation - They have specific formulae, assumptions and simplifications so that the most common
situations like a rectangular room with artificial lighting can be accurately analyzed and are generally used for
the analysis of artificial lighting in software (Ochoa, Aries and Hensen 2010).
2) Raytracing - It is a rendering technique where 3-dimensional objects are rendered by casting rays in a scene,
tracing a path of light through pixels in an image plane and simulating what happens when the light hits virtual
objects (Whitted 1980) . These are generally grouped according to the direction from where the rays are coming
- from the light source (forward tracing), from the observer's eye (backward tracing), from both the light source
and observer (bi-directional tracing) (Ochoa, Aries and Hensen 2010). In backward raytracing, only a few rays of
light from a light source, which can see a specific portion of the object, render it which is seen through a pixel in
a virtual picture plane. It Is possible to create realistic images with this technique with effects such as reflections
29
and shadows (Whitted 1980). The disadvantage is that it is a slow technique, and these are view-dependant,
hence, to change the view of a space, recalculation is necessary (Ochoa, Aries and Hensen 2010) (Fig. 2-1) (Refer
to Section 5.1.5 for more discussion).
Figure 2-1: Raytracing Process (Wikipedia Images)
3) Radiosity - It is a rendering technique for scenes with surfaces that reflect diffuse light (Goral et al. 1984). It
considers light leaving a light source and then being reflected diffusely a few times and then hitting the eye
(Dudka 2013). In this technique, all the rays from a light source hit a surface on the object and transfer the light
energy. This surface then acts as the light source and sends out a portion of the light energy to other surfaces
and this process goes on till all the surfaces have been reached. These are view-independent and hence
walkthroughs are possible. This is generally used for calculations (Ochoa, Aries and Hensen 2010) (Fig. 2-2).
Figure 2-2: Radiosity Process ((Ochoa, Aries and Hensen 2010)
4) Integrative approaches - Two or more algorithms are combined. Eg. Radiance uses both backward raytracing
and radiosity (Ochoa, Aries and Hensen 2010).
5) Calculation aids like Monte Carlo method where the expected value is assumed to be correct and then a
number of expected outcomes are averaged for a solution. It requires a lot of samples and often have accuracy
issues (Ochoa, Aries and Hensen 2010).
30
Table 2-1:Non-exhaustive summary of analysis tools and their properties
There are two ways a software can be validated, against real data gathered in existing buildings (or a scaled
physical model) or against data gathered in laboratory settings (Ochoa, Aries and Hensen 2012). The first
method has the limitations of the kinds of tests that can be performed, difficulty in replicating surface properties
and geometries, and that sky conditions are variable. The second method, on the other hand, ignores normal
human interventions and requires special equipment that can be difficult to acquire.
The International Commission on Illumination (CIE) proposed a series of test cases to be used as benchmarks
(CIE 2006). These test cases try to include the widest range of possible variations simulation software might
encounter, including both artificial and natural light, and diverse types of surfaces (Ochoa, Aries and Hensen
Tool Algorithms
used
Purpose Advantages Disadvantages Availability
Ecotect/Radiance
(Ochoa, Aries and
Hensen 2010)
- backward
raytracing
- scene
radiance
General
Purpose
- capable of serious
research and
environmental
consequences
- powerful capacity
of solving lighting
problems.
- has been validated
by CIE (CIE 2006).
- unfriendly user
interface
- requires extensive
knowledge and
expertise
- requires advanced
knowledge of material
properties.
Ecotect, the
energy
analysis tool
cannot be
purchased
from
Autodesk
since March
20, 2015
whereas
Radiance,
the lighting
analysis
plugin is
free.
AGi32(Ochoa, Aries
and Hensen 2010)
- direct
calculation
- radiosity
- limited
raytracing
- Luminaire
design
- daylight
integration
- competent electric
lighting analysis.
- has been validated
by CIE (CIE 2006).
- modeling is
complicated and time-
consuming
- Has to be
purchased
Rhino/Diva - backward
raytracing
- Daylighting
calculation
- whole
building
analysis
- code
compliance
testing
- potentially
combining a
powerful modeling
tool and accurate
lighting simulation
algorithms
- has not been validated
yet according to CIE
standards
- Has to be
purchased
Rhino/Grasshopper
/Honeybee
- backward
raytracing
- Daylighting
calculation
- whole
building
analysis
- potentially
combining a
powerful modeling
tool and accurate
lighting simulation
algorithms
- has not been validated
yet according to CIE
standards
- Has to be
purchased
31
2012). The CIE designed a number of simple tests with limited parameters that would test the different aspects
of lighting simulation based on the different kinds of light propagation (IEA SHC Task 31/ IEA ECBCS Annex 29,
2005).
The tests include (IEA SHC Task 31/ IEA ECBCS Annex 29, 2005):
1) Artificial direct lighting - point sources
2) Artificial direct lighting - area light sources
3) Artificial lighting - experimental reference data
4) Daylighting - luminous flux conservation
5) Directional transmittance of clear glass
6) Direct daylighting - unglazed opening
7) Direct daylighting - glazed opening
8) Direct daylighting with external mask
9) Indirect lighting - diffuse reflection
10) Indirect lighting - diffuse reflection with internal obstructions
11) Indirect lighting - diffuse inter-reflections
It is not possible to get completely similar numerical results when two or more software packages are compared
against each other or against other data because of differences in algorithms used, human errors in modeling,
etc.. CIE estimated that an acceptable range would be 10% for average illuminance calculations and 20% for
measured point values (Ochoa, Aries and Hensen 2012). The same range has been adopted for the current
comparison.
Lighting simulation has also developed from being static to dynamic over time. Annual weather files and
dynamic shade elements are being recently incorporated. Table 6-6 in Section 6.2 includes an investigation of
which of the software packages chosen can do this.
The methods used for input in lighting simulation software have ranged from text files to CAD input over the
years. Recently, direct data input from reality has been made possible through High Dynamic Range (HDR)
photography (Ochoa, Aries and Hensen 2012). The current study compares the different types of input and
output possible in each of the chosen software packages.
Lighting analysis can produce two types of outputs - quantitative output and qualitative output (Ochoa, Aries
and Hensen 2010):
Quantitative - It includes calculation results that are text only, calculation grids that present point-by-point
information (Fig. 2-3).
Qualitative - These can include interactive renderings, graphical data (isocontours, falsecolor images, etc. for
luminance, illuminance)(Fig. 2-3).
32
Figure 2-3: Examples of Quantitative (left) and Qualitative output (right)
Some knowledge of lighting techniques and trends is required to interpret the results (Hong et al. 2000). There is
a need for adequate data visualization and interpretation with large amounts of simulation results (Glaser et al.
2004) (Table 2-2).The kinds of output interpretation available have been studied in the current comparison.
Criteria Classification Means Values Examples
Type of output Quantitative - Text file - Illuminance
- Luminance
- Glare
- Daylight Factor
- tabular format
(for use in post processing
like Excel sheets for
calculation, etc.)
Qualitative - On screen
rendering
- Image file
- High dynamic
range (HDR)
pictures
- image
- falsecolor
- isocontours
- luminance assessment
- user preferences
- spatial lighting quality
- medical uses
Output
interpretation
By model - Very few offer
interpretation
- indicate code compliance
By model user - Lighting experts - considerable expertise
needed
- building physics
researches, physicists
- Non-Lighting
experts
- interpretation on what
results mean
- architects who dedicate
primarily to design
Table 2-2: Output in lighting simulation(Ochoa, Aries and Hensen 2012)
There is also a high demand for the integration of lighting simulation with whole building analysis to find out
how much daylighting can help in reducing the need for electrical power.
Lighting software can be used at different stages of design: early design, design development, code compliance,
and building commissioning.
33
a. Early design stage - This phase requires minimal detail and computation time and mostly uses design
intuitions rather than calculated decisions. Some examples of tools include Lightsolve, a virtual
Heliodon, NewFacades, the Virtual Lighting Simulator, Sustarc, Helios, Sunscapes and solar masks
(Ochoa, Aries and Hensen 2006).
b. Design development stage - As fundamental issues like orientation, height of building, massing, façade
structure, size and shape of openings are decided, sophisticated and detailed lighting simulation is
possible in a number of tools, including the ones currently being compared (Ochoa, Aries and Hensen
2012).
c. Compliance with building codes - As the design of the building gradually becomes finalized, code
compliance is calculated in a number of tools like Rhino/Diva, and EnergyPro (Ochoa, Aries and Hensen
2012.
d. Building commissioning and operation - There are very few tools for simulation of the actual operation
of a building. Occupancy and user interaction with control elements are difficult to simulate in most
lighting analysis tools. Daysim is one tool that analyzes different strategies using control elements
(Ochoa, Aries and Hensen 2012).
2.1 Pathological cases for evaluation of lighting simulation tools
Pathological tests are defined to test the accuracy of a software. They are extreme situations set up in a way
such that the outcome can be predicted based on laws of physics. If the results do not match, the failure mode
can be determined and the algorithm corrected.
A previous comparative study between four lighting simulation software packages, AGi32, Desktop Radiance,
Lightscape and 3DSMax was conducted using tests on the following lighting concepts: specular reflection for
artificial and natural lighting, reflectance, and transmittance (Fig.2-4) (Namburi 2006).
Figure 2-4:Summary Chart
Some drawbacks of these studies:
1) The tests did not focus on either artificial or natural lighting, and the testing pattern was irregular.
34
2) The simulation software packages are compared against each other and first principles calculations only on
the basis of accuracy of output. Issues like interoperability, input types, render qualities, types of metrics
calculated, ability to do whole building analysis, and code compliance were not a part of the study.
3) The test results were a little vague in defining the limits of accuracy and the charts give no indication as to
when and where the software packages should be used.
2.2 Discussion of test cases and parameters while evaluating software
The feasibility of developing a scoring system for lighting simulation tools like the IEA BESTEST for thermal
simulation software packages states the need for going beyond the CIE test cases to create an easier to
understand "consumer guide."Parameters like time and difficulty of learning, ease of building an input file, run-
time, accuracy, quality of render, user assessment of their usability, etc. should be made a part of comparative
studies (Donn et al. 2007).
Parameters that are missing in the CIE evaluation (Donn et al. 2007) and that are included are as follows:
1) Import and export options
2) Analysis plugins used
3) Modeling complexity
4) Required familiarity
5) Availability/cost
6) Types of sky recognized
7) Types of render algorithms used
8) Types of metrics calculated
9) Types of code compliancy tests possible
10) Speed of render
11) Types of output visualizations possible
12) Clarity of Graphical User Interface(GUI).
2.3 A narrative on development of daylight analysis metrics
Leadership in Energy and Environmental Design (LEED) is a green building certification program started by the
U.S. Green Building Council (USGBC) that recognizes the best sustainable strategies in the building industry. To
earn the certification, buildings have to satisfy some prerequisites and then earn points based on the strategies
used. Strategies are classified into different categories, one of them being daylight. The points that can be
earned based on daylighting have been defined differently over the years and the definition of daylight analysis
metrics have changed gradually with time, becoming stricter and more complicated (Fig. 2-5).
35
Figure 2-5: A brief history of LEED simulation-based daylight credit compliance (Jakubiec 2014)
IES-LM-83 is a method approved by the Illumination Engineering Society (IES) Daylight Metrics Committee that
resulted in the creation of two new metrics for daylit spaces (Jakubiec 2014) -
1) Spatial Daylight Autonomy (sDA
300lx,50%
) (Fig. 2-6 and 2-7) -
sDA300lux is the percentage of occupied that gets atleast 300lux from daylight alone.
sDA300lux,50% is the percentage of floor area that receives atleast 300lux more than 50% of the
occupied hours (Jakubiec 2014).
According to LEED v4:
1) sDA300lux,50% > 55% of floor area – 2 LEED points
2) sDA300lux,50% > 75% of floor area – 3 LEED points
2) Annual Sunlight Exposure (ASE
1000lx,250h
) (Fig. 2-8) -
ASE1000lux,250h is the percentage of floor area that has greater than 1000lux greater than 250 hours (Jakubiec
2014).
According to LEED v4:
Requirement for Points - ASE1000lux,250h <10%
36
It attempts to standardize many unknowns in a typical simulation process (Jakubiec 2014).
1) Shading Controls - To be closed when more than 2% of a space's floor area receives 1000lx of illuminance
from direct sunlight.
2) Occupancy - 8 AM to 6 PM, 365 days a year (10*365 = 3650 hours a year)
3) It assumes default values for materials whose reflectances are often not known.
- 30% vertical exterior surfaces reflectance
- 20% trees reflectance
- 10% ground reflectance
- 20% floor reflectance
- 50% wall reflectance
- 70% ceiling reflectance
- 50% furniture reflectance
4) Workplane height ( 30" Above Finished Floor ) and sensor spacing.
5) Modeling guidelines (furniture, window mullions, context) - fixed furniture to be modeled 36" above the floor
6) Simulation parameters (light reflections, simulation detail)
37
Figure 2-6: sDA Calculation Logic (Jakubiec 2014)
38
Figure 2-7: Graphical Representation of sDA (Jakubiec 2014)
39
Figure 2-8: Calculation Logic of ASE (Jakubiec 2014)
2.4 Software testing
A comparative study between 3DSMax, Ecotect, Radiance, Daysim, Diva and Su2ds, the tools in maximum use in
Brisbane, Australia, compared the accuracy of the tools in calculating metrics like Daylight Factor (DF), Daylight
Autonomy (DA), Useful Daylight Illuminance (UDI), using measured horizontal illuminances (Panitz, Garcia-
Hansen 2013). The software programs were also compared on the basis of their ability to calculate the new
climate based metrics. This study finally led to several useful charts of comparison between the software based
on the types of sky that could be used, the render algorithms used, the metrics that could be calculated, the
speed of calculation, the types of input and output possible (Fig. 2-9). This chart is helpful for users to
understand which software to use based on the application. The comparative charts in Section 6.2 are loosely
based on the following chart.
40
Figure 2-9:Comparative charts used in study (Panitz, Garcia-Hansen 2013)
2.5 Conclusion
Background study of the comparisons and validations of different software programs led to the decision of
including additional parameters, such as the types of sky that could be used, the render algorithms used, the
metrics that could be calculated, the speed of calculation, the types of input and output possible, which relate
directly to the users besides testing accuracy of output and the ability of the software to mimic the physical laws
of light. Code compliance and ability to calculate climate based metrics are other important issues that have
been included.
41
Chapter 3
The methodology of software validation and comparison
3.0 Inference from background research
The background research has contributed considerably to the selection of the pathological cases and the need
for testing additional parameters, such as the types of sky that could be used, the render algorithms used, the
metrics that could be calculated, the speed of calculation, the types of input and output possible, code
compliancy and ability to calculate climate based metrics, which relate directly to the users besides testing
accuracy of output and the ability of the software to mimic the physical laws of light. It paved the way towards
the formation of the evaluation charts. The methodology followed was the selection of tests followed by
building of the physical model and gathering data, simulating in software and comparing the data followed by
the creation of a comparative chart (Fig. 3-1).
Figure 3-1: Workflow
3.1 Selection of Pathological Cases
Initially, there were to be five tests, three of them to be based on some of the tests defined by the CIE for
validation of lighting simulation tools. They included the Directional Transmission of Clear Glass and Diffuse
42
Reflection Test. An additional test was the Verification of Daylight Factor under different sky conditions. These
tests were deleted from the study since the first two were difficult to replicate in a physical model within an
academic scope. The third one was deleted since it did not fall under the domain of experiments that would test
the physics of light.
The final three tests that were chosen were the specular reflection test, the reflection test, and the luminous
flux conservation test.
3.1.0 Specular reflection test
The base case is arectangular room 40' x 20' x 20' with 3 unglazed openings 2' x 2' each on top (Namburi 2006)
(Fig. 3-2).
Figure 3-2:Specular reflection test model layout plan and 3D view
Glazing has been omitted on purpose to keep the number of parameters to a minimum. The openings are
covered and then are opened one at a time to allow a direct ray of sunlight to hit the center of the north wall.
According to the solar gnomon for the city of Los Angeles (Fig. 3-3), this condition happens on December 15th at
12:00 PM by first rotating the building such that the East-West axis is at 45 degrees to the north line, then by
rotating it such that it is at 90 degrees to the north line, and finally by rotating it such that it is at 135 degrees to
the north line to get three different azimuth angles.
Figure 3-3: Solar Gnomon for Southern California (LBNL 1997)
43
For Test 1, just the first opening was kept open, and the longer axis of the building was kept at 45° to the E-W
axis for the sun's ray to hit the center of the north wall (Fig. 3-4).
Figure 3-4: Test 1: A open
For Test 2, B is open and the longer axis of the building is at 90° to the E-W axis (Fig. 3-5).
Figure 3-5: Test 2: B ope
For Test 3, C is open and the longer axis of the building is at -45° to the E-W axis (Fig.3-6).
Figure 3-6: Test 3: C open
All of the interior surfaces had almost the similar values for properties like reflectance, specularity and
roughness, etc. (Table 3-1) except the north wall and the floor that were specular and the ceiling. It was
anticipated that provided everything was followed correctly, the light from the opening should hit the center of
the north wall each time at a different angle and be reflected to form a bright patch on the floor in three
different positions.
44
Testing
Terminology
North Wall South Wall West Wall East Wall Floor Ceiling
Reflectance 95 49.7 49.7 49.7 96.7 80
Specularity 95 0 0 0 0 0
Roughness 2 0 0 0 0 0
Color Bleed n/a n/a n/a n/a n/a n/a
Surface
Orientation
Interior Interior Interior Interior Interior Interior
Transmittance 0 0 0 0 0 0
Calculation
Type
Luminance/
Illuminance
n/r n/r n/r Luminance/
Illuminance
n/r
Table 3-1: Surface properties (Namburi 2006); Notes: n/a - Not Applicable, n/r - Not required
The physical model was built at a scale of 3/8" = 1' with foam core(Fig. 3-7). The specular surface was made with
silver plastic sheet. The specularity of the surface was measured with a device called a Statistical Novogloss. The
novogloss is a device that comes with a piece of glass with known reflectivity and specularity for the purposes of
calibration. It tells you the specularity of any material when placed over it (Fig. 3-8).
Figure 3-7: Physical model for specularity test
Figure 3-8: Gardco Statistical Novogloss
45
The model was photographed using a fish eye lens at four different exposure settings (1/2, 1/15, 1/125 and
1/1000) (Fig. 3-9). The four photographs are then combined in Photolux software , which can create a luminance
map from digital pictures (Photolux website) to create a false color image which gives the luminances on all the
surfaces and gives an image of how the light hits the center of the north wall and is reflected on the floor as is
evident from the hotspots (Fig. 3-10).
Figure 3-9: Images captured at 4 different exposures using a fish eye lens
Figure 3-10: Example of an image generated by photography from physical model using a fish-eye lens and Photolux software
3.1.1 Reflection test
The base case is a rectangular room 40' x 20' x 20' with 1 unglazed opening 2' x 2' on top (Namburi 2006)
(Fig. 3-11).
46
Figure 3-11:Reflection test model layout plan and 3D view
Glazing has been omitted on purpose to keep the number of parameters to a minimum. The opening allows a
direct ray of sunlight to hit the center of the north wall on December 15th at 12:00 PM such that the East-West
axis of the building is at 90 degrees to the north line.
Exitance is the total luminous flux density leaving a surface measured in lumens per square foot whereas
luminance is the luminous flux density leaving a surface in a particular direction measured in lumen per square
foot (Schiler 1992). The concept of reflectivity states that a surface is reflective if the light, after hitting the
surface bounces off in all directions such that the angle of incidence is equal to the angle of reflection (Namburi
2006). For a 100% reflective surface, the illuminance at the point where the light first hits is equal to the
exitance at that point. For a 50% reflective surface, the illuminance is equal to double the exitance (Namburi
2006).
All of the interior surfaces had almost the similar values for properties like reflectance, specularity and
roughness, etc. (Tables 3-2 and 3-3) except the north wall and the floor which were highly reflective. It was
anticipated that provided everything was followed correctly, the light from the opening should hit the center of
the north wall to form a bright patch at which point the illuminance and exitance would be measured. For
software packages that do not measure exitance, like Ecotect/Radiance, Rhino/DIVA and
Rhino/Grasshopper/Honeybee, luminance would be measured since we are concerned with the light level at the
first point before it's directionality makes a difference.
Testing
Terminology
North Wall South
Wall
West Wall East Wall Floor Ceiling
Reflectance 100 1 1 1 100 1
Specularity 0 0.5 0.5 0.5 0 0.5
Roughness 0 15 15 15 0 15
Color Bleed n/a n/a n/a n/a n/a n/a
Surface
Orientation
Interior Interior Interior Interior Interior Interior
47
Transmittance 0 0 0 0 0 0
Calculation
Type
Illuminance/
Exitance
n/r n/r n/r n/r n/r
Table 3-2: Surface properties for a 100% reflective north wall (Namburi 2006); Notes: n/a - Not Applicable, n/r - Not required
Testing
Terminology
North Wall South
Wall
West Wall East Wall Floor Ceiling
Reflectance 50 1 1 1 100 1
Specularity 0 0.5 0.5 0.5 0 0.5
Roughness 0 15 15 15 0 15
Color Bleed n/a n/a n/a n/a n/a n/a
Surface
Orientation
Interior Interior Interior Interior Interior Interior
Transmittance 0 0 0 0 0 0
Calculation
Type
Illuminance/
Exitance
n/r n/r n/r n/r n/r
Table 3-3: Surface properties for a 50% reflective north wall (Namburi 2006); Notes: n/a - Not Applicable, n/r - Not required
The physical model (Fig. 3-12) was built at a scale of 3/8" = 1'with foam core. Since, a 100% reflective surface is
not possible in reality, the reflective surface was made with white foam core for the 100% reflective surface and
a white foam core painted grey for the 50% reflective surface.
Figure 3-12: Physical model for 100% reflection test
Similar to the specular reflection test, pictures were taken using the fish eye lens and false color images were
created in Photolux software. It was not possible to measure exitance in a physical model within an academic
scope, hence, the false color images have been used for a pictorial reference alone.
48
3.1.2 Luminous Flux Conservation test
The base case is a rectangular room 12' x 12' x 10' with a 1 unglazed opening on top, progressing in size from
3'x3' to 6'x6' to 9'x9' and finally to 12'x12' (Maamari 2005) (Fig. 3-13). The importance of this test lies in the fact
that if the luminous flux conservation is not taken into account in software simulations, there will be an error in
the illuminance level inside the space.
Figure 3-13:Luminous flux conservation test model layout plan and 3D view
The luminous flux conservation theory states that the total direct luminous flux reaching the different interior
surfaces ( ɸ
i
) at a given time is equal to the total flux at the opening surface ( ɸ
o
) (Maamari 2005).
Incident Flux ɸ
o
= E
o
x S
o
where E
o
is the average illuminance at the opening surface and S
o
is the area of the opening surface.
ɸ
n
is the luminous flux reaching the internal surface n.
Therefore, ɸ
i
= Ԑ ɸ
n
The interior surfaces will have 0% reflectance and the ground should have 0% reflectance.
No physical model was built for this test since at a scale of 3/8" = 1' which is the scale at which the models for
the specular reflection test and the reflection test have been built, the illuminance meter would itself interfere
with the test by absorbing some of the flux. Additionally, it would be a laborious process to measure and
calculate the average illuminances of all surfaces and it could not be compared against the values generated in
the software packages as the intensity of sunlight would vary. Instead, the software simulation results will be
compared against each other.
3.2 Selection of additional parameters and comparative charts
In addition to the experiments testing the ability of the software packages to simulate the different physical
properties of light accurately, some additional parameters have been compared after inferring their importance
from literature reviews. Typical comparative charts have been created (Tables 3-4, 3-5, and 3-6). The charts are
49
based on the types of charts used in the study mentioned in Section 2.4. The advantages and disadvantages of
each software based on a specific parameter are listed in these charts so that the knowledge can be used to
determine a specific software for a specific application.
Software Export
Options
Import
Options
Analysis
Plug-ins
Modeling
Complexity
Required
Familiarity
Availability/
Cost
AGi32
Ecotect/
Radiance
Rhino/
Diva
Rhino/
Grasshopper
/Honeybee
Table 3-4: Analysis Software Interoperability Observations
Software Editing
Imported
Geometry
Model
Editing
Interface
Analysis
Interface
GUI
AGi32
Ecotect/
Radiance
Rhino/
Diva
Rhino/
Grasshopper
/Honeybee
Table 3-5: Analysis Software Modeling Observations
Software Sky models
recognized
Render
Algorithms
used
Speed of
Render; Quality
of Render
Metrics Calculated;
Whole Building
Analysis
Code
Compliance
Testing
AGi32
Ecotect/
Radiance
Rhino/
Diva
Rhino/
Grasshopper
/Honeybee
Table 3-6: Analysis Software Metrics Observations
50
3.3 Conclusion
The conclusion after the data collection, discussion and experiences while testing leads to the final accuracy
summary chart (Table 3-7, 3-8 and 3-9) and additional notes that has been designed to give an overview in a
glance.
Software Specular Reflection Test Notes
AGi32
Ecotect/Radiance
Rhino/DIVA
Rhino/Grasshopper/Honeybee
Table 3-7: Comparative chart for specular reflection test
Software Luminous Flux Conservation
Test
Notes
AGi32
Ecotect/Radiance
Rhino/DIVA
Rhino/Grasshopper/Honeybee
Table 3-8: Comparative chart for reflection test
Software Reflection Test Notes
AGi32
Ecotect/Radiance
Rhino/DIVA
Rhino/Grasshopper/Honeybee
Table 3-9: Comparative chart for luminous flux transmission test
51
Chapter 4
Simulation results of daylight analysis tests
4.0 Executive summary of methodology
The first step for the comparative study was the selection of software. The four software programs chosen were
AGi32, Ecotect/Radiance, Rhino/DIVA and Rhino/Grasshopper/Honeybee. AGi32 is used in almost all the lighting
design firms in North America and its recent release, Version 15.3 has incorporated several tools for daylight
modeling. A survey conducted in 2004 among users of daylighting analysis software showed that in 42 of the
analysis software named by the users, over 50% were ones that used Radiance (Reinhart and Fitz 2006).
Rhino3D is a powerful modeling tool used by architects for design. DIVA and Honeybee are two of the plugins for
Rhino. DIVA can be used by itself or through Grasshopper, which is a visual programming tool that can be used
with Rhino while Honeybee needs Grasshopper to run. Both DIVA and Honeybee make several kinds of energy
analysis possible. The possibility of accurate daylight simulation in Rhino can potentially be very advantageous.
The next step was the selection of three pathological tests; specular reflection test, reflection test, and
luminous flux conservation test were chosen.
Additionally, parameters like the clarity of the GUI, time and training required for full use, quality of scene
renders, ability to calculate climate-based metrics and code compliance, and interoperability with other
software have been compared across the same software programs, and the results listed to create a number of
comparative charts at the end.
4.1 Specular reflection test
A surface can be called specular if the light after hitting the surface retains the reflected image and does not
scatter (Namburi 2006). The anticipation of the test was that at a given time on a given day if a ray of light enters
through a single opening and hits a highly specular surface, it should be reflected to form a similar spot on
another surface such that the angle of incidence is equal to the angle of reflection.
4.1.1 Results of physical model test
The physical model test faced certain challenges and had to be repeated thrice (Refer to Section 5.1.1). The final
outputs show what was anticipated and are a good reference for comparing the software outputs.
52
Test 1: A Open
The model was tilted so that the EW axis was at 45° to true north and was photographed using a fish eye lens at
four different exposure settings (1/2, 1/15, 1/125 and 1/1000) with aperture size 3.8. It can be seen from the
images that the north wall is almost behaving like a mirror and there are multiple reflections (Fig. 4-1).
Figure 4-1: Fish eye images taken at 1/2, 1/15, 1/125 and 1/1000 (from left to right) exposure settings for Test1.
The four photographs are then combined in Photolux software to create a false color image (Fig. 4-2). The false
color image shows that the luminance levels at the center of the north wall and the floor are 215 cd/m
2
and
12400 cd/m
2
respectively.
Figure 4-2: False color image created in Photolux software showing luminance for Test1.
Test 2: B Open
The model was tilted so that the EW axis was at 0° to true north and was photographed using a fish eye lens at
four different exposure settings (1/2, 1/15, 1/125 and 1/1000) with aperture size 3.8. It can be seen from the
images that the north wall is almost behaving like a mirror and there are multiple reflections (Fig. 4-3).
(Mirror) North Wall
Hotspots on wall and floor
53
Figure 4-3:Fish eye images taken at 1/2, 1/15, 1/125 and 1/1000 (from left to right) exposure settings for Test2.
The four photographs are then combined in Photolux software to create a false color image (Fig. 4-4). The false
color shows the luminance levels at the center of the north wall and the floor are 282 cd/m
2
and 11400 cd/m
2
respectively.
Figure 4-4: False color created in Photolux software showing luminances for Test2.
Test 3: C Open
The model was tilted so that the EW axis was at -45° to true north and was photographed using a fish eye lens at
four different exposure settings (1/2, 1/15, 1/125 and 1/1000) with aperture size 3.8. It can be seen from the
images that the north wall is almost behaving like a mirror and there are multiple reflections (Fig. 4-5).
Figure 4-5:Fish eye images taken at 1/2, 1/15, 1/125 and 1/1000 (from left to right) exposure settings for Test3.
(Mirror) North Wall
Hotspots on wall and floor
54
The four photographs are then combined in Photolux software to create a false color image(Fig. 4-6). The false
color shows the luminance levels at the center of the north wall and the floor are 292 cd/m
2
and 1680 cd/m
2
respectively.
Figure 4-6:False color created in Photolux software showing luminances for Test3.
4.1.2 Results of simulation in AGi32
A number of runs were simulated in AGi32 (Refer to Section 5.1.2). The final outputs do not show the expected
results and upon exploration of the tool, it has been concluded that it is not possible to get the anticipated
results in Agi32.
Test 1: A Open
The final output in Agi32 shows a distinct spot at the center of the north wall as expected but does not show its
reflection on the floor, which is incorrect. Upon using raytracing, the north wall behaves like a mirror and a
reflection of the original ceiling opening can be seen (Fig. 4-7). But there is no light reflected from the north
(mirror) wall onto the floor. (Fig. 4-7).
(Mirror) North Wall
Hot spot and its reflection
at the bottom of West wall
55
Figure 4-7: Renderings in AGi32 using radiosity (left) and raytracing (right) for Test1
The false color shows a wide spread high illuminance with indistinct edges which was not anticipated. The
luminance map shows a hotspot of 334 cd/m
2
at the center of the north wall and 34.7 cd/m
2
at the center of
the floor (Fig. 4-8). There is no hotspot where the reflection onto the floor should be.
Figure 4-8: Illuminance false color (left) and luminance false color (right) for Test1
The illuminance at the center of the north wall is 1914 fc and that on the floor is 10.3 fc (Fig. 4-9). Again, there is
no hotspot where the reflection on the floor should be arriving.
Figure 4-9: Calculation points on north wall (left) and floor (right) for Test1
56
Test 2: B Open
The final output in Agi32 shows a distinct spot at the center of the north wall as expected but does not show its
reflection on the floor. Upon using raytracing, the north wall behaves like a mirror and a reflection of the ceiling
opening can be seen (Fig. 4-10). There is, however, still no hotspot on the floor, which is incorrect.
Figure 4-10: Renderings in AGi32 using radiosity (left) and raytracing (right) for Test2
The false color shows a wide spread high illuminance with indistinct edges, which was not anticipated. The
luminance map shows a hotspot of 555 cd/m
2
at the center of the north wall 88.3 cd/m
2
at the center of the
floor(Fig. 4-11). It shows no bounce on the floor.
Figure 4-11:Illuminance false color (left) and luminance false color (right) for Test2
The illuminance at the center of the north wall is 3185 fc and that on the floor is 26.2 fc (Fig. 4-12).
Figure 4-12:Calculation points on north wall (left) and floor (right) for Test2
57
Test 3: C Open
The final output in Agi32 shows a distinct spot at the center of the north wall as expected but does not show its
reflection on the floor. Upon using raytracing, the north wall behaves like a mirror and a reflection of the
opening can be seen (Fig. 4-13).
Figure 4-13:Renderings in AGi32 using radiosity (left) and raytracing (right) for Test3
The false color shows a wide spread high illuminance with indistinct edges that was not anticipated. The
luminance map shows a hotspot of 398 cd/m
2
at the center of the north wall and 95.7 cd/m
2
at the center of the
floor (Fig. 4-14).
Figure 4-14:Illuminance false color (left) and luminance false color (right) for Test3
The illuminance at the center of the north wall is 2286 fc and that on the floor is 9.5 fc (Fig. 4-15). Even the
raytracing does not result in spot on the floor.
Figure 4-15:Calculation points on north wall (left) and floor (right) for Test3
58
4.1.3 Results of simulation in Ecotect/Radiance
A number of runs were simulated in Ecotect/Radiance (Refer to Section 5.1.3). The final outputs do show the
expected results and has been concluded that it is possible to get the anticipated results in Ecotect/Radiance.
Test 1: A Open
The results of the test show a distinct spot on the north wall and the floor as anticipated when the illuminance
calculation is done. The internal specular reflections are taken into account as Radiance uses the raytracing
method for rendering. However, the north wall behaves like a mirror in the luminance calculation and a
reflection of the opening and the spot on the floor can be seen and the spot at the center of the north wall
cannot be seen, which is correct (Fig. 4-16).
Figure 4-16: Test 1 renderings in Ecotect/Radiance - illuminance (left) and luminance (right)
The illuminance false color shows that the illuminance level at the location of the spot on the north wall is
37655.6 lux and that on the floor is 34660.3 lux. The luminance levels at these two locations are 4.3 cd/m
2
and
10485.7 cd/m
2
respectively (Fig. 4-17).
59
Figure 4-17:Test 1 false color diagrams in Ecotect/Radiance - illuminance (left) and luminance (right)
Test 2: B Open
The results of the test show a distinct spot on the north wall and the floor as anticipated when the illuminance
calculation is done. The internal specular reflections are taken into account as Radiance uses the raytracing
method for rendering. However, the north wall behaves like a mirror in the luminance calculation and a
reflection of the opening and the spot on the floor can be seen and the spot at the center of the north wall
cannot be seen, which is correct (Fig. 4-18). Due to a slight difference in altitude, the light hits the north wall a
little higher than the center and is reflected in the opposite direction almost where the south wall starts. The
reflection of the reflected spot on the south wall is seen at the bottom of the north wall in the luminance image.
Figure 4-18: Test 2 renderings in Ecotect/Radiance - illuminance (left) and luminance (right)
60
The illuminance false color shows that the illuminance level at the location of the spot on the north wall is
56531.0 lux and that on the floor is 5287.9 lux. The luminance levels at these two locations are 8342 cd/m
2
and
1580.6 cd/m
2
respectively (Fig. 4-19).
Figure 4-19:Test 2 false color diagrams in Ecotect/Radiance - illuminance (left) and luminance (right)
Test 3: C Open
The results of the test show a distinct spot on the north wall and the floor as anticipated when the illuminance
calculation is done. The internal specular reflections are taken into account as Radiance uses the raytracing
method for rendering. However, the north wall behaves like a mirror in the luminance calculation and a
reflection of the opening and the spot on the floor can be seen and the spot at the center of the north wall
cannot be seen (Fig. 4-20). This is as it should be.
61
Figure 4-20:Test 3 renderings in Ecotect/Radiance - illuminance (left) and luminance (right)
The illuminance false color shows that the illuminance level at the location of the spot on the north wall is
42435.3 lux and that on the floor is 34641.1 lux. The luminance levels at these two locations are 3.5 cd/m
2
and
10524.2 cd/m
2
respectively (Fig. 4-21).
Figure 4-21: Test 3 false color diagrams in Ecotect/Radiance - illuminance (left) and luminance (right)
4.1.4 Results of simulation in Rhino/DIVA
A number of runs were simulated in Rhino/DIVA (Refer to Section 5.1.4). The final outputs show inconsistent
performance and it has been concluded that it is not possible to get the anticipated results in Rhino/DIVA.
62
Test 1: A Open
It is not possible to create illuminance visualizations with DIVA. The luminance visualization behave like those
from Ecotect/Radiance. Thus, it can be assumed that the illuminance results would have shown the spot on the
floor as anticipated. The internal specular reflections are taken into account as Radiance uses the raytracing
method for rendering. The north wall behaves like a mirror and a reflection of the opening and the spot on the
floor can be seen, which is correct (Fig. 4-22). The focal length of the camera has been decreased intentionally to
see as much the floor and ceiling as much as possible.
Figure 4-22: Test 1 luminance rendering in Rhino/DIVA
The illuminance grid-based calculation show the illuminance levels at the center of the north wall and at the
spot of the reflection on the floor to be 40 lux and 11 lux respectively. The luminance false color image shows
that the luminance levels at the point where the opening is reflected to be 2416 cd/m
2
(Fig. 4-23).
Figure 4-23: Test 1 luminance false color rendering in Rhino/DIVA
Test 2: B Open
It is not possible to create illuminance visualizations with DIVA. The luminance visualization behave like those
from Ecotect/Radiance. Thus, it can be assumed that the illuminance results would have shown the spot on the
floor as anticipated. The internal specular reflections are taken into account as Radiance uses the raytracing
63
method for rendering. However, the north wall behaves like a mirror and a reflection of the opening can be seen
(Fig. 4-24). The focal length of the camera has been decreased intentionally to see as much the floor and ceiling
as much as possible. Due to a slight difference in altitude, the light hits the north wall a little higher than the
center and is reflected in the opposite direction almost where the south wall starts. The reflection of the
reflected spot on the south wall is seen at the bottom of the north wall in the luminance image.
Figure 4-24: Test 2 luminance rendering in Rhino/DIVA
The illuminance grid-based calculation show the illuminance levels at the center of the north wall and at the
spot of the reflection on the floor to be 3261 lux and 14 lux respectively. The luminance false color shows that
the luminance level at the location of the spot on the north wall is 5395.021 cd/m
2
and on the floor to be less
than 10 cd/m
2
(Fig. 4-25).
Figure 4-25:Test 2 luminance false color rendering in Rhino/DIVA
64
Test 3: C Open
It is not possible to create illuminance visualizations with DIVA. The luminance visualization behave like those
from Ecotect/Radiance. Thus, it can be assumed that the illuminance results would have shown the spot on the
floor as anticipated. The internal specular reflections are taken into account as Radiance uses the raytracing
method for rendering. However, the north wall behaves like a mirror and a reflection of the opening and the
spot on the floor can be seen and the spot at the center of the north wall cannot be seen, which is correct (Fig.
4-26). The focal length of the camera has been decreased intentionally to see as much the floor and ceiling as
much as possible.
Figure 4-26:Test 3 luminance rendering in Rhino/DIVA
The illuminance grid-based calculation show the illuminance levels at the center of the north wall and at the
spot of the reflection on the floor to be 47 lux and 18 lux respectively. The luminance false color shows that the
luminance level at the center of the north wall is very low and that on the floor is approximately 550 cd/m
2
(Fig.
4-27).
Figure 4-27:Test 3 luminance false color rendering in Rhino/DIVA
65
4.1.5 Results of simulation in Rhino/Grasshopper/Honeybee
A number of runs were simulated in Rhino/Grasshopper/Honeybee (Refer to Section 5.1.5). The final outputs do
show the expected results and it has been concluded that it is possible to get the anticipated results in
Rhino/Grasshopper/Honeybee.
Test 1: A Open
The results of the test show a distinct spot on the north wall and the floor as anticipated when the illuminance
calculation is done. The internal specular reflections are taken into account as Radiance uses the raytracing
method for rendering. However, the north wall behaves like a mirror in the luminance calculation and a
reflection of the opening and the spot on the floor can be seen and the spot at the center of the north wall
cannot be seen, which is correct (Fig. 4-28). The focal length of the camera has been decreased intentionally to
see as much the floor and ceiling as much as possible.
Figure 4-28:Test 1 renderings in Honeybee - illuminance (left) and luminance (right)
The illuminance false color shows that the illuminance level at the location of the spot on the north wall is
46757.25 lux and that on the floor is 101.96 lux. The luminance levels at these two locations are 22.77 cd/m
2
and 56.37 cd/m
2
respectively (Fig. 4-29).
Figure 4-29: Test 1 false color renderings in Honeybee - illuminance (left) and luminance (right)
66
Test 2: B Open
The results of the test show a distinct spot on the north wall and the floor as anticipated when the illuminance
calculation is done. The internal specular reflections are taken into account as Radiance uses the raytracing
method for rendering. Due to a slight difference in altitude, the light hits the north wall a little higher than the
center and is reflected in the opposite direction almost where the south wall starts. The reflection of the
reflected spot on the south wall is seen at the bottom of the north wall in the luminance image (Fig. 4-30).
However, a reflection of the opening cannot be seen, which is incorrect.
Figure 4-30:Test 2 renderings in Honeybee - illuminance (left) and luminance (right)
The illuminance false color shows that the illuminance level at the location of the spot on the north wall is
281.23 lux and that on the floor is 324.71lux. The luminance levels at these two locations are 13.23cd/m
2
and
1784.28 cd/m
2
respectively (Fig. 4-31).
Figure 4-31: Test 2 false color renderings in Honeybee - illuminance (left) and luminance (right)
Test 3: C Open
The results of the test show a distinct spot on the north wall and the floor as anticipated when the illuminance
calculation is done. The internal specular reflections are taken into account as Radiance uses the raytracing
method for rendering. However, the north wall behaves like a mirror in the luminance calculation and a
reflection of the opening and the spot on the floor can be seen and the spot at the center of the north wall
cannot be seen, which is correct (Fig. 4-32).
67
Figure 4-32: Test 3 renderings in Honeybee - illuminance (left) and luminance (right)
The illuminance false color shows that the illuminance level at the location of the spot on the north wall is
213.27 lux and that on the floor is 123.63 lux. The luminance levels at these two locations are 22.46 cd/m
2
and
48.60 cd/m
2
respectively (Fig. 4-33).
Figure 4-33: Test 2 false color renderings in Honeybee - illuminance (left) and luminance (right)
4.1.6 Conclusion of specular reflection test
The most important aspect of this test is the visual aspect. As such, Ecotect/Radiance and
Rhino/Grasshopper/Honeybee pass the test while AGi32 does not, and it cannot be determined with certainty
that Rhino/DIVA passes the test or not.
The illuminance values recorded show a widespread discrepancy among the software packages for all the tests.
Only Ecotect/Radiance maintains some consistency across all the tests (Fig. 4-34, 4-35 and 4-36). For the
illuminance values, the software packages are not compared against the physical model as only the luminance
information was available for the physical model testing.
68
Figure 4-34: Illuminance values in footcandles recorded for Test1 in the four software packages.
Figure 4-35:Illuminance values in footcandles recorded for Test2 in the four software packages.
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
AGi32 Ecotect/Radiance Rhino/DIVA Rhino/Honeybee
NorthWall
Floor
0
1000
2000
3000
4000
5000
6000
AGi32 Ecotect/Radiance Rhino/DIVA Rhino/Hobeybee
NorthWall
Floor
Illuminance
(fc)
Illuminance Values for NW and Floor for Test1
Illuminance
(fc)
Illuminance Values for NW and Floor for Test2
69
Figure 4-36: Illuminance values in footcandles recorded for Test 3 in the four software packages.
The luminance values recorded show a widespread discrepancy among the software packages for all the tests.
Only Ecotect/Radiance comes close to the values recorded in the physical model for all the tests (Fig. 4-37, 4-38
and 4-39).
Figure 4-37: Luminance values in cd/m
2
recorded for Test 1 in the physical model and four software packages
0
500
1000
1500
2000
2500
3000
3500
4000
4500
AGi32 Ecotect/Radiance Rhino/DIVA Rhino/Hobeybee
NorthWall
Floor
0
2000
4000
6000
8000
10000
12000
NorthWall
Floor
Illuminance
(fc)
Illuminance Values for NW and Floor for Test3
luminance
(cd/m
2
)
Luminance Values for NW and Floor for Test1
70
Figure 4-38:Luminance values in cd/m
2
recorded for Test 2 in the physical model and four software packages
Figure 4-39:Luminance values in cd/m
2
recorded for Test 3 in the physical model and four software packages
0
2000
4000
6000
8000
10000
12000
14000
NorthWall
Floor
0
2000
4000
6000
8000
10000
12000
NorthWall
Floor
luminance
(cd/m
2
)
Luminance Values for NW and Floor for Test3
luminance
(cd/m
2
)
Luminance Values for NW and Floor for Test2
71
4.2 Reflection test
The anticipation of the test was that at a given time on a given day if a ray of light enters through a single
opening and hits a 100% reflective surface, the illuminance at that point should equal the exitance, which is the
total luminous flux density leaving a surface, at that point and for a 50% reflective surface, the illuminance at
that point should equal twice the exitance at that point.
4.2.1 Results of physical model test
Test 1: 100% Reflective wall
The model was photographed using a fish eye lens at four different exposure settings (1/2, 1/15, 1/125 and
1/1000) with aperture size 3.8 (Fig. 4-40). White foam core was used as the 100% reflective north wall, although
it is recognized that this is only an approximation. It can be seen that the hotspot at the center of the north wall
becomes more distinct with decreasing the exposure setting.
Figure 4-40:Fish eye images taken at 1/2, 1/15, 1/125 and 1/1000 (from left to right) exposure settings.
The four photographs are then combined in Photolux software to create a false color image (Fig. 4-41). The false
color shows the luminance levels at the center of the north wall is 10000 cd/m
2
.
Figure 4-41:False Color created in Photolux software showing luminances.
72
Test 2: 50% Reflective wall
The model was photographed using a fish eye lens at four different exposure settings (1/2, 1/15, 1/125 and
1/1000) with aperture size 3.8 (Fig. 4-42). Grey paper on foam core was used as the 50% reflective north wall.
Figure 4-42:Fish eye images taken at 1/2, 1/15, 1/125 and 1/1000 (from left to right) exposure settings.
The four photographs are then combined in Photolux software to create a false color image (Fig. 4-43). The false
color shows the luminance levels at the center of the north wall is 5440 cd/m
2
.
Figure 4-43: False Color created in Photolux software showing luminances.
4.2.2 Results of simulation in AGi32
The final outputs do show the expected results and upon exploration of the tool, it has been concluded that it is
possible to get the anticipated results in Agi32.
Test 1: 100% Reflective wall
The calculation values for the north wall shows the illuminance at the center of the wall to be 3171 fc and the
exitance to be 3139 fc (Fig. 4-44). This is just a 1% decrease and it can be safely concluded that AGi takes into
account the reflectance values of the surfaces.
73
Figure 4-44: Illuminance levels on the north wall (top) and exitance levels on the north wall (bottom).
The rendering and the false color shows the luminance level at the center of the north wall to be 10000 cd/m
2
and thus behaves like the physical model (Fig. 4-45).
Figure 4-45: Radiosity rendering and luminance false color in AGi32 for a 100% reflective wall test.
Test 2: 50% Reflective wall
The calculation values for the north wall shows the illuminance at the center of the wall to be 3171 fc and the
exitance to be 1592 fc (Fig. 4-46). According to what was anticipated for a 50% reflective wall, the exitance
should be 1585.5 fc. The test results show a 0.4% increase and it can be safely concluded that AGi takes into
account the reflectance values of the surfaces.
74
Figure 4-46:Illuminance levels on the north wall (top) and exitance levels on the north wall (bottom).
The rendering and the false color shows the luminance level at the center of the north wall to be 5000 cd/m
2
and thus behaves like the physical model (Fig. 4-47).
Figure 4-47:Radiosity rendering and luminance false color in AGi32 for a 50% reflective wall test.
4.2.3 Results of simulation in Ecotect/Radiance
The final outputs do show the expected results and upon exploration of the tool, it has been concluded that it is
possible to get the anticipated results in Ecotect/Radiance.
Test 1: 100% Reflective wall
It is not possible to calculate exitance in Ecotect/Radiance hence illuminance and luminance plots will be
compared.Using rule of thumb relationships, exitance is the total luminous flux density leaving a surface
measured in lumens per square foot whereas luminance is the luminous flux density leaving a surface in a
particular direction measured in lumens per square foot (Schiler 1992). Since the importance lies at the point
75
where the light hits the center of the north wall, where the light path has not yet made a difference, luminance
can be used for the same comparison. According to the formula (Eq. 4-1),
Luminance = Illuminance x reflectance (Schiler 1992),
L = E x 1……………(Eq. 4-1)
The Ecotect/Radiance calculations show that the illuminance and the luminance at the center of the north wall
are 17204.6 lux and 17941.8 cd/m
2
(Fig. 4-48). This is a 4.2% increase and it can be safely concluded that
Ecotect/Radiance does take into account the reflectance of surfaces. It should be noted that the illuminance and
luminance distributions on the rest of the surfaces look different since the rest of the surfaces have almost 0
reflectance value.
Illuminance Plot
Luminance Plot
Illuminance Contour Lines
Luminance Contour Lines
76
Illuminance FalseColor
Luminance FalseColor
Figure 4-48: Comparison between Illuminance and Luminance distributions for a 100% reflective surface in Ecotect/Radiance
Test 2: 50% Reflective wall
According to the formula (Eq. 4-2),
Luminance = Illuminance x reflectance (Schiler 1992),
L = E x 50/100
L = E/2…………….(Eq. 4-2)
The Ecotect/Radiance calculations show that the illuminance and the luminance at the center of the north wall
are 14222.9 lux and 7247.2 cd/m
2
(Fig. 4-49). According to the formula, for a 50% reflective surface, the
luminance at the center of the north wall should be 7114.95 cd/m
2
. The calculation shows an increase of 1.8%
and it can be safely concluded that Ecotect/Radiance does take into account the reflectance of surfaces. It
should be noted that the illuminance and luminance distributions on the rest of the surfaces look different since
the rest of the surfaces have almost 0 reflectance value.
Illuminance Plot
Luminance Plot
77
Illuminance Contour Lines Luminance Contour Lines
Illuminance FalseColor
Luminance FalseColor
Figure 4-49: Comparison between Illuminance and Luminance distributions for a 50% reflective surface in Ecotect/Radiance
4.2.4 Results of simulation in Rhino/DIVA
The final outputs do not show the expected results and it has been concluded that Rhino/DIVA does not take
into account the reflectivity of surfaces.
Test 1: 100% Reflective wall
It is not possible to render illuminance visualizations in Rhino/DIVA, hence grid-based illuminance values and
luminance values determined from the visualizations will be compared (Eq. 4-1).
The calculations show that the illuminance and the luminance at the center of the north wall are 311 lux and
396 cd/m
2
(Fig. 4-50). This is a 21.4% increase and for this scenario, Rhino/DIVA does not take into account the
reflectance of surfaces.
78
Figure 4-50: Illuminance Plot and Luminance rendering in Rhino/DIVA for 100% reflective surface
Test 2: 50% Reflective wall
It is not possible to render illuminance visualizations in Rhino/DIVA, hence grid-based illuminance values and
luminance values determined from the visualizations will be compared (Eq. 4-2).
The calculations show that the illuminance and the luminance at the center of the north wall are 300 lux and
284 cd/m
2
(Fig. 4-51). According to Equation 4.2, the luminance value at the center of the north wall should be
150 cd/m
2
. This is a 47.1% increase and thus, Rhino/DIVA does take into account the reflectivity of surfaces.
Figure 4-51:Illuminance Plot and Luminance render in Rhino/DIVA for 50% reflective surface
4.2.5 Results of simulation in Rhino/Grasshopper/Honeybee
The final outputs do not show the expected results and upon exploration of the tool, it has been concluded that
it is not possible to get the anticipated results in Rhino/Grasshopper/Honeybee.
Test 1: 100% Reflective wall
It is not possible to calculate exitance in Rhino/Grasshopper/Honeybee, hence illuminance and luminance plots
will be compared (Eq. 4-1).
79
The Rhino/Grasshopper/Honeybee calculations show that the illuminance and the luminance at the center of
the north wall are 257.95 lux and 3.82 cd/m
2
(Fig. 4-52). This is a 98.5% increase and it can be safely concluded
that Rhino/Grasshopper/Honeybee does not take into account the reflectance of surfaces. It should be noted
that the illuminance and luminance distributions on the rest of the surfaces look different since the rest of the
surfaces have almost 0 reflectance value.
Illuminance Plot
Luminance Plot
Illuminance Contour Lines
Luminance Contour Lines
Illuminance FalseColor
Luminance FalseColor
Figure 4-52: Comparison between Illuminance and Luminance distributions for a 100% reflective surface in Rhino/Grasshopper/Honeybee
Test 2: 50% Reflective wall
The Rhino/Grasshopper/Honeybee calculations show that the illuminance and the luminance at the center of
the north wall are 252.93 lux and 1.95 cd/m
2
(Fig. 4-53). For a 50% reflective surface, the luminance at the
center of the north wall should be 126.4 cd/m
2
(Eq. 4-2). The calculation shows an increase of 98.4% and it can
be safely concluded that Rhino/Grasshopper/Honeybee does not take into account the reflectance of surfaces. It
should be noted that the illuminance and luminance distributions on the rest of the surfaces look different since
the rest of the surfaces have almost 0 reflectance value.
80
Illuminance Plot
Luminance Plot
Illuminance Contour Lines
Luminance Contour Lines
Illuminance FalseColor
Luminance FalseColor
Figure 4-53: Comparison between Illuminance and Luminance distributions for a 50% reflective surface in Rhino/Grasshopper/Honeybee
4.2.6 Conclusion of reflection test
AGi32 and Ecotect/Radiance take into account the reflectivity of the surfaces for all cases whereas Rhino/DIVA
and Rhino/Grasshopper/Honeybee do not (Fig. 4-54).
81
Figure 4-54: Percentage of deviation for all the software packages for the reflectivity tests.
4.3 Luminous Flux Conservation test
If at a given time on a given day if a ray of light enters through a single opening, the total luminous flux at the
opening surface should equal the total luminous flux reaching the different interior surfaces.
4.3.1 Results of simulation in AGi32
A number of tests were simulated in AGi32 (Refer to Section 5.3.1). The final outputs do not show the expected
results for the first two tests and show it for the last two tests and upon exploration of the tool, it has been
concluded that it is not possible to get the anticipated results in Agi32 in all cases.
Test 1: 3x3 Opening
The average illuminance, as given directly by the software and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 0.29 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 0.29 x 144
= 41.76 lumen
0
20
40
60
80
100
120
100% Reflective NW
50% Reflective NW
% Deviation
(upto 20%
allowed by
CIE)
Percentage of deviation shown by software in reflectivity tests
Ideal situation:
no deviation
82
2) Floor
Average illuminance - 34.82 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 34.82 x 144
= 5014.08 lumen
3) Walls - 18.85 + 26.43 + 225.74 + 24.46 fc
Average illuminance - 295.48 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 295.48 x 120
= 35457.6 lumen
Total Luminous Flux on all the interior surfaces combined = 41.76 + 5014.08 + 35457.6 = 40513.44 lumens
4) Opening
Average illuminance - 4623 fc
Area - 2x2 = 4 feet
Luminous Flux - Average illuminance x Area
= 4623 x 4
= 18492 lumen
% Deviation = (40513.44 - 18492)/40513.44 x 100
= 54.35%
Test 2: 6x6 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 1.12 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 1.12 x 144
= 161.28 lumen
2) Floor
Average illuminance - 136.03 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 136.03 x 144
= 19588.32 lumen
83
3) Walls - 76.65 + 109.87 + 933.98 + 101.33 fc
Average illuminance - 1221.83 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 1221.83 x 120
= 146619.6 lumen
Total Luminous Flux on all the interior surfaces combined = 161.28 + 19588.32 + 146619.6 = 166369.2 lumens
4) Opening
Average illuminance - 4623 fc
Area - 3x3 = 9 feet
Luminous Flux - Average illuminance x Area
= 4623 x 9
= 41607 lumen
% Deviation = (166369.2 - 41607)/166369.2 x 100
= 75%
Test 3: 9x9 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 4.97 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 4.97 x 144
= 715.68 lumen
2) Floor
Average illuminance - 292.04 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 292.04 x 144
= 42053.76 lumen
3) Walls - 177.32 + 264.31 + 1875 + 242.70 fc
Average illuminance - 2559.33 fc
84
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 2559.33 x 120
= 307119.6 lumen
Total Luminous Flux on all the interior surfaces combined = 715.68 + 42053.76 + 307119.6 = 349889.04 lumens
4) Opening
Average illuminance - 4617 fc
Area - 9x9 = 81 feet
Luminous Flux - Average illuminance x Area
= 4617 x 81
= 373977 lumen
% Deviation = (349889.04 - 373977)/349889.04 x 100
= 6.4%
Test 4: 12x12 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Floor
Average illuminance - 488.61 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 488.61 x 144
= 70359.84 lumen
2) Walls - 328.26 + 597.3 + 3749 + 472.2 fc
Average illuminance - 5146.76 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 5146.76 x 120
= 617611.2 lumen
Total Luminous Flux on all the interior surfaces combined = 70359.84 + 617611.2 = 687971.04 lumens
3) Opening
Average illuminance - 4617 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 4617 x 144
= 664848 lumen
% Deviation = (687971.04 - 664848)/687971.04 x 100 = 3.3%
85
It is however interesting to note that the percentage of deviation decreases significantly as the size of the
opening increases (Fig. 4-55). This is possibly due to the fact that as the size of the opening increases, the
resolution and the number of points on it increase and as such the reading on it becomes more accurate.
Figure 4-55: Graph showing percentage of deviation in AGi32 for all the tests.
4.3.2 Results of simulation in Ecotect/Radiance
The test was conducted in Ecotect/Radiance with successful results for the first test and unsuccessful results for
the other three. The calculation of average illuminance on the surfaces in Ecotect/Radiance was done by hand
(Refer to Section 5.3.2). It was concluded that Ecotect/Radiance does not conserve luminous flux at most times.
Test 1: 3x3 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 0.37 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 0.37 x 144
= 53.28 lumen
0
10
20
30
40
50
60
70
80
Test 1 3x3 Test 2 6x6 Test 3 9x9 Test 4 12x12
Luminous Flux Conservation in AGi32
% Deviation
%
Deviation
Ideal situation:
no deviation
86
2) Floor
Average illuminance - 169 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 169 x 144
= 24336 lumen
3) Walls - 5.26 + 5.75 + 8.29 + 7.02fc
Average illuminance - 26.32 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 26.32 x 120
= 3158.4 lumen
Total Luminous Flux on all the interior surfaces combined = 53.28 + 24336 + 3158.4 = 27547.68 lumens
4) Opening
Average illuminance - 7638.31 fc
Area - 2x2 = 4 feet
Luminous Flux - Average illuminance x Area
= 7638.31 x 4
= 30553.24 lumen
% Deviation = (30553.24 - 27547.68)/30553.24 x 100
= 9.8%
Test 2: 6x6 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 1.75 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 1.75 x 144
= 252 lumen
2) Floor
Average illuminance - 371.35 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 371.35 x 144
= 534744 lumen
87
3) Walls - 59.05 + 66.54 + 89.09 + 74.41 fc
Average illuminance - 289.09 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 289.09 x 120
= 34690.8 lumen
Total Luminous Flux on all the interior surfaces combined = 252 + 534744 + 34690.8 = 88417.2 lumens
4) Opening
Average illuminance - 8313.83 fc
Area - 3x3 = 9 feet
Luminous Flux - Average illuminance x Area
= 8313.83 x 9
= 74824.47 lumen
% Deviation = (88417.2 - 74824.27)/88417.2 x 100
= 15.3%
Test 3: 9x9 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 10.03 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 10.03 x 144
= 1444.32 lumen
2) Floor
Average illuminance - 5641.37 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 5641.37 x 144
= 812357.28 lumen
3) Walls - 160.58 + 184.44 + 242.17 + 208.73 fc
Average illuminance - 795.92 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 795.92 x 120
= 95510.4 lumen
88
Total Luminous Flux on all the interior surfaces combined = 1444.32 + 812357.28 + 95510.4 = 909312 lumens
4) Opening
Average illuminance - 8313.64 fc
Area - 9x9 = 81 feet
Luminous Flux - Average illuminance x Area
= 8313.64 x 81
= 673404.84 lumen
% Deviation = (909312 - 673404.84)/909312 x 100
= 25.9%
Test 4: 12x12 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Floor
Average illuminance - 7546.03 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 7546.03 x 144
= 1086628.32 lumen
2) Walls - 469.39 + 2253.45 + 6921.77 + 7447.09 fc
Average illuminance - 17091.7 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 17091.7 x 120
= 2051004 lumen
Total Luminous Flux on all the interior surfaces combined = 1086628.32 + 2051004 = 3137632.32 lumens
3) Opening
Average illuminance - 8313.83 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 8313.83 x 144
= 1197191.52 lumen
% Deviation = (3137632.32 - 1197191.52)/3137632.32 x 100 = 61.8%
The percentage of deviation does not change much with the size of the opening except for the last test when
the opening is the size of the entire ceiling (Fig. 4-56). This should happen the other way around.
89
Figure 4-56: Graph showing percentage of deviation in Ecotect/Radiance for all the tests.
4.3.3 Results of simulation in Rhino/DIVA
The test was conducted in Rhino/DIVA with unsuccessful results for all the tests. The average illuminances were
calculated by hand (Refer to Section 5.3.3). It was concluded that Rhino/DIVA does not conserve luminous flux at
anytime.
Test 1: 3x3 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 0 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 0. x 144
= 0lumen
2) Floor
Average illuminance - 4.44 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 4.44 x 144
= 639.36 lumen
0
10
20
30
40
50
60
70
Test 1 3x3 Test 2 6x6 Test 3 9x9 Test 4 12x12
Luminous Flux Conservation in Ecotect/Radiance
% Deviation
%
Deviation
Ideal situation:
no deviation
90
3) Walls - 17.22 + 4.33 + 3.22 + 3.77 fc
Average illuminance - 28.54 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 28.54 x 120
= 3424.8 lumen
Total Luminous Flux on all the interior surfaces combined = 639.36 + 3424.8 = 4064.16 lumens
4) Opening
Average illuminance - 529 fc
Area - 2x2 = 4 feet
Luminous Flux - Average illuminance x Area
= 529 x 4
= 2116 lumen
% Deviation = (4064.16 - 2116)/4064.16 x 100 = 47.9%
Test 2: 6x6 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 0.208 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 0.208 x 144
= 29.95 lumen
2) Floor
Average illuminance - 17.55 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 17.55 x 144
= 2527.2 lumen
3) Walls - 220.22 + 17.77 + 13.44 + 16.55 fc
Average illuminance - 267.98 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 267.98 x 120
= 32157.6 lumen
Total Luminous Flux on all the interior surfaces combined = 29.95 + 2527.2 + 32157.6 = 34714.75 lumens
91
4) Opening
Average illuminance - 529 fc
Area - 3x3 = 9 feet
Luminous Flux - Average illuminance x Area
= 529 x 9
= 4761 lumen
% Deviation = (34714.75 - 4761)/34714.75 x 100
= 86.2%
Test 3: 9x9 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 0.79 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 0.79 x 144
= 113.76 lumen
2) Floor
Average illuminance - 38 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 38 x 144
= 5472 lumen
3) Walls - 327.77 + 41.22 + 28.55 + 37.55 fc
Average illuminance - 435.09 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 435.09 x 120
= 52210.8 lumen
Total Luminous Flux on all the interior surfaces combined = 113.76 + 5472+ 52210.8 = 57796.56 lumens
4) Opening
Average illuminance - 529.33 fc
Area - 9x9 = 81 feet
Luminous Flux - Average illuminance x Area
= 529.33 x 81
= 42875.73 lumen
92
% Deviation = (57796.56 - 42875.73)/57796.56 x 100
= 25.8%
Test 4: 12x12 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Floor
Average illuminance - 62.66 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 62.66 x 144
= 9023.04 lumen
2) Walls - 468.88 + 69.77 + 41.22 + 57.33 fc
Average illuminance - 637.2 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 637.2x 120
= 76464 lumen
Total Luminous Flux on all the interior surfaces combined = 9023.04 + 76464 = 85487.04 lumens
3) Opening
Average illuminance - 529.66 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 529.66 x 144
= 76271.04 lumen
% Deviation = (85487.04 - 76271.04)/85487.04 x 100 = 10.7%
The percentage of deviation decreases with an increase in the size of the opening (Fig. 4-57). This is possibly due
to the fact that as the size of the opening increases, the resolution and the number of points on it increase and
as such the reading on it becomes more accurate.
93
Figure 4-57:Graph showing percentage of deviation in Rhino/DIVA for all the tests.
4.3.4 Results of simulation in Rhino/Grasshopper/Honeybee
The test was conducted in Rhino/Grasshopper/Honeybee with unsuccessful results for all the tests. It was
concluded that Rhino/Grasshopper/Honeybee does not conserve luminous flux at anytime
Test 1: 3x3 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 0.91 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 0.91 x 144
= 131.04 lumen
2) Floor
Average illuminance - 13.02 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 13.02 x 144
= 1874.88 lumen
0
10
20
30
40
50
60
70
80
90
100
Test1 3x3 Test2 6x6 Test3 9x9 Test4 12x12
Luminous Flux Conservation in Rhino/DIVA
% Deviation
%
Deviation
Ideal situation:
no deviation
94
3) Walls - 37.65 + 13.39 + 11.14 + 12.5 fc
Average illuminance - 74.68 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 74.68 x 120
= 8961.6 lumen
Total Luminous Flux on all the interior surfaces combined = 131.04 + 1874.88 + 8961.6 = 10967.52 lumens
4) Opening
Average illuminance - 5140.535 fc
Area - 2x2 = 4 feet
Luminous Flux - Average illuminance x Area
= 5140.535 x 4
= 20562.14 lumen
% Deviation = (20562.14 - 10967.52)/20562.14 x 100
= 46.6%
Test 2: 6x6 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 3.5 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 3.5 x 144
= 504 lumen
2) Floor
Average illuminance - 50.74 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 50.74 x 144
= 7306.56 lumen
3) Walls - 1756.71 + 56.52 + 45.22 + 52.16 fc
Average illuminance - 1910.61 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 1910.61 x 120
= 229273.2 lumen
Total Luminous Flux on all the interior surfaces combined = 504 + 7306.56 + 229273.2 = 237083.76 lumens
95
4) Opening
Average illuminance - 5140.79 fc
Area - 3x3 = 9 feet
Luminous Flux - Average illuminance x Area
= 5140.79 x 9
= 46267.11 lumen
% Deviation = (237083.76 - 46267.11)/237083.76 x 100
= 80.4%
Test 3: 9x9 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Ceiling
Average illuminance - 7.66 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 7.66 x 144
= 1103.04 lumen
2) Floor
Average illuminance - 109.96 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 109.96 x 144
= 15834.24 lumen
3) Walls - 3556.9 + 132.8 + 99.7 + 122.9 fc
Average illuminance - 3912.3 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 3912.3 x 120
= 469476 lumen
Total Luminous Flux on all the interior surfaces combined = 1103.04 + 15834.24 + 469476= 486413.28 lumens
4) Opening
Average illuminance - 5140.5 fc
Area - 9x9 = 81 feet
Luminous Flux - Average illuminance x Area
= 5140.5 x 81
= 416380.5 lumen
96
% Deviation = (486413.28 - 416380.5)/486413.28 x 100
= 14.3%
Test 4: 12x12 Opening
The average illuminance and luminous flux on all surfaces were as follow:
1) Floor
Average illuminance - 185.45 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 185.45 x 144
= 26704.8 lumen
2) Walls - 6844.7 + 299.3 + 150.04 + 193.5 fc
Average illuminance - 7487.54 fc
Area - 12x10 = 120 feet
Luminous Flux - Average illuminance x Area
= 7487.54 x 120
= 898504.8 lumen
Total Luminous Flux on all the interior surfaces combined = 26704.8 + 898504.8 = 925209.6 lumens
3) Opening
Average illuminance - 5140.49 fc
Area - 12x12 = 144 feet
Luminous Flux - Average illuminance x Area
= 5140.49 x 144
= 740230.56 lumen
% Deviation = (925209.6 - 740230.56)/925209.6 x 100 = 19.9%
The percentage of deviation follows no particular order with the change in the size of the opening (Fig. 4-58).
However, it shows a significant decrease with the increase in the size of the opening. This is possibly due to the
fact that as the size of the opening increases, the resolution and the number of points on it increase and as such
the reading on it becomes more accurate.
97
Figure 4-58: Graph showing percentage of deviation in Rhino/Grasshopper/Honeybee for all the tests.
4.3.5 Conclusion of Luminous flux conservation test
The percentages of deviation recorded is erratic for all the software packages except Ecotect/Radiance, which
shows a regular pattern with the changing size of the opening (Fig. 4-59).
Figure 4-64: Percentage of deviation shown by the software packages with increasing size of opening
0
10
20
30
40
50
60
70
80
90
Test 1 3x3 Test 2 6x6 Test 3 9x9 Test 4 12x12
Luminous Flux Conservation in Rhino/Grasshopper/Honeybee
% Deviation
0
10
20
30
40
50
60
70
80
90
100
AGi32 Ecotect/Radiance Rhino/DIVA Rhino/Hobeybee
Test1
Test2
Test3
Test4
Ideal situation:
no deviation
%
deviation
%
deviation
Percentage of deviation shown by all software in all tests
Ideal situation:
no deviation
98
Chapter 5
Evaluation and critique of simulation results
5.0 Executive summary of the tests
The software packages performed differently in the different tests. While Agi32 performed well in the
reflectivity tests, it did not do so in the specular reflection or the luminous flux conservation test.
Ecotect/Radiance performed the best with passing the specular reflection test and the reflectivity test.
Rhino/DIVA did not give accurate results for any of the three tests while Rhino/Grasshopper/Honeybee gave
accurate results for only the specular reflection test.
5.1 Specular reflection test
The anticipation of the test was that at a given time on a given day if a ray of light enters through a single
opening and hits a highly specular surface, it should be reflected to form a similar spot on another surface such
that the angle of incidence is equal to the angle of reflection.
5.1.1 Discussion of physical model test
The physical model was made of white foam core of 2" thickness at a scale of 3/8" = 1'. A number of materials
were studied to find the most specular one for the north wall using a gloss meter (Fig. 5-1). The units are in Gloss
Units (GU). The gloss meter can measure a range of 0 - 1000 GU at 60° and 0 - 2000 GU at 20°.
Figure 5-1: From left to right - a)White foam core (4.6), b)Black foam core (1.0), c)Grey paper (2.7), d) Stainless Steel plate (301.1), e)
Double sided gold plastic (546.0), f) Silver plastic sheet (722.1), g) Aluminum sheet (544.6)
The particular gloss meter used conforms to BS EN ISO 2813:2000 as one of the many international standards
and that used in the United States. According to this standard, the gloss value of a surface is equal to 100
multiplied to the ratio of the luminous flux reflected by a surface to that reflected by a standard glass piece
having a refractive index of 1567 at a wavelength of 5867 nm for a specified reflection angle (60 used) in
specular fashion, measured in gloss units and is influenced by the roughness, texture and structure of a surface.
"It is not permitted to interpret and express gloss values as % reflection" (ISO 2814:2014).
99
Consequently, the silver plastic sheet was used for the north wall but the exact value for its specularity was
approximated in the software simulations as 95%. Additionally, since a solar gnomon was used to simulate the
date and time when a direct ray from the sun would enter through each of the openings and strike the center of
the north wall, on different days, the intensity of light was different in the physical model and the software
simulations. As such, the illuminance values recorded in the physical model is significantly different from the
ones seen in the software simulations and for the purposes of the specular reflection test, the physical models
are referred to for a pictorial evidence only and also for seeing the luminance values at the center of the north
wall and the floor.
Different materials were chosen for the different surfaces (Table 5-1). Light leaks were prevented by using
electrical black tape and three solar gnomons were pasted on top of the ceiling for the three different north
lines in the three different tests (Fig. 5-2).
Testing
Terminology
North Wall
Silver Plastic
Sheet
South Wall
Grey Paper
West Wall
Grey Paper
East Wall
Grey Paper
Floor
White Foam
Core
Ceiling
White Foam
Core
Reflectance 95 49.7 49.7 49.7 96.7 80
Specularity 95 0 0 0 0 0
Roughness 2 0 0 0 0 0
Table 5-1: Surface properties Notes: n/a - Not Applicable, n/r - Not required
Figure 5-2: Physical model for the specular reflection test
The model was photographed using a fish eye lens with aperture size 3.8 at four different exposure settings
(1/2, 1/15, 1/125 and 1/1000). The four photographs are then combined in Photolux software to create a false
color image which gives the luminances on all the surfaces.
5.1.2 Discussion of simulation in AGi32
According to the surface properties stated in Section 3.1.0 Table 3-1, the corresponding values in Agi32
software were used (Table 5-2) (Fig. 5-3).
100
Testing
Terminology
AGI32
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 0.95 .5 .5 .5 0.96 0.8
Specularity Specularity 0.95 0 0 0 0 0
Roughness n/a n/a n/a n/a n/a n/a
Color Bleed Color Bleed 1 1 1 1 1 1
Surface
Orientation
Surface Type Single Sided Single
Sided
Single
Sided
Single
Sided
Single Sided Single
Sided
Transmittance Transmittance n/a n/a n/a n/a n/a n/a
Calculation
Type
Luminance/
Illuminance
n/r n/r n/r Luminance/
Illuminance
n/r
Table 5-2: Surface properties in AGi32 Notes: n/a - Not Applicable, n/r - Not required
Figure 5-3: Test model setup in Agi32
Calculation points were set up in a 1x1 grid for exitance on the north wall and illuminance on the floor (Fig. 5-4).
Figure 5-4: Test model setup with calculation points
For interior daylighting studies in AGi32, the following steps are necessary (AGi32 V15.3, Help - Contents and
Index - Daylighting):
1) An opening was given in the space through which daylight can come in and a "planar" object defined.
101
2) The definition of the planar object could be - a) 20 - Daylight transition glass (for transparent glazing), b) 21 -
Daylight transition glass (for diffuse glazing), and c) 22 - Daylight transition opening (for an unglazed opening).
The object can be defined in the surface edit window (Fig. 5-5). The important point to note here is that a planar
object had to be modeled and defined even for an unglazed opening.
Figure 5-5: Defining the planar object in the surface edit window
3) There are two kinds of studies that can be conducted - a) daylighting (for one particular date and time) and b)
daylight study (for multiple days or months).
4) The "Daylight" option was checked off (Fig. 5-6) and the necessary parameters of site, true north, date and
time and sky condition were selected (Fig. 5-7). EPW weather files can be downloaded from the EnergyPlus
website for the irradiation values for Perez all weather sky. The default North in AGi32 is the top of the screen
and is considered 90°. It is measured positively counterclockwise from North such that West is 180° and so on
and so forth.
Figure 5-6: Daylight option in the toolbar
102
Figure 5-7: The daylighting parameters window.
Test 1: A Open
The model was set up with the corresponding surface properties and calculation points. The north was set to
45°.The first run did not show any light entering the interior of the box (Fig. 5-8)
Figure 5-8: Run 1 - No light in the interior
For the second run, the north wall was designated as a "Daylight exterior surface" in the surface edit window,
which is what is recommended when daylight studies are conducted on an exterior surface. This time a spot was
seen at the center of the north wall but there was no reflection on the floor (Fig. 5-9).
103
Figure 5-9: Run 2 - Hazy spot at the center of north wall
It is interesting to note that even though there was no visible reflection on the floor, the calculation values
showed a definite hotspot on the north wall and floor, even though the floor hotspot was in a different position
than that anticipated (Fig. 5-10).
Figure 5-10: Hotspot at the center of the north wall (left) and floor (right)
For the third and final run, the technical support team at Lighting Analysts, Inc. who developed AGi32 were
contacted. According to them, the "Daylight exterior surface" option is to be used only when calculating the
daylight levels on an exterior surface, and therefore that is the wrong approach to take. AGi32 has two surface
types for an opaque surface - single-sided and double-sided. A single-sided surface behaves as an interior
surface if it's normal is pointing inwards and like an exterior surface if it's normal is pointing towards the outside.
A double-sided surface acts as both an exterior surface and an interior surface. In an interior daylighting
calculation, AGi32 needs a surface separator through which daylight can enter a building. This is the planar
object placed in the opening on the ceiling. AGi32 sees this object and treats it as a source of light. Thus, if there
was an opening but no correctly defined “skylight” object, daylight would not come into the interior. Again, if
there was no opening and a planar object with daylight transition properties was made floating in the interior,
light would be emitted from it. Daylight transition surfaces do not have the double/single-sided feature since
they are meant to receive daylight on one side and emit it from the other side. Again, if any of the surfaces were
given "Daylight exterior surface" properties, they could be directly interacting with the sunlight provided that
104
they were made "Double-sided" surfaces instead of the default "Single-sided" surface, which affects the
direction towards which it can interact with light. Single-sided daylight exterior surfaces can interact with
daylight provided their normal is directed towards the exterior (Lighting Analysts, Inc. 2015).
For the specular reflection test, a planar object would have to be added having daylight transition opening
properties. However, it will not work unless the default normal direction of the object is flipped from top to
bottom facing the interior of the room since the direction of the normal dictates the direction in which flux
leaves the surface when sunlight passes through it (Fig. 5-11) (Lighting Analysts, Inc. 2015).
Figure 5-11: Flipping the surface normal of the planar object to direct light towards the interior.
Taking the above points into account and making the necessary changes, the third run yielded slightly better
results (Fig. 5-12). Hotspots were found at the center of the north wall and the floor, but a spot on the floor was
still missing (Fig. 5-13, 5-14).
Figure 5-12: Run 3 - Spot missing on the floor
105
Figure 5-13: Run 3 - North wall exitance values
Figure 5-14: Run 3 - Floor illuminance values
The reason behind the spot missing on the floor is that AGi32 uses the radiosity rendering technique which
considers all surfaces as lambertian (diffuse). Radiosity assumes that light is reflected equally in all directions
which is not what happens with a specular surface. The raytrace rendering technique is better suited for that.
However, even with a raytrace, a surface having high reflectance and high specularity appears to be washed out
with light. Therefore, while running a raytrace on AGi32, a surface should be given a high value of specularity
and a low value of reflectance for it to perform as desired (Fig. 5-15) (Lighting Analysts, Inc. 2015).
It is clear from the images that this is not true. There is still no spot on the floor. And a spot has been added to
the north wall that would show ONLY if the surface is lambertian. They are attempting to create an image that
they think displays high specularity and have shown, instead, that they do not know what it should look like. It
should look like the left image, but with a spot on the floor, which is then also reflected in the mirror wall.
106
Figure 5-15: Raytrace with high reflectivity and specularity (left) and raytrace with low reflectivity and high specularity (right)
An additional point to note is that more refined renderings can be created if the mesh level is increased and
"adaptive subdivision" is enabled (Fig. 5-16). This means that each surface will be broken down into a larger
number of surfaces that will each receive light, making the resulting image more refined. The only disadvantage
is that run-time will increase.
Figure 5-16: Rendering with no adaptive subdivision and default mesh levels (left) and rendering with high adaptive subdivision and
increased mesh levels (right) (Lighting Analysts, Inc. 2015)
Test 2 and 3 were conducted like Test 1. The north was set to the default 90°in the first case and 135° in the
second case.
5.1.3 Discussion of simulation in Ecotect/Radiance
According to the surface properties stated in Section 3.1.0 Table 3-1, the corresponding values in
Ecotect/Radiance software were used (Table 5-3) (Fig. 5-17).
107
Testing
Terminology
Ecotect/
Radiance
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 95 50 50 50 96 80
Specularity Shininess 95 0 0 0 0 0
Roughness Roughness 2 0 0 0 0 0
Color Bleed Color Bleed n/a n/a n/a n/a n/a n/a
Surface
Orientation
Surface Type Interior Interior Interior Interior Interior Interior
Transmittance Transparency 0 0 0 0 0 0
Calculation
Type
Luminance/
Illuminance
n/r n/r n/r Luminance/
Illuminance
n/r
Table 5-3: Surface Properties in Ecotect/Radiance Notes: n/a - Not Applicable, n/r - Not required
Figure 5-17: Model setup for Test1 in Ecotect/Radiance
A number of steps have to be followed sequentially for correct simulation in Ecotect/Radiance:
1) The Ecotect file has to be saved in the "C/Radiance" folder without any spaces in its name since Radiance
gives an error if there are spaces in the filename.
2) Different material .RAD files have to be created for the different surfaces and saved in the "C/Radiance"
folder along with the Ecotect file. These files have to be created following the conventions of making Radiance
files.
3) Layers for the different materials have to be created in Ecotect and they have to be named exactly like the
material .RAD files for Radiance to work correctly.
4) It is slightly complicated to make material .RAD files in Ecotect. They can be opened and edited in Microsoft
NotePad. A number of usual materials are discussed (http://www.artifice.com/radiance/rad_materials.html).
108
1) plastic - used for most everyday non-metallic materials
void plastic "name"
0
0
5 red green blue specularity roughness
# red, green, blue are reflectance - values greater than 0.9 are not usually realistic.
# Specularity greater than 0.1 is usually not realistic.
# Roughness varies from 0=perfectly smooth, to 0.5=perfectly rough.
# Roughness greater than 0.4 is usually not realistic.
2) metal
void metal "name"
0
0
5 red green blue specularity roughness
# Specularity greater than 0.9 is typical.
# Roughness greater than 0.2 is usually not realistic.
3) glass
void glass "name"
0
0
3 redTransmission greenTransmission blueTransmission
# A transmission value of 0.96 is typical, for standard 88% transmissivity glass.
4) trans
void trans "name"
0
0
3 redTransmission greenTransmission blueTransmission specularity roughness transmissivity
transmittedSpecular
# Used for translucent materials. Transmissivity is the fraction of light that gets through the material.
transmittedSpecular is is what gets through without being diffusely scattered.
5) mirror
void mirror "name"
0
0
3 redReflectance greenReflectance blueReflectance
109
5) The normals of all the surfaces which determine which direction they are facing and consequently whether
they will be behaving like an exterior surface or an interior surface, have to be checked in Ecotect and flipped to
face the interior.
6) The default north is in the positive y axis direction in Ecotect and it has to be changed accordingly.
Once the model is setup, the radiance calculation is selected and a number of options like calculation type, sky
type, date, and time can be selected (Fig. 5-18).
Figure 5-18:Radiance options selection panel
The last step requires a number of things to be checked off (Fig. 5-19). However, if done once the settings
usually remain as selected till the next time something is changed. It is advisable to check the settings each time
a calculation is run.
Figure 5-19:Radiance options selection panel
110
For the first run, the north wall was given a material definition of an opaque material.
void metal NorthWall
0
0
5 0.95 0.95 0.95 0.95 0.02
The resulting simulations showed the spot at the center of the wall and the reflection of the opening but there
was no reflected spot on the floor (Fig. 5-20).
Figure 5-20: Run 1 for Test 1,2,3 (from left to right) in Ecotect/Radiance
For the next run, the definition of the north wall was changed to that of a mirror.
void mirror NorthWall
0
0
3 0.99 0.99 0.99
The resulting illuminance simulations showed the anticipated results (Fig. 5-21). This means that the word
"mirror" triggers off the effect that a highly specular and reflective opaque north wall should have had. This is a
bug in the system.
111
Figure 5-21: Run2 for Test 1,2,3 (from left to right) in Ecotect/Radiance, illuminance maps.
The luminance simulations, however, performed differently (Fig. 5-22).
Figure 5-22:Run2 for Test 1,2,3 (from left to right) in Ecotect/Radiance, luminance maps.
Radiance uses backward raytracing rendering technique so it takes into account the specularity of surfaces.
However, to get more detailed visualizations at the cost of increased run time, the number of indirect reflections
can be increased from 2 to 5 in the Radiance options selection panel (Fig. 5-23).
112
Figure 5-23:Radiance options selection panel
Radiance also has a "Radiance Control Panel" which can be opened to get details of any tests run on a particular
Ecotect model and images can be opened from different tests (Fig. 5-24).
Figure 5-24: Radiance Control Panel
113
5.1.4 Discussion of simulation in Rhino/DIVA
According to the surface properties stated in Section 3.1.0 Table 3-1, the corresponding values in Rhino/DIVA
software were used (Table 5-4) (Fig. 5-25).
Testing
Terminology
Rhino/DIVA
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 95 50 50 50 96 80
Specularity Shininess 95 0 0 0 0 0
Roughness Roughness 2 0 0 0 0 0
Surface
Orientation
Surface Normal Interior Interior Interior Interior Interior Interior
Calculation
Type
Luminance/
Illuminance
n/r n/r n/r Luminance/
Illuminance
n/r
Table 5-4: Surface Properties in Rhino/DIVA Notes: n/a - Not Applicable, n/r - Not required
Figure 5-25: Model setup for Test1 in Rhino
A number of steps have to be followed sequentially for correct simulation in Rhino/DIVA:
1) A ground plane has to be included.
2) Glazing has to be modeled as a single surface.
3) Layers for the different materials have to be created in Rhino.
4) Default materials can be added and custom materials can be created and added to the material.rad file saved
in C:\DIVA\Daylight folder.
5) The normals of all the surfaces have to be checked and flipped to face the interior.
6) The default north is in the positive y axis direction and it has to be changed accordingly.
Once the model is setup, running simulations is a 4-step process (Fig. 5-26).
114
Figure 5-26: Workflow in DIVA
The location involves selecting an EnergyPlus weather (EPW) file. Adding nodes involves selection of a particular
or multiple surfaces and adding the spacing between nodes and the offset distance of the nodes from the
surface. It is very important to flip the normals of the surfaces to face the interior before adding the nodes. This
ensures that the surfaces behave like interior surfaces and do not interact directly with the sun. The same
materials that were used for the specular reflection test in Ecotect/Radiance were used for the tests in
Rhino/DIVA. Once the materials are added to the base material.rad file in the C:\DIVA\Daylight folder, they
become available in the default material list that comes up when the "Materials" option is selected (Fig. 5-27).
Figure 5-27: The materials tab in Rhino/DIVA
115
A number of simulations are available in DIVA (Fig. 5-28). An illuminance rendering is not one of them, which
was a problem in understanding the capabilities of the software for the specular reflection test.
Figure 5-28: Simulation options tab in DIVA
The first run did not show any spot on the wall or on the floor (Fig. 5-29).
Figure 5-29: Test1 Run1 in Rhino/DIVA
116
For the first run, the Perez custom sky was selected. The diffuse horizontal irradiance is the illumination received
per square area of a surface, not directly from the sun but from what is scattered by the atmosphere and the
direct normal irradiance is the radiation received per square area of a surface directly from the sun when the
surface is perpendicular to the current position of the sun (3TIER website). In AGi32, the diffuse horizontal
irradiance and direct normal irradiance are automatically filled in by the software on selecting the EPW file and
the date and time. In DIVA, these have to be added by the user (Fig. 5-30). Also they have to be converted from
Watts/m
2
to Watts/square feet.
Figure 5-30: Sky type parameters in AGi32 (top) and DIVA (bottom)
117
For the second run, the diffuse horizontal irradiance and direct normal irradiance were added. The result was
that the luminance distribution was comparable to the one given by Ecotect/Radiance (Fig. 5-31). However, It
was still not determined whether DIVA takes into account the specularity of surfaces.
Figure 5-31: Test1 Luminance distributions in Ecotect/Radiance (left) and DIVA (right)
For the third run, grid based illuminance simulation was run to determine if there were any hot spots at the
center of the north wall and the right side of the floor. The results did not show any such spots (Fig. 5-32).
Figure 5-32: Illuminance plots for Test1 on floor (left) and north wall (right)
DIVA cannot create false color plots for orthographic projections or for illuminance distributions. It is still being
developed and has some way to go. Tests 2 and 3 were also conducted in a similar way.
5.1.5 Discussion of simulation in Rhino/Grasshopper/Honeybee
According to the surface properties stated in Section 3.1.0 Table 3-1, the corresponding values in
Rhino/Honeybee software were used (Table 5-5) (Fig. 5-33).
118
Testing
Terminology
Rhino/Honey
bee
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 95 50 50 50 96 80
Specularity Shininess 95 0 0 0 0 0
Roughness Roughness 2 0 0 0 0 0
Surface
Orientation
Surface Normal Interior Interior Interior Interior Interior Interior
Calculation
Type
Luminance/
Illuminance
n/r n/r n/r Luminance/
Illuminance
n/r
Table 5-5: Surface Properties in Rhino/Honeybee Notes: n/a - Not Applicable, n/r - Not required
Figure 5-33: Model setup for Test1 in Rhino
A number of steps have to be followed sequentially for correct simulation in Rhino/Honeybee:
1) The model has to be set to meters and lux units.
2) All the surfaces have to be defined in Honeybee. Radiance materials can be added in the same place. Rhino
layers do not matter.
3) The default north is in the positive y axis direction and it has to be changed accordingly.
Running simulations in Honeybee is a 7-step process (Fig. 5-34).
119
Figure 5-34: Workflow in Honeybee
The same materials that were used for the specular reflection test in Ecotect/Radiance and Rhino/DIVA were
used for the tests in Rhino/Grasshopper/Honeybee(Fig. 5-35). The whole GUI is very transparent and it is very
easy to understand where the process is going wrong or if there is a mistake in setting up the simulation.
Figure 5-35: Defining surface-wise geometry and materials in Honeybee
Adding the sky file requires adding an EPW file and selecting the time and day (Fig. 5-36)
120
Figure 5-36: Adding sky
A number of simulation recipes are available - image based and grid based. For the specular reflection test, the
image based simulation recipe was used (Fig. 5-37). Named Rhino views can be selected and simulation type, i.e.
illuminance or luminance has to be selected. Additional Radiance parameters can be added.
Figure 5-37: Image based analysis recipe in Honeybee
121
The recipe is run by adding the defined geometry and the defined analysis recipe and turning the toggle button
to "True" (Fig. 5-38). Care should be taken to either turn the toggle button back to "False" before closing or
locking the grasshopper canvas before opening a file to prevent the modules from running as soon as a file is re-
opened.
Figure 5-38: Running the analysis
There are a number of ways of visualizing the images produced. Every time a simulation is run, a folder is
created with all the HDR images in the C:\Ladybug folder. An HDR to GIF convertor can also be used to
automatically save the image as a GIF file which can then be opened in regular image viewers. However, a new
plugin called EmbryoViz has been developed which allows the image to be seen in the Grasshopper canvas itself
(Fig. 5-39). Care should be taken however, to disconnect the EmbryoViz module every time the file is closed to
avoid the Grasshopper file from crashing.
122
Figure 5-39: EmbryoViz module for visualizing images in the Grasshopper canvas
False color images can be created through another module that converts the image received after running the
simulation to a false color (Fig. 5-40).
Figure 5-40: Module for creating a false color image in Honeybee
The first illuminance visualization showed the spot at the center of the north wall but not on the floor (Fig. 5-41).
123
Figure 5-41: Test1 run1 in Honeybee
For the second run, radiance parameters were added in which the ambient bounce was raised to 4 from the
default option of 1. The spots at the center of the north wall and the floor were visible now (Fig. 5-42).
Figure 5-42: Increasing the ambient bounce (top) and the resultant visualization of Test1 run2 in Honeybee
124
Honeybee uses Radiance for the actual calculation which uses backward raytracing rendering technique, where
a ray of light from each point on the model goes towards the source of light. If a point on the model cannot see
the source of light, it will not be illuminated or rendered. If the ambient bounce is one, the path of light from a
point to the light source will be terminated at the first obstruction. Increasing it ensures that it bounces multiple
times and thus reaches the source of light (Fig. 5-43). It is also necessary, in order to have ANY specular
simulation.
Figure 5-43: With ab set to 1, the path of light from the point on the table would have terminated at the first contact with the horizontal
blind. Raising it ensures that it bounces and reached the source of light (Reinhart 2010).
Tests 2 and 3 were also conducted in a similar way.
5.2 Reflection test
The anticipation of the test was that at a given time on a given day if a ray of light enters through a single
opening and hits a100% reflective surface, the illuminance at that point should equal the exitance at that point
and for a 50% reflective surface, the illuminance at that point should equal twice the exitance at that point.
5.2.1 Discussion of physical model test
The physical model was made of foam core of 2" thickness at a scale of 3/8" = 1'. Different materials were
chosen for the different surfaces (Table 5-6, 5-7). Light leaks were prevented by using electrical black tape and
three (Fig. 5-44).
125
Testing
Terminology
North Wall
White Foam
Core
South Wall
Black Foam
Core
West Wall
Black Foam
Core
East Wall
Black Foam
Core
Floor
White Foam
Core
Ceiling
Black Foam
Core
Reflectance 100 1 1 1 100 1
Specularity 0 0.5 0.5 0.5 0 0.5
Roughness 0 15 15 15 0 15
Table 5-6: Surface Properties for 100% reflection test; Notes: n/a - Not Applicable, n/r - Not required
Testing
Terminology
North Wall
Grey Paper
South Wall
Black Foam
Core
West Wall
Black Foam
Core
East Wall
Black Foam
Core
Floor
White Foam
Core
Ceiling
Black Foam
Core
Reflectance 50 1 1 1 100 1
Specularity 0 0.5 0.5 0.5 0 0.5
Roughness 0 15 15 15 0 15
Table 5-7: Surface Properties for 50% reflection test; Notes: n/a - Not Applicable, n/r - Not required
Figure 5-44: Physical model for the 100% reflection test
The model was photographed using a fish eye lens with aperture size 3.8 at four different exposure settings
(1/2, 1/15, 1/125 and 1/1000). The four photographs are then combined in Photolux software to create a false
colorimage which gives the luminances on all the surfaces.
5.2.2 Discussion of simulation in AGi32
According to the surface properties stated in Section 3.1.1 Table 3-2 and Table 3-3, the corresponding values in
Agi32 software were used (Table 5-8, 5-9) (Fig. 5-45).
126
Testing
Terminology
AGI32
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 1 0 0 0 1 0
Specularity Specularity 0 0 0 0 0 0
Roughness n/a n/a n/a n/a n/a n/a
Color Bleed Color Bleed 1 1 1 1 1 1
Surface
Orientation
Surface Type Single Sided Single
Sided
Single
Sided
Single
Sided
Single Sided Single
Sided
Transmittance Transmittance n/a n/a n/a n/a n/a n/a
Calculation
Type
Exitance/
Illuminance
n/r n/r n/r n/r n/r
Table 5-8: Surface Properties for 100% reflection test in AGi32; Notes: n/a - Not Applicable, n/r - Not required
Testing
Terminology
AGI32
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 0.5 0 0 0 0 0
Specularity Specularity 0 0 0 0 0 0
Roughness n/a n/a n/a n/a n/a n/a
Color Bleed Color Bleed 1 1 1 1 1 1
Surface
Orientation
Surface Type Single Sided Single
Sided
Single
Sided
Single
Sided
Single Sided Single
Sided
Transmittance Transmittance n/a n/a n/a n/a n/a n/a
Calculation
Type
Exitance/
Illuminance
n/r n/r n/r n/r n/r
Table 5-9: Surface Properties for 50% reflection test in AGi32; Notes: n/a - Not Applicable, n/r - Not required
Figure 5-45: Model setup in AGi32
Conducting the reflection test in AGi32 was very straightforward and the anticipated results were achieved.
127
5.2.3 Discussion of simulation in Ecotect/Radiance
According to the surface properties stated in Section 3.1.1 Table 3-2 and Table 3-3, the corresponding values in
Ecotect/Radiance software were used (Table 5-10, 5-11) (Fig. 5-46).
Testing
Terminology
Ecotect/
Radiance
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 100 1 1 1 100 1
Specularity Shininess 0 0.5 0.5 0.5 0 0.5
Roughness Roughness 0 15 15 15 0 15
Color Bleed Color Bleed n/a n/a n/a n/a n/a n/a
Surface
Orientation
Surface Type Interior Interior Interior Interior Interior Interior
Transmittance Transparency 0 0 0 0 0 0
Calculation
Type
Luminance/
Illuminance
n/r n/r n/r n/r n/r
Table 5-10: Surface Properties in Ecotect/Radiance for 100% reflection test; Notes: n/a - Not Applicable, n/r - Not required
Testing
Terminology
Ecotect/
Radiance
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 50 1 1 1 100 1
Specularity Shininess 0 0.5 0.5 0.5 0 0.5
Roughness Roughness 0 15 15 15 0 15
Color Bleed Color Bleed n/a n/a n/a n/a n/a n/a
Surface
Orientation
Surface Type Interior Interior Interior Interior Interior Interior
Transmittance Transparency 0 0 0 0 0 0
Calculation
Type
Luminance/
Illuminance
n/r n/r n/r n/r n/r
Table 5-11: Surface Properties in Ecotect/Radiance for 50% reflection test; Notes: n/a - Not Applicable, n/r - Not required
128
Figure 5-46: Model setup in Ecotect/Radiance
The model was setup as discussed in Section 5.1.3 and the north wall was defined as an opaque material.
void metal NorthWall_100
0
0
5 0.99 0.99 0.99 0.0 0.00
void metal NorthWall_50
0
0
5 0.5 0.5 0.5 0.0 0.00
Simulation of the test was straightforward and the anticipated results were achieved.
5.2.4 Discussion of simulation in Rhino/DIVA
According to the surface properties stated in Section 3.1.1 Table 3-2 and Table 3-3, the corresponding values in
Rhino/DIVA software were used (Table 5-12, 5-13) (Fig. 5-47).
Testing
Terminology
Rhino/DIVA
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 100 1 1 1 100 1
Specularity Shininess 0 0.5 0.5 0.5 0 0.5
Roughness Roughness 0 15 15 15 0 15
Surface
Orientation
Surface Normal Interior Interior Interior Interior Interior Interior
Calculation
Type
Luminance/
Illuminance
n/r n/r n/r n/r n/r
Table 5-12: Surface Properties in Rhino/DIVA for 100% reflection test; Notes: n/a - Not Applicable, n/r - Not required
129
Testing
Terminology
Rhino/DIVA
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 50 1 1 1 100 1
Specularity Shininess 0 0.5 0.5 0.5 0 0.5
Roughness Roughness 0 15 15 15 0 15
Surface
Orientation
Surface Normal Interior Interior Interior Interior Interior Interior
Calculation
Type
Luminance/
Illuminance
n/r n/r n/r n/r n/r
Table 5-13: Surface Properties in Rhino/DIVA for 50% reflection test; Notes: n/a - Not Applicable, n/r - Not required
Figure 5-47: Model setup in Rhino
The model was setup as discussed in Section 5.1.4. However, since DIVA does not give either a grid based
luminance or exitance calculation option, the luminance visualization had to be opened in a software called
WXFalseColor which opens HDR pictures and gives point illuminances (Fig. 5-48).
Figure 5-48: WXFalseColor window
Simulation of the test was straightforward.
130
5.2.5 Discussion of simulation in Rhino/Grasshopper/Honeybee
According to the surface properties stated in Section 3.1.1 Table 3-2 and Table 3-3, the corresponding values in
Rhino/Honeybee software were used (Table 5-14, 5-15) (Fig. 5-49).
Testing
Terminology
Rhino/Honey
bee
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 100 1 1 1 100 1
Specularity Shininess 0 0.5 0.5 0.5 0 0.5
Roughness Roughness 0 15 15 15 0 15
Surface
Orientation
Surface Normal Interior Interior Interior Interior Interior Interior
Calculation
Type
Luminance/
Illuminance
n/r n/r n/r n/r n/r
Table 5-14: Surface Properties in Rhino/Honeybee for 100% reflection test; Notes: n/a - Not Applicable, n/r - Not required
Testing
Terminology
Rhino/Honey
bee
Terminology
North Wall South
Wall
West
Wall
East
Wall
Floor Ceiling
Reflectance Reflectance 50 1 1 1 100 1
Specularity Shininess 0 0.5 0.5 0.5 0 0.5
Roughness Roughness 0 15 15 15 0 15
Surface
Orientation
Surface Normal Interior Interior Interior Interior Interior Interior
Calculation
Type
Luminance/
Illuminance
n/r n/r n/r n/r n/r
Table 5-15: Surface Properties in Rhino/Honeybee for 50% reflection test; Notes: n/a - Not Applicable, n/r - Not required
Figure 5-49: Model setup in Rhino
131
The model was setup as discussed in Section 5.1.5. Grid based analysis recipe was chosen which involves the
creation of test points and connecting it to the recipe along with sky file path (Fig. 5-50). Similar to the image
based module, the simulation type can be chosen.
Figure 5-50: Grid based simulation module in Honeybee
There are a number of ways of visualizing analysis grid values. A color patch, node values, false color, text file are
all among the options that can be selected (Fig. 5-51).
Figure 5-51: Setting up the visualization of a grid based analysis in Honeybee (top) and some ways of visualization (bottom)
Simulation of the test was straightforward.
5.3 Luminous Flux Conservation test
At a given time on a given day if a ray of light enters through a single opening, the total luminous flux at the
opening surface should equal the total luminous flux reaching the different interior surfaces. No physical model
was built for this test since the inclusion of the measuring devices would corrupt the measurement totals, it
would be a laborious process to measure and calculate the average illuminances of all surfaces and it could not
be compared against the values generated in the software packages as the intensity of sunlight would vary.
5.3.1 Discussion of simulation in AGi32
According to the surface properties discussed in Section 3.1.2, all the surfaces had 0.01 reflectance (Fig. 5-52).
132
Figure 5-52: Model setup for Test1 in AGi32
Conducting the luminous flux conservation test in AGi32 was straightforward (Fig. 5-53). The first run showed
the average illuminance on the external surface of the unglazed panel to be very low. Flipping the normal to face
the outside only while calculating fixed the problem (Fig. 5-54). This is because in AGi32, the direction in which
an object is facing determines how it interacts with the light.
Figure 5-53: Setting calculation points on all surfaces (left) and on the panel (right)
Figure 5-54: Illuminance levels on panel facing inside (left) and facing outside (right)
All the tests were conducted in the same way.
5.3.2 Discussion of simulation in Ecotect/Radiance
According to the surface properties stated in Section 3.1.2, all the surfaces had 0% reflectance (Fig. 5-55).
133
Figure 5-55:Model setup in Ecotect/Radiance for Test 1
The model was setup as discussed in Section 5.1.3 and all the surfaces were defined in the same way.
void plastic Wall
0
0
5 0.01 0.01 0.01 0.00 0.02
However, unlike the other tests in Ecotect/Radiance, actual illuminance values were required for this test.
Hence, sensor points were added to each surface in turn to get illuminance values and they were imported into
the model (Fig. 5-56).
Figure 5-56: Imported sensor point values on a wall for Test 1
Each test had numerous values for all the surfaces. These values were exported to Excel for calculating the
average footcandles for each surface for each test (Fig. 5-57).
134
Figure 5-57:Data Chart created for each test in Excel
All the tests were conducted in the same way.
5.3.3 Discussion of simulation in Rhino/DIVA
According to the surface properties stated in Section 3.1.2, all the surfaces had 0% reflectance (Fig. 5-58).
Figure 5-58: Test1 setup in Rhino
The model was setup as discussed in Section 5.1.3 and all the surfaces were given the same definitions as was
given for Ecotect/Radiance. Point-in-time illuminance calculations were run and the average illuminance was
calculated (Fig. 5-59).
Figure 5-59: Illuminance values on some surfaces for Test1 as calculated in DIVA
All the tests were conducted in the same way.
135
5.3.4 Discussion of simulation in Rhino/Grasshopper/Honeybee
According to the surface properties stated in Section 3.1.2, all the surfaces had 0% reflectance (Fig. 5-60).
Figure 5-60: Test1 setup in Rhino
The model was setup as discussed in Section 5.1.4 and all the surfaces were given the same definitions as was
given for Ecotect/Radiance. Illuminance calculations were run after setting up a grid-based analysis recipe as
discussed in Section 5.2.5. Text files were created and average illuminances were calculated (Fig. 5-61).
Figure 5-61: Illuminance value text file for the outer surface of the aperture.
All the tests were conducted in the same way.
136
Chapter 6
Conclusion: Comparative charts
6.0 Conclusions
Comparative charts have been designed to form an easy and quick reference guide for lighting designers and
students of lighting and day lighting to select a software based on their use. Levels of accuracy have been
included based on the pathological tests.
6.1 Comparative charts for tests
The accuracy of the software packages based on the pathological tests have been tabulated with instructions on
the best situation in which the software can be used (Tables 6-1, 6-2, 6-3). A red dot means the software has
failed in that particular test whereas a green dot means it has passed.
Software Specular Reflection Test Notes
AGi32 Uses radiosity rendering technique and hence
cannot consider specularity of surfaces. Not to be
used when looking for glare in interior space
calculations.
Ecotect/Radiance Uses backward raytracing rendering technique
which is accurate.
Rhino/DIVA Cannot produce an illuminance visualization, hence
the accuracy cannot be determined. Calculated
illuminance values differ widely from the other
software packages.
Rhino/Honeybee Uses backward raytracing rendering technique
which is accurate.
Table 6-1: Comparative chart for specular reflection test
Software Reflection Test Notes
AGi32 Absolutely accurate. Should be used when
analyzing a space with different reflectivities.
Ecotect/Radiance Absolutely accurate. Should be used when
analyzing a space with different reflectivities.
Rhino/DIVA Not accurate.
Rhino/Honeybee Not accurate.
Table 6-2: Comparative chart for reflection test
137
Software Luminous Flux
Conservation Test
Notes
AGi32 Accuracy increases with increasing size of the
opening.
Ecotect/Radiance Accuracy decreases with increasing size of the
opening.
Rhino/DIVA Accuracy increases with increasing size of the
opening.
Rhino/Honeybee Accuracy increases with increasing size of the
opening.
Table 6-3: Comparative chart for luminous flux transmission test
6.2 Comparative charts for additional parameters
Some additional parameters were discussed in Section 3.2. These parameters of the various software packages
were noted while conducting the pathological tests and have been compiled into charts (Table 6-4, 6-5 and 6-6).
Software Export
Options
Import
Options
Analysis
Plug-ins
Modeling
Complexity
Required
Familiarity
Availability/
Cost
AGi32 .DWG CAD, LaiDex None Medium Medium $895 one time.
Ecotect/
Radiance
Radiance,
EnergyPlus,
DOE and
GBXML; also
image files like
JPEG, BMP,
GIF.
Can import
model data
from CAD,
Sketchup,
analysis
data from
Radiance,
GBXML.
Radiance,
Daysim,
Daysimps,
EnergyPlus,
DOE, WinAir4
for CFD.
Modeling in
Ecotect is
easy but
setting the
Radiance
parameters
requires a
certain
knowledge
base.
High Previosuly
$200/year for
Ecotect from
Autodesk.Has been
discontinued from
March 20, 2015.
Rhino/
Diva
The Rhino
model can be
exported to a
variety of 2d
and 3d
formats.
Rhino can
import dwg
and 3d
formats.
Radiance and
Daysim
Rhino
modeling is
tricky for new
users.
High $995 for Rhino +
$575 for DIVA (one
time)
Rhino/
Honeybee
The Rhino
model can be
exported to a
variety of 2d
and 3d
formats.
Rhino can
import dwg
and 3d
formats.
Radiance,
Daysim, Open
Studio,
EnergyPlus
(Only Radiance
and Daysim
have been
developed at
present).
High High $995 for Rhino
(one time).
Grasshopper,
Ladybug,
Honeybee,
Radiance, Daysim
are all free
downloads.
Table 6-4: Analysis Software Interoperability Observations
138
Software Editing
Imported
Geometry
Model
Editing
Interface
Analysis
Interface
GUI
AGi32 Not
supported
Easy but with
few options
Does not have too many visualization
options for presentations
Easy to use.
Comprehensive
tutorials are
available in the Help
section.
Ecotect/Radiance Supported Easy Does not have too many visualization
options for presentations
Easy to use.
Rhino/Diva Supported Requires
familiarity
Has multiple options for visualization
but lacks the false color option for
grid based calculations.
Rhino modeling
requires some
practice. DIVA has a
very simple and
clear GUI.
Comprehensive
tutorials are
available online.
Rhino/ Honeybee Supported Requires
familiarity
Has multiple options for visualization. Both Rhino
modeling and the
visual programming
in Grasshopper
requires some
practice.
Comprehensive
tutorials are
available online.
Table 6-5: Analysis Software Modeling Observations
Software Sky models
recognized
Render
Algorithms
used
Speed of
Render; Quality
of Render
Metrics Calculated;
Whole Building
Analysis
Code
Compliance
Testing
AGi32 15 recognized
CIE types and
Perez all
weather
Radiosity and
limited
raytracing
Medium
depending on the
complexity of the
model; Medium
Luminance,
illuminance,
exitance, Daylight
Factor; No
None
Ecotect/Radiance Sunny,
Intermediate,
Cloudy and
Uniform.
Backward
raytracing
Medium
depending on the
complexity of the
model; High
Luminance,
illuminance, daylight
factor and sky
component; Yes
None
Rhino/Diva CIE clear (with
and without
sun), overcast,
intermediate
(with and
without sun)
and Perez all
weather.
Backward
raytracing
Fast; Low Visualization -
Luminance, dynamic
timelapse, radiation
and glare (annual
and point-in-time);
Grid based - Daylight
factor, point-in-time
illuminance, cliamte-
based and radiation;
Only for single zones.
LEED v44,
NECHPS IEQ
P2, MACHPS
EO.C2
Rhino/ Honeybee Climate based,
cumulative,
Backward
raytracing
Fast; High Illuminance,
radiation, luminance,
None
139
standard CIE,
average sky
and sky with
certain
illuminance
level
daylight factor,
dynamic climate
based analysis, dgp;
Through Ladybug in
the same model and
grasshopper file
Table 6-6: Analysis Software Metrics Observations
6.2 Future work in the field of analysis of daylighting in software
The field of daylight analysis will only grow in the future due to stricter codes and an increased awareness of the
need for daylighting in the design community. A number of further studies can be conducted in this field.
A) Based on the comparison of AGi32, Ecotect/Radiance, Rhino/DIVA and Rhino/Honeybee
1) Each of the tests can be further validated with more data. The same software packages can be retested if and
when their newer versions come out, especially DIVA and Honeybee which are still under development.
2) Further research can be based on fixing the bugs found, like why the luminous flux is not conserved or why
Ecotect, Diva and Honeybee give different results even though they are all based on Radiance and are using the
same sky type and material files.
3) The software packages can be tested with additional tests like - a) directional transmittance of clear glass, b)
determination of daylight factor under different sky conditions.
4) The software packages can also be tested for other properties of light like color rendering, absorption, and
difference in the results on using different sky types.
5) The software packages that performed well can be tested with additional parameters like different glazing
types and complex geometries to further understand their limits.
6) Comparative studies can also be conducted on the availability and quality of tutorials available for the
software packages (Namburi 2006) and the responsiveness of the different technical teams in responding to
technical issues faced by users.
7) Similar tests can be performed using the DIVA plugin for Grasshopper to investigate the options available and
the accuracy.
8) To study the Radiance file path for Ecotect/Radiance, Rhino/DIVA and Rhino/Grasshopper/Honeybee to see
what parameters are being lost in translation.
B) Based on analysis of daylighting in software
1) Existing buildings that were analyzed using any of the tested software can be studied for their real-time
performance, which can then be compared to the analysis results.
140
2) Comparative studies can be performed on the ability of the software to analyze dynamic shading options, the
integration of artificial lighting with daylighting, the use of control systems and whole building analysis.
3) Radiance analysis can be compared against Daysim analysis.
4) Similar studies can be performed with other software packages like Revit FalseColor, ElumTools, Dialux,
Sefaira and LightStanza.
6.3 Conclusion
Analysis of daylighting in software simulations still has a long way to go. Performances of software packages and
accuracy of results remain inconsistent, as was shown through the specular reflection, reflectivity and luminous
flux conservation tests. AGi32 shows potential and can be used for studies in the early stages of design.
However, the tool should mostly be used for artificial lighting analysis. Based on the findings it can be said that
Ecotect/Radiance is still the most powerful tool when it comes to daylight analysis and the decision of Autodesk
to phase it out removes the best tool from the marketplace. DIVA needs to offer certain important options like
illuminance visualizations and orthographic false color images. Honeybee does show potential and should be
studied and tested again when it is more developed. The comparison of additional parameters has given
valuable insights into the different applications of each software, especially for the use of lighting designers. The
intention of the research is not only to demonstrate the limitations faced by each of the software packages but
to remind the users the need for understanding the fundamental concepts of anything that is being tested and
to be skeptical when analyzing results and not to trust the software blindly. It is important to always carry out
small tests and be aware of the limitations.
141
References
Chapter 1
1) Baker, Nick, and Koen Steemers. Daylight Design of Buildings. London: James & James, 2002.
2) Bass, Michael. Handbook of Optics Volume II - Devices, Measurements and Properties.2nd ed. Vol. 2.McGraw-Hill, 1995.
24-47.
3) "DIVA for Rhino." DIVA for Rhino. Accessed March 23, 2015. http://diva4rhino.com/.
4) "Ecotect Analysis - Sustainable Building Design Software - Autodesk." Ecotect Analysis - Sustainable Building Design
Software - Autodesk. Accessed March 23, 2015. http://usa.autodesk.com/ecotect-analysis/.
5) "Sensory Reception: Human Vision: Structure and Function of the Human Eye." In Encyclopædia Britannica. Vol. 27. 1987.
Jackson, Gregory R., C. Owsley, and G. McGwin Jr. Aging and Dark Adaptation. 1998.
6) Foster, R.g., I. Provencio, D. Hudson, S. Fiske, W. Grip, and M. Menaker. "Circadian Photoreception in the Retinally
Degenerate Mouse (rd/rd)." Journal of Comparative Physiology A, no. 169 (1991): 39-50.
7)Geebelen, Benjamin. "Daylighting Computation Methods." Accessed March 23, 2015.
https://lirias.kuleuven.be/bitstream/123456789/160607/1/kviv_geebelen.pdf -.
8) Hastings, Robert. Passive Solar Commercial and Institutional Buildings: A Sourcebook of Examples and Design Insights.
Chichester: J. Wiley, 1994.
9)Hygge, Staffan, and Hans Allan Löfberg. "Post Occupancy Evaluation of Daylight in Buildings." International Energy
Agency, 1999.
10)Inanici, Mehlika. "Application Of The State-Of-The-Art Computer Simulation And Visualization In Architectural Lighting
Research." IBPSA Conference, 2001
11) Kota, S and Haberl, J, Historical Survey Of Daylighting Calculations Methods And Their Use In Energy
12) "Ladybug Honeybee." Food4Rhino. Accessed November 14, 2014. http://www.food4rhino.com/project/ladybug-
honeybee.Mistrick, Richard. "DAYSIM Development Penn State’s Contributions." August 13, 2013. Accessed November
14, 2014. http://www.radiance-online.org/community/workshops/2013-golden-co/Mistrick_Daysim-Development.pdf.
13) Lewy A J, Kern H A, Rosenthal N E, Wehr T A (1982). Bright artificial light treatment of a manic-depressive patient with a
seasonal mood cycle. Am J Psychiatry 139: 1496–1498.
14) "Daylight Harvesting Tutorial." Lighting Control & Design. Accessed November 14, 2014.
http://www.lightingcontrols.com/design/onlinetraining/pdfs/lcd_dhlessons_2of8.pdf.
15) "Lighting Design Software by Lighting Analysts." Lighting Design Software by Lighting Analysts. Accessed March 23,
2015. http://www.agi32.com/.
16) Lillesand, Thomas M., and Ralph W. Kiefer. Remote Sensing and Image Interpretation. New York: Wiley, 1979.
17) Lipman, Jonathan, and Frank Lloyd Wright. Frank Lloyd Wright and the Johnson Wax Buildings. New York: Rizzoli, 1986.
18) Mahone Group, Heschong. "HeschongMahone Group, Inc. - Daylighting and Productivity." HeschongMahone Group, Inc.
- Daylighting and Productivity. Accessed November 14, 2014. http://h-m-g.com/projects/daylighting/summaries on
daylighting.htm.
142
19) Namburi, Varma. "COMPARATIVE EVALUATION OF LIGHTING DESIGN SOFTWARE PROGRAMS." January 1, 2006.
Accessed November 14, 2014.
https://arch.usc.edu/sites/default/files/mbs/papers/varma_namburi_(2006)_comparative_evaluation_of_lighting_desi
gn_software_programs1242157896882.pdf.
20) Ochoa, Carlos, Myriam Aries, and Jan Hensen. "CURRENT PERSPECTIVES ON LIGHTING SIMULATION FOR BUILDING
SCIENCE." IBPSA-NVL 2010 Event. Accessed January 1, 2014.
http://www.bwk.tue.nl/bps/Hensen/publications/10_ibpsa-nvl_ochoa.pdf.
21) Photometry Unit Chart - Schiler, Marc. "Concepts, Terms, and Basic Physics." In Simplified Design of Building Lighting,
13-15.Wiley, 1992.
22) Radiometry Unit Chart - Palmer, James. "Radiometry and Photometry FAQ." Radiometry vs. Photometry FAQ. Accessed
November 14, 2014. http://fp.optics.arizona.edu/Palmer/rpfaq/rpfaq.htm.
23) Reinhart, C 2014, Daylighting Handbook 1.
24) Rosenthal, N. E., D. A. Sack, J. C. Gillin, A. J. Lewy, F. K. Goodwin, Y. Davenport, P. S. Mueller, D. A. Newsome, and T. A.
Wehr (1980). "Seasonal Affective Disorder: A Description of the Syndrome and Preliminary Findings With Light
Therapy." Archives of General Psychiatry: 72-80.
25) Rosenthal N E, Sack D A, Gillin J C, et al. (1984). Seasonal affective disorder. A description of the syndrome and
preliminary findings with light therapy. Arch Gen Psychiatry 41: 72–80.
26) Schiler, Marc, Daylight Harvesting, The Gas Company, Downey, CA, 1995, 682-86/87-96 4/86
27) Steffy, Gary R. Architectural Lighting Design. 3rd ed. New York: John Wiley, 2008.
28) Suk, Jae Yong. "Absolute and Relative Glare Factor Method." 2014, 4.
29) Terman M, Terman J S, Quitkin F M, et al. (1989). Light therapy for seasonal affective disorder. A review of efficacy.
Neuropsychopharmacology 2: 1–22.
30)2013 California Building Code: California Code of Regulations, Title 24, Part 2, Volume 1. Whittier, Calif.: ICBO, 2013.
31) Ubbelohde,, Susan. "Comparative Evaluation of Four Daylighting Software Programs." 1988.
http://aceee.org/files/proceedings/1998/data/papers/0327.PDF.
Chapter 2
1)CIE, Commission Internationale de L’Ecleraige, 2006. Test Cases to Assess the Accuracy of
Lighting Computer Programs, Publication 171. Vienna, Austria: CIE
2)Donn, M, Xu, D, Harrison, D and Maamari, F, Using Simulation Software Calibration Tests As A Consumerguide –
A Feasibility Study Using Lighting Simulation Software
http://www.academia.edu/1036172/USING_SIMULATION_SOFTWARE_CALIBRATION_TESTS_AS_A_CONSUMER_GUIDE_A_
FEASIBILITY_STUDY_USING_LIGHTING_SIMULATION_SOFTWARE
3) Glaser, D.C., Feng, O., Voung, J. and Xiao, L., 2004. Towards an algebra for lighting simulation.
Building and Environment, 39, 895- 903.
4) Goral, C. M., Torrance, K. E., Greenberg, D. P. and Battaile, B., 1984. Modelling the interaction
143
of light between diffuse surfaces. In: Proceedings 11th annual conference on Computer
graphics and interactive techniques, SIGGRAPH, 213–222
5) Hong, T., Chou, S.K., Bong, T.Y., 2000. Building simulation: An overview of developments and
information sources. Building and Environment, 35, 347-361.
6) Jakubiec, J. Alstan. "Introduction to Simulating LEEDv4 Daylighting Credit Compliance and IES-LM-83." DIVA for Rhino.
Accessed October 2, 2014. http://solemma.net/DIVADay/2014/11 - DIVA Day 2014 - LEEDv4 Panel Discussion.pdf.
7) Kniss, J., Premoze, S., Hansen, C., Shirley, P. and McPherson, A., 2003. A model for volume
lighting and modelling. IEEE Transactions on Visualization and Computer Graphics,
9(2), 150-162.
8) Kota, S and Haberl, J, Historical Survey Of Daylighting Calculations Methods And Their Use In Energy
9) Maamari, F., M. Fontoynont, and N. Adra. "Application of the CIE Test Cases to Assess the Accuracy of Lighting Computer
Programs." Energy and Buildings: 869-77. Accessed March 23, 2015.
http://www.ibp.fraunhofer.de/content/dam/ibp/de/documents/Kompetenzen/waermetechnik/2009_05_2_MaamariI
EAReportBenchmarks2005tcm45-81671.pdf.
10) Modelling, Editing and Host Programs for Analysis: Observations - Panitz, Konrad, and Veronica Garcia-Hansen.
"Daylighting Design and Simulation : Ease of Use Analysis of Digital Tools for Architects." QUT EPrints. January 1, 2013.
Accessed November 14, 2014. http://eprints.qut.edu.au/58526/.
11) Ochoa, Carlos, Myriam Aries, and Jan Hensen. "CURRENT PERSPECTIVES ON LIGHTING SIMULATION FOR BUILDING
SCIENCE." IBPSA-NVL 2010 Event. Accessed January 1, 2014.
http://www.bwk.tue.nl/bps/Hensen/publications/10_ibpsa-nvl_ochoa.pdf.
12) Ochoa, Carlos E., MyriamB.c. Aries, and Jan L.m. Hensen. "State of the Art in Lighting Simulation for Building Science: A
Literature Review." Journal of Building Performance Simulation 5 (2012): 209-233.
13) Plant, C.G.H. and Archer, D.W., 1973. A computer model for lighting prediction. Energy and
Buildings, 8, 351-361.
14) "RRV - Radiosity Renderer and Visualizer (C , OpenGL)." Dudka.cz:. Accessed March 23, 2015. http://dudka.cz/rrv.
15) Svendenius, N., and Pertola, P., 1995. Searching for useful lighting design software: Developing
technical specifications based on real needs. In: Right Light Proceedings: the 3rd
European Conference on Energy-Efficient Lighting, IAEEL, NewCastle-upon-Tyne,
United Kingdom, 19-24.
16)Whitted, Turner. "An Improved Illumination Model for Shaded Display." Communications of the ACM, 1980, 343-49.
Chapter 3
1) "Application of the CIE Test Cases to Assess the Accuracy of Lighting Computer Programs." April 1, 2005. Accessed
November 14, 2014.
http://www.ibp.fraunhofer.de/content/dam/ibp/de/documents/Kompetenzen/waermetechnik/2009_05_2_MaamariI
EAReportBenchmarks2005tcm45-81671.pdf.
2)Namburi, Varma. "COMPARATIVE EVALUATION OF LIGHTING DESIGN SOFTWARE PROGRAMS." January 1, 2006. Accessed
November 14, 2014.
https://arch.usc.edu/sites/default/files/mbs/papers/varma_namburi_(2006)_comparative_evaluation_of_lighting_desi
gn_software_programs1242157896882.pdf.
144
3) "PHOTOLUX LUMINANCE, Photometric Measurement System or a Photoluminance Meter, a Compound of a Processing
Software and a Calibrated Digital Camera Provided with an Fish-eye Lens, Photo Luminance." PHOTOLUX Software
Allows Producing Luminance Maps from Digital Pictures Taking into Account the Camera Calibration Curves. Accessed
March 23, 2015. http://www.photolux-luminance.com/index.php/le-programme-software/photolux-v3-2.
4) Schiler, Marc. "Concepts, Terms, and Basic Physics." In Simplified Design of Building Lighting, 13-15.Wiley, 1992.
5) Tips for Daylighting with Windows. Berkeley, Calif.: Program, 1997.
Chapter 4
1) Namburi, Varma. "COMPARATIVE EVALUATION OF LIGHTING DESIGN SOFTWARE PROGRAMS." January 1, 2006. Accessed
November 14, 2014.
https://arch.usc.edu/sites/default/files/mbs/papers/varma_namburi_(2006)_comparative_evaluation_of_lighting_desi
gn_software_programs1242157896882.pdf.
2) Reinhart, Christoph, and Annegret Fitz. "Findings from a Survey on the Current Use of Daylight Simulations in Building
Design." Energy and Buildings, 2006, 824-35. Accessed March 23, 2015.
http://web.stanford.edu/group/narratives/classes/08-09/CEE215/ReferenceLibrary/Daylighting/Daylighting Use
Study.pdf.
3) Schiler, Marc. "Concepts, Terms, and Basic Physics." In Simplified Design of Building Lighting, 13-15.Wiley, 1992.
Chapter 5
1) "3TIER Support." What Is Diffuse Horizontal Irradiance? Accessed March 23, 2015.
http://www.3tier.com/en/support/solar-prospecting-tools/what-diffuse-horizontal-irradiance-solar-prospecting/.
2) Accessed March 23, 2015. https://www.iso.org/obp/ui/#iso:std:iso:2814:ed-1:v1:en.
3) Reinhart, Christoph. "Tutorial on the Use of Daysim Simulations for Sustainable Design." 2010. Accessed March 23, 2015.
https://dl.dropboxusercontent.com/u/16228160/Daysim3.0.Tutorial.pdf.
Chapter 6
1) Namburi, Varma. "COMPARATIVE EVALUATION OF LIGHTING DESIGN SOFTWARE PROGRAMS." January 1, 2006. Accessed
November 14, 2014.
https://arch.usc.edu/sites/default/files/mbs/papers/varma_namburi_(2006)_comparative_evaluation_of_lighting_desi
gn_software_programs1242157896882.pdf.
145
Bibliography
The bibliography consists of all the books, online resources and papers perused during the course of the project and
includes even those sources which have not been cited.
1) Altmann, K and Apian-Bennewitz, P, Report on an Investigation of the Application andLimits of Currently Available
Programme Types for Photorealistic Rendering of Light and Lighting inArchitecture - The Kimbell Art Museum as a Case
Study for Lightscape, Radiance and 3D-Studio MAX
2) Amarpreet, S 2003, Conclusions on the lighting software program (Ecotect). “A study of daylighting techniques and their
energy implications using designer friendly simulation software”http://www.sbse.org/awards/docs/2003/Sehti.pdf.
3) Christakou, E and Amorim, D, Daylighting Simulation: Comparison of Software for Architect’s Utilization.
4) Donn, M, Xu, D, Harrison, D and Maamari, F, Using Simulation Software Calibration Tests As A Consumerguide –
A Feasibility Study Using Lighting Simulation Software
http://www.academia.edu/1036172/USING_SIMULATION_SOFTWARE_CALIBRATION_TESTS_AS_A_CONSUMER_GUIDE_A_
FEASIBILITY_STUDY_USING_LIGHTING_SIMULATION_SOFTWARE
5) Estes, J, Schreppler, S and Newsom, T, Daylighting Prediction Software: Comparative Analysis And Application
6) Harvey Bryan, Sayyed Mohammed Autif, “Lighting/Daylighting Analysis: A Comparision” School of Architecture, Arizona
State Univeristy, Tempe, AZ. 85287-1605
7) Hitchcock, R.J. 1995, Advancing Lighting and Daylighting Simulation: The transition from analysis to design aid
tools,http://www.btech.lbl.gov/papers/37285.pdf.
8) IES Lighting Handbook, 2013
9) Koenigsberger 1998, Manual of Tropical Housing and Building, Orient Longman
10) Kota, S and Haberl, J, Historical Survey Of Daylighting Calculations Methods And Their Use In Energy
11) Larson, G and Shakespeare, R ,Rendering with Radiance, Morgan Kaufmann Publishers.
12) Light Guide, Light Loss Factor, http://www.lightsearch.com/resources/lightguides/lightloss.html, 2006.
13) Namburi, V 2006, Comparative Evaluation Of Lighting Design Software Programs For Analysis Of Daylighting In
Buildings,
http://arch.usc.edu/sites/default/files/mbs/papers/varma_namburi_%282006%29_comparative_evaluation_of_lightin
g_design_software_programs1242157896882.pdf
14) Reinhart, C and Ibarra, D, Daylight Factor Simulations – How Close Do Simulation Beginners ‘Really’ Get?
15) Reinhart, Daylighting Course
http://ocw.mit.edu/courses/architecture/4-430-daylighting-spring-2012/
16) Reinhart, C 2014, Daylighting Handbook 1.
17) Santos, C and Bastos, L, A Methodology For Daylight Systems Evaluation In Open-Plan Offices Using Desktop Radiance
Software.
146
18) Schiler, M. 1992, Simplified Design of Building Lighting, John Wiley and sons
19) Schiler, Marc, Daylight Harvesting, The Gas Company, Downey, CA, 1995, 682-86/87-96 4/86
20) Steffy, Gary R. Architectural Lighting Design. 2nd ed. John Wiley, 2002.
21) Ubbelohde, S, Comparative Evaluation of Four Daylighting Software Programs
http://aceee.org/files/proceedings/1998/data/papers/0327.PDF
22) https://www.youtube.com/watch?v=M4MA82_CPPA&list=PLkjfDmSc5OryXkWSt57ltJFU4qXD5ss1v&index=1
Abstract (if available)
Abstract
The use of daylighting to supplement and offset artificial lighting requirements in interior spaces is an important step in achieving California’s current and future energy goals. The analysis of daylighting is not only difficult for lighting designers, but also the results are not always accurate. Four popular lighting analysis software programs (AGi32, Ecotect/Radiance, Rhino/Diva, and Rhino/Grasshopper/Honeybee) failed to provide results that matched the data gathered in physical models and from calculations based on the laws of physics for the Specular Reflection Test, Reflectance Test (Namburi 2006) and Luminous Flux Conservation Test (CIE 171:2006). The results recorded for most of the simulations have shown more than 20% variation (the threshold for validation tests according to CIE), between the software packages and real data. ❧ Additional parameters, such as the clarity of the Graphical User Interface (GUI), time and training required for full use, quality of scene renderings, ability to calculate climate-based metrics and code compliance, and interoperability with other software have been compared across the same software programs and the results tabulated. The tabulation also lists the critical nuances for daylighting simulation in each of the four software programs. This together with a narrative describing the change of metrics used in digital daylight simulation will help designers analyze their buildings more accurately. ❧ Hypothesis: AGi32, Ecotect/Radiance, Rhino/Diva, and Rhino/Grasshopper/Honeybee cannot consistently and accurately simulate the physical properties of daylight for Specular Reflection, Reflection, and Luminous Flux Transmission Test.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Simulation-based electric lighting control algorithm: integrating daylight simulation with daylight-linked lighting control
PDF
Daylight and health: exploring the relationship between established daylighting metrics for green building compliance and new metrics for human health
PDF
Simplified acoustic simulation - Rutabaga Acoustics: a Grasshopper plug-in for Rhino
PDF
Visualizing architectural lighting: creating and reviewing workflows based on virtual reality platforms
PDF
Comparing visual comfort metrics for fourteen spaces using simulation-based luminance mapping
PDF
Double skin façades performance: effects on daylight and visual comfort in office spaces
PDF
Multi-domain assessment of a kinetic facade: determining the control strategy of a kinetic façade using BIM based on energy performance, daylighting, and occupants’ preferences; Multi-domain asse...
PDF
Effective light shelf and form finding: development of a light shelf design assistant tool using parametric methods
PDF
Exterior glare simulation: understanding solar convergence from concave facades using heat maps
PDF
Performative shading design: parametric based measurement of shading system configuration effectiveness and trends
PDF
Evaluation and development of solar control screens: using daylight simulation to improve the performance of facade solar control screens
PDF
Energy efficient buildings: a method of probabilistic risk assessment using building energy simulation
PDF
The intelligent control strategy of kinetic façades for daylight and energy performance: evaluating the daylight effect of adaptive systems based on parametric workflow
PDF
Environmentally responsive buildings: multi-objective optimization workflow for daylight and thermal quality
PDF
Comparative evaluation of lighting design software programs for daylighting in buildings
PDF
Absolute glare and relative glare factors: predicting and quantifying levels of interior glare and exterior glare caused by sunlight and daylight
PDF
Integrating non-visual effects of lighting in the evaluation of electrical lighting designs for commercial offices
PDF
Daylighting in Riyadh and Los Angeles: comparison of cultural factors in potential market penetration
PDF
Digital tree simulation for residential building energy savings: shading and evapotranspiration
PDF
Acoustics simulation for stadium design using EASE: analyzing acoustics and providing retrofit options for the Los Angeles Memorial Coliseum
Asset Metadata
Creator
Banerjee, Sebanti
(author)
Core Title
Daylight prediction: an evaluation of daylighting simulation software for four cases
School
School of Architecture
Degree
Master of Building Science
Degree Program
Building Science
Publication Date
04/21/2015
Defense Date
03/23/2015
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
daylighting,OAI-PMH Harvest,Radiance,simulation
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Schiler, Marc (
committee chair
), Kensek, Karen M. (
committee member
), Noble, Douglas (
committee member
)
Creator Email
sebantib@usc.edu,sebantib1989@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-556106
Unique identifier
UC11298956
Identifier
etd-BanerjeeSe-3360.pdf (filename),usctheses-c3-556106 (legacy record id)
Legacy Identifier
etd-BanerjeeSe-3360.pdf
Dmrecord
556106
Document Type
Thesis
Format
application/pdf (imt)
Rights
Banerjee, Sebanti
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
daylighting
Radiance
simulation