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
/
Visualizing thermal data in a building information model
(USC Thesis Other)
Visualizing thermal data in a building information model
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
VISUALIZING THERMAL DATA IN A BUILDING INFORMATION MODEL
by
Qianqian Fan
Committee Members:
Karen M. Kensek
Joon-Ho Choi, Marc Schiler
A Thesis Presented to the
FACULTY OF THE USC SCHOOLE OF ARCHITECTURE
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF BUILDING SCIENCE
MAY 2016
Copyright 2016 Qianqian Fan
i
Acknowledgements
First and foremost, I would like to express my gratitude to my thesis committee chair,
Professor Karen Kensek, who has offered great help on this thesis project. Many
obstacles have been encountered during the research in the past year, and she has offered
excellent ideas and great directions along the way. Her creative and critical way of
thinking, the positive attitude facing difficulties, and her passion for work will always
inspire and encourage me even after I graduate.
I also would like to express my appreciation to my committee members, Professor Joon-
Ho Choi, and Professor Marc Schiler. They have offered great support, inspiring ideas,
and cheering encouragement for me on this project. I really appreciate your help.
Last but not least, I would like to thank my family and all of my friends, especially the
lovely MBS group. I had a wonderful time in this two years and learned a lot. Thank you
so much for your help and support.
ii
Abstract
In order to be able to meet rigorous energy standards, serious attention needs to be given
to correcting problems with recently constructed buildings and retrofitting older buildings.
One step is determining where facades might not be performing as digitally modelled. A
literature review was undertaken to evaluate and then select several methods that are able
to convert infrared thermal colors (surface temperature) taken by a digital FLIR camera
to corresponding real thermal resistance values (R-value). These algorithms were tested
on a physical model, and the algorithm with more trust worthy result was selected to
continue with. A series of 2D thermal and digital photos of the building were taken, and
Autodesk ReCap was used to generate a 3D point cloud and mesh model encoding the
thermal colors at the point level. The next step was to align the Revit model with the
thermal information and have Revit detect those colors and use Dynamo to convert colors
to corresponding thermal data. The final output was a protocol which was able to achieve
a false color building model carrying information of different thermal properties of the
envelope. The protocol could help engineers make better retrofitting decision, help
owners save energy consumption, and provide data to energy programs that could
ultimately allow occupants to have a higher comfort level. The process encountered many
obstacles where compromises were made resulting in the conclusion that this
methodology can be applied only to cases under certain conditions. A protocol for
identifying those cases and using those methods is presented. Future work would include
more rigorous validation of the methodology established.
Keywords: thermal data, infrared camera, R-value, BIM
iii
Table of Contents
Acknowledgements .............................................................................................................. i
Abstract ........................................................................................................................... ii
List of Tables....................................................................................................................... v
List of Figures .................................................................................................................... vi
Chapter 1 Introduction ....................................................................................................... 1
1.1 Background........................................................................................................ 2
1.1.1 Building Information Modeling and Energy Issues ................................ 2
1.1.2 Building Commissioning and Retrofit .................................................... 5
1.1.3 Thermal Properties Testing of Building Facades .................................... 8
1.1.4 Hypothesis Statement.............................................................................11
1.2 Significance of Study .......................................................................................11
1.3 Terms ............................................................................................................... 12
1.3.1 Thermal resistance ................................................................................ 12
1.3.2 Infrared camera ..................................................................................... 13
1.3.3 Building Information Modeling (BIM)................................................. 13
1.3.4 Building Energy Modeling.................................................................... 15
1.3.5 Point Cloud ........................................................................................... 15
1.4 Study Scope ..................................................................................................... 16
1.4.1 Factors affecting building energy prediction ........................................ 16
1.4.2 Test method adoption ............................................................................ 17
1.4.3 Locate thermal images on models......................................................... 17
1.5 Objectives and Deliverables ............................................................................ 18
1.5.1 Objectives.............................................................................................. 18
1.5.2 Deliverables .......................................................................................... 18
Chapter 2 Previous Work: Background and Literature Review ....................................... 19
2.1 Creating Thermal Building Models ................................................................. 19
2.1.1 Hybrid Laser Scanning System.................................................................. 19
2.1.2 2D Images to 3D Models ........................................................................... 22
2.2 Convert Thermal Images to Thermal Resistances ........................................... 25
2.2.1 Algorithm I ................................................................................................. 26
2.2.2 Algorithm II................................................................................................ 27
2.2.3 Algorithm III .............................................................................................. 28
2.2.4 Algorithm IV .............................................................................................. 30
2.2.5 Algorithm V ............................................................................................... 31
2.3 Chapter Summary ............................................................................................ 32
iv
Chapter 3 Methodology.................................................................................................... 33
3.1 Thermal Images to R-values Method Selection and V alidation ...................... 34
3.2 Build up Thermal Building Models................................................................. 40
3.3 R-value Detection and Visualization ............................................................... 43
3.4 Chapter Summary ............................................................................................ 45
Chapter 4 Process and Data.............................................................................................. 46
4.1 Lab Tests for Algorithm V alidation ................................................................. 46
4.2 ReCap Tests ..................................................................................................... 49
4.3 Manually Assign Thermal Photos.................................................................... 51
4.4 View in BIM .................................................................................................... 53
4.5 Create Colors Database ................................................................................... 54
4.6 Chapter Summary ............................................................................................ 60
Chapter 5 Analysis and Results........................................................................................ 62
5.1 Algorithm Selection for R-value Calculation .................................................. 62
5.2 Dynamo Script for Thermal Data Visualization in BIM ................................. 65
5.2.1 Thermal Photos Assignment ................................................................. 65
5.2.2 Colors Database Creation...................................................................... 66
5.2.3 Surface Temperatures Reading ............................................................. 67
5.2.4 R-values Calculation ............................................................................. 68
5.2.5 Thermal Data Visualization .................................................................. 69
5.3 Restrictions and Limits .................................................................................... 70
5.4 Chapter Summary ............................................................................................ 71
Chapter 6 Conclusions and Future Works ........................................................................ 73
6.1 Overall Significance ........................................................................................ 73
6.2 From Thermal Image to Thermal Resistance .................................................. 74
6.3 Thermal Data Visualization Protocol .............................................................. 75
6.4 Future Works Referring to Thermal Models ................................................... 76
6.5 Further Application.......................................................................................... 78
6.6 Chapter Summary ............................................................................................ 81
Bibliography...................................................................................................................... 82
Appendices........................................................................................................................ 85
Appendix A Complete data and results of lab test (algorithm validation) ................ 85
Appendix B The Complete Dynamo Script ............................................................ 100
v
List of Tables
Table 3.1 Selected algorithms to be assessed............................................................. 34
Table 3.2 R-TECH Insulation sheet R-value (Insulfoam, 2015)................................ 35
Table 3.3 Tested sample material ............................................................................... 36
Table 4.1 Sample raw data from test (Unit: ℃) ......................................................... 47
Table 4.2 Sample calculation results for material 1 (Unit: m
2
℃/W) ......................... 48
Table 5.1 Notional R-value of material 1................................................................... 63
Table 5.2 Notional R-value of material 2................................................................... 64
Table A.1 Data of lab test - Material 1 ....................................................................... 85
Table A.2 Result of algorithm validation - Material 1 ............................................... 89
Table A.3 Data of lab test - Material 2 ....................................................................... 92
Table A.4 Result of algorithm validation - Material 2 ............................................... 96
vi
List of Figures
Figure 1.1 Infrared photo of studs in a wall ................................................................11
Figure 1.2 Sample infrared/ thermal image ............................................................... 13
Figure 1.3 Sample Building Information Model........................................................ 14
Figure 1.4 Material Properties of Wall Layers ........................................................... 14
Figure 1.5 Sample material properties information in Revit ..................................... 14
Figure 1.6 Sample building energy modeling ............................................................ 15
Figure 1.7 Point Cloud of A Building Model............................................................. 16
Figure 3.1 Overall Methodology Framework ............................................................ 33
Figure 3.2 Insulated box............................................................................................. 35
Figure 3.3 R-TECH Insulation Sheet (Insulfoam, 2015) ........................................... 35
Figure 3.4 "R-TECH" and "FOAMULAR" being tested at the same time................ 36
Figure 3.5 1 Air conditioner and 2 heaters................................................................. 37
Figure 3.6 Tripod with sensors................................................................................... 37
Figure 3.7 IButton, adapter, HOBO & light bulb ...................................................... 38
Figure 3.8 Light bulbs used to adjust simulated indoor temperature......................... 38
Figure 3.9 FLIR E8 infrared camera (FLIR website, 2015) ...................................... 40
Figure 3.10 The small building model being tested in ReCap ................................... 41
Figure 3.11 Building D of School of Cinematic Arts in USC .................................... 42
Figure 3.12 Building information model of Building D built in Revit ...................... 42
Figure 3.13 Sample thermal photos taken around Building D................................... 43
Figure 3.14 Vision of final prototype ......................................................................... 44
vii
Figure 4.1 Examples of thermal photos during test ................................................... 47
Figure 4.2 Project from digital images ...................................................................... 50
Figure 4.3 Project from thermal photos ..................................................................... 50
Figure 4.4 Project from Photoshop thermal photos ................................................... 50
Figure 4.5 Project from digital and thermal photos ................................................... 51
Figure 4.6 Divide model surface using "create parts" in Revit.................................. 52
Figure 4.7 Sample thermal photo before and after cropping ..................................... 52
Figure 4.8 Error reported when applying " Element.OverrideColorInView " node .. 53
Figure 4.9 "Rainbow" color range creation- method 1 .............................................. 55
Figure 4.10 "Rainbow" color range creation- method 2 ............................................ 55
Figure 4.11 Failure of looking up colors in "rainbow" database ............................... 56
Figure 4.12 Colors by ARGB (alpha, red, green, and blue components) .................. 57
Figure 4.13 Thermal photos in "rainbow", "iron", and "grey" mode ......................... 58
Figure 4.14 "Iron" color range creation- method 1 .................................................... 58
Figure 4.15 "Iron" color range creation- method 2 .................................................... 58
Figure 4.16 Failure of looking up colors in "iron" database ...................................... 59
Figure 4.17 "Grey" color range creation- method 1................................................... 59
Figure 4.18 Success of looking up colors in "grey" database .................................... 60
Figure 4.19 Method of building up thermal model (Xs stand for unsuccessful
attempts).............................................................................................................. 61
Figure 5.1 R-values calculated from test- material 1 ................................................. 63
Figure 5.2 R-values calculated from test- material 2 ................................................. 64
Figure 5.3 The complete Dynamo script.................................................................... 65
viii
Figure 5.4 Dynamo script of assign thermal photos to building surfaces.................. 66
Figure 5.5 Thermal textures on surfaces shown in Dynamo...................................... 66
Figure 5.6 "Grey" colors database creation ............................................................... 67
Figure 5.7 Index of colors in database were returned in a list ................................... 67
Figure 5.8 Math logic from index numbers to corresponding temperatures.............. 68
Figure 5.9 Equation to calculate R-values from temperatures................................... 69
Figure 5.10 R-values plotting back onto building model surfaces ............................ 70
Figure 5.11 Degradation or leakages indication ........................................................ 70
Figure 5.12 Overall workflow of revised protocol .................................................... 72
Figure B.1 The complete Dynamo script ................................................................. 100
Figure B.2 Thermal photo assignment ..................................................................... 101
Figure B.3 "Grey" colors database creation ............................................................. 101
Figure B.4 Index of colors in database were returned in a list ................................ 102
Figure B.5 Math logic from index numbers to corresponding temperatures ........... 102
Figure B.6 R-values calculation............................................................................... 103
Figure B.7 Thermal data visualization ..................................................................... 104
1
Chapter 1 Introduction
Due to increasing concern for environmental and energy issues in both voluntary
standards and mandatory codes, stricter requirements for lowering building energy
consumption have been implemented. These are not only aiming for newly constructed
buildings, but also including existing buildings when doing renovation. They need to
meet the updated energy codes. For those planning to carry out retrofitting, even if some
of the buildings were designed as exceptional energy saving buildings originally,
differences between the ideal design and post-construction are inevitable, including
building physical characteristic (shell), operation method, and occupants’ schedule
(system). There is always a gap existing between the predicted building energy
performance and the real situation in operation and maintenance phase.
In order to be able to meet the rigorous energy standards, serious attention should be give
to retrofitting. Before retrofitting, a thorough inspection is imperative for owners to be
informed of the current situation of the buildings. Then that data discovered should be
encoded into a 3D model. These real data, including the thermal properties of the
building envelope, may be different from that in the original 3D model due to
construction deviation, gradual degradation, influences caused by manual operation, etc.
The possible, sometimes even huge differences, make it necessary to update real
properties of post-construction buildings and import the real values into the thermal or
energy models, which will make the energy simulation result more accurate and reliable.
Good visualization will make it more intuitive for the owner to understand. If the related
properties of the corresponding elements in building information models can be updated
2
by associating them with real measurements, they could be easier to handle and be useful
for understanding current conditions of existing buildings and how they relate to energy
performance.
1.1 Background
1.1.1 Building Information Modeling and Energy Issues
Climate changes and resources depletion issues are becoming increasingly serious
concerns around the world recently. The building industry accounts for up to 40% of the
total energy use, 25% of water use, and 40% of global resources according to the
statistical data from United Nations Environment Programme- Sustainable Buildings and
Climate initiative (UNEP-SBCI, 2015). The significant carbon footprint has a strong
negative impact on our living planet, leading to the problem including global warming,
lack of fossil energy, and the deterioration of living environment for both human beings
and other creatures around (Hong et al., 2015).
Therefore, it becomes more and more common to perform building energy modeling in
the early design phase and to carry on systematic commissioning process regularly
afterwards for a building. Different sustainable design alternatives could be compared in
terms of building energy efficiency from the beginning of design, and the properties and
systems can be managed or retrofitted appropriately and in time to secure building
performance.
The popularization of building energy modeling encourages architects to design more
3
sustainable buildings with lower carbon emission during their whole life cycle. The
comparison results during this design process will help building owners and designers
make wiser decisions about what kind of sustainable design strategies should be selected
and adopted. Trade-off is carried out for both shell and system design of the building,
considering building envelope, windows, lighting, HV AC system etc. Besides its
contribution to the sustainability, essentially, building energy modeling simulates the
situation when buildings are occupied, thus helping designers develop ing the system
design based on that certain assumed status to satisfy people's comfort demands. In
addition, building energy modeling also plays an important part when rating the energy
consumption level of a buildings. Evaluations are done both in design phase according to
the construction drawings and documents and in some cases after construction is
complete when a certain building is under regular operation.
In the architecture, engineering, and construction (AEC) industry, the development of
building information modeling (BIM) offers a revolutionary solution for both
architectural projects design and building energy modeling process. Building information
modeling (BIM) is an emerging technology in the industry compared to computer-aided
design (CAD). It has been popularized rapidly and there is a considerable increase in the
use of BIM in AEC industry in recent years. BIM provides technical support for the
project plan optimization and scientific decision, including geometry design, built
environment design, energy consumption analysis, economy analysis, cost estimation,
quality monitoring, safety analysis, inspection and simulation of the whole process, etc.
(Riaz et al., 2014).
4
There are several obvious advantages of the application of BIM. First of all, BIM tools
help architects design a building using 3D graphics directly instead of 2D drawings using
traditional CAD and enable the shareholders to better visualize the development of their
building projects all over the different construction phases in the model, or even more
directly, through simulated tour animations created from BIM tools (Kensek and Noble,
2014). Second, it allows both architects and engineers start the project based on a
platform which is much more accurate, detailed and easy to operate, thus significantly
raising the working efficiency. Third, BIM tools are able to carry various related
information of every single elements during design process. For example, a wall or a roof
can be modeled in the BIM tools as an object with layers of materials with different
thermal properties. Also, it is very convenient to change parameters to control the
dimensions of the BIM objects and spaces with the geometry updating automatically in
the whole project based on any change. Most importantly, BIM can be used in various
stages of a project, such as planning, investigation, design, construction, commissioning,
operation and maintenance and so on, so that makes it possible for the various building
design information from multiple disciplines to be integrated within one model and
enables all the participants sharing data in the same multi-dimensional building
information model during the whole life cycle of building (Kim et al, 2015). This
contributes significantly to the trade-off of design strategies in the early design phase and
can easily realize the coordination and the detection of possible clashes among different
disciplines, before real construction process, which are not available in the traditional
CAD tool.
5
In many cases, models carrying thermal properties data of building elements are exported
into energy modeling software to predict heating and cooling loads and help with the
decision making in early stage design. However, the same building information models
should not be used when doing commissioning or retrofitting for post-construction
buildings. Since properties of building elements may be different from the original design
values, data need to be updated as real performance status at that time during
commissioning and before retrofitting.
1.1.2 Building Commissioning and Retrofit
To decrease energy consumption of buildings, a green design with various sustainable
design strategies is significantly important as a good start. Applications of both passive
design strategies such as natural ventilation, daylighting, solar energy uses, etc. and
active design strategies including heat pump, chilled beams, forced-air HV AC systems etc.
are theoretically minimizing the building energy consumption from the very beginning.
Nevertheless, people are beginning to realize that commissioning, monitoring, diagnosing
and retrofitting during operation and maintenance phases after building construction are
even more important (Haasl and Heinemeier, 2006). It is difficult or impossible to
operate a building as its original design status, since deviation exists ubiquitously
between before and after construction among material properties, system settings, and
schedule operations. So following up with commissioning is necessary to ensure that the
construction meets the related requirements basis of design and construction documents,
and the building is calibrated and perform as owner expected.
6
ASHRAE Standard 202-2013, The Commissioning Process for Buildings and Systems,
and ASHRAE Guideline 0, The Commissioning Process define commissioning as: "A
quality-focused process for enhancing the delivery of a project. The process focuses upon
verifying and documenting that all of the commissioned systems and assemblies are
planned, designed, installed, tested, operated, and maintained to meet the Owner's Project
Requirements." According to "California Commissioning Guide: New Buildings" and
"California Commissioning Guide: Existing Buildings," "benefits of commissioning
include reduced energy use, lower operating costs, reduced contractor callbacks, better
building documentation, improved occupant productivity, and verification that the
systems perform in accordance with the owner’s project requir ements" (Haasl and
Heinemeier, 2006).
When some apparent defects are observed and are not able to be fixed by adjustment of
basic system parametric settings, the building needs retrofitting. In real projects, on-site
test and measurement are carried out to diagnose the possible flaws, together with
observations from building management system with sensors. Sometimes, problems can
be revealed directly from the energy bill, and the search scope can be limited to a certain
zone if submeters were installed in advance. A diagnostic report is created that
summarizes issues and includes charts of the testing data, a list of existing possible
problems and suggestions for retrofit methods. In very few cases, engineers will try to
develop a detailed energy model of an existing building, and test different retrofit
strategies in energy simulation tools and make final decisions based on comparison, just
7
as engineers and designers tried to optimize design scheme at the very beginning of the
project before the building was constructed. However, this method has a lot of challenges,
like how to match the energy model with the existing building, since most of properties
are from the design drawings and construction documents. Construction defects and
degradations are often not taken into consideration, which will cause inaccurate
results. Also assuming the envelope is constructed and performs exactly as expected
properties, what if the simulated energy consumption doesn't match the energy bill? Even
taking the system and schedule into consideration only, various factors are influencing
the final result simultaneously, which interact each other. Manually modeling and
calibration are not impossible, but complicated and time consuming. Hence, it's going to
be challenging to build up a detailed energy model for an existing building requiring
retrofit.
In addition, continuous monitoring is also essential in case of any possible unexpected
deterioration of physical properties and system or equipment errors. Currently, BIM is
being widely used both during building design process and in operation monitoring or
retrofitting phases. Generally speaking, BIM tools have the ability to integrate various
information from different disciplines and categories, and those related information can
be transferred to building energy simulation tools for analysis and making adjustment
accordingly (Kim et al., 2015). And in most of the cases today, in BIM-based energy
modeling process, thermal properties of building elements are directly derived from
gbXML-based BIM database (Ham and Golparvar-Fard, 2015). Therefore, if related
properties of the BIM elements can be updated by associating real measurements with
8
their corresponding elements in gbXML file, especially if the measurements are easy to
handle, this method will aid in diagnosing and visualizing construction.
1.1.3 Thermal Properties Testing of Building Facades
According to a report by the National Institute of Standards and Technology (NIST), "one
of the top measurement challenges for energy-efficient building is measuring material
aging" (NIST, 2010). Material aging influences the performance of envelope insulation
significantly and is strongly related to the space heating and cooling, which account for
48.8% and 24.8% of the total usage in residential and commercial buildings respectively
(U.S. DOE, 2010). Building materials are likely to deteriorate gradually at different rates,
and as a result, the real value of thermal resistance of building envelope is likely to be
lower than its original notional values, and the result will probably be different across the
surface area of a wall or roof. So instead of directly deriving from the notional values
built in modeling software based on industry standard, trying to measure or detect the
thermal properties distribution among the envelope of an existing building is highly
recommended.
However, the current method of doing energy modeling is not encouraging engineers to
use a distribution of the different thermal resistances along an architecture component
like a wall or roof. Instead, an average and constant value is adopted. That leads to
inaccuracy and inability to be applied as a basis or proof to conduct related building
diagnosing and retrofitting work. The U.S. General Services Administration also
highlighted “accurate modeling of as -built conditions” as one of the top challenges for
9
BIM-based energy modeling (USGSA, 2012). So considering both the construction
deviation of the envelope with selected materials and the degradation effects during the
operation period, testing and monitoring the thermal properties of an existing building
envelope is necessary and important.
There are two ways to test the thermal properties of the envelope, "destructive" and "non-
destructive" (Ham and Golparvar-Fard, 2015). One destructive method is to break the
wall and take samples to the lab. Testing is conducted under certain environmental
condition control, and due to the controllability, results are accurate and reliable.
However, destructive way is not applicable for an existing building under operation and
will continuously be under use. So in general cases, nondestructive way is adopted in the
on field inspection.
It is common and traditional to apply thermocouple and heat flux sensors to measure
thermal resistance of a piece of wall. By measuring heat flow and the temperature
differences on indoor and outdoor surfaces, the R-value is calculated. This method
requires monitoring, usually more than 72 hours when the temperature is under steady
state conditions, otherwise it may take "more than 7 days" sometimes (Albatici and
Tonelli, 2010). Moreover, because of the contact of sensors and surface of the wall when
measuring, convection properties has already been changed. As a result, the calculated R-
value is not convincing enough due to the unavoidable interference during experiment.
It is possible to do the test without contact using thermography. "Thermography is
10
defined as the process of detecting and measuring heat variations emitted by an object
under inspection and transforming identified changes into visible imagery" (Eads, 2000).
The infrared camera was first developed and applied by US military for night vision.
Instead of detecting visible light as regular camera, an infrared camera is a device that
can create a picture by detecting infrared radiation from heat sources. In infrared images,
different colors represent a range of corresponding temperatures of the objects. Infrared
cameras are widely used in "building diagnostics," "energy auditing and inspection,"
"property and facility management," "HVAC and plumbing," "moisture and restoration,"
etc. in the AEC industry (FLIR website, 2015).
An infrared camera can help quickly see what is going on under a surface (Figure 1.1).
Building envelope air leakage and infiltration, missing or damaged insulation of facade,
existing thermal bridges influencing energy efficiency can all be detected in a quick scan
and visualized intuitively, plus issues with floor radiant heating system, structure defects
and inappropriate water flow (FLIR website, 2015). An infrared camera makes
diagnosing more reliable by generating thermal pictures fast and clear to reveal possible
problems of system with complex design and envelope of large areas.
11
Figure 1.1 Infrared photo of studs in a wall
To make it simpler and easier to measure real R-value for existing buildings, there are
several studies which have figured out a certain way to convert temperatures to R-values,
based on complicated algorithms or data integration regression.
1.1.4 Hypothesis Statement
In post-construction buildings, deviation exists between the real performance of the
materials and how they are expected to behave. In order to predict energy performance of
an existing building or before carrying out retrofitting, testing and monitoring the thermal
properties of the envelope quickly is significant and necessary. It is possible to obtain R-
value of the envelope from quick measurements with an infrared camera and simple
temperature sensors. The thermal images would be put on the building information model
for visualization, and with the help of Dynamo plug-in, Revit would be able to detect
thermal colors and convert them to thermal data.
1.2 Significance of Study
In the current simulation process, it is assumed that each building element has a constant
12
property (such as R-value) as designed, which is derived directly from the BIM tools'
default database. However, the unexpected possible changes due to inappropriate
construction process, building materials gradual deterioration, the general changes of
climate environment, as well as the energy cost fluctuation and related policy influences
need to be taken into consideration. Therefore, the building energy model requires
correction according to the real changes, in order to achieve a more reliable energy
consumption analysis. Otherwise, the estimated result may be widely different from the
as-is operation building energy consumption and may even mislead the retrofit methods
adopted and cause the waste of energy or increasing occupant dissatisfaction with the
indoor thermal comfort level. It will be helpful to generate a building information model
with clear visualization of data for owners and stakeholders to allow them make wiser
decision of whether retrofit is needed in time. The diagnosing and treatment of the
buildings can lower energy costs.
1.3 Terms
1.3.1 Thermal resistance
Thermal resistance is a thermal property reflecting the ability of resisting heat transfer. It
is the reciprocal of thermal conductance which is measured by the amount of heat flow
going through a unit area of a material with a specified temperature difference between
two surfaces. Usually, thermal resistance is reported by R-value, with the units
(ft2· ° F· hr)/Btu in imperial units or (m2K)/W in SI units. The values depend on property
of materials. (Tritt, 2004)
13
1.3.2 Infrared camera
An infrared camera (also called an thermographic camera or thermal imaging camera) is
a device that creates picture by detecting infrared radiation of heat sources instead of
detecting visible light as regular camera. In infrared/ thermal images, different colors
represent a range of corresponding temperature of the objects. Infrared cameras were first
developed and applied by US military using for night vision, and are widely used for
diagnosing building problems and inspecting for energy auditing (FLIR website, 2015)
(Figure 1.2).
Figure 1.2 Sample infrared/ thermal image
1.3.3 Building Information Modeling (BIM)
Building information modeling is a process of integrating information from different
disciplines and design phases into one single model through the whole life cycle of a
building (Figure 1.3). The model contains not only 3D physical geometry information
from design phase, but also functional properties including scheduling as the 4th
dimension, costing as the 5th dimension, and even operation data as the 6th dimension
(Figure 1.4, 1.5). It is a "shared knowledge resource" of "digital representation" (National
BIM Standards - United States, 2015).
14
Figure 1.3 Sample Building Information Model
Figure 1.4 Material Properties of Wall Layers
Figure 1.5 Sample material properties information in Revit
15
1.3.4 Building Energy Modeling
Building energy modeling is computer based simulation of a building focusing on energy
consumption (Figure 1.6). It helps engineers and building owners to predict the energy
use and cost of various kinds of energy related items, such as space cooling and heating,
lighting, domestic hot water usage, and to size mechanical equipment. Building energy
modeling also helps architects and engineers to do the trade off among various
sustainable design strategies in both pre-design stage and deep retrofits (Department of
Energy, 2015).
Figure 1.6 Sample building energy modeling
1.3.5 Point Cloud
A point cloud is a set of points located by XYZ coordinates, which are generated by 3D
laser scanners.. By scanning the surface of the object, a set of points are detected with
different distance between scanner and the object. Those points are exported as a whole
to represent the shape of a target object (Hammoudi et al., 2010) (Figure 1.6).
16
Figure 1.7 Point Cloud of A Building Model
1.4 Study Scope
The update and visualization of thermal properties of an envelope in a BIM-based model
were focused on. By evaluating previous studies, an algorithm was selected for the
conversion from infrared thermal image to corresponding real R values. By locating the
thermal image at the right place, integrated with other thermal condition parameters,
Revit is able to detect thermal colors and convert them to thermal data correspondingly.
The methodology covered heat transfer theories, physical apparatus measurement,
Building information modeling, and application of programming.
1.4.1 Factors affecting building energy prediction
There are various factors affecting building energy prediction. For large scale buildings,
when a building is occupied, both the physical properties and system operation related
parameters will be under control of building management system. If something abnormal
is observed, engineers or technicians will be notified to make judgments and diagnose the
possible problems according to the system records. However small scale buildings, such
as residential houses, are not monitored by complete building management systems in
most of the cases. So a regularly quick inspection is necessary to guarantee the building
17
performance with possible retrofit in time. Among those parameters during inspection,
for an existing building, the one being concerned is the real thermal resistance (R-value)
distribution of the envelope. Since it heavily related to the space cooling and heating
energy section, which account for a large part of the total building energy consumption..
1.4.2 Test method adoption
To test the real R-values, nondestructive methods should be adopted for occupied
buildings. The traditional method of installing thermocouples and heat flux sensors takes
a long time to complete a single measurement. In order to achieve a faster process, a new
technique is adopted, which is called thermography. The application of infrared camera
makes it possible to detect the surface temperature distribution of the envelope at a glance,
and it is fairly easy to operate even for non-skilled users. Combining with the
environmental conditions surrounded then, thermal resistance can be calculated based on
some algorithms.
1.4.3 Locate thermal images on models
Another key problem is how to locate the thermal images at the right place on the
building information model. Previous studies with the application of laser scan are trying
to develop a completely new model for a certain space based on the huge point cloud
generated by laser scanner. Or, in some cases, UA V (unmanned aerial vehicle) with GPS
is getting involved in order to capture accurate dimensions (V an Der Heijde et al., 2015).
Due to the limitation of equipment accessibility, a easier method to locate thermal images
is adopted by simply matching the digital image, thermal image, and Revit model
18
together.
1.5 Objectives and Deliverables
1.5.1 Objectives
There are three main objectives:
1. Evaluate previous research projects on converting infrared images to thermal
resistances (R-value) and validate the selected algorithms on a physical model.
2. Develop a protocol which can achieve a quick measurement of thermal properties of
building envelop an intuitive visualization of thermal resistance.
1.5.2 Deliverables
The final deliverable would be a protocol to achieve thermal data visualization in
building information models. An algorithm would be developed and validated for
converting thermal images from an infrared camera to R-values, and a Dynamo script
was written to achieve a false color building information model that displays thermal
properties of the envelope.
19
Chapter 2 Previous Work: Background and Literature Review
Related background literature review is introduced, summarized and critiqued in this
chapter. Based on the hypothesis statement, key problems are summarized by two topics,
which are how to create thermal building models and how to convert thermal images to
thermal resistances. In terms of creating thermal models, laser scanning system is
introduced, and the technique of 2D images to 3D models is presented. Literature review
are done to list 5 algorithms which can achieve calculation of R-value.
2.1 Creating Thermal Building Models
In order to achieve the goal of updating thermal information of the building information
model, an original building information model is required to start with. There are several
ways to create building models. The most common is to use BIM software and input the
building components “manually” according to the designed geometry. A second method
is to create the model by taking a 3D scan of a building. A point cloud model representing
the shape and dimensions of the surface of a target building is then generated according
to the actual dimensions of the existing building. The point cloud model obtained is then
translated into a 3D building information model.
In either case, the thermal information has to be registered with the model so that the data
is correctly located and related to the building geometry.
2.1.1 Hybrid Laser Scanning System
There have been several studies that accomplished generating 3D thermal models using
20
the laser scanning technique. In the building industry, "Cho and Wang are among the first
to introduce a 3D thermal modeling method of an residential building faç ade using a
Hybrid LIDAR system" (Ham, and Golparvar-Fard, 2013). The proposed system was a
combination of a laser scanner and a thermal camera pointed at the building envelope.
The system can generate 3D point cloud models with each individual point carrying
information of the corresponding temperature values and thermal colors generated from
the infrared camera. The developed system was tested both in lab and on a residential
building in field (Cho and Wang, 2011). The researchers also developed a hybrid light
detection and ranging system, adding in window detection algorithms due to the failure of
collecting geometric data from transparent objects by previous system (Wang et al., 2013).
A procedure was proposed for regenerating the "thermographies" by using MATLAB. By
referring to a set of control points, they registered "thermographies" in the cloud of
points. And by comparing with another "textured 3D point cloud". they rectified the
deviations of the "thermographies" model, thus generating a model which is able to show
the heat distribution based on point level (Lagü ela et al., 2011).
A project called ThermalMapper constructed precise thermal 3D models of indoor
environments (Borrmann et al., 2012) The researchers set up a robot equipped with a 3D
laser scanner, a thermal camera, and a color camera in one of their studies. They gathered
the data from the sensors above from different locations and joined them into one high-
precise 3D model with the help of the automatic co-calibration of the sensors for indoor
environment. In a later study, 3D scan data sets were successfully recorded for both
21
inside environment and outside building facade of residential buildings using a Riegl VZ-
400 and a Optris PI IR camera (Borrmann et al., 2012).
Another new methodology of visualizing thermal performance of the existing buildings
on web-based geospatial programs obtained the building geometry using 3D point cloud
generated by a hybrid thermal laser scanner, converted the data set into a "geographic
coordinate system," and presented the corresponding information in KMZ 3D model in
Google Earth (Im et al., 2012). The motivation is to enable occupants, owners, and
retrofitting decision makers to visualize the measured building energy performance in an
"easily accessible web-based geospatial program," such as Google Earth, thus simplifying
the process of energy consumption audition and building retrofitting. Integrating
information on Google Earth platform makes it possible to expand the thermal models
visualization into a town-scope or even a city-scope to broader audience.
The methods described above can create an accurate thermal model by assigning thermal
information to every single point. However, considering the high price of laser scanner, it
is not a accessible, common equipment for most of the people. Even professional
building commissioning and retrofitting agencies need to make careful decisions before
investing a large amount of money on the equipment. So this leads to the idea of figuring
out an algorithm or software which can realize the conversion from 2D photos to 3D
models and generate point clouds accordingly using "2D scanning" instead of 3D laser
scanner.
22
2.1.2 2D Images to 3D Models
Infrared Concepts Corporation (ICC) proposed a 2D non-destructive testing thermal
mapping methodology for roof leaking repair by aerial surveying. The result has been
validated according to ASTM guidelines, and the results indicated an accuracy of 95% to
100% (ICC, 2015). But the method requires manual matching and mapping thermal and
digital images, which is hard-loading and time-consuming, let alone for matching thermal
images with digital ones for 3D modeling. Although some tricks can help, such as taking
notes or voice recording related information when collecting 2D images, or applying the
thermal camera with built in digital camera which allows picture-in-picture function to
locate images.
Youngjib Ham and Mani Golparvar-Fard from University of Illinois at Urbana-
Champaign have done a lot of work on this subject. They simplified the required
equipment and presented a new method to automatically build up thermal models by
matching 2D thermal and digital images captured by a single thermal camera. The
methodology they proposed was to build up the model geometry of existing building
using "a new image-based 3D reconstruction pipeline which consists of Graphic
Processing Unit (GPU)-based Structure-from-Motion (SfM) and Multi-View Stereo
(MVS) algorithms." Then, use "3D thermal modeling algorithm" to generate a
corresponding thermal point cloud model. The final deliverable called "3D spatio-thermal
model" is created by combining the 3D building model and thermal point cloud model
(Ham and Golparvar-Fard 2013). In later research, they moved further to propose a new
method for updating R-values of building elements in building information model. R-
23
values are calculated at the level of 3D points, and a single R-value was derived for each
element by calculating average. Finally, they automatically update the thermal data of the
corresponding BIM elements in gbXML file (Ham and Golparvar-Fard 2014). The
protocol is really amazing with excellent algorithms. However, an average is calculated
and applied for each building element, instead of keeping the original value distribution
at point level. Diagnosis of the position of leakage and degradation of material is the
purpose of inspection after all. If an average is calculated, some leakages of that element
might be ignored, since the problematic part is merged with normal parts and a slightly
low value is shown rather than obvious difference.
Several software programs exist to convert 3D images to 3D models -- photogrammetry.
ReCap developed by Autodesk is widely applied in the market right now. Autodesk
ReCap is a software program that works with data and images collected from laser
scanners or digital cameras. One can also use it to view and edit 3D point clouds
generated by a 3D laser scanner. Autodesk has successfully developed "photo-to-3D
model" function on Autodesk ReCap 360 web API. The embedded algorithms are able to
recognize and rearrange the location of images, merge the overlap parts, and register
them to a complete 3D mesh model or 3D points cloud model (Autodesk ReCap, 2015).
With this tool, 3D models can be generated in true dimensions based 2D digital photos
gathered on field. If it works as described, it is capable of creating 3D models of existing
buildings without the use of a 3D scanner.
Concerning this issue, help was seek from professionals in field via emails. Mr. Gustav
24
Fagerströ m from Buro Happold New Y ork office and Mr. Matt Jezyk and Mr Aniruddha
Deodhar from Autodesk were interviewed and offered some really valuable and helpful
ideas (Gustav Fagerströ m and Matt Jezyk, e-mail messages to author, August 29, 2015)
(Aniruddha Deodhar, e-mail messages to author, August 31, 2015).
R. V an Der Heijden, A.Tai, and G. Fagerströ m research focused on thermal and texture
mapping for a building. To deal with the large amount of photos and thermal data
detected, they assembled digital camera, thermal sensors and GPS reader on a UA V. By
controlling the UAV flying around the target building, all the graphic and numeric
information were captured simultaneously, including the thermal temperatures in the
photo, a set of XYZ coordinates of the position of camera, and a vector indicating the
direction of camera at the moment of snapping. Through coordination of thermal data
"normalized in RGB color scale" and geometry information and orientation vectors, with
CAD mesh models all together, the team successfully mapped the information onto Rhino
model with the help of Grasshopper plug-in (V an Der Heijden et al. 2015).
Jezyk has mentioned about the work SkyCatch is doing might be helpful for reference or
inspiration (Matt Jezyk, e-mail messages to author, December 4, 2015). After checking
their website, some work referring to incorporating thermal images and
environmental sensor data into their acquisition process has been found. And their
methodology is using UA V collecting digital photos, and collaborating with Autodesk
BIM360 to monitor construction sites (SkyCatch, 2015). The basic concept of idea is
similar to the one in Fagerströ m's research.
25
An excellent solution to achieve 2D to 3D would be software platform like Autodesk
ReCap, which allows users to get 3D model by just simply inputting 2D images. An
automatic process and the clear user interface are presented, and everything else is taking
care by the software. In terms of the application of UAV , if conditions permits, and a
drone is available on hand, the concept of creating a coordinate system is scientific and
rational. The crux is the process dealing with a huge number of raw data to figure out the
certain correspondence between XYZ points of camera and XYZ points of building
surfaces. Nevertheless, if no UAV available, manually registering 2D images with 3D
model can yet be regard as a feasible solution.
2.2 Convert Thermal Images to Thermal Resistances
The traditional way of measuring thermal resistance is using HFM (heat flow meters).
Based on the basic conduction equation, Q=UA△T ,measure Q by installing a heat flux
sensor and thermal couple on the two sides of the wall in a testing period of time.
However, due to the contact of sensors and the wall surface, the conduction and
convection status are disturbed and not able to perform as usual, so it is not accurate
enough to reflect the real value of Q. Also, in order to get a relatively convincing result,
HFM method always require a long time measurement to ensure the conduction is steady
and that indicates a certain number of days, or at least 72 hours continuous measurement
for a single test while the minimum difference between inside and outside temperature is
10 to15 °C (Albatici and Tonelli 2010).
26
A non-contact method to get the thermal resistance (R-value) of a building envelope
could be used instead. Infrared cameras have been used as efficient tools during building
diagnosing and retrofitting. The quick operation allows engineers, occupants, and owners
to notify the leakage and make decisions on how to conduct further retrofit. Infrared
cameras can read the surface temperature by demonstrating both the values and the
corresponding thermal colors. To investigate the application of infrared camera in
measuring R-values, researchers have come up with several solutions to convert
temperature to R-values.
2.2.1 Algorithm I
One of the simplest algorithm came out from mechanical engineers (Peter Simmonds,
pers. comm.). For radiant heating system for residential buildings, heat is released from
hot water or high temperature steam in pipes under floor and warm up the space with the
air flow going up and circulating through the whole room. Based on the fundamental heat
transfer regulation, when designing, engineers calculate the heat transfer considering
convection effect between the heated floor and the cold space, using Equation (1):
Equation (1)
where is the convective heat transfer coefficient which can be adopted from related
design standards. Combining with Equation (2)
A Equation (2)
in this case, building envelope equivalent to the heated floor in radiant heating system.
Based on the same concept, by simply measuring inside temperature, outside temperature,
surface temperature, thermal resistance (R-value) can be calculated as Equation (3).
27
Equation (3)
2.2.2 Algorithm II
The thermal power (heat) lost from the walls surface can be calculated as the combination
of E – thermal power (heat) dissipated by radiation, and H – thermal power (heat)
dissipated by convection (Albatici and Tonelli, 2010). Based on Stefan–Boltzman Law
for grey body radiation as Equation (4)
Equation (4)
ɛ is the surface integral emissivity, is the Stefan–Boltzman constant, T is the surface
temperature. And considering the sensible heat flux for convection as Equation(5)
Equation (5)
where c is the convective heat transfer coefficient, Ts is the surface temperature,
and Tair is the air temperature.
After wiping out the existing numbers in the original equation from the paper, thermal
resistance (R-value) can be calculated as Equation (6)
Equation(6)
where Tin is inside air temperature, Tout is outside temperature, and Ts is the surface
temperature of the element, v is wind speed.
Compared with the traditional way of measuring R-value with heat flux meters, this
method of applying an infrared camera is much more convenient and requires much less
28
time. When measuring with heat flux meters, it takes more than 10 hours to get a rather
steady and reliable result, and it is not able to tell the distribution of different thermal
resistance among the large area of an entire building envelope. In contrast, the proposed
method can complete the whole process simply within two to three hours.
However, the experiment has limitations due to the severe requirement for the
environmental conditions. The measurement can be done only during night time after
sunset to avoid the influence of direct solar radiation. And in order to get a more accurate
result, wind speed must be controlled less than 1m/s to avoid convective phenomena out
of control, and temperature difference between inside and outside should be larger than
10-15 ℃ to allow a significant effect of heat transfer. So the best time for measurement is
early in the morning around 3-4 am, when the temperature difference between indoor and
outdoor environment reach the peak value of the day and under a relative steady
condition (Albatici and Tonelli, 2010).
2.2.3 Algorithm III
A system was developed that can successfully generate "3D spatio-thermal models" of
buildings by measuring the actual thermal resistance using information from thermal
images, surrounding temperature, and the reflected apparent temperature to get R-value
(Ham and Golparvar-Fard, 2015).
In general, thermal resistance can be defined as the temperature difference between
indoor and outdoor surfaces which results in a unit of heat flows through the area of
29
building surface in unit time, as Equation (7):
Equation (7)
In this equation, the total amount of heat transfer can be calculated by summing up the
heat transfer contributions of convection (Equation 8) and radiation (Equation 9) between
the indoor and outdoor conditions of the building.
Equation(8)
Equation (9)
So, the thermal resistance can be calculated using Equation (10)
Equation(10)
Besides Tinside, air, Toutside, air, Tinside, wall, convective , and as mentioned in
algorithm I, there is another parameter involved in this equation, Tinside, reflected which is
called reflected temperature. "Reflected temperature is the amount of infrared radiation
reflected to the camera lens from heat producing objects in the viewing area" (Infrared
training center, 2015). It is measured using a small crumpled aluminum foil (low
emissivity and high reflectivity) target placed on the inspection areas before the
thermographic inspection according to ASTM E1862 and ISO 18434.
An assumption in this method is the existence of a single reflected temperature for the
whole interior surfaces in the building environment, which allows the amount of heat
transfer caused by thermal radiation to be calculated by Equation (9). Inside wall
29
30
temperature and reflected temperature are measured with the use of infrared camera;
Inside and outside air temperatures are measured using a thermometer; and the
convective heat transfer coefficients ( con) are adopted from ISO 6946.
The test was focusing on the interior environment, the measuring instruments were set
indoors. Considering the building as a whole instead of single rooms, and viewing the
entire facade from outside rather than indoor, the equation should be modified aiming for
this research as Equation (11):
Equation (11)
2.2.4 Algorithm IV
Another research on this infrared thermography to thermal resistance topic was done and
concluded with a slightly different equation considering thermal radiation an thermal
convection as Equation (12) and Equation (13) respectively (Fokaides and Kalogirou,
2011).
Equation(12)
Equation(13)
Hence, U-Value could be calculated by the following relationship:
Equation(14)
31
The algorithm was validated through lab tests and obtained data with the error of results
within an acceptable range of 10%-20%, and suggested a minimum difference between
inside and outside air temperature as 10 ° C, and taking 3-4 hours to measure a set of data.
The advantages of this method are significant because it takes less time than traditional
methods and is more accurate (Fokaides and Kalogirou, 2011). Taking radiation effects
into consideration is not a common practice in traditional heat flux meter method, and
that is definitely contributing to the final result to be more correct theoretically. Current
errors appeared in the test actually count more on the rough operation process, rather than
the overall method. There is still large potential to improve the accuracy for infrared
thermography to thermal resistance method by optimizing sensors and procedures.
Modify the final equation as the same format, this algorithm can be concluded as
Equation (15):
Equation (15)
2.2.5 Algorithm V
In another study, researchers tried to measure thermal resistance using a lab-based
destructive method according to the law of heat conduction, also known as Fourier’s law .
The theoretical basis is that the heat transfer rate is inversely proportional to the
temperature change gradient. The experimental facilities include a high sensitivity
thermal camera, a Peltier module, a thermal tank, a reference sample, and the studied
sample. The consequent relationship of measuring conductivity of the studied sample is
32
shown in Equation (16):
Equation (16)
where "d, , k are respectively the thickness, the diffusion length and the thermal
conductivity and subscript r denotes the reference sample." By measuring the thermal
conductivity of the reference known material, the unknown thermal conductivity k is able
to be deduced from the equation (Boué and Holé , 2012).
The fifth method is a destructive one, requiring test sample from the inspected material
(Boué and Holé , 2012). Considering the object of test is existing buildings, it is not wise
to sample the exterior walls. Furthermore, since this would be a whole scale inspection
for the entire building, it is impossible to destruct and take random samples everywhere
around the building. Therefore, the first four non-destructive methods were selected as
algorithms to be assessed. Those four algorithms would be tested on a physical model in
a controlled lab environment with the same sets of experiment data. And conclusion
would be drawn to decide the algorithm with more trust worthy result, which would be
adopted to continue with in the following procedures.
2.3 Chapter Summary
After the background literature review, due to the limit access to laser scanner and UAV,
as general public, manually building up thermal model was selected for the protocol.
Among algorithms regarding to the topic how to convert thermal images to thermal
resistances, four algorithms were selected to continue with the lab tests for validation.
33
Chapter 3 Methodology
A literature review was undertaken to find and then select several methods that are able
to convert infrared thermal colors (surface temperature) taken by a digital FLIR camera
to corresponding real thermal resistance values (R-value). These algorithms were tested
on a physical model, and the algorithm with more trust worthy result was selected to
continue with. A series of 2D thermal and digital photos of the building were taken.
Autodesk ReCap was used to generate a 3D point cloud and mesh model encoding the
thermal colors at the point level. Thermal photos were manually assigned to sub-divided
model surfaces divided into parts. Then Revit was able to detect those colors and convert
colors to thermal data correspondingly using Dynamo plug-in (Figure 3.1). The final
output was a protocol which was able to achieve a false color building model carrying
information of different thermal properties of the envelope. The protocol could help
engineers make better retrofitting decision, help owners save energy consumption, and
provide data to energy programs that could ultimately allow occupants to have a higher
comfort level.
Figure 3.1 Overall Methodology Framework
34
3.1 Thermal Images to R-values Method Selection and Validation
There are two measurement methods that can be used to quantify actual thermal
properties: destructive methods and non-destructive methods. It is not realistic to destruct
the existing wall and get samples to be tested. So instead, non-destructive methods are
chosen to obtain the real R value of the envelope. However, there is no such a sensor or
apparatus which is able to detect the objects' thermal resistance (R-value) directly.
Algorithms indicating the certain corresponding relations between the thermal images
taken by infrared camera and the real thermal resistance (R-value) of the object was
selected to be validated. Previous researches are done based on the heat transfer laws
combining with related database collection and analysis summarize regression. Four
algorithms/ methods were selected to be tested (Table 3.1).
Table 3.1 Selected algorithms to be assessed
Algorithm Equation Reference
Algorithm I
Interview of mechanical
engineers (Peter Simmonds, pers.
comm.)
Algorithm II
Albatici and Tonelli, 2010
Algorithm III
Ham and Golparvar-Fard, 2015
Algorithm IV
Fokaides and Kalogirou, 2011
To test the selected algorithms, a physical model simulation chamber was used made by
Alshiddi mimicking a building model (Alshiddi, 2015). The insulated box’s wall
35
thickness is 4.5 inch in four sides of the box, and the internal chamber’s size is 16" x 16"
x 16" (Figure 3.2). The material is foam thermal insulation sheets “R -TECH” (Figure 3. 3,
Table 3.2).
Figure 3.2 Insulated box
Table 3.2 R-TECH Insulation sheet R-value (Insulfoam, 2015)
R- Value R-TECH Test Method
Warranted R-V alues @ 20 years
4.8/inch
4.4/inch
ASTM C518
@40˚ F
@75˚ F
Published R-V alue (Thermal Resistance)
4.8/inch
4.4/inch
ASTM C518
@40˚ F
@75˚ F
Figure 3.3 R-TECH Insulation Sheet (Insulfoam, 2015)
36
The material being tested were "R-TECH" and "FOAMULAR" (Table 3.2), which were
installed as two sides of the insulated test box at the same time during tests, mimicking
real situation that the envelope of a building with different thermal properties (Figure 3.4).
Table 3.3 Tested sample material
Test Material R- Value/inch
R-TECH 4.4 hr ft
2
F/ Btu 0.78 m
2
C/W
FOAMULAR 5.0 hr ft
2
F/ Btu 0.88 m
2
C/W
Figure 3.4 "R-TECH" and "FOAMULAR" being tested at the same time
The physical model testing was carried out at the thermal environmental control lab
located at the basement of Watt Hall, School of Architecture, University of Southern
California. There is one air conditioner and two heaters in the lab connecting to a central
control system (Figure 3.5). A tripod was assembled with temperature, relative humidity,
CO2 level sensors in the lab, and data could be automatically recorded in the program
once hitting run (Figure 3.6). Both the control of equipments and the data recording were
operated by the software LabVIEW.
37
Figure 3.5 1 Air conditioner and 2 heaters
Figure 3.6 Tripod with sensors
To guarantee the accuracy of temperature measurement, HOBOs and iButtons were set at
the same time both inside the model and outside. Simulated outside model temperature
was controlled by air conditioner and heaters in the lab. A 11W and a 60W light bulb
were used respectively to adjust the simulated indoor temperature (Figure 3.7, 3.8).
38
Figure 3.7 IButton, adapter, HOBO & light bulb
Figure 3.8 Light bulbs used to adjust simulated indoor temperature
The original test plan was to set several conditions with steady temperature points with
the control of combination of inside and outside temperatures. However, for the
simulated indoor temperature, the light bulb made the temperature keep raising without a
limit. With the 60W light bulb, the chamber was heated up so fast that had to be turned
off in case of melting or fire. The same problem happened to the heater, as a result, there
was no literally steady point during test. To think this problem from another aspect,
fluctuating temperature actually makes more sense in real situations. Since the purpose of
the entire protocol is to give occupants or owners a quick glance of thermal resistance of
39
the building envelope, the on field test may be done at any time point in real application.
At the moment of measuring, it is not secured that the space is under steady thermal
environment rather than with the heater or A/C just turning on/off. In that case, the
measurement is actually recording one temperature at a certain point when the
temperature is changing. Therefore, in order to achieve a better user experience
afterwards, the result will be more trust worthy to test algorithms under changing
environment with fluctuating temperatures.
The final test plan for simulating the outside temperature was to turn on the air
conditioner first to cool down the simulated outside temperature for an hour, then turn off
the air conditioner and turn on the first heater for an hour. The third time period was to
turn on the second heater for an extra half an hour. As for simulated inside temperature
control, keeping the same outside temperature change cycle, test was done three separate
times with no light bulb, 11W light bulb, and 60 W light bulb inside the chamber.
Temperatures were monitored every one minute by both HOBOs and iButtons, and every
10-11 seconds by the Tripod. Thermal pictures were shot at random time during the entire
test duration using FLIR E8 infrared camera (Figure 3.9). Data and conclusions of the
physical model testing are presented in chapter 4 and 5.
40
Figure 3.9 FLIR E8 infrared camera (FLIR website, 2015)
3.2 Build up Thermal Building Models
Once the algorithm was chosen the next step was to build up thermal model and import
color information of thermal images into a visualization tool. BIM was selected as the
preferred platform because BIM is a collaborative way of working and promotes efficient
method of information management throughout the building lifecycle. Then if thermal
properties can be updated in BIM, information can be further used by energy engineers
when doing analysis. And the 3D property of BIM makes it more intuitive and clear to
view the entire model.
There are two approaches to create thermal models: attach the thermal information as the
entire model or break down the model and attach thermal photos by pieces. As discussed
in chapter 2.1, due to the limited access to laser scanner and UA V for general public, it
was decided to create a methodology based on simple temperature sensors and a
handheld infrared camera (200 to 5,000 dollars). Although the accuracy would not be that
ideal due to the low granularity of cheap infrared camera, the logic of the protocol would
still be a less expensive alternative to help engineers, house owners, and clients have a
general concept of potential problematic parts of the entire building.
41
ReCap 360 web API was tested considering the building as a whole. A small building
model was used (Figure 3.10). The infrared camera FLIR E8 can be set in a mode that
allows the infrared image and regular digital one being taken at the same time at one shot.
A series of thermal photos and regular digital pictures of a small scale building model
were taken. Photos were taken every 10-15º around the model, and some more photos
were shoot to show the relatively complicated details to support the 3D imaging process.
Tests were recorded in detailed process in chapter 4.2.
Figure 3.10 The small building model being tested in ReCap
For manually assigning thermal information, Building D of School of Cinematic Arts in
University of Southern California was selected as the building for case study (Figure
3.11). It has two floors and a mezzanine in between. Besides the medium size and
relatively brief and clear floor plan of the building, the main reason of selecting this
building is that it is on campus and easy to access. 2D drawings of the building including
detailed floor plans and elevations were offered by Facilities Management Services of
University of Southern California. A 3D building information model was created first in
Revit according to dimensions and materials of building elements available in 2D
drawings (Figure 3.12).
42
Figure 3.11 Building D of School of Cinematic Arts in USC
Figure 3.12 Building information model of Building D built in Revit
Thermal photos were taken by infrared camera FLIR E8 when walking around the target
building. A series of pictures were taken with overlap in order to be aligned on the
building information model at proper locations. Photos were taken while circum
navigating the building from 8 pm to 9 pm. During the day time, the exterior walls are
gaining a large amount of radiant heat under sunshine, so taking measurement a few
hours after sunset allows heat stored within walls to release and allows surface
temperature to cool down to minimize the impact of radiant heat. On the other hand,
during this time period, most of the rooms are still occupied, thus indoor temperatures
can be considered as air conditioned temperature in calculations, which was one of the
assumptions of the whole protocol.
43
Figure 3.13 Sample thermal photos taken around Building D
After both the 3D building information model and thermal photos were ready, a Dynamo
script was then written to assign thermal photos to the 3D model to match thermal data
with building geometry. The Dynamo script is shown in detail in chapter 5.2.
3.3 R-value Detection and Visualization
After successfully importing thermal images at the correct location to match with the
exterior wall view from BIM model, the Dynamo plug-in was used to have Revit detect
those colors and convert colors to thermal data to R-values using the chosen algorithm.
First, a database of colors within the color range of thermal picture was created in
Dynamo. Since the temperature range was set as 10 to 24 degree centigrade spanning 14
degree centigrade when shooting thermal photos, the color range image was divided
evenly into 1400 colors, representing difference of 0.01 degree centigrade among each
color.
After the color database was ready, the next procedure was to read the colors from
preprocessed thermal photos and matching colors with the database. The number of
sample pixels read from each photo at the specified grid points was set as 30 x 30 on X
44
and Y axes, according to the real size it was representing. As a result, a list containing
900 color pixels was created for each thermal photo. Then each color pixel was looked up
through the database to find the matched thermal color (surface temperature) respectively.
Then a list of surface temperatures read from each surface part was generated and input
into the equation selected after the algorithm validation to calculate the list of
corresponding R-values. At last, R-values were plotted back onto the surfaces of the
entire building model to indicate the degradation situation of envelope material and the
leakages that needed to be fixed.
The original vision of final prototype is a complete building information model in Revit
with R-values of enveloped being visualized (Figure 3.14).
Figure 3.14 Vision of final prototype
45
3.4 Chapter Summary
The methodology includes four sections: algorithm selection from literature review,
algorithm validation through lab tests, attach thermal information to building information
model, and the process to achieve R-value visualization. The first two sections are
theoretical basis, and the last two sections are the exploration to develop the protocol.
Four algorithms were selected and validated in the lab environment to build up the
theoretical basis. However, many obstacles were encountered during the process o f
developing the protocol. Some concession has to be made, and the final protocol could
still achieve the proposed hypothesis but with some assumptions and limitations. Detailed
process and the problems met while developing the protocol are recorded in chapter 4,
and the complete Dynamo script of the protocol is presented and explained in chapter 5.
46
Chapter 4 Process and Data
This chapter presents test data for the validation of the algorithms. Detailed processes of
some of the following tests are also shown, including the attempts at creating 3D thermal
models using Autodesk ReCap, difficulties met when assigning thermal photos and
viewing them in the BIM, and different methods of creating a color database in Dynamo.
4.1 Lab Tests for Algorithm Validation
Four algorithms were tested on two kinds of material. Surface temperatures were
measured by an infrared camera FLIR E8 by shooting thermal photos of the model
(Figure 4.1). As both inside and outside temperatures were changing, inside temperatures
were monitored by HOBOs and iButtons, and outside temperatures were recorded by
HOBOs, iButtons, and a tripod with temperature sensors connecting to LabVIEW
software. After comparison among those three temperature recorders, the data from the
HOBOs was adopted. Since the difference was up to 1 degree centigrade among three
kinds of sensors, taking an average among all the sensors was not going to make sense.
The HOBOs had been validated before, but not the iButtons and temperature sensors
installed on the tripod, so the average between the two HOBOs set inside the chamber
and the average between two HOBOs outside the model were determined as the inside
and outside temperature respectively instead. The raw data from the results of the lab
experiment were tabulated (Table 4.1, Appendix A). Among the according result samples
of calculation, data within ± 20% range were shown with an orange background, and cells
containing results within ± 30% range were colored light orange (Table 4.2, Appendix A).
47
Figure 4.1 Examples of thermal photos during test
Table 4.1 Sample raw data from test (Unit: ℃)
Time
HOBO
1
HOBO
2
T
inside
HOBO
3
HOBO
4
T
outside
T
surface
T
reflected
12/01/15 06:07
PM
27.76 27.96 27.86 25.03 24.55 24.79 25.60 27.50
12/01/15 06:08
PM
27.96 28.26 28.11 25.13 24.64 24.88 25.60 27.50
12/01/15 06:15
PM
29.45 29.85 29.65 26.00 25.42 25.71 26.70 28.40
12/01/15 06:16
PM
29.65 30.05 29.85 26.10 25.51 25.81 26.70 28.40
12/01/15 06:15
PM
29.45 29.85 29.65 26.00 25.42 25.71 27.00 28.40
12/01/15 06:16
PM
29.65 30.05 29.85 26.10 25.51 25.81 27.00 28.40
12/01/15 06:25
PM
31.47 31.88 31.68 27.08 26.39 26.73 27.90 26.80
12/01/15 06:26
PM
31.68 32.09 31.88 27.17 26.39 26.78 27.90 26.80
12/01/15 06:27
PM
31.88 32.19 32.03 27.27 26.49 26.88 27.90 26.80
12/01/15 06:25
PM
31.47 31.88 31.68 27.08 26.39 26.73 28.10 26.80
12/01/15 06:26
PM
31.68 32.09 31.88 27.17 26.39 26.78 28.10 26.80
12/01/15 06:27
PM
31.88 32.19 32.03 27.27 26.49 26.88 28.10 26.80
12/01/15 06:36
PM
33.43 33.85 33.64 27.96 27.17 27.57 28.40 27.90
12/01/15 06:37
PM
33.54 33.95 33.74 28.06 27.17 27.62 28.40 27.90
12/01/15 06:36
PM
33.43 33.85 33.64 27.96 27.17 27.57 28.70 27.90
12/01/15 06:37
PM
33.54 33.95 33.74 28.06 27.17 27.62 28.70 27.90
12/01/15 06:45
PM
34.69 35.12 34.90 28.66 27.76 28.21 29.30 28.60
12/01/15 06:46
PM
34.80 35.22 35.01 28.66 27.76 28.21 29.30 28.60
12/01/15 06:55
PM
35.97 36.40 36.19 29.55 28.56 29.05 29.90 29.60
12/01/15 06:56
PM
36.08 36.51 36.30 29.65 28.66 29.15 29.90 29.60
48
Table 4.2 Sample calculation results for material 1 (Unit: m
2
℃/W)
From a cursory review of the sample data, the second and third algorithms seemed
performing better with more results falling within the acceptable range. The complete
results are summarized in chapter 5.1.
Time I-1 II-1 III-1 IV-1
12/01/15 06:07 PM 0.95 0.70 0.17 0.23
12/01/15 06:08 PM 1.13 0.83 0.19 0.24
12/01/15 06:15 PM 0.99 0.73 0.21 0.30
12/01/15 06:16 PM 1.13 0.83 0.23 0.31
12/01/15 06:15 PM 0.76 0.56 0.20 0.31
12/01/15 06:16 PM 0.85 0.62 0.21 0.32
12/01/15 06:25 PM 1.06 0.76 0.98 0.46
12/01/15 06:26 PM 1.14 0.82 1.11 0.48
12/01/15 06:27 PM 1.26 0.91 1.14 0.50
12/01/15 06:25 PM 0.90 0.65 0.85 0.39
12/01/15 06:26 PM 0.97 0.70 0.95 0.41
12/01/15 06:27 PM 1.06 0.76 0.97 0.42
12/01/15 06:36 PM 1.82 1.31 1.17 0.99
12/01/15 06:37 PM 1.96 1.40 1.30 1.03
12/01/15 06:36 PM 1.34 0.96 0.95 0.67
12/01/15 06:37 PM 1.41 1.01 1.04 0.69
12/01/15 06:45 PM 1.53 1.09 1.02 0.80
12/01/15 06:46 PM 1.56 1.11 1.04 0.82
12/01/15 06:55 PM 2.11 1.49 1.10 1.40
12/01/15 06:56 PM 2.39 1.69 1.30 1.52
49
4.2 ReCap Tests
To assign thermal information to the model using the Autodesk ReCap 360 web API, four
projects were created. The first 3D model was created by importing 41 regular 2D digital
photos taken around the testing object; the 3D model was generated successfully from
these (Figure 4.2). The second attempt tried to create a project with 41 thermal photos,
but the system reported a failure of creation as a result indicating that the program could
not recognize thermal photos with false colors (Figure 4.3). The false colors of the
background might be the problem as the software program could not tell which part
belongs to the object and which part is actually the background. In the third try, the
background part of each thermal image was removed using Photoshop, and the processed
images were imported into ReCap. This did not work (Figure 4.4). But since thermal
photos and regular digital photos were taken at the same shoots that presenting exactly
the same view, it seemed possible to use the regular digital ones to create the 3D model
and somehow associate the thermal images as the texture map inside ReCap. Then, a new
project was tried to set up with 41 regular digital photos and 41 thermal photos input
together. 3D model was successfully built up, but only with the use of digital pictures,
and thermal photos listed down at the bottom were not stitched actually, which means
thermal data was still not attached (Figure 4.5).
50
Figure 4.2 Project from digital images
Figure 4.3 Project from thermal photos
Figure 4.4 Project from Photoshop thermal photos
51
Figure 4.5 Project from digital and thermal photos
The conclusion was drawn that thermal information attachment cannot be achieved with
the current ReCap 360 web API. But by revising the algorithms, the software program
could in the future potentially be able to recognize thermal pictures and achieve assigning
thermal colors to the digital 3D model at the exact same locations of points. Therefore, to
continue with the thermal information reading and R-value visualization, manually
matching thermal data to the model by parts was selected as the final plan.
4.3 Manually Assign Thermal Photos
It was intended to use “full size” false color images from the FLIR camera to map to the
surfaces onto the 3D model. However, the camera images could only be used for portions
of the wall, and the color images did not work. Hence, pre-processing of both 3D model
and image maps was necessary.
For the model, the "create parts" command in Revit was used to divide the exterior wall
surface into 3 meter x 3 meter squares. For walls, the default action of this demand in
52
Revit is to break the entire stacked wall into layers as separate parts instead of breaking
the surface of walls into parts. In order to divide the wall surfaces, reference lines needed
to be applied when selecting the target surface as reference working plane (Figure 4.6).
Infrared photos also needed to be preprocessed by cropping to show exact every 3 meter
x 3 meter square with texture of thermal colors (Figure 4.7). This required a large amount
of work since they are manually recognized, organized, and aligned.
Figure 4.6 Divide model surface using "create parts" in Revit
Figure 4.7 Sample thermal photo before and after cropping
Once building information model with divided wall surface and cropped thermal photos
were all ready, Dynamo plug-in of Revit was used to assign thermal photos to wall
surfaces of the model correspondingly. The Dynamo script is shown in detail in chapter
5.2.
53
4.4 View in BIM
The false colored building model can only be seen within Dynamo rather than Revit.
Because "Display.BySurfaceColor" node is able to accept an array of colors, then images
containing various colors can be shown on surfaces in Dynamo. However,
"Element.OverrideColorInView" command, which was originally designed to import
colors and override the colors in the view in Revit, is not able to take an array of colors
data and override on a single element. Instead, it can only accept a single value color on
each element (Figure 4.8).
Figure 4.8 Error reported when applying " Element.OverrideColorInView " node
Some research was done trying to solve this problem, and some people in the Dynamo
discuss group offered help by suggesting to use adaptive panels instead (Smith, 2015)
(Autodesk, 2015). This method works because each panel is assigned a certain color.
However, in terms of the building model in this case, every image is covering 3 m x 3 m
area, which is already quite small scope with detailed and complicated color texture.
Therefore adaptive panels are not a good choice. The crux of this problem is to fix the
program code of "Element.OverrideColorInView" node referring to
54
"Display.BySurfaceColor" node to make it able to accept array of colors data. If one
could display a set of pixels on each element in Revit, a better visualization of thermal
conditions of the building could be achieved in the future.
4.5 Create Colors Database
Two methods of generating the color range database were tried: reading the scale range
diagram from screenshot of infrared photos, and creating the scale through basic RGB
color components manually.
The "rainbow" mode (named for the spectrum of false colors that were used) was tried
first. A screenshot of the spectrum color range was captured and imported into Dynamo
"Image.ReadFromFile" node. As mentioned in chapter 3.3, the colors database should
contain 1400 colors. In this approach, however, the solution of screenshot was not high
enough. If one set 1400 pixels as sample number, more than 60% of the colors would be
read as the same in the list. Therefore, the method of expansion was adopted to generate
the colors database. 140 colors were read directly from the screenshot color range first,
then linear values were inserted in between RGB values respectively to expand the colors
list to 1400 colors (Figure 4.9).
55
Figure 4.9 "Rainbow" color range creation- method 1
The second method, "create the scale through basic RGB color components", as
mentioned above, was tested to obtain the entire detailed color database with 1400 co lors
in only one step. Seven colors were picked and generated by RGB values mimicking the
"rainbow" color range. White, red, orange, yellow, green, blue, and black were input and
placed evenly at the positions along the range. 1400 colors were read from the gradient
spectrum directly (Figure 4.10).
Figure 4.10 "Rainbow" color range creation- method 2
56
The methodology was reasonable and convincing enough, and it seemed everything was
going along the correct way according to the protocol described in chapter 3.3. However,
the result of following test running indicated these two "rainbow" color database were not
qualified for the coming calculation.
The logic was to read the colors from preprocessed thermal photos and match colors with
the database. If the color pixel read from thermal photos was able to find a match within
database, the index number of that color in the database list would be returned; if a color
read from thermal photos was not matching any of the combination in the database, the
program would return "-1" as the index number of the list to indicate false or illegal
instruction. Unfortunately, for the sample thermal photo being tested, neither of the
"rainbow" color database created by two methods could recognize over 10% of the pixel
colors read from the target thermal photo (Figure 4.11).
Figure 4.11 Failure of looking up colors in "rainbow" database
Assuming alpha is constant at 255, and is not a variable, then a certain color was decided
57
by a set of three values of red, green, and blue. Each component has 256 values from 0 to
255, even one digit being read differently, the combination would not match the existing
number sets listed within color database (Figure 4.12). 1400 colors were not enough to
cover the entire "rainbow" color range.
Figure 4.12 Colors by ARGB (alpha, red, green, and blue components)
To address this issue, the complexity of the color range had to be reduced. In the view
mode setting of the infrared camera, "iron" mode was applied instead of "rainbow"
(Figure 4.13). In contrast to the "rainbow" mode with seven signature colors. the color
range of "iron" mode has four ranging from white to orange to purple to black.
Theoretically, by reducing the complexity of color range, the deviation between the
colors from the created database and the colors read from thermal photos would be
reduced. Using the same method used in the "rainbow" mode, both "read the scale range
diagram from screenshot of infrared photos," and "create the scale through RGB color
components manually" were tested and run (Figure 4.14, 4.15). But the results turned out
that the recoganization rates were still below 20% (Figure 4.16).
58
Figure 4.13 Thermal photos in "rainbow", "iron", and "grey" mode
Figure 4.14 "Iron" color range creation- method 1
Figure 4.15 "Iron" color range creation- method 2
59
Figure 4.16 Failure of looking up colors in "iron" database
Another solution of using a "grey" range was tested as a method of last resort. Since the
signature colors for the range were reduced to only white and black, there was not much
difference between using the screenshot color range or manually created white to black
gradient (Figure 4.17). When importing the screenshot white to black color range, the
database was able to match more than 80% of the extracted colors from thermal photos.
While by generating white to black gradient manually, the database containing 1400
colors was able to cover more than 90% of the result, which appeared as a better practice
and was chosen as the final plan (Figure 4.18).
Figure 4.17 "Grey" color range creation- method 1
60
Figure 4.18 Success of looking up colors in "grey" database
"Grey" mode was chosen, although the "grey" thermal pictures would not look as colorful
and beautiful as "rainbow" mode. To continue with the following calculation and
visualization, steps in chapter 4.3 were redone to assign processed "grey" mode thermal
photos at the correct location of exterior wall surface parts. After the 100 pixel colors on
each surface part looking up matches through the white to black gradient color database
with 1400 color, a list of index numbers of database was generated, indicating colors of
each pixel.
4.6 Chapter Summary
Through the process of exploring the feasibility of the protocol, some problems were
encountered when doing tests. It was found that the original methodology roadmap was
full of potholes, some avoidable, others requiring a re-thinking of the process. These
issues fell into three categories: problems with the ReCap software, limitations of the
61
FLIR camera, and the lack of needed features in Dynamo and Revit.
These three categories of problems led to a revision in the original methodology: a Revit
model was created without the use of ReCap with the surfaces split manually, and grey
scale infrared images were used (Figure 4.19). The project continued with these
modifications.
Figure 4.19 Method of building up thermal model (Xs stand for unsuccessful attempts)
62
Chapter 5 Analysis and Results
This chapter covers the results of the algorithm validation, which are plotted as scatter
diagrams, and the final choice of equation to achieve from temperatures to thermal
resistances. The complete Dynamo script is presented in detailed from assigning thermal
photos to R-value visualization, and limits of the entire protocol are analyzed.
5.1 Algorithm Selection for R-value Calculation
The algorithms were calculated respectively according to the same sets of temperatures
data. For material 1, the notional thermal resistance is 4.4 hr ft
2
F/ Btu per inch, and the
sheet being tested is 1.5 inch thick, so the notional R-value of the test sample is 6.6 hr ft
2
F/ Btu in IP unit or 1.16 m
2
C/W in SI unit. Giving it a ± 20% range, the acceptable R-
value range of material 1 would be 0.93~1.39 m
2
C/W (Table 5.1). The results after
calculation were summarized in the scatter diagrams, in which the light purple
rectangular is covering the range of values which are acceptable (Figure 5.1).
The results of algorithm I and algorithm II were not quite as good, but algorithm III and
IV looked better. In algorithm I, most of the spots fell out of the purple box very far away.
As for algorithm II, the situation improved a little bit, but still not convincing enough to
be selected. Algorithm III got much better, with most of the points covered by the light
purple box, and others although not fell inside the ± 20% range, but centralized very well
and only a few point exceed ± 30% range. In terms of algorithm IV , the results closely
centered around the qualified range, but the accuracy was relatively not as good.
63
Table 5.1 Notional R-value of material 1
R-TECH
6.6 hr ft
2
F/ Btu
1.16 m
2
C/W (0.93~1.39) (± 20% )
Figure 5.1 R-values calculated from test- material 1
For the second material, the notional thermal resistance is 5.0 hr ft
2
F/ Btu per inch, and
the sheet being tested is inch thick, so the notional R-value of the test sample is 5 hr ft
2
F/
Btu in IP unit, or 0.88 m
2
C/W in SI unit. Giving it a ± 20% range, the acceptable R-value
range of material 1 would be 0.70~1.06 m
2
C/W (Table 5.2). The results after calculation
were presented in scatter diagrams, in which the light purple rectangular is covering the
range of values which can be accepted (Figure 5.2).
The results of algorithm I and II were random and far from reasonable. The third and the
fourth were closer and more trust worthy, especial algorithm III, with most of the points
64
actually fell into the light purple rectangular indicating a ± 20% range. Algorithm IV also
performed well, and results were really centralized around the accurate value.
Table 5.2 Notional R-value of material 2
FOAMULAR
5 hr ft
2
F/ Btu
0.88 m
2
C/W (0.70~1.06) (± 20% )
Figure 5.2 R-values calculated from test- material 2
After the comparison among all of these percentages of whether points were falling into
the acceptable area or not, the third algorithm was selected and its equation would be
applied as theoretical basis for the following R-value visualization (Equation 11).
R
T
in
T
out
T
s
T
out
T
out
4
T
ref
4
T
in
— inside temperature
T
out
— outside temperature
65
T
s
— surface temperature
α — convective heat transfer coefficients (ISO 6946)
T
ref
— use a small crumpled aluminum foil (low emissivity and high reflectivity) placed
on the inspection areas to measure
ε — thermal emissivity ( match object ε with measurement a piece of black tape ε = 0.95)
σ — Stefan–Boltzmann constant
5.2 Dynamo Script for Thermal Data Visualization in BIM
The entire Dynamo script consists of five sections: thermal photos assignment, colors
database creation, surface temperatures reading, R-values calculation, and thermal data
visualization (Figure 5.3).
Figure 5.3 The complete Dynamo script
5.2.1 Thermal Photos Assignment
With the Revit model and proper sized thermal photos ready, a Dynamo script was
written to achieve thermal information registration based on parts level (Figure 5.4).
Images were read from their file paths, and pixels were extracted from each image at the
66
density of 1 pixel per 9 square centimeters (~1 in
2
). An array of colors representing by
RGB color components was generated as a result. Then surfaces of the model in Revit
were selected in corresponding order, thus plotting each pixel at correct spots and
displaying thermal textures appropriately on each surface (Figure 5.5).
Figure 5.4 Dynamo script of assign thermal photos to building surfaces
Figure 5.5 Thermal textures on surfaces shown in Dynamo
5.2.2 Colors Database Creation
"Grey" mode was selected after several tests. The method, "create the scale through basic
RGB color components", was chosen to obtain the entire detailed color database with
1400 colors in one step. Since the signature colors for the range were reduced to only
white and black, a white to black gradient was created manually. 1400 colors were read
from the gradient evenly and directly to generate the colors database (Figure 5.6).
67
Figure 5.6 "Grey" colors database creation
5.2.3 Surface Temperatures Reading
After the color database was ready, the next procedure was reading colors from pre-
processed thermal photos and matching colors with the database. The number of sample
pixels read from each photo at the specified grid points was set as 30 x 30 on X and Y
axes. As a result, a list containing 900 color pixels was created for each thermal photo.
Then the list was flattened and chopped into 900 lists containing 1 color per list in order
to be able to look up matching colors in the color database respectively. If the color pixel
read from thermal photos was matching one of the color within database, the index of the
number of the color in color database list would be returned (Figure 5.7).
Figure 5.7 Index of colors in database were returned in a list
68
Because the temperature range was set as 10 to 24 degrees centigrade covering 14
degrees centigrade when shooting thermal photos, the grey color range database
contained 1400 colors, representing difference of 0.01 degree centigrade among each
color. A simple math was done to link index number 0~ 1399 to temperature range
10.00~24.00 degree centigrade as Equation (17) (Figure 5.8).
Ts n 0 01 10 Equation (17)
Then a list of surface temperature read from the specified 30 x 30 grid points on each
surface part was generated.
Figure 5.8 Math logic from index numbers to corresponding temperatures
5.2.4 R-values Calculation
According to the validated equation (11): R
T
in
T
out
T
s
T
out
T
out
4
T
ref
4
, the Dynamo script
was written to calculate the list of corresponding R-values (Figure 5.9). In the darker pink
box on the left, T
in,
T
out,
T
ref
are the values to be input, and T
s
is the list of temperatures
69
read from the thermal pictures. The output from the step would be the list of R-values
calculated from the specified 30 x 30 grid points on each surface part.
Figure 5.9 Equation to calculate R-values from temperatures
5.2.5 Thermal Data Visualization
The R-values of specified 30 x 30 grid points on each surface part were then plotted back
onto the surface to indicate the degradation situation of the envelope material and the
leakages that need to be fixed (Figure 5.10). A darker color means a relatively high R-
value and a lighter color indicates degradation or leakages (Figure 5.11). A list of thermal
images and surfaces can be matched using this protocol, and a complete model of a single
room or a small scale building can be created with visualization of R-values.
70
Figure 5.10 R-values plotting back onto building model surfaces
Figure 5.11 Degradation or leakages indication
5.3 Restrictions and Limits
As previously mentioned, the results of the lab testing led to the choice of Ham and
Golparvar-Fard’s algorithm. In hindsight, the calculation of R-value with the requirement
of knowing the interior and exterior temperatures was problematical given the choice of
building. A large, open warehouse would have been a better test case for the first round
or a single room would have been a better scope of test.
71
Other limitations introduced restrictions that had to be applied to the entire protocol. As
in the case studied, was assumed as 4.0 according to ISO 6946. was estimated equal
to 0.9, as general infrared inspection estimated (FLIR website, 2015). They were
considered to be the same on every surface in order to be processed all at once in batches.
The same happened to indoor temperatures and reflected temperatures. Theoretically, to
be precise enough, there are differences between temperatures of rooms or zones, because
it is very possible that air conditioning temperatures are set as differently from zone to
zone. The reflected temperature should be different depending on the location of camera
and the surrounding environment when shooting thermal photos.
Therefore, to test a whole building, a few assumptions needed to be made: the indoor
temperatures of each room should be considered the same as the air conditioned
temperatures; the outdoor temperatures around the building should be considered as a
single value; "reflected temperature", convective heat transfer coefficients , and thermal
emissivity should be constant for all surfaces.
5.4 Chapter Summary
Algorithm R
T
in
T
out
T
s
T
out
T
out
4
T
ref
4
was validated in a lab environment as being able to
calculate thermal resistance (R-value) using temperatures and some basic parameters and
coefficients. As some obstacles were encountered in the protocol developing process,
several adjustments and compromises had to be made. The final protocol was presented
in the revised Dynamo script that would apply to cases under certain conditions (Figure
5.12).
72
Figure 5.12 Overall workflow of revised protocol
73
Chapter 6 Conclusions and Future Works
In this chapter, conclusions are summarized and future work for developing the protocol
are suggested. Also, other protocols are criticized and limitations are explained; specific
situations to apply these methods are given. Further exploration of the application of this
topic is discussed.
6.1 Overall Significance
Due to the general degradation of material and possible construction deviation, a gap
always exists between the predicted building energy consumption and the actual
performance. To prevent a possibly high energy cost and to help owners be aware of
some degradation, leakages, infiltration, and mildew, it is necessary for existing buildings
to conduct a thorough inspection regularly. In order to achieve a more trust worthy energy
analysis of existing building, especially before big retrofitting, it is important to update
the building energy model with real measured properties instead of deriving values
directly from software tools' default databases.
Since building information modeling is widely used as a platform for building
information integration and helps with communication between building model and
energy simulation software, combination of updated real properties with BIM tool is
significant and may potentially be further used.
Infrared images can help detect problems that are invisible to the naked eyes but can
actually be treated for energy savings. Thermal resistance (R-value) can be calculated via
74
this non-contact method using thermal photos from infrared camera. And a building
information model with good visualization of real measured thermal data can be more
intuitive and clear to view, which will help owners and stakeholders make wiser decision
of whether retrofit is needed in time.
So far, the proposed protocol can be applied to small scale buildings or single spaces. A
false colored model can be generated with visualization of R-values in Dynamo, in which
deterioration of envelope materials, possible leakages, and some hidden problems can be
revealed.
6.2 From Thermal Image to Thermal Resistance
After running tests to validate the algorithms, the third algorithm was selected, which was
able to convert temperature values to thermal resistance with a relatively small error
(Ham and Golparvar-Fard, 2015). The equation is equation (11):
R
T
in
T
out
T
s
T
out
T
out
4
T
ref
4
For further explorations on algorithms, equations can be derived starting from the basic
heat transfer laws, and the number of parameters may be reduced. Then based on the
larger amount of data collected, one might be able to propose a more accurate algorithm
using data regression.
For the validation in the lab, more data could be recorded by adjusting temperatures at
wider ranges. As for recording temperatures, considering the deviation among devices,
75
different temperature sensors should be used to record temperature changes
simultaneously. An average can be taken as the final recorded value for every time spot.
This was the original reason why HOBOs, iButtons, and a tripod with temperature
sensors were used at the same time during the tests. However, due to lack of calibration
of devices, this idea finally failed coming to fruition and only data from HOBOs were
adopted.
For the entire test process, once the algorithm is determined, besides experiment under
lab environment, on site tests with full sized walls should also be done to further validate
the method. The calculated R-values from the infrared method should be compared with
results from other methods as well, such as simulated computerized prediction,
measurement by traditional heat flux meters on site, even lab-based destructive method
(eg. on an abandoned warehouse) that was eliminated from the beginning.
6.3 Thermal Data Visualization Protocol
For the 3D thermal model creation and R-value visualization in BIM parts, the revised
methodology used the manual assignment of thermal data rather than automatic metrics,
and many steps of the methodology were compromised due to limitations of the FLIR
camera and the lack of needed features in Dynamo and Revit. This led to a final protocol
that was not as easy, clear, or efficient as the original proposal. Since the proposed
methodology was constrained by several limitations that were discovered during the
research process, it is concluded that the use of a laser scanner to integrate thermal data
with geometry model on points level is far more accurate as has been accomplished by
76
other researchers (Ham, and Golparvar-Fard, 2013) (Cho and Wang, 2011) (Wang et al.,
2013) (Lagü ela et al., 2011) (Borrmann et al., 2012) (Im et al., 2012).
The protocol can be further developed by writing a Dynamo node for properly dealing
with a list of photos and elements through the same algorithm. Thus the protocol could
automatically match a list of thermal photos to a list of wall surfaces respectively, and
after R-values calculation is completed, thermal data would then be able to be plotted
back onto the list of surfaces respectively instead of all this being done manually. Then, a
complete model of a single room or a small scale building can be created with
visualization of R-values.
The current protocol could be used for a large open warehouse, a small scale building, or
a single space to explore material degradation and house leakage inspection. The
methodology could be improved and produce better results by having ReCap interpret
false color images properly, getting finer granularity of the FLIR camera image, and
writing specific nodes in Dynamo.
6.4 Future Works Referring to Thermal Models
To assign thermal information to building models, there are basically four approaches so
far: generating point cloud model with thermal data assigned using laser scanner,
manually attaching thermal textures to divided model surfaces, locating thermal pictures
based on the XYZ coordinate data collected by UA V (uninhabited aerial vehicle), and
automatically building up model using the ReCap software.
77
After several attempts and tests of various methods, the application of laser scanner
appears as the best choice. A laser scanner is able to build up actual sized building model
by scanning the object building from outside or generate the point cloud model of an
entire room space scanning from inside. When combined with an infrared camera, it is
able to collect surface temperatures of each individual points. Since data are processed on
points level, indoor temperatures and reflected temperatures can be input and assigned
respectively on different points as well. Then thermal resistance (R-value) can be
calculated on points level to indicate a much more trust worthy and detailed distribution,
which can be significant future work. The major downside is the cost of the laser scanner.
If using an UA V , the XYZ coordinate system should be established and both locations
and vector data of each photo shooting should be recorded. Photos are located according
to points, so there is no problem with overlapping photos. The advantage of this method
is that the angle of each photo can be fixed using recorded vectors, thus eliminating the
angular deviation of photos due to perspective view. Future work would be the
development of an automatic process dealing with a huge number of raw data to figure
out the certain correspondence between XYZ points of camera and XYZ points of
building surfaces, plus angular correction. If a drone is available on hand, navigation
around the building using UA V to collect data and establishing an XYZ coordinate
system to deal with data will definitely help a lot, especially when the target building is a
large sized one or with complex design.
78
Nevertheless, if no laser scanner or UA V available, manually registering 2D images with
3D model can yet be regard as a feasible solution. Photos cannot overlap since they are
assigned to divided surfaces individually. Pictures are required to be cropped to fit the
size of target surfaces and angles are adjusted manually. This preprocessing requires a lot
of manpower. Especially when a large number of 2D images were captured without geo-
tagged information attached, it is very difficult to figure out the location of each single
image and conduct further integration and fusion work. Hence, manually assigning
thermal information only applies to small sized building or single room/ space.
The best solution to build up thermal building models is to achieve 2D to 3D function of
thermal photos based on a software platform like Autodesk ReCap, which allows users to
get 3D model by just simply inputting 2D digital images. By revising the algorithms, the
software program could in the future potentially be able to recognize thermal pictures and
achieve assigning thermal colors to the digital 3D model at the exact same locations of
points. Then a 3D point cloud and mesh model encoding the thermal colors can be
generated at the point level, plus the associated rectification and the perspective angels
fixing. In the future, an automatic process and the clear user interface could be presented.
All users need to do is simply importing a series of 2D thermal and digital photos of the
building and everything else is taking care by the software.
6.5 Further Application
The topic of visualizing thermal data in building information models is of great
significance, and there are more related works can be done. Besides the potential
79
solutions discussed above aiming at the limitations of this protocol, broader and deeper
development can be make for further applications.
Once thermal properties can be calculated at each point or each small piece of area, one
can assign those values back into the Revit model properties on points level. In Revit, the
model is displayed by elements, so each element should have a single value of one
thermal parameter in order to appear in the properties. However, if an average is
calculated, some leakages of that element might be ignored, since the problematic part is
merged with normal parts and a slightly low value is shown rather than obvious
difference. Therefore, a better idea would be dividing the walls into pieces according to
calculated R-values. Thus no average is calculated and each point is keeping its originally
"measured" thermal resistance in properties.
In addition to diagnosing the degradation and leakages, this will also help with energy
analysis of the building. Since actual thermal resistance values have been updated in
Revit file, when transferring to other building energy modeling software, those values
would be updated through gbXML file being exported by Revit. Then, based on the
updated building energy model, the change of entire building energy consumption and
influence on energy bill can be analyzed. This could be done without leaving Revit in
Green Building Studio or other software that has plug-ins to Revit as well. The intent
would be to discover the differences between before and after the R-values update.
Furthermore, based on the same theory, a hand held tool/sensor for reading R-values of
80
building envelope can be developed. It can be transformed from an infrared camera. As
the algorithm needs indoor and outdoor temperatures input, if being able to connect with
the existing building manage system, a real-time R-value measuring tool can be achieved.
Then a quick measurement of R- value can be done instead of application of heat flow
meters collecting data for several hours or days.
Besides applications of combining thermal information and BIM, it is also instructive to
expand the vision to the combination of real data and BIM. Especially since many tiny
sensors are not expensive now, a wireless sensor network can be established with various
sensors monitoring indoor environment. In a more data driven building operations and
maintenance condition, almost all aspects of indoor environment can be monitored and
controlled based on real-time data from sensors, such as indoor temperatures, relative
humidity, air quality (carbon dioxide and pollutant levels), lighting environment, and
acoustic comfort.
For each sub-item, data from continuously monitored sensors can somehow be integrated
with real-time building management systems in a building information model, then a
smart built environment can be achieved. The good visualization in BIM will help
occupants be aware of the status of the surrounding environment. And suggestions may
be able to made to the occupants through the user interface of building management
system or through an affiliated mobile application. Then, occupants can set the mode to
adjust the environment manually or let the system perform autonomous intelligent actions.
Occupants' reaction to the system suggestion is actually giving the system feedback of
81
their own acceptable comfort zone. As a result, the control system can adjust the default
threshold and customize the built environment smartly, thus satisfying people's demand,
improving building efficiency, and saving energy at the same time.
6.6 Chapter Summary
In post-construction buildings, it is possible to obtain the approximate R-value of the
envelope from quick measurements with an infrared camera within certain limitations for
the methodology described. The thermal images could be put on the building information
model for visualization, and with the help of the Dynamo plug-in, Revit is able to detect
thermal colors and convert them to thermal data, even plot thermal data back onto the
building model. But the building should be small or single room space, and the revised
protocol appears not as convenient as originally proposed. Some future work can be done
to further improve this protocol. Further exploration on this topic is interesting and
significant for building industry. Buildings degrade over time, sometimes in ways not
easily noticeable by the owners. Infrared images can help detect problems that can be
treated for energy savings.
82
Bibliography
Albatici, Rossano, and Arnaldo M. Tonelli. 2010. “Infrared Thermovision Technique for
the Assessment of Thermal Transmittance Value of Opaque Building Elements on Site.”
Energy and Buildings 42 (11): 2177-83.
Alshiddi, Fahad. 2015. "Pre-cast Concrete Envelopoes in Hot-humid Climates: Examing
Envelops to Reduce Cooling Load and Elecrtical Consumption." Master of Building
Science thesis, University of Southern California, US.
ASTM E1862-97. 2002. Standard test methods for measuring and compensating for
reflected temperature using infrared imaging radiometers.
Autodesk. 2014. "Dynamo 0.7.X." Accessed February 4, 2016. http://archi-
lab.net/downloads/revitdynamo/
Autodesk ReCap. 2015. "How ReCap Works." https://recap.autodesk.com/
Borrmann, Dorit, Andreas Nü chter, Marija Dakulovic, Ivan Maurovic, Ivan Petrovic,
Dinko Osmankovic, and Jasmin V elagic. 2012. "The Project ThermalMapper- Thermal
3D Mapping of Indoor Environments for Saving Energy."Proceedings of the 10th
International IFAC Symposium on Robot Control (SYROCO '12), V olume 10, Part 1,
Dubrovnik, Croatia, September 2012
Borrmann, Dorit, Jan Elseberg, and Andreas Nü chter. 2012. "Thermal 3D Mapping of
Building Faç ades." Proceedings of the 12th Conference on Intelligent Autonomous
Systems (IAS '12), Jeju Island, Korea, June 2012
Boué, Christine, and Stéphane Holé. 2012. “Infrared Thermography Protocol for Simple
Measurements of Thermal Diffusivity and Conductivity.” Infrared Physics & Tech nology
55 (4): 376-79.
Cho, Y ., C. Wang. 2011. "3D thermal modeling for existing buildings using hybrid
LIDAR system." Proceedings of the 2011 ASCE International Workshop on Computing
in Civil Engineering ASCE, Miami, Florida.
Department of Energy. 2015. "Building Energy Modeling."
http://energy.gov/eere/buildings/building-energy-modeling
Eads, L.G. 2000. "Thermography." ASHRAE Journal 42 (2000): p. 51.
FLIR. 2015. "Thermal Imaging for Building Diagnostics."
http://www.flir.com/instruments/building/display/?id=49417
FLIR system. 2010. "Thermographic measurement techniques, FLIR Reporter
Professional, Professional Edition." 133-137.
Fokaides, Paris A., and Soteris A. Kalogirou. 2011. “Application of Infrared
83
Thermography for the Determination of the Overall Heat Transfer Coefficient (U-V alue)
in Building Envelopes.” Applied Energy 88 (12): 4358 -65.
Haasl, T., and K. Heinemeier. 2006. "California Commissioning Guide: New Buildings"
and "California Commissioning Guide: Existing Buildings". California Commissioning
Collaborative.
Ham, Y oungjib, and Mani Golparvar-Fard. 2015. “Mapping Actual Thermal Properties to
Building Elements in gbXML-Based BIM for Reliable Building Energy Performance
modeling.” Automation in Construction 49 (January): 214 -24.
Ham, Y oungjib, and Mani Golparvar-Fard. 2014. “Updating R -V alues of BIM Elements
Using 3D Thermography for Accurate Building Energy Performance Simulation.” The
Sixth International Conference on Computing in Civil and Building Engineering, Orlando,
Florida, June 23-25.
Ham, Y oungjib, and Mani Golparvar-Fard. 2013. “An Automated Vision -Based Method
for Rapid 3D Energy Performance Modeling of Existing Buildings Using Thermal and
Digital Imagery.” Advanced Engineering Informatics 27 (3): 395 -409.
Hammoudi, Karim, Fadi Dornaika, Bahman Soheilian, and Nicolas Paparoditis. 2010.
“Extracting Wire -Frame Models of Street Facades From 3D Point Clouds and the
Corresponding Cadastral Map.” Iaprs XXXVIII: 91 -96.
Hong, Taehoon, Choongwan Koo, Jimin Kim, Minhyun Lee, and Kwangbok Jeong. 2015.
“A Review on Sustainable Construction Management Strategies for Monitoring,
Diagnosing, and Retrofitting the Building’s Dynamic Energy Performance: Focused on
the Operation and Maintenance Phase.” Applied Energy 155 (October): 671 -707.
Im, H.J., M. Gai, C. Wang, Y . Cho. 2012. "Hybrid approach to visualize building energy
information model in geospatial application programs." Construction Research Congress,
(2012):1262-1270
Infrared Concepts Corporation. 2015. "Thermal Mapping & Surveying."
http://www.iccthermalmapping.com/
Infrared Training Center. 2015. "T Reflected and T Ambient."
http://www.infraredtraining.com/community/boards/thread/213/
Infrared training center. 2015. "Message Board Thread - T Reflected and T
Ambient." http://www.infraredtraining.com/community/boards/thread/213/
Insulfoam. 2015. " Energy Efficient Do-It-All Insulation for Residential Applications. "
http://insulfoam.com/energy-efficient-do-it-all-insulation-for-residential-applications/
ISO 18434-1. 2008. Condition monitoring and diagnostics of machines - thermography -
part 1: general procedures.
84
ISO 6946. 2007. Building components and building elements - Thermal resistance and
thermal transmittance - Calculation method.
Kensek, Karen M., and Douglas Noble. 2014. Building Information Modeling: BIM in
Current and Future Practice.
Kim, Min-Koo, Jack C.P. Cheng, Hoon Sohn, and Chih-Chen Chang. 2015. “A
Framework for Dimensional and Surface Quality Assessment of Precast Concrete
Elements Using BIM and 3D Laser Scanning.” Automati on in Construction 49 (January):
225-38.
Lagüela, S., J. Martínez, J. Armesto, and P. Arias. 2011. “Energy Efficiency Studies
through 3D Laser Scanning and Thermographic Technologies.” Energy and Buildings 43
(6): 1216–21.
National BIM Standards - United States. 2015. "What is a BIM?"
https://www.nationalbimstandard.org/faqs#faq1
NIST. 2010. "Measurement Science Roadmap for Net-Zero Energy Buildings." National
Institute of Standards and Technology (NIST) Technical Note 1660.
Riaz, Zainab, M. Arslan, Adnan K. Kiani, and Salman Azhar. 2014. “CoSMoS: A BIM
and Wireless Sensor Based Integrated Solution for Worker Safety in Confined Spaces.”
Automation in Construction 45 (September): 96-106.
SkyCatch. 2015. "What we do." https://www.skycatch.com/
Smith, Lilli. 2015. "Dynamo 0.8.1 Now in Technicolor! " Last modified April 16.
http://dynamobim.org/dynamo-0-8-1-now-in-technicolor/
Tritt, Terry M.. 2004. Thermal Conductivity: Theory, Properties, and Applications.
UNEP-SBCI. 2015. "Why Buildings."
http://www.unep.org/sbci/AboutSBCI/Background.asp
U.S. DOE. 2010. 2010 U.S. DOE Buildings Energy Databook, U.S. Department of
Energy.
U.S. General Services Administration. 2012. GSA BIM Guide for Energy Performance.
V an Der Heijden, R, A.Tai, and G. Fagerströ m. 2015. "V olatile Data Mining: A proof of
concept for the creation of data-enhanced represen-tations of the built environment using
consumer grade drone." Proceedings of the 20th International Conference of the
Association for Computer-Aided Architectural Design Research in Asia CAADRIA 2015
Wang, C., Y . Cho, and M. Gai. 2014. "As-Is 3D thermal modeling for existing building
envelopes using a hybrid LIDAR system." Journal of Computing in Civil Engineering
27(6):645-656.
85
Appendices
Appendix A Complete data and results of lab test (algorithm validation)
Table A.1 Data of lab test - Material 1
HOBO 1 HOBO 2 T inside HOBO 3 HOBO 4 T outside T surface T reflected
1 36.95 38.49 37.72 28.06 28.46 28.26 26.2 27.7
2 38.71 40.30 39.51 28.06 28.46 28.26 26.2 27.7
3 47.50 60.16 53.83 26.88 27.37 27.12 28.2 29.9
4 48.69 63.76 56.23 26.88 27.47 27.17 28.2 29.9
5 39.39 44.83 42.11 27.96 28.06 28.01 29.4 28.9
6 39.39 44.83 42.11 27.96 28.06 28.01 29.4 27.2
7 39.39 44.83 42.11 27.96 28.06 28.01 29.4 27.3
8 39.39 44.83 42.11 27.96 28.06 28.01 29.4 27.1
9 39.39 44.83 42.11 27.96 28.06 28.01 29.4 27.7
10 39.39 44.83 42.11 27.96 28.06 28.01 29.4 29.8
11 39.05 43.97 41.51 28.06 28.06 28.06 29.4 28.9
12 39.05 43.97 41.51 28.06 28.06 28.06 29.4 27.2
13 39.05 43.97 41.51 28.06 28.06 28.06 29.4 27.3
14 39.05 43.97 41.51 28.06 28.06 28.06 29.4 27.1
15 39.05 43.97 41.51 28.06 28.06 28.06 29.4 27.7
16 39.05 43.97 41.51 28.06 28.06 28.06 29.4 29.8
17 38.71 43.24 40.98 28.16 28.16 28.16 29.4 28.9
18 38.71 43.24 40.98 28.16 28.16 28.16 29.4 27.2
19 38.71 43.24 40.98 28.16 28.16 28.16 29.4 27.3
20 38.71 43.24 40.98 28.16 28.16 28.16 29.4 27.1
21 38.71 43.24 40.98 28.16 28.16 28.16 29.4 27.7
22 38.71 43.24 40.98 28.16 28.16 28.16 29.4 29.8
23 38.38 42.52 40.45 28.26 28.26 28.26 29.4 28.9
24 38.38 42.52 40.45 28.26 28.26 28.26 29.4 27.2
25 38.38 42.52 40.45 28.26 28.26 28.26 29.4 27.3
26 38.38 42.52 40.45 28.26 28.26 28.26 29.4 27.1
27 38.38 42.52 40.45 28.26 28.26 28.26 29.4 27.7
28 38.38 42.52 40.45 28.26 28.26 28.26 29.4 29.8
29 39.39 44.83 42.11 27.96 28.06 28.01 28.4 29.8
30 39.05 43.97 41.51 28.06 28.06 28.06 28.4 29.8
31 38.71 43.24 40.98 28.16 28.16 28.16 28.4 29.8
32 38.38 42.52 40.45 28.26 28.26 28.26 28.4 29.8
33 39.39 44.83 42.11 27.96 28.06 28.01 27.5 28.9
34 39.39 44.83 42.11 27.96 28.06 28.01 27.5 29.8
35 39.05 43.97 41.51 28.06 28.06 28.06 27.5 28.9
36 39.05 43.97 41.51 28.06 28.06 28.06 27.5 29.8
86
37 38.71 43.24 40.98 28.16 28.16 28.16 27.5 28.9
38 38.71 43.24 40.98 28.16 28.16 28.16 27.5 29.8
39 38.38 42.52 40.45 28.26 28.26 28.26 27.5 28.9
40 38.38 42.52 40.45 28.26 28.26 28.26 27.5 29.8
41 36.40 38.71 37.56 28.95 28.75 28.85 29.4 28
42 36.40 38.71 37.56 28.95 28.75 28.85 29.4 28.6
43 36.19 38.27 37.23 29.05 28.75 28.90 29.4 28
44 36.19 38.27 37.23 29.05 28.75 28.90 29.4 28.6
45 34.59 35.65 35.12 29.65 29.25 29.45 28.9 29.2
46 34.37 35.44 34.90 29.75 29.25 29.50 28.9 29.2
47 27.76 27.96 27.86 25.03 24.55 24.79 25.6 27.5
48 27.96 28.26 28.11 25.13 24.64 24.88 25.6 27.5
49 29.45 29.85 29.65 26.00 25.42 25.71 26.7 28.4
50 29.65 30.05 29.85 26.10 25.51 25.81 26.7 28.4
51 29.45 29.85 29.65 26.00 25.42 25.71 27 28.4
52 29.65 30.05 29.85 26.10 25.51 25.81 27 28.4
53 31.47 31.88 31.68 27.08 26.39 26.73 27.9 26.8
54 31.47 31.88 31.68 27.08 26.39 26.73 27.9 27
55 31.68 32.09 31.88 27.17 26.39 26.78 27.9 26.8
56 31.68 32.09 31.88 27.17 26.39 26.78 27.9 27
57 31.88 32.19 32.03 27.27 26.49 26.88 27.9 26.8
58 31.88 32.19 32.03 27.27 26.49 26.88 27.9 27
59 31.47 31.88 31.68 27.08 26.39 26.73 28.1 26.8
60 31.68 32.09 31.88 27.17 26.39 26.78 28.1 26.8
61 31.88 32.19 32.03 27.27 26.49 26.88 28.1 26.8
62 33.43 33.85 33.64 27.96 27.17 27.57 28.4 27.9
63 33.43 33.85 33.64 27.96 27.17 27.57 28.4 27.4
64 33.54 33.95 33.74 28.06 27.17 27.62 28.4 27.9
65 33.54 33.95 33.74 28.06 27.17 27.62 28.4 27.4
66 33.43 33.85 33.64 27.96 27.17 27.57 28.7 27.9
67 33.43 33.85 33.64 27.96 27.17 27.57 28.7 27.4
68 33.54 33.95 33.74 28.06 27.17 27.62 28.7 27.9
69 33.54 33.95 33.74 28.06 27.17 27.62 28.7 27.4
70 34.69 35.12 34.90 28.66 27.76 28.21 29.3 28.6
71 34.69 35.12 34.90 28.66 27.76 28.21 29.3 28.7
72 34.80 35.22 35.01 28.66 27.76 28.21 29.3 28.6
73 34.80 35.22 35.01 28.66 27.76 28.21 29.3 28.7
74 35.97 36.40 36.19 29.55 28.56 29.05 29.9 29.6
75 35.97 36.40 36.19 29.55 28.56 29.05 29.9 29.2
76 36.08 36.51 36.30 29.65 28.66 29.15 29.9 29.6
77 36.08 36.51 36.30 29.65 28.66 29.15 29.9 29.2
78 37.27 37.71 37.49 30.76 29.65 30.21 31.6 30.9
79 37.38 37.82 37.60 30.86 29.75 30.31 31.6 30.9
87
80 38.05 38.60 38.32 30.05 29.05 29.55 28.5 30.3
81 38.05 38.60 38.32 30.05 29.05 29.55 28.5 28.6
82 38.05 38.60 38.32 30.05 29.05 29.55 28.5 28.8
83 38.05 38.60 38.32 30.05 29.05 29.55 28.5 28.2
84 38.16 38.71 38.44 29.95 28.95 29.45 28.5 30.3
85 38.16 38.71 38.44 29.95 28.95 29.45 28.5 28.6
86 38.16 38.71 38.44 29.95 28.95 29.45 28.5 28.8
87 38.16 38.71 38.44 29.95 28.95 29.45 28.5 28.2
88 38.05 39.28 38.66 28.06 28.36 28.21 27.4 28.9
89 38.16 39.39 38.77 28.06 28.36 28.21 27.4 28.9
90 38.05 39.28 38.66 28.06 28.36 28.21 26.7 28.9
91 38.16 39.39 38.77 28.06 28.36 28.21 26.7 28.9
92 37.38 40.19 38.78 27.57 28.16 27.86 26.4 26.8
93 37.38 40.19 38.78 27.57 28.16 27.86 26.4 26.3
94 37.38 40.19 38.78 27.57 28.16 27.86 26.4 26.7
95 37.38 40.19 38.78 27.57 28.16 27.86 26.4 28.3
96 37.49 40.19 38.84 27.47 28.06 27.76 26.4 26.8
97 37.49 40.19 38.84 27.47 28.06 27.76 26.4 26.3
98 37.49 40.19 38.84 27.47 28.06 27.76 26.4 26.7
99 37.49 40.19 38.84 27.47 28.06 27.76 26.4 28.3
100 37.49 40.30 38.90 27.47 27.96 27.71 26.4 26.8
101 37.49 40.30 38.90 27.47 27.96 27.71 26.4 26.3
102 37.49 40.30 38.90 27.47 27.96 27.71 26.4 26.7
103 37.49 40.30 38.90 27.47 27.96 27.71 26.4 28.3
104 37.60 40.53 39.07 27.47 27.86 27.67 26.4 26.8
105 37.60 40.53 39.07 27.47 27.86 27.67 26.4 26.3
106 37.60 40.53 39.07 27.47 27.86 27.67 26.4 26.7
107 37.60 40.53 39.07 27.47 27.86 27.67 26.4 28.3
108 37.38 40.19 38.78 27.57 28.16 27.86 26.7 26.8
109 37.38 40.19 38.78 27.57 28.16 27.86 26.7 26.3
110 37.38 40.19 38.78 27.57 28.16 27.86 26.7 26.7
111 37.38 40.19 38.78 27.57 28.16 27.86 26.7 28.3
112 37.49 40.19 38.84 27.47 28.06 27.76 26.7 26.8
113 37.49 40.19 38.84 27.47 28.06 27.76 26.7 26.3
114 37.49 40.19 38.84 27.47 28.06 27.76 26.7 26.7
115 37.49 40.19 38.84 27.47 28.06 27.76 26.7 28.3
116 37.49 40.30 38.90 27.47 27.96 27.71 26.7 26.8
117 37.49 40.30 38.90 27.47 27.96 27.71 26.7 26.3
118 37.49 40.30 38.90 27.47 27.96 27.71 26.7 26.7
119 37.49 40.30 38.90 27.47 27.96 27.71 26.7 28.3
120 37.60 40.53 39.07 27.47 27.86 27.67 26.7 26.8
121 37.60 40.53 39.07 27.47 27.86 27.67 26.7 26.3
88
122 37.60 40.53 39.07 27.47 27.86 27.67 26.7 26.7
123 37.60 40.53 39.07 27.47 27.86 27.67 26.7 28.3
89
Table A.2 Result of algorithm validation - Material 1
I II III IV
1 1.15 0.83 0.83 0.58
2 1.37 0.99 0.99 0.68
3 6.21 4.47 1.34 1.94
4 7.08 5.09 1.50 2.14
5 2.54 1.81 1.34 1.68
6 2.54 1.81 1.40 0.78
7 2.54 1.81 1.48 0.81
8 2.54 1.81 1.33 0.76
9 2.54 1.81 1.93 0.93
10 2.54 1.81 0.90 1.80
11 2.51 1.79 1.34 1.64
12 2.51 1.79 1.33 0.76
13 2.51 1.79 1.40 0.78
14 2.51 1.79 1.26 0.73
15 2.51 1.79 1.83 0.90
16 2.51 1.79 0.89 1.76
17 2.58 1.84 1.41 1.65
18 2.58 1.84 1.25 0.74
19 2.58 1.84 1.32 0.76
20 2.58 1.84 1.18 0.71
21 2.58 1.84 1.70 0.88
22 2.58 1.84 0.90 1.77
23 2.67 1.90 1.49 1.65
24 2.67 1.90 1.17 0.72
25 2.67 1.90 1.23 0.74
26 2.67 1.90 1.11 0.69
27 2.67 1.90 1.59 0.86
28 2.67 1.90 0.92 1.79
29 9.05 6.48 1.21 1.50
30 9.89 7.08 1.21 1.46
31 13.30 9.51 1.26 1.46
32 21.46 15.35 1.32 1.45
33 6.90 4.97 2.01 1.44
34 6.90 4.97 1.16 0.95
35 6.00 4.32 1.94 1.34
36 6.00 4.32 1.12 0.90
37 4.86 3.49 1.89 1.23
38 4.86 3.49 1.08 0.83
39 4.02 2.89 1.84 1.13
40 4.02 2.89 1.04 0.77
90
41 3.98 2.83 1.25 0.86
42 3.98 2.83 2.41 1.30
43 4.19 2.97 1.18 0.84
44 4.19 2.97 2.26 1.28
45 2.57 1.82 1.56 1.45
46 2.24 1.59 1.31 1.32
47 0.95 0.70 0.17 0.23
48 1.13 0.83 0.19 0.24
49 0.99 0.73 0.21 0.30
50 1.13 0.83 0.23 0.31
51 0.76 0.56 0.20 0.31
52 0.85 0.62 0.21 0.32
53 1.06 0.76 0.98 0.46
54 1.06 0.76 0.80 0.51
55 1.14 0.82 1.11 0.48
56 1.14 0.82 0.90 0.54
57 1.26 0.91 1.14 0.50
58 1.26 0.91 1.09 0.57
59 0.90 0.65 0.85 0.39
60 0.97 0.70 0.95 0.41
61 1.06 0.76 0.97 0.42
62 1.82 1.31 1.17 0.99
63 1.82 1.31 1.43 0.68
64 1.96 1.40 1.30 1.03
65 1.96 1.40 1.41 0.70
66 1.34 0.96 0.95 0.67
67 1.34 0.96 1.11 0.51
68 1.41 1.01 1.04 0.69
69 1.41 1.01 1.11 0.53
70 1.53 1.09 1.02 0.80
71 1.53 1.09 0.94 0.86
72 1.56 1.11 1.04 0.82
73 1.56 1.11 0.96 0.88
74 2.11 1.49 1.10 1.40
75 2.11 1.49 1.69 0.97
76 2.39 1.69 1.30 1.52
77 2.39 1.69 2.20 1.03
78 1.31 0.91 0.76 0.76
79 1.41 0.98 0.85 0.79
80 2.08 1.48 1.04 0.61
81 2.08 1.48 0.91 1.84
82 2.08 1.48 1.04 1.49
83 2.08 1.48 0.74 1.49
91
84 2.36 1.67 1.04 0.65
85 2.36 1.67 1.04 2.05
86 2.36 1.67 1.20 1.64
87 2.36 1.67 0.83 1.64
88 3.23 2.32 1.47 0.91
89 3.27 2.35 1.49 0.91
90 1.73 1.25 1.05 0.58
91 1.75 1.26 1.07 0.58
92 1.87 1.35 0.93 1.36
93 1.87 1.35 0.75 1.71
94 1.87 1.35 0.89 1.46
95 1.87 1.35 1.32 0.67
96 2.03 1.47 1.03 1.45
97 2.03 1.47 0.82 1.84
98 2.03 1.47 0.98 1.56
99 2.03 1.47 1.31 0.70
100 2.13 1.54 1.08 1.50
101 2.13 1.54 0.86 1.93
102 2.13 1.54 1.03 1.62
103 2.13 1.54 1.31 0.71
104 2.25 1.63 1.16 1.57
105 2.25 1.63 0.91 2.03
106 2.25 1.63 1.10 1.70
107 2.25 1.63 1.33 0.74
108 2.35 1.70 1.04 2.10
109 2.35 1.70 0.82 1.59
110 2.35 1.70 0.98 2.35
111 2.35 1.70 1.54 0.81
112 2.60 1.88 1.15 2.30
113 2.60 1.88 0.90 1.72
114 2.60 1.88 1.09 2.60
115 2.60 1.88 1.53 0.85
116 2.76 1.99 1.23 2.43
117 2.76 1.99 0.94 1.79
118 2.76 1.99 1.16 2.76
119 2.76 1.99 1.53 0.87
120 2.95 2.14 1.32 2.59
121 2.95 2.14 1.00 1.88
122 2.95 2.14 1.24 2.95
123 2.95 2.14 1.54 0.90
92
Table A.3 Data of lab test - Material 2
HOBO1 HOBO2 T inside HOBO3 HOBO4 T outside T surface T reflected
1 41.93 43.97 42.95 27.96 28.46 28.21 30.1 28.7
2 41.93 43.97 42.95 27.96 28.46 28.21 30.1 28.8
3 41.93 43.97 42.95 27.96 28.46 28.21 30.1 28.5
4 43.60 46.08 44.84 27.86 28.46 28.16 30.1 28.7
5 43.60 46.08 44.84 27.86 28.46 28.16 30.1 28.8
6 43.60 46.08 44.84 27.86 28.46 28.16 30.1 28.5
7 41.93 43.97 42.95 27.96 28.46 28.21 31.4 28.7
8 41.93 43.97 42.95 27.96 28.46 28.21 31.4 28.8
9 41.93 43.97 42.95 27.96 28.46 28.21 31.4 28.5
10 43.60 46.08 44.84 27.86 28.46 28.16 31.4 28.7
11 43.60 46.08 44.84 27.86 28.46 28.16 31.4 28.8
12 43.60 46.08 44.84 27.86 28.46 28.16 31.4 28.5
13 45.08 48.16 46.62 27.86 28.46 28.16 31.4 28.7
14 45.08 48.16 46.62 27.86 28.46 28.16 31.4 28.8
15 45.08 48.16 46.62 27.86 28.46 28.16 31.4 28.5
16 40.65 43.00 41.82 26.98 27.37 27.17 29.2 28.1
17 40.65 43.00 41.82 26.98 27.37 27.17 29.2 26.2
18 40.65 43.00 41.82 26.98 27.37 27.17 29.2 27.9
19 40.65 43.00 41.82 26.98 27.37 27.17 29.2 25.9
20 40.65 43.00 41.82 26.98 27.37 27.17 29.2 26
21 41.81 44.46 43.13 26.98 27.37 27.17 29.2 28.1
22 41.81 44.46 43.13 26.98 27.37 27.17 29.2 26.2
23 41.81 44.46 43.13 26.98 27.37 27.17 29.2 27.9
24 41.81 44.46 43.13 26.98 27.37 27.17 29.2 25.9
25 41.81 44.46 43.13 26.98 27.37 27.17 29.2 26
26 42.88 45.96 44.42 26.98 27.37 27.17 29.2 26.2
27 42.88 45.96 44.42 26.98 27.37 27.17 29.2 27.9
28 42.88 45.96 44.42 26.98 27.37 27.17 29.2 25.9
29 42.88 45.96 44.42 26.98 27.37 27.17 29.2 26
30 43.97 48.69 46.33 26.98 27.37 27.17 29.2 26.2
31 43.97 48.69 46.33 26.98 27.37 27.17 29.2 25.9
32 43.97 48.69 46.33 26.98 27.37 27.17 29.2 26
33 45.20 52.44 48.82 26.88 27.37 27.12 29.2 26.2
34 45.20 52.44 48.82 26.88 27.37 27.12 29.2 25.9
35 45.20 52.44 48.82 26.88 27.37 27.12 29.2 26
36 40.65 43.00 41.82 26.98 27.37 27.17 29.4 28.1
37 40.65 43.00 41.82 26.98 27.37 27.17 29.4 26.2
38 40.65 43.00 41.82 26.98 27.37 27.17 29.4 27.9
39 40.65 43.00 41.82 26.98 27.37 27.17 29.4 28.2
40 40.65 43.00 41.82 26.98 27.37 27.17 29.4 25.9
41 40.65 43.00 41.82 26.98 27.37 27.17 29.4 26
42 41.81 44.46 43.13 26.98 27.37 27.17 29.4 28.1
43 41.81 44.46 43.13 26.98 27.37 27.17 29.4 26.2
44 41.81 44.46 43.13 26.98 27.37 27.17 29.4 27.9
45 41.81 44.46 43.13 26.98 27.37 27.17 29.4 28.2
46 41.81 44.46 43.13 26.98 27.37 27.17 29.4 25.9
47 41.81 44.46 43.13 26.98 27.37 27.17 29.4 26
93
48 42.88 45.96 44.42 26.98 27.37 27.17 29.4 28.1
49 42.88 45.96 44.42 26.98 27.37 27.17 29.4 26.2
50 42.88 45.96 44.42 26.98 27.37 27.17 29.4 27.9
51 42.88 45.96 44.42 26.98 27.37 27.17 29.4 28.2
52 42.88 45.96 44.42 26.98 27.37 27.17 29.4 25.9
53 42.88 45.96 44.42 26.98 27.37 27.17 29.4 26
54 43.97 48.69 46.33 26.98 27.37 27.17 29.4 26.2
55 43.97 48.69 46.33 26.98 27.37 27.17 29.4 27.9
56 43.97 48.69 46.33 26.98 27.37 27.17 29.4 25.9
57 43.97 48.69 46.33 26.98 27.37 27.17 29.4 26
58 45.20 52.44 48.82 26.88 27.37 27.12 29.4 26.2
59 45.20 52.44 48.82 26.88 27.37 27.12 29.4 27.9
60 45.20 52.44 48.82 26.88 27.37 27.12 29.4 25.9
61 45.20 52.44 48.82 26.88 27.37 27.12 29.4 26
62 40.65 43.00 41.82 26.98 27.37 27.17 29.7 28.1
63 40.65 43.00 41.82 26.98 27.37 27.17 29.7 26.2
64 40.65 43.00 41.82 26.98 27.37 27.17 29.7 27.9
65 40.65 43.00 41.82 26.98 27.37 27.17 29.7 28.2
66 40.65 43.00 41.82 26.98 27.37 27.17 29.7 25.9
67 40.65 43.00 41.82 26.98 27.37 27.17 29.7 26
68 41.81 44.46 43.13 26.98 27.37 27.17 29.7 28.1
69 41.81 44.46 43.13 26.98 27.37 27.17 29.7 26.2
70 41.81 44.46 43.13 26.98 27.37 27.17 29.7 27.9
71 41.81 44.46 43.13 26.98 27.37 27.17 29.7 28.2
72 41.81 44.46 43.13 26.98 27.37 27.17 29.7 25.9
73 41.81 44.46 43.13 26.98 27.37 27.17 29.7 26
74 42.88 45.96 44.42 26.98 27.37 27.17 29.7 28.1
75 42.88 45.96 44.42 26.98 27.37 27.17 29.7 26.2
76 42.88 45.96 44.42 26.98 27.37 27.17 29.7 27.9
77 42.88 45.96 44.42 26.98 27.37 27.17 29.7 28.2
78 42.88 45.96 44.42 26.98 27.37 27.17 29.7 25.9
79 42.88 45.96 44.42 26.98 27.37 27.17 29.7 26
80 43.97 48.69 46.33 26.98 27.37 27.17 29.7 28.1
81 43.97 48.69 46.33 26.98 27.37 27.17 29.7 26.2
82 43.97 48.69 46.33 26.98 27.37 27.17 29.7 27.9
83 43.97 48.69 46.33 26.98 27.37 27.17 29.7 28.2
84 43.97 48.69 46.33 26.98 27.37 27.17 29.7 25.9
85 43.97 48.69 46.33 26.98 27.37 27.17 29.7 26
86 45.20 52.44 48.82 26.88 27.37 27.12 29.7 28.1
87 45.20 52.44 48.82 26.88 27.37 27.12 29.7 26.2
88 45.20 52.44 48.82 26.88 27.37 27.12 29.7 27.9
89 45.20 52.44 48.82 26.88 27.37 27.12 29.7 28.2
90 45.20 52.44 48.82 26.88 27.37 27.12 29.7 25.9
91 45.20 52.44 48.82 26.88 27.37 27.12 29.7 26
92 47.50 60.16 53.83 26.88 27.37 27.12 30.7 26.8
93 47.50 60.16 53.83 26.88 27.37 27.12 30.7 29.1
94 47.50 60.16 53.83 26.88 27.37 27.12 30.7 28.4
95 47.50 60.16 53.83 26.88 27.37 27.12 30.7 29.9
96 48.69 63.76 56.23 26.88 27.47 27.17 30.7 26.8
97 48.69 63.76 56.23 26.88 27.47 27.17 30.7 29.1
94
98 48.69 63.76 56.23 26.88 27.47 27.17 30.7 28.4
99 48.69 63.76 56.23 26.88 27.47 27.17 30.7 29.9
100 50.73 70.77 60.75 26.98 27.47 27.22 30.9 27.4
101 50.73 70.77 60.75 26.98 27.47 27.22 30.9 27.1
102 50.73 70.77 60.75 26.98 27.47 27.22 30.9 26.3
103 51.72 73.83 62.77 26.98 27.47 27.22 30.9 27.4
104 51.72 73.83 62.77 26.98 27.47 27.22 30.9 27.1
105 51.72 73.83 62.77 26.98 27.47 27.22 30.9 26.3
106 52.01 75.33 63.67 26.98 27.47 27.22 30.9 27.4
107 52.01 75.33 63.67 26.98 27.47 27.22 30.9 27.1
108 52.01 75.33 63.67 26.98 27.47 27.22 30.9 26.3
109 51.30 73.58 62.44 26.98 27.47 27.22 30.9 27.4
110 51.30 73.58 62.44 26.98 27.47 27.22 30.9 27.1
111 51.30 73.58 62.44 26.98 27.47 27.22 30.9 26.3
112 39.39 44.83 42.11 27.96 28.06 28.01 28.7 29.8
113 39.05 43.97 41.51 28.06 28.06 28.06 28.7 29.8
114 38.71 43.24 40.98 28.16 28.16 28.16 28.7 29.8
115 38.38 42.52 40.45 28.26 28.26 28.26 28.7 29.8
116 36.40 38.71 37.56 28.95 28.75 28.85 29.7 30.7
117 36.40 38.71 37.56 28.95 28.75 28.85 29.7 28
118 36.19 38.27 37.23 29.05 28.75 28.90 29.7 30.7
119 36.19 38.27 37.23 29.05 28.75 28.90 29.7 28
120 34.59 35.65 35.12 29.65 29.25 29.45 30.2 30.5
121 34.59 35.65 35.12 29.65 29.25 29.45 30.2 29.2
122 34.37 35.44 34.90 29.75 29.25 29.50 30.2 30.5
123 34.37 35.44 34.90 29.75 29.25 29.50 30.2 29.2
124 31.47 31.88 31.68 27.08 26.39 26.73 28 27.3
125 31.47 31.88 31.68 27.08 26.39 26.73 28 26.8
126 31.47 31.88 31.68 27.08 26.39 26.73 28 27
127 31.68 32.09 31.88 27.17 26.39 26.78 28 27.3
128 31.68 32.09 31.88 27.17 26.39 26.78 28 26.8
129 31.68 32.09 31.88 27.17 26.39 26.78 28 27
130 31.88 32.19 32.03 27.27 26.49 26.88 28 27.3
131 31.88 32.19 32.03 27.27 26.49 26.88 28 26.8
132 31.88 32.19 32.03 27.27 26.49 26.88 28 27
133 31.47 31.88 31.68 27.08 26.39 26.73 28.2 26.8
134 31.47 31.88 31.68 27.08 26.39 26.73 28.2 27
135 31.68 32.09 31.88 27.17 26.39 26.78 28.2 26.8
136 31.68 32.09 31.88 27.17 26.39 26.78 28.2 27
137 31.88 32.19 32.03 27.27 26.49 26.88 28.2 26.8
138 31.88 32.19 32.03 27.27 26.49 26.88 28.2 27
139 33.43 33.85 33.64 27.96 27.17 27.57 29.5 27.9
140 33.43 33.85 33.64 27.96 27.17 27.57 29.5 27.4
141 33.54 33.95 33.74 28.06 27.17 27.62 29.5 27.9
142 33.54 33.95 33.74 28.06 27.17 27.62 29.5 27.4
143 34.69 35.12 34.90 28.66 27.76 28.21 30 28.6
144 34.69 35.12 34.90 28.66 27.76 28.21 30 28.7
145 34.80 35.22 35.01 28.66 27.76 28.21 30 28.6
146 34.80 35.22 35.01 28.66 27.76 28.21 30 28.7
147 35.97 36.40 36.19 29.55 28.56 29.05 31 29.6
95
148 35.97 36.40 36.19 29.55 28.56 29.05 31 29.2
149 35.97 36.40 36.19 29.55 28.56 29.05 31 29
150 36.08 36.51 36.30 29.65 28.66 29.15 31 29.6
151 36.08 36.51 36.30 29.65 28.66 29.15 31 29.2
152 36.08 36.51 36.30 29.65 28.66 29.15 31 29
153 37.27 37.71 37.49 30.76 29.65 30.21 32.5 30.8
154 37.38 37.82 37.60 30.86 29.75 30.31 32.5 30.8
155 38.05 39.28 38.66 28.06 28.36 28.21 27.7 25.8
156 38.05 39.28 38.66 28.06 28.36 28.21 27.7 25.6
157 38.05 39.28 38.66 28.06 28.36 28.21 27.7 26.4
158 38.16 39.39 38.77 28.06 28.36 28.21 27.7 25.8
159 38.16 39.39 38.77 28.06 28.36 28.21 27.7 25.6
160 38.16 39.39 38.77 28.06 28.36 28.21 27.7 26.4
161 37.38 40.19 38.78 27.57 28.16 27.86 27.2 26.3
162 37.38 40.19 38.78 27.57 28.16 27.86 27.2 25.7
163 37.38 40.19 38.78 27.57 28.16 27.86 27.2 25.6
164 37.38 40.19 38.78 27.57 28.16 27.86 27.2 26.1
165 37.49 40.19 38.84 27.47 28.06 27.76 27.2 26.3
166 37.49 40.19 38.84 27.47 28.06 27.76 27.2 25.7
167 37.49 40.19 38.84 27.47 28.06 27.76 27.2 25.6
168 37.49 40.19 38.84 27.47 28.06 27.76 27.2 26.1
169 37.49 40.30 38.90 27.47 27.96 27.71 27.2 26.3
170 37.49 40.30 38.90 27.47 27.96 27.71 27.2 25.7
171 37.49 40.30 38.90 27.47 27.96 27.71 27.2 25.6
172 37.49 40.30 38.90 27.47 27.96 27.71 27.2 26.1
173 37.60 40.53 39.07 27.47 27.86 27.67 27.2 26.3
174 37.60 40.53 39.07 27.47 27.86 27.67 27.2 25.7
175 37.60 40.53 39.07 27.47 27.86 27.67 27.2 25.6
176 37.60 40.53 39.07 27.47 27.86 27.67 27.2 26.1
96
Table A.4 Result of algorithm validation - Material 2
I II III IV
1 1.95 1.38 1.43 0.95
2 1.95 1.38 1.35 0.98
3 1.95 1.38 1.60 0.88
4 2.15 1.52 1.55 1.06
5 2.15 1.52 1.47 1.10
6 2.15 1.52 1.73 0.99
7 1.15 0.81 0.95 0.52
8 1.15 0.81 0.92 0.53
9 1.15 0.81 1.02 0.50
10 1.29 0.91 1.04 0.58
11 1.29 0.91 1.01 0.60
12 1.29 0.91 1.12 0.56
13 1.42 1.00 1.15 0.65
14 1.42 1.00 1.11 0.66
15 1.42 1.00 1.24 0.62
16 1.81 1.29 1.11 1.02
17 1.81 1.29 1.09 0.59
18 1.81 1.29 1.21 0.95
19 1.81 1.29 0.97 0.55
20 1.81 1.29 1.01 0.56
21 1.97 1.41 1.20 1.12
22 1.97 1.41 1.19 0.64
23 1.97 1.41 1.31 1.03
24 1.97 1.41 1.06 0.60
25 1.97 1.41 1.10 0.61
26 2.13 1.52 1.28 0.69
27 2.13 1.52 1.42 1.12
28 2.13 1.52 1.14 0.65
29 2.13 1.52 1.18 0.66
30 2.36 1.69 1.42 0.77
31 2.36 1.69 1.27 0.72
32 2.36 1.69 1.32 0.73
33 2.61 1.87 1.62 0.86
34 2.61 1.87 1.44 0.81
35 2.61 1.87 1.50 0.82
36 1.64 1.18 1.04 0.90
37 1.64 1.18 1.03 0.54
38 1.64 1.18 1.13 0.84
39 1.64 1.18 1.00 0.93
40 1.64 1.18 0.92 0.51
41 1.64 1.18 0.95 0.52
42 1.79 1.28 1.14 0.98
43 1.79 1.28 1.12 0.59
44 1.79 1.28 1.23 0.92
45 1.79 1.28 1.09 1.02
46 1.79 1.28 1.00 0.56
47 1.79 1.28 1.04 0.57
97
48 1.94 1.39 1.23 1.06
49 1.94 1.39 1.21 0.64
50 1.94 1.39 1.33 0.99
51 1.94 1.39 1.18 1.10
52 1.94 1.39 1.08 0.60
53 1.94 1.39 1.12 0.61
54 2.15 1.54 1.34 0.71
55 2.15 1.54 1.48 1.10
56 2.15 1.54 1.20 0.67
57 2.15 1.54 1.25 0.68
58 2.38 1.71 1.53 0.80
59 2.38 1.71 1.62 1.23
60 2.38 1.71 1.37 0.75
61 2.38 1.71 1.42 0.77
62 1.45 1.04 0.96 0.76
63 1.45 1.04 0.95 0.49
64 1.45 1.04 1.04 0.72
65 1.45 1.04 0.93 0.79
66 1.45 1.04 0.86 0.46
67 1.45 1.04 0.88 0.47
68 1.58 1.13 1.05 0.83
69 1.58 1.13 1.03 0.53
70 1.58 1.13 1.13 0.79
71 1.58 1.13 1.01 0.86
72 1.58 1.13 0.93 0.50
73 1.58 1.13 0.96 0.51
74 1.71 1.22 1.13 0.90
75 1.71 1.22 1.12 0.58
76 1.71 1.22 1.22 0.85
77 1.71 1.22 1.09 0.93
78 1.71 1.22 1.01 0.54
79 1.71 1.22 1.04 0.55
80 1.90 1.35 1.26 1.00
81 1.90 1.35 1.24 0.64
82 1.90 1.35 1.35 0.94
83 1.90 1.35 1.21 1.03
84 1.90 1.35 1.12 0.61
85 1.90 1.35 1.16 0.62
86 2.11 1.50 1.38 1.12
87 2.11 1.50 1.41 0.72
88 2.11 1.50 1.49 1.06
89 2.11 1.50 1.33 1.15
90 2.11 1.50 1.27 0.68
91 2.11 1.50 1.32 0.69
92 1.87 1.33 1.66 0.73
93 1.87 1.33 1.05 1.14
94 1.87 1.33 1.25 0.97
95 1.87 1.33 0.89 1.41
96 2.06 1.46 1.80 0.80
97 2.06 1.46 1.17 1.25
98
98 2.06 1.46 1.39 1.07
99 2.06 1.46 0.99 1.55
100 2.28 1.62 2.14 0.96
101 2.28 1.62 2.18 0.92
102 2.28 1.62 1.69 0.82
103 2.42 1.72 2.27 1.02
104 2.42 1.72 2.31 0.97
105 2.42 1.72 1.80 0.86
106 2.48 1.76 2.32 1.05
107 2.48 1.76 2.37 1.00
108 2.48 1.76 1.84 0.89
109 2.39 1.70 2.24 1.01
110 2.39 1.70 2.29 0.96
111 2.39 1.70 1.78 0.86
112 5.11 3.65 1.10 1.58
113 5.25 3.75 1.09 1.54
114 5.92 4.23 1.12 1.54
115 6.90 4.92 1.17 1.53
116 2.57 1.82 0.63 0.96
117 2.57 1.82 1.07 0.67
118 2.61 1.85 0.62 0.94
119 2.61 1.85 1.01 0.65
120 1.89 1.33 0.63 1.20
121 1.89 1.33 1.28 0.65
122 1.93 1.36 0.64 1.20
123 1.93 1.36 1.20 0.64
124 0.98 0.70 0.60 0.55
125 0.98 0.70 0.91 0.42
126 0.98 0.70 0.76 0.46
127 1.05 0.76 0.66 0.58
128 1.05 0.76 1.02 0.44
129 1.05 0.76 0.84 0.49
130 1.15 0.83 0.76 0.61
131 1.15 0.83 1.05 0.46
132 1.15 0.83 1.00 0.51
133 0.84 0.61 0.79 0.36
134 0.84 0.61 0.67 0.39
135 0.90 0.65 0.88 0.38
136 0.90 0.65 0.74 0.41
137 0.98 0.70 0.90 0.39
138 0.98 0.70 0.87 0.43
139 0.79 0.56 0.63 0.36
140 0.79 0.56 0.70 0.31
141 0.81 0.58 0.67 0.37
142 0.81 0.58 0.70 0.32
143 0.93 0.66 0.72 0.44
144 0.93 0.66 0.68 0.46
145 0.95 0.67 0.73 0.45
146 0.95 0.67 0.69 0.47
147 0.92 0.64 0.66 0.45
99
148 0.92 0.64 0.83 0.39
149 0.92 0.64 0.88 0.37
150 0.97 0.68 0.72 0.46
151 0.97 0.68 0.93 0.40
152 0.97 0.68 0.87 0.38
153 0.79 0.55 0.58 0.38
154 0.83 0.58 0.63 0.39
155 5.14 3.69 0.68 0.83
156 5.14 3.69 0.64 0.76
157 5.14 3.69 0.87 1.13
158 5.19 3.73 0.69 0.84
159 5.19 3.73 0.64 0.77
160 5.19 3.73 0.88 1.14
161 4.12 2.97 0.97 1.43
162 4.12 2.97 0.75 1.00
163 4.12 2.97 0.72 0.95
164 4.12 2.97 0.88 1.25
165 4.91 3.54 1.07 1.53
166 4.91 3.54 0.81 1.05
167 4.91 3.54 0.78 1.00
168 4.91 3.54 0.97 1.33
169 5.43 3.92 1.13 1.59
170 5.43 3.92 0.85 1.08
171 5.43 3.92 0.82 1.03
172 5.43 3.92 1.02 1.37
173 6.13 4.42 1.21 1.67
174 6.13 4.42 0.90 1.12
175 6.13 4.42 0.86 1.06
176 6.13 4.42 1.09 1.44
100
Appendix B The Complete Dynamo Script
Figure B.1 The complete Dynamo script
100
101
Figure B.2 Thermal photo assignment
Figure B.3 "Grey" colors database creation
101
102
Figure B.4 Index of colors in database were returned in a list
Figure B.5 Math logic from index numbers to corresponding temperatures
102
103
Figure B.6 R-values calculation
103
104
Figure B.7 Thermal data visualization
104
Abstract (if available)
Abstract
In order to be able to meet rigorous energy standards, serious attention needs to be given to correcting problems with recently constructed buildings and retrofitting older buildings. One step is determining where facades might not be performing as digitally modelled. A literature review was undertaken to evaluate and then select several methods that are able to convert infrared thermal colors (surface temperature) taken by a digital FLIR camera to corresponding real thermal resistance values (R-value). These algorithms were tested on a physical model, and the algorithm with more trust worthy result was selected to continue with. A series of 2D thermal and digital photos of the building were taken, and Autodesk ReCap was used to generate a 3D point cloud and mesh model encoding the thermal colors at the point level. The next step was to align the Revit model with the thermal information and have Revit detect those colors and use Dynamo to convert colors to corresponding thermal data. The final output was a protocol which was able to achieve a false color building model carrying information of different thermal properties of the envelope. The protocol could help engineers make better retrofitting decision, help owners save energy consumption, and provide data to energy programs that could ultimately allow occupants to have a higher comfort level. The process encountered many obstacles where compromises were made resulting in the conclusion that this methodology can be applied only to cases under certain conditions. A protocol for identifying those cases and using those methods is presented. Future work would include more rigorous validation of the methodology established.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
A BIM-based visualization tool for facilities management: fault detection through integrating real-time sensor data into BIM
PDF
Building bridges: filling gaps between BIM and other tools in mechanical design
PDF
Streamlining sustainable design in building information modeling: BIM-based PV design and analysis tools
PDF
Developing environmental controls using a data-driven approach for enhancing environmental comfort and energy performance
PDF
Building energy performance estimation approach: facade visual information-driven benchmark performance model
PDF
Bridging performance gaps by occupancy and weather data-driven energy prediction modeling using neural networks
PDF
Developing a data-driven model of overall thermal sensation based on the use of human physiological information in a built environment
PDF
BIM+AR in architecture: a building maintenance application for a smart phone
PDF
Enhancing thermal comfort: data-driven approach to control air temperature based on facial skin temperature
PDF
Data visualization in VR/AR: static data analysis in buildings
PDF
CFD visualization: a case study for using a building information modeling with virtual reality
PDF
Building information modeling: guidelines for project execution plan (PxP) for India
PDF
Exploration for the prediction of thermal comfort & sensation with application of building HVAC automation
PDF
A simplified building energy simulation tool: material and environmental properties effects on HVAC performance
PDF
Energy simulation in existing buildings: calibrating the model for retrofit studies
PDF
Economizer performance and verification: effect of human behavior on economizer efficacy and thermal comfort in southern California
PDF
Energy use intensity estimation method based on building façade features by using regression models
PDF
Impact of occupants in building performance: extracting information from building data
PDF
Pushing the solar envelope: determining solar envelope generating principles for sites with existing buildings
PDF
Environmental adaptive design: building performance analysis considering change
Asset Metadata
Creator
Fan, Qianqian
(author)
Core Title
Visualizing thermal data in a building information model
School
School of Architecture
Degree
Master of Building Science
Degree Program
Building Science
Publication Date
04/22/2016
Defense Date
03/21/2016
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
BIM,infrared camera,OAI-PMH Harvest,R-value,thermal data
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Kensek, Karen M. (
committee chair
), Choi, Joon-Ho (
committee member
), Schiler, Marc (
committee member
)
Creator Email
fqqsunny@hotmail.com,qfan@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c40-240013
Unique identifier
UC11277856
Identifier
etd-FanQianqia-4353.pdf (filename),usctheses-c40-240013 (legacy record id)
Legacy Identifier
etd-FanQianqia-4353.pdf
Dmrecord
240013
Document Type
Thesis
Format
application/pdf (imt)
Rights
Fan, Qianqian
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
BIM
infrared camera
R-value
thermal data