ELECTROCHROMIC GLASS THERMAL COMFORT

Proposals for Improving Existing Workflow

By

Yihong Cheng

Presented to the
FACULTY OF THE
SCHOOL OF ARCHITECTURE
UNIVERSITY OF SOUTHERN CALIFORNIA
In partial fulfillment of the
Requirements of degree
MASTER OF BUILDING SCIENCE

August 2017

Copyright 2017 Yihong Cheng

1




ProQuest Number: 10820231




All rights reserved

INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.






ProQuest 10820231

Published by ProQuest LLC (2018 ). Copyright of the Dissertation is held by the Author.


All rights reserved.
This work is protected against unauthorized copying under Title 17, United States Code
Microform Edition © ProQuest LLC.


ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
COMMITTEE

Douglas Noble, FAIA, Ph.D

Associate Professor
USC School of Architecture
[email protected]
(213) 740-2723

Kyle Konis, Ph.D., AIA

Assistant Professor
USC School of Architecture
[email protected]

Karen M. Kensek, LEED AP BD+C, Assoc. AIA

Assistant Professor
USC School of Architecture
[email protected]
(213)740-2081

2
 Acknowledgement

First and foremost, I would like to express my deepest gratitude to my committee chair, Professor
Douglas Noble, my committee members, Professor Karen Kensek and Professor Kyle Knois. They
have offered excellent advisory on this project and provided great ideas and helped in ensuring the
research direction in the right way. Their positive attitude towards work, dedication in teaching,
creative way of thinking, and paying attention to details, will always benefit my life. I really appreciate
their valuable advice on improving the research and my writing skills.

Thank you Arcadia, Inc, Giroux Glass and Sage Glass kindly donated the necessary materials for the
physical full size glass unit fabrication. It was fun doing such a big model for study.

Last but not least, thank you, the lovely MBS family. Without the support from each MBS member, I
would not complete this thesis project while enjoying such a wonderful time.

3
 Abstract
Windows, as one of the least energy efficient components of a building, are responsible for up to 40 percent of the
total energy cost including heating, cooling and lighting, and over 30 percent of this energy is lost through poor
efficient building components (DOE, 2006). Electrochromic glass (EC) is one of the state-of-art technologies and can
substantially reduce the energy consumptions of buildings (View Glass, 2013). Unlike traditional glass, EC glass
dynamically changes its tinting states based on occupant needs. When the glass tint level changes, the glass properties
(e.g. Tv, SHGC) change accordingly. Previous studies have been done on electrochromic glass energy simulations
(Sullivan et al, 1994, Pease et al, 2010, Lee et al, 2013). However, most of the existing softwares don’t have the
capability of adding the four tinting states to the EC glass or applying a specific tinting schedule, which makes the
simulation result inaccurate and not reliable.

The research objective is to find a better way to conduct thermal comfort simulations for EC glass windows. Occupant
comfort is a critical condition that may eventually be a market driver for new energy-saving technologies (Lee et al.,
2006). EC glass manufactures, designers and building owners can get a better understanding of how the occupants
feel in a space with EC glass installed. Also by comparing the thermal comfort results with traditional glass installed
in the same space, they can see how EC glass can help to improve occupants comfort level.

Workflows have been developed in COMFEN, DesignBuilder, CBE Thermal Comfort Model, and Grasshopper (with
Ladybug and Honeybee plugins). The 71T lab at National Lawrence Berkeley Lab (NLBL) was chosen to serve as the
simulation prototype. The lab is modeled in software as two identical rooms with the same configurations except for
window types: the test case with EC glass and the reference room with low-e glass. The EC control strategy used in
the simulation process is automatic daylighting control. Simulation process and results are evaluated and compared,
and the workflows in CBE Comfort Model and Grasshopper seem to indicating more reliable results than the
workflows in COMFEN, DesignBuilder.

As a supplementary part of the thesis, a full size unit with two pieces of electrochromic glass was assembled in the
shop of Arcadia Inc. (an aluminum window and façade system manufacturer and re-seller). Learning the fabrication
process is an important approach to understanding the glass system because it involves installation details and wire
works which are all different from conventional glass.

Keywords: Electrochromic Glass, thermal comfort, EC glass tinting states, tinting schedule, solar radiation impacts,
software simulation, physical model fabrication

Hypothesis
Simulation workflows in CBE Thermal Comfort Model and Grasshopper (with Ladybug and Honeybee plugins) which
take multiple main factors (e.g. four tinting states, solar load impacts, glass tinting schedule) into consideration seem
to indicating improved reliabilities.

4
 Contents
Acknowledgement .........................................................................................................................................................3
Abstract .........................................................................................................................................................................4
Hypothesis .....................................................................................................................................................................4
List of Figures .................................................................................................................................................................8
List of Tables ..................................................................................................................................................................9
Chapter 1 Introduction ................................................................................................................................................10
1.1 An introduction to Electrochromic Glass ...................................................................................................11
1.1.1 What is electrochromic glass .................................................................................................................11
1.1.2 How does electrochromic work .............................................................................................................11
1.1.3 Key benefits of electrochromic glass .....................................................................................................12
1.1.4 Drawbacks of electrochromic glass .......................................................................................................14
1.2 Basic terminologies about electrochromic glass ........................................................................................14
1.2.1 IGU .........................................................................................................................................................14
1.2.2 Frame Cable ...........................................................................................................................................15
1.2.3 Pigtail .....................................................................................................................................................15
1.2.4 Tinting States .........................................................................................................................................16
1.2.5 Control Strategies ..................................................................................................................................16
1.2.6 Glass Switching Speed ...........................................................................................................................18
1.3 Thermal Comfort ........................................................................................................................................19
1.3.1 Thermal Comfort Metrics ......................................................................................................................19
1.3.2 Thermal Comfort Importance ................................................................................................................20
1.4 Abstract, Hypothesis, and Objectives ........................................................................................................20
1.4.1 Abstract..................................................................................................................................................20
1.4.2 Hypothesis .............................................................................................................................................21
1.4.3 Objectives ..............................................................................................................................................21
1.5 Significance of Study ..................................................................................................................................21
1.6 Chapter Structure .......................................................................................................................................21
1.7 Summary ...................................................................................................................................................22
Chapter 2: Previous Work: Background and Literature Review ..................................................................................23
2.1 General Electrochromic Glass Study History ..............................................................................................23
2.2 Previous Electrochromic Glass Energy Performance Studies .....................................................................23
2.2.1 Energy simulation in BEopt ....................................................................................................................23
2.2.2 Energy simulation in eQuest ..................................................................................................................23
5
 2.2.3 Energy simulation in EnergyPlus ............................................................................................................24
2.2.4 Energy Performance Evaluation through Monitored Data ....................................................................24
2.2.5 Summary ................................................................................................................................................25
2.3 Previous Thermal Comfort Study ...............................................................................................................25
2.3.1 Thermal Comfort Near Glass Facades....................................................................................................25
2.3.2 Post Occupancy Study............................................................................................................................25
2.3.3 Thermal Comfort Simulation Software in Practice ................................................................................26
2.4 Limitations of Thermal Comfort Study on EC Glazing ................................................................................26
2.4.1 Parameters limitation ............................................................................................................................26
2.4.2 Human Subject Test Limitation ..............................................................................................................26
2.4.3 Time Consuming ....................................................................................................................................27
2.4.4 High Cost ................................................................................................................................................27
2.5 Physical Model Fabrication ........................................................................................................................27
2.6 Summary ....................................................................................................................................................29
Chapter 3: Methodology .............................................................................................................................................30
3.1 Workflow Overview ...................................................................................................................................30
3.2 Case Study Room........................................................................................................................................30
3.3 Workflow Development .............................................................................................................................31
3.3.1 COMFEN.................................................................................................................................................31
3.3.2 DesignBuilder .........................................................................................................................................32
3.3.3 Ladybug and Honeybee .........................................................................................................................34
3.3.4 CBE Thermal Comfort Tool ....................................................................................................................37
3.4 Inputs .........................................................................................................................................................41
3.4.1 Weather Data ........................................................................................................................................41
3.4.2 Thermal Comfort Parameters ................................................................................................................41
3.4.3 Occupant Location .................................................................................................................................41
3.4.4 Tinting Schedule ....................................................................................................................................41
3.5 Workflow Evaluation ..................................................................................................................................42
3.5.1 Solar Load Impacts .................................................................................................................................42
3.5.2 EC Glass Tinting States ...........................................................................................................................42
3.5.3 EC glass Tinting Schedule .......................................................................................................................42
3.6 Summary ....................................................................................................................................................42
Chapter 4: Results ........................................................................................................................................................43
4.1 Results from Each Workflow ......................................................................................................................43

6
 4.1.1 COMFEN.................................................................................................................................................43
4.1.2 DesignBuilder .........................................................................................................................................46
4.1.3 Ladybug and Honeybee .........................................................................................................................47
4.1.4 CBE Thermal Comfort Model .................................................................................................................51
4.2 Conclusion ..................................................................................................................................................54
Chapter 5: Discussion ..................................................................................................................................................55
5.1 Result analysis of Each Workflow...............................................................................................................55
5.1.1 COMFEN.................................................................................................................................................55
5.1.2 DesignBuilder .........................................................................................................................................55
5.1.3 Ladybug and Honeybee .........................................................................................................................56
5.1.4 CBE Thermal Comfort Model .................................................................................................................60
5.2 Workflow Evaluation ..................................................................................................................................62
5.2.1 Solar Load Impacts .................................................................................................................................62
5.2.2 EC Glass Tinting States ...............................................................................................................................62
5.2.3 EC Glass Tinting Schedule ......................................................................................................................63
5.2.4 Other Evaluation Factors .......................................................................................................................63
5.3 Study Limitations........................................................................................................................................64
5.4 Summary ....................................................................................................................................................64
Chapter 6: Conclusions and Future Work ....................................................................................................................65
6.1 Result Reliabilities ......................................................................................................................................65
6.2 Workflow Developed .................................................................................................................................65
6.3 Future Work ...............................................................................................................................................66
6.4 Summary ....................................................................................................................................................67
Reference.....................................................................................................................................................................68
Appendix X ...................................................................................................................................................................71
X-1: Illuminance Calculation Code in Grasshopper .................................................................................................71
X-2: Thermal Comfort Simulation Code in Grasshopper .........................................................................................72

7
 List of Figures
Figure 1 Electrochromic Glass (View Inc, 2017) ........................................................................................................ 11
Figure 2 EC glass in Clear, Intermediate and Tinted state (Burdis et al, 2007) ........................................................... 12
Figure 3 Dynamic Insulating Glass Unit (SageGlass Architect Brochure, 2016) ........................................................ 15
Figure 4 Frame Cable .................................................................................................................................................. 15
Figure 5 IGU Pigtail (SageGlass Pigtail Cut Sheet, 2017) .......................................................................................... 16
Figure 6 Multiple Panes and Zoning Design (Pease et al, 2010) ................................................................................. 17
Figure 7 An Example of Control Strategy (Sage Glass Control System, 2016) .......................................................... 18
Figure 8 Glazing characteristics (Pease et al, 2010) .................................................................................................... 23
Figure 9: Glass Unit Materials from Different Company ............................................................................................ 27
Figure 10 EC Glass Unit Fabrication Process.............................................................................................................. 28
Figure 11 Small Scale EC Window Installation (Lee et al, 2006) ............................................................................... 29
Figure 12 EC Glass Exhibition .................................................................................................................................... 29
Figure 13 Workflow Overview .................................................................................................................................... 30
Figure 14 Floor Plan and Section of Test Model ......................................................................................................... 31
Figure 15 COMFEN User Interface............................................................................................................................. 32
Figure 16 DeisgnBuilder User Interface ...................................................................................................................... 33
Figure 17 Output Fanger PMV in DesignBuilder ........................................................................................................ 34
Figure 18 PMV Model in Ladybug and Honeybee (Mackey, 2012) ........................................................................... 35
Figure 19 Honeybee Window Shade Generator .......................................................................................................... 35
Figure 20 Model Four States EC Glass in Grasshopper .............................................................................................. 36
Figure 21 PMV and Adaptive Comfort Recipe ........................................................................................................... 36
Figure 22 Workflow in Grasshopper ........................................................................................................................... 37
Figure 23 CBE Thermal Comfort Model User Interface ............................................................................................. 38
Figure 24 CBE Thermal Comfort Model Software Installation Folder ....................................................................... 39
Figure 25 EC Glass Properties in WINDOW 7.5 ........................................................................................................ 39
Figure 26 Editing the “DelphiGlassType” ................................................................................................................... 40
Figure 27 Solar Load Setting in CBE Thermal Comfort Model .................................................................................. 40
Figure 28 Sun Path Code ............................................................................................................................................. 40
Figure 29 Annual Average Thermal Comfort .............................................................................................................. 44
Figure 30 Spring Average Thermal Comfort ............................................................................................................... 44
Figure 31 Summer Average Thermal Comfort ............................................................................................................ 45
Figure 32 Autumn Average Thermal Comfort ............................................................................................................ 45
Figure 33 Winter Average Thermal Comfort .............................................................................................................. 46
Figure 34 Models in Revit and DesignBuilder ............................................................................................................ 46
Figure 35 Test Room PMV Results on Winter Solstice Day in DesignBuilder .......................................................... 47
Figure 36 Reference Room PMV Results on Winter Solstice Day in DesignBuilder ................................................. 47
Figure 37 Test Room PMV Results on Summer Solstice Day in DesignBuilder ........................................................ 47
Figure 38 Reference Room PMV Results on Summer Solstice Day in DesignBuilder ............................................... 47
Figure 39 Illuminance Calculation Code in Grasshopper ............................................................................................ 48
Figure 40 Thermal Comfort Simulation Code in Grasshopper .................................................................................... 50
Figure 41 Models in CBE Thermal Comfort Model .................................................................................................... 51
Figure 42 Thermal Comfort Results in CBE Thermal Comfort Model ....................................................................... 53
Figure 43 Comfort and Sensation Scale ...................................................................................................................... 54
Figure 44 Sun Path ...................................................................................................................................................... 56
Figure 45 Illuminance Distribution on Winter Solstice Day ....................................................................................... 57
Figure 46 Illuminance Distribution on Summer Solstice Day ..................................................................................... 57
Figure 47 Sensor Locations in Case Study VS in Real-life ......................................................................................... 58
Figure 48 PMV Comparison on Winter Solstice Day in Grasshopper ........................................................................ 59
Figure 49 PMV Comparison on Summer Solstice Day in Grasshopper ...................................................................... 59
Figure 50 Overall Comfort on Winter Solstice Day .................................................................................................... 60
Figure 51 Overall Sensation at Winter Solstice Day ................................................................................................... 61
Figure 52 Overall Comfort on Summer Solstice Day .................................................................................................. 61
8
Figure 53 Overall Sensation on Summer Solstice Day ................................................................................................ 62
Figure 54 Workflow Overview .................................................................................................................................... 65

List of Tables

Table 1 Applicable LEED Credits for EC Glass (SageGlass LEED Brochure, 2017) ................................................ 14
Table 2 Sage Glass Tinting State (Sage Glass Performance & Acoustic Data, 2016) ................................................. 16
Table 3 PMV Sensation Scale (ASHRAE Standard 55, 2004) .................................................................................... 19
Table 4 Minimum Annual Energy Savings for SAGE Dynamic Glazing (Pease et al, 2010) ..................................... 24
Table 5 Questions Related to Thermal Comfort (Lee et al, 2013) ............................................................................... 25
Table 6 EC glass in COMFEN .................................................................................................................................... 32
Table 7 EC Glass Tinting State Threshold (Ardakan, 2015) ....................................................................................... 41
Table 8 Glass Used in COMFEN ................................................................................................................................ 43
Table 9 Annual Energy Results from COMFEN ......................................................................................................... 46
Table 10 Glass Properties Used in DesignBuilder ....................................................................................................... 47
Table 11 Winter Solstice Day Illuminance and EC Glass States ................................................................................. 49
Table 12 Summer Solstice Day Illuminance and EC Glass States .............................................................................. 49
Table 13 Test Room PMV Results in Ladybug and Honeybee ................................................................................... 50
Table 14 Reference Room PMV Results in Ladybug and Honeybee .......................................................................... 51
Table 15 Solar Azimuth and Altitude .......................................................................................................................... 52
Table 16 Thermal Comfort Results at Winter Solstice Day in CBE Thermal Comfort Model ................................... 52
Table 17 Thermal Comfort Results at Summer Solstice Day in CBE Thermal Comfort Model ................................. 53
Table 18 PMV Results in DesignBuilder .................................................................................................................... 55
Table 19 Illuminance Results in Case Study VS in Real-life ...................................................................................... 58
Table 20 Workflow Comparison ................................................................................................................................. 63

9
 Chapter 1 Introduction

Buildings account for approximately 40 percent of all energy consumption in the US (DOE, 2016). Windows, as one
of the least energy efficient components of a building, are responsible for up to 40 percent of the total energy cost,
including heating, cooling and lighting, and over 30 percent of this energy is lost through poor efficient building
components (DOE, 2016). Electrochromic glass (EC) is one of state-of-art technologies and can substantially reduce
the energy consumptions of buildings (View Glass, 2013). EC glass can block near infrared solar radiation passing
through in summer and allow light coming through in winter, substantially reducing energy consumption.

For building envelopes, conventional glazing only offers fixed properties (e.g. Tv, SHGC, %Rb Int., %Tsol, %Tdw-
K) and single control of lighting passing through it. EC glass offers a better solution given a wider range of visible
lighting transmittance (Lampert, 2003). Because of its dynamic character, EC glass needs a specific control algorithm
to control which state the glass should be at a certain time of the day, and therefore the tinting schedule varies. Lots
of research had been on studying about the property, energy performance, visual comfort of EC glass, however thermal
comfort is an area that has not been well studied. Thermal comfort should be paid more attention to because it
influences work productivity and human health, which might be the most convincing market driver in the future.

Despite that EC glass has been in markets and been developed for decades, it is still not widely used in building
envelopes. EC glass has many unique components compared to conventional glass. This chapter demonstrates the
basic concepts and terminologies about EC glass, so as to get a better understanding of how the glass work and
potential problem areas. Knowing about the key benefits and drawbacks of EC glass is also important for the decision-
making process for the application of the product. There are a few types of electrochromic glass on the market,
including SageGlass, View Inc, Smartglass, Polytronix, etc. Both SageGass and View Inc. are manufactures
specialized in developing tintable electrochromic glass for use in building windows. Many descriptions are derived
from their product brochure and more information can be found on their websites: www.sageglass.com and
www.viewglass.com.

10
1.1 An introduction to Electrochromic Glass
Electrochromic glass (EC) definition and basic principles of how the glass work will be introduced. Key benefits and
drawbacks of EC glass is also elaborated in detail which are helpful to people who do not know much about EC glass.

1.1.1 What is electrochromic glass

EC glass can change its light transmission properties when given a stimulus of voltage, then the glass will respond to
the voltage and changes its state, blocking or admitting some wavelengths of light. In this way, electrochromic glass
can control the amount of light passing through the window (Figure 1) (Woodford, 2017).

Figure 1 Electrochromic Glass (View Inc, 2017)

1.1.2 How does electrochromic work

According to the description by Chris Woodford, ordinary windows only have a single pane of glass, while double
glazed windows are made from two glass panes that are separated by an air gap so as to prevent outside heat and sound
(Woodford, 2017). EC glass and low-E glass windows have much more sophisticated coatings with multiple layers.
As for EC glass, the coating has five ultra-thin layers on its inside surface: a separator in the middle, two electrodes
on either side of the separator, and two transparent electrical contact layers on both side of the electrodes. Basically,
the work principle is when voltage is applied across the coating, lithium ions travel between the two electrodes through
the separator. The lithium ions stay in the innermost electrode when most light can go through the glass, which is how
the glass achieve in its clear state. The ions remain in this layer until a voltage is applied and then the ions migrate
from the innermost layer to outermost layer. In this case, the outermost layer will reflect more light making the glass
opaque. Power is not needed to maintain electrochromic glass in its clear or dark state, only needed when change the
ions from one state to the other (Woodford, 2017).

No matter it is clear or tinted state, visible and the near infrared radiation incident on the glass can transmit through
the window into the building interior space. Films in EC glass can be tinted over from fully clear to fully colored and

11
held between these two states. The film absorbs a certain amount of the incident solar radiation depending on the tint
level. The more it gets tinted, more light gets absorbed. The absorption of the energy makes the EC panes heat up, and
the heat is preferentially radiated back out of the space as a result (Figure 2). Note that the visible transmission never
goes down to zero (Tvis=1% when fully tinted) so that the outside view is always maintained (Burdis et al, 2007).

Figure 2 EC glass in Clear, Intermediate and Tinted state (Burdis et al, 2007)

1.1.3 Key benefits of electrochromic glass

According to a new market research report, the total EC glass market is expected to reach $2.59 billion by 2020
(Markets and Markets, 2015). The following benefits were summarized by View Glass (View Inc, 2017) and the
applicable LEED credits were provided by Sage Glass (Sage Glass, 2017). These characteristics are the main reason
why more and more people choose to use EC glass in their buildings.

12
 Peak cooling load reduction: Electrochromic glass can block more than 90 percent of sun light during the
peak cooling demand period. Reduction in peak cooling energy use can result in reduced HVAC equipment
size and simplify the system in a sense compared to traditional glass windows (View Inc, 2017).

 Annual energy savings: Electrochromic glass windows can reduce energy consumption by blocking
excessive solar gain in hot summer and allowing s passive heat gain in cold winter. Intermediate states have
additional lighting energy reduction by allowing for optimal daylighting (View Inc, 2017).

 Greater architecture design freedom: Designers can use more glass while still meeting the performance
requirement of building energy codes and standards. Architecture form can be more flexible and more
transparent given design freedom (View Inc, 2017).

 Increased occupant comfort and productivity: In addition to improved visual and thermal comfort, occupants
can enjoy the unobstructed outside views in all tinting states. More daylight and better regulation of indoor
temperatures have been shown to increase productive by up to 5% (View Inc, 2017)

 Reduction or elimination of internal blinds: Typically, buildings have interior solar control devices such as
mini blinds or roller shades. The management of these devices is often left to the user to mitigate glare and
enhance privacy. Research has showed that in many cases the blinds remain closed for extended time, which
will result in excessive interior lighting energy consumption and passive solar heat gain in winter.
Electrochromic glass can be controlled automatically based on interior or exterior conditions (such as lighting
illuminance, temperature), which is a very good solution for the above problem (View Inc, 2017).

 Fading Reduction: UV radiation brought by daylight can cause fading. When tinted, electrochromic glass can
block almost all UV rays so that the lifetime of the interior equipment, furnishings and fixtures will be
increased (View Inc, 2017).

 Achieving green building certifications: Benefits of electrochromic glass, such as reduced energy
consumption, user controllability, improved thermal and visual comfort can help buildings in achieving
multiple credits for green building certifications like LEED. According to SageGlass, projects going for
LEED certification can earn credits in 5 categories (Table 1) (SageGlass, 2017).

Categories Sub-categories Credits

Energy and Atmosphere Optimize energy performance 18 credit
Sustainable Sites Light pollution reduction 1 credit
Innovation Building innovation 5 credits
Material Resources Building life-cycle impact reduction 5 credits
option 4, whole building LCA
Building product disclosure & 2 credits
optimization environmental product
declarations
Building product disclosure & 1 credit
optimization sourcing of raw materials
Building product disclosure & 2 credits
optimization material ingredients
Indoor Environmental Thermal comfort
Quality 1 credit

Daylighting 3 credits
Quality views 1 credits
Acoustic performance 1 credits

13
 Table 1 Applicable LEED Credits for EC Glass (SageGlass LEED Brochure, 2017)

1.1.4 Drawbacks of electrochromic glass

Despite all the benefits that EC glass can provide, there are some general issues, daylight issue and energy issue that
need to be taken account of when applying EC glass in a building. Understanding those issues is beneficial for the
decision making process for users or clients.

 General Issues: The sophisticated coatings make EC glass more expensive than ordinary glass. Also due to
the complexity composition of electrochromic glass, there are also questions about how durable the glass is
compared with conventional glazing. Besides, not as most people might have expected, instead of taking
several seconds, EC glass may need several minutes to change from one state to another (Woodford, 2015).

 Daylight quality Issue: In order to reduce the human discomfort caused by solar glare at the perimeter zones
of the floor plate, EC glass is set to tint to its lowest level of visual transmittance when a high level of daylight
is incident on the window, especially when sunlight is present. While this strategy maintains the light level
within optimal range, normally between 100 lux to 2000 lux (NOAO, 2015) for the perimeter zones, it can
make the inner zones of the floor darker than the perimeter zone (Ardakan, 2015).

 Energy Issue: Electrochromic glass window might still cause noticeably disturbing glare to users even when
switched to its lowest tinting level. Under this circumstance, other shading devices might be used to avoid
the glare issue. However, either the tinted state of EC glass or the deployment of these shading devices may
substantially increase lighting energy use because in many cases occupants tend to not open the shades after
the period of discomfort has passed (Rubin et al, 1978).

1.2 Basic terminologies about electrochromic glass

Unlike conventional glass, some unique components (e.g. frame cable, pigtail) about EC glass have to be explained.
EC glass, as a switchable glazing, has different tinting states and gets tinted when receiving a command. The control
strategies are illustrated and how fast the glass with switch is also discussed.

1.2.1 IGU
The dynamic insulating glass unit (IGU) are assembled with multiple layers of glass (commonly consist of two layers),
filled with air or a noble gas, such as argon or krypton inside, acting as a thermal barrier to increase energy efficiency
(Figure 3). The noble gases are likely to conduct less heat through the IGU, providing a further insulation. The EC
glass coating and deployment of noble gas also improve the window’s overall u-value, which is a measurement of the
heat transmission through a building part. The IGU is assembled with a prewired connection to the intelligent control
system, and with high level of precision and process control in state-of-the-art dynamic glass manufacturing facilities
(View Inc, 2017). With larger panels of glass available in IGUs, the application of EC glass is broader which can be
used in curtain walls, skylights, atria etc.

14
 Figure 3 Dynamic Insulating Glass Unit (SageGlass Architect Brochure, 2016)

1.2.2 Frame Cable

The IGUs are connected to the control system by frame cable (Figure 4). To be more specifically, frame cable attaches
to the IGU pigtail on one end, routes through the framing system, and connects to the control system. The lengths and
conductor configurations can be ordered in a variety of sizes to meet specific installation requirements (SageGlass
Frame Cable, 2017).

Figure 4 Frame Cable

1.2.3 Pigtail
Each IGU is connected to the frame cable by another cable called pigtail which extends from one edge of the IGU
(Figure 5). The pigtail stores an electronic serial number and provides the electrical connection to every single IGU.
It also includes IGU-specific data that can be used for system start-up, commissioning and troubleshooting. The pigtail
can be manufactured extending from multiple locations as needed for flexibility in installation (SageGlass Pigtail Cut
Sheet, 2017).

15
 Figure 5 IGU Pigtail (SageGlass Pigtail Cut Sheet, 2017)

1.2.4 Tinting States

The visible light transmission of EC glass can be varied from 1% (fully tinted state) to 60% (clear state). A standard
Sage glazing can also be programmed to have two intermediate states which has 18% and 6% visible light transmission
respectively. When fully clear or tinted states are not necessary, these two intermediate states often provide optimal
shading effects, and this range of variability of tinting states provide many new design considerations for the architects
and building owners (Sage Glass Control System, 2016). Based on project daylighting design needs, the EC glass
system can be fully customizable and programmed to designated specifications (Sage Glass Product Guide, 2017).
The performance specifications of a standard EC glazing of Sage Glass include Tvis, SHGC, U-factor, %Rb
Int., %Tsol, % Tuv, %Tdw-K (Table 2). “Visible transmittance (Tvis) represents the amount of visible light that passes
through the glazing material. Solar Heat Gain Coefficient (SHGC) is defined as the fraction of incident solar radiation
that enter the space through entire window assembly as heat gain. U-factor is the standard way to qualify overall heat
flow” (Window Technologies).

Table 2 Sage Glass Tinting State (Sage Glass Performance & Acoustic Data, 2016)

1.2.5 Control Strategies

When it receives a command from the control system, EC glass changes its tinting states. Electronics are housed in
control panels and are pre-programmed with the algorithms that carry out the commands which can best meet the
project requirements. The algorithms are often determined by the building’s designer, and the control systems have
advanced tools to manage the following factors, such as daylighting, color rendering, energy usage, glare incident on
occupants and interior light levels. For optimum performance, there are several control strategies can be used to
determine what tinting state the glass should be, including daylight control, schedule control and glare control, etc.
(SageGlass Control System, 2016).

 Daylight Control
This control strategy is for maximum natural daylight which control the dimmable electric lighting and natural light
coming through the windows synergistically. A daylight sensor can control the tint level depending on sky conditions,
sun positions. When the sky is overcast, the glass is more likely to be set to clear state and admit more daylight to
16
obtain better illumination for the interior space. When the sky is clear, the daylight sensor can tint the glass to block
certain daylight into the space to achieve the same illuminance target, which at the same time can reduce direct solar
gain space (Pease et al, 2010).

 Schedule Control
EC window can be controlled according to time and desired solar energy passing into the space. The basic principle
is limit solar gain in summer which may reduce the air conditioning systems load and allow needed solar gain during
the winter so as to reduce the heating energy consumption. The seasonal variation along with daylighting controls can
help to achieve a better energy performance with EC glazing space (Pease et al, 2010).

 Glare Control
“Glare is a visual sensation caused by excessive and uncontrolled brightness” (Lighting Research Center, 2017). Glare
causes visual discomfort, and the direct solar irradiation improves contrast at the work plane which will affect occupant
work efficiency. Luckily, EC glazing can directly be controlled to reduce glare. Under direct sun or when exposed to
severe reflected light, the glass will tint to fully tinted state. Typically, this control strategy allows users to tint the
offending window zone while allowing daylighting coming into the space through other panes (Pease et al, 2010).

Studies at Lawrence Berkeley National Laboratory (LBNL) have shown that the daylighting control algorithm can
yield best overall annual energy by substantially reducing electric lighting requirements. Studies also indicate that
when daylighting is not in priority, control based on space cooling load yields the best performance by reducing solar
heat gain. For control strategy based on total solar radiation, the performance varies according to the window size .
As small to moderate window size have smaller solar gain, a larger illuminance set-point (the threshold when EC glass
switch from one state to another state) range is best since it allows increased illuminance without much cooling load
needed. While for larger window sizes which itself can provide adequate daylight, a smaller illuminance set-point
range is best because it helps to reduce unwanted solar gain and requires less cooling loads (Sullivan et al., 1994)

Zoning allows specific panes to be tinted as required for glare and heat control, while allows others to remain clear to
provide more neutral light feeling in the space (Pease et al, 2010). Other than the above control strategy, zoning
design can be applied depending on building orientation, occupant location and needs, and interior design (Figure 6).

Figure 6 Multiple Panes and Zoning Design (Pease et al, 2010)

17
Daylighting control strategies as shown in the example that combine automatic and manual controls can help to best
manage sunlight (Figure 7). The system is highly customized and can be programmed to meet the different needs of
any project. When the sun rises and shines directly on the east elevation at a low incidence angle, the east façade is
set to tint to reduce glare, and the rest of glass is set to maximize daylighting. With the sun pass overhead in the early
afternoon, glass on the east and south elevations tints to intermediate states to achieve specific light levels. In the late
afternoon, direct glare becomes a problem in certain areas, so the west window need to be fully tinted to block direct
sun light coming through, while the glass elsewhere in the building need to be clear for maximum daylighting. As the
sun keep going down to sunset time, the top windows transition to clear mode to harvest as much daylight as possible.
During nighttime, windows can be programed to either fully clear state if outside view is desired or fully tinted state
to reduce light pollution at night (SageGlass Control System, 2016).

Sunrise Early Afternoon Late Afternoon

Sunset Nighttime

Figure 7 An Example of Control Strategy (Sage Glass Control System, 2016)

1.2.6 Glass Switching Speed

According to a design guide for EC windows early market done by Lee et al in LBNL, EC glass switching speed
depends on the following factors, such as ambient temperature, glass size and distance between the bus bars. Normally,
a higher outside temperature results in a faster glass transitioning speed. Smaller panes transition faster than larger
panes, and larger panes with a center bus bar will transit faster than those panes without a center bus bar. However,
instead of transitioning from fully clear to fully tinted state, EC glass often changes in smaller increments in daily use,
for example from one state to the next lighter or darker state, so in this case the transition takes less time (Lee et al,
2006).

18
1.3 Thermal Comfort
ASHRAE Standard 55 defines thermal comfort as “the state of mind which expresses satisfaction with the thermal
environment and is assessed by subjective evaluation.” (ASHRAE Standard 55, 2004). Maintaining occupants thermal
comfort meeting the standard is one of the most important goals of heating, ventilation and air conditioning (HVAC)
designers. Energy efficient buildings are only efficient when the occupants are actually feeling comfortable. If not,
they will take alternative means such as turning on the space heaters or air conditioners, which could be substantially
worse than normal HVAC systems (Autodesk sustainability workshop, 2017).

1.3.1 Thermal Comfort Metrics

 PMV and PPD
Predicted mean vote (PMV) refers to a thermal comfort scale that ranging from Cold (-3) to Hot (+3) and was originally
developed by Fanger’s thermal comfort model which was developed based on testing a large number of people in a
climate chamber at different locations (Fanger, 1970). The Fanger PMV model is a mathematic model representing
the relationship between the determining factors. There are six primary factors that directly affect thermal comfort
and can be grouped into two categories: environmental factors including air temperature, mean radiant temperature,
air speed and relative humidity, and physiological factors that include metabolic rate, clothing insulation. The results
relate thermal comfort factors to heat balance principles and produce a scale from -3 to 3 (Table 3). According to the
current comfort scale standard in the US, ASHRAE standard 55: -0.5<PMV<+0.5 is considered as acceptable thermal
environments and zero is the ideal value (ASHRAE Standard 55, 2004).

Value Sensation

-3 Cold

-2 Cool

-1 Slightly Cool

0 Neutral

1 Slightly Warm

2 Warm

3 Hot

Table 3 PMV Sensation Scale (ASHRAE Standard 55, 2004)

Predicted percentage of dissatisfied people (PPD) is another expression of PMV and is used to predict the percentage
of occupants that dissatisfied with the thermal conditions. If PMV moves further from 0 to a negative or positive value,
PPD increases accordingly. The recommended PPD range acceptable is less than 10% people dissatisfied for the
thermal environment (Autodesk Sustainable Workshop, 2016). As it is impossible to please all people, the minimum
number cannot be 0%. According to the PPD index, even when PMV is zero, i.e., when the average individuals in the
space are perfectly comfortable, 5% of the occupant will still be dissatisfied with the thermal environment (Webb,
2012).

However, the chamber for creating the PMV model did not have any windows, so there was no sun or shortwave
radiation in the space (Fanger, 1970). The Fanger PMV model does not account for sunlight hitting the human body
and warming it up, which is always the source of thermal discomfort near windows. As the PMV model underestimates
thermal discomfort when an occupant is in direct sunlight, it is not a good metric for evaluating thermal comfort
studies near windows. So when it comes to the thermal comfort study for rooms with EC glass installed, it is critical
to see if solar radiation is being considered in the evaluation process.
19
  Sensation and Comfort Model
Human subject tests have shown that skin and core temperature and thermal sensation and comfort have strong
relationships, which have been put into equations (Zhang, 2003) and implemented into the Berkeley Comfort Model
software. “This model is capable of accessing local comfort for different body parts which is influenced by radiant
heat exchange through the window, solar gain, air motion, and non-uniform air temperatures, and integrates local
comfort levels to yield whole body thermal comfort” (Zhang et al. 2004). The output from the algorithms are thermal
sensation (warm or cold) and thermal comfort (comfortable or uncomfortable). Overall comfort and overall sensation
both use a scale of -4.0 (very cold or very uncomfortable) to +4.0 (very hot or very comfortable).

1.3.2 Thermal Comfort Importance

Thermal satisfaction with the environment is very important because it influences productivity in work and human
health. For example, when people are satisfied with their thermal environment workers are more productive, students
learn better and patients heal faster (Sage Glass Occupant Comfort, 2017). Thermal comfort is a critical part to most
building owners and might be most convincing market driver when it comes to adopt new energy saving technologies
like EC glass glazing (Sullivan, et al., 1994). Human thermal comfort is a multidisciplinary subject which requires an
exploration of both physical and psychosocial factors and is strongly related to occupant satisfaction and building
energy use and is one of most fundamental aspects of indoor environmental quality (Webb, 2012).

As occupant experiences air temperature, humidity, radiant temperature, air velocity, and shortwave radiation (heat
from the sun), while metabolic rates, and clothing levels a bit differently based on their physiology and state, thermal
comfort is highly subjective which makes it difficult to measure (Autodesk sustainability workshop, 2016). Human
subject tests on thermal is time consuming and typically it is very hard to do a yearly test. Monitored data is
overwhelming and complicated, which makes it hard to draw a convincing conclusion based on the test. Software
simulation is an alternative method to field testing and can be used for thermal comfort analysis. Using simulation
tools early in the design process can help to understand the impacts of EC glazing and the impacts can be understood
and adopted or rejected in the design process of a high-performance building (Haglund, 2010).

1.4 Abstract, Hypothesis, and Objectives

1.4.1 Abstract
Windows, as one of the least energy efficient components of a building, are responsible for up to 40 percent of the
total energy cost including heating, cooling and lighting, and over 30 percent of this energy is lost through poor
efficient building components (DOE, 2006). Electrochromic glass (EC) is one of the state-of-art technologies and can
substantially reduce the energy consumptions of buildings (View Glass, 2013). Unlike traditional glass, EC glass
dynamically changes its tinting states based on occupant needs. When the glass tint level changes, the glass properties
(e.g. Tv, SHGC) change accordingly. Previous studies have been done on electrochromic glass energy simulations
(Sullivan et al, 1994, Pease et al, 2010, Lee et al, 2013). However, most of the existing softwares don’t have the
capability of adding the four tinting states to the EC glass or applying a specific tinting schedule, which makes the
simulation result inaccurate and not reliable.

The research objective is to find a better way to conduct thermal comfort simulations for EC glass windows. Occupant
comfort is a critical condition that may eventually be a market driver for new energy-saving technologies (Lee et al.,
2006). EC glass manufactures, designers and building owners can get a better understanding of how the occupants
feel in a space with EC glass installed. Also by comparing the thermal comfort results with traditional glass installed
in the same space, they can see how EC glass can help to improve occupants comfort level.

Workflows have been developed in COMFEN, DesignBuilder, CBE Thermal Comfort Model, and Grasshopper (with
Ladybug and Honeybee plugins). The 71T lab at National Lawrence Berkeley Lab (NLBL) was chosen to serve as the
simulation prototype. The lab is modeled in software as two identical rooms with the same configurations except for
window types: the test case with EC glass and the reference room with low-e glass. The EC control strategy used in
the simulation process is automatic daylighting control. Simulation process and results are evaluated and compared,
20
and the workflows in CBE Comfort Model and Grasshopper seem to indicating more reliable results than the
workflows in COMFEN, DesignBuilder.

As a supplementary part of the thesis, a full size unit with two pieces of electrochromic glass was assembled in the
shop of Arcadia Inc. (an aluminum window and façade system manufacturer and re-seller). Learning the fabrication
process is an important approach to understanding the glass system because it involves installation details and wire
works which are all different from conventional glass.

1.4.2 Hypothesis
Simulation workflows in CBE Thermal Comfort Model and Grasshopper (with Ladybug and Honeybee plugins) which
take multiple main factors (e.g. four tinting states, solar load impacts, glass tinting schedule) into consideration seems
to indicating improved reliabilities.

1.4.3 Objectives
EC glass manufactures, designers, and building owners need to get a better understanding of how the occupant feel in
a space with EC glass installed, given that using electrochromic glass will highly increase building envelope cost .
Also by comparing with the thermal comfort results with traditional glass installed in the same space, they can see
how EC glass help to improve the interior comfort level. The objective can be divided into the following parts:
.
 Develop workflows in different software for electrochromic glass thermal comfort simulation.
 Evaluate the workflows and find limitations and advantages in each software.
 Fabricate a physical electrochromic glass unit in factory and the model can be used for future study.

The research results will help electrochromic developers improve product designs and offer other stakeholders, such
as designers, utilities, building owners, and consumers, an informed understanding of the thermal comfort benefits of
EC glazing.

1.5 Significance of Study

Some stakeholders perceive the cost and operational barriers making EC products not worthy of short-term investment,
but EC façade is an active element that can influence the building as well as human performance (Sullivan et al. 1994).
EC window’s energy performance and occupant comfort, which are crucial to the market success, should be given
more consideration (Lee, 2006). However, EC window systems have various glass properties (e.g. Tv, SHGC) in each
state, while the glass libraries of conventional analytical tools are not sufficient enough for energy or thermal comfort
simulations considering all four tinting states of the glass. As the material is still in development, the future market
viability of the EC windows depends in part on the ability to predict performance while the product attributes can be
readily changed. Although the primary focus of the research is on thermal comfort study for EC glass, the proposed
workflows might be beneficial to other related performance studies (e.g. energy performance, visual performance)
and helpful for the market success of EC glazing.

1.6 Chapter Structure

Chapter 1 Introduction. Basic concepts and terminologies about electrochromic glass are introduced, which is aimed
for readers who do not know much about EC glazing. Thermal comfort importance is emphasized ,and the research
hypothesis, objective and significance are also demonstrated.

Chapter 2 Literature Review. Previous general study and research related on electrochromic glass energy performance
simulations are summarized. Thermal comfort metrics are discussed and solar load impacts on thermal comfort is
investigated. Limitations on thermal comfort study are identified and an EC glass physical model is fabricated, and
the process is recorded.

21
Chapter 3 Methodology. Case study rooms are elaborated. Workflows are developed in COMFEN, DesignBuilder,
Grasshopper (with Ladybug and Honeybee plugins), and the evaluation method is discussed. Input data for the
workflows is also explained.

Chapter 4 Results. Specific input values are given in each workflow. Case studies results and findings for the test
room and reference room through all the workflows are recorded.

Chapter 5 Discussion. Results from each workflow are discussed and evaluated. The workflows using different
software are compared.

Chapter 6 Conclusions and Future Work. The whole research is summarized, and conclusions from results analysis
are stated. What should be done to improve the proposed workflow and other possible workflows are provided as
future work.

1.7 Summary
This chapter introduced the basic knowledge and terminologies of EC glass. Thermal comfort concept, study
importance and metrics including PMV, PPD, overall sensation and overall comfort were discussed. When it comes
to EC glass and thermal comfort, study importance along with objective were explained. Previous studies related to
EC glass and thermal comfort will be discussed in the following chapter.

22
 Chapter 2: Previous Work: Background and Literature Review
Previous related electrochromic glass study is summarized in this chapter. Researches (Lee et al., 2000, 2005, and
2013) had been done studying the energy performance and visual comfort in buildings with electrochromic glass (EC)
installed, however, thermal comfort is an area that has not been well studied. On the other hand, thermal comfort
simulation methods for buildings with conventional static glass are summarized, which is a very important for
identifying the limitations for EC glass study. Also, as a background study, a physical model was assembled in the
shop of Arcadia Inc. to learn from the fabrication process.

2.1 General Electrochromic Glass Study History

In 1969, the first switchable electrochromic device was exhibited in the laboratory (Deb, 1969). Accord to a summary
by Lee et al, EC glass size, switching range and cycling capabilities gradually evolved in the ensuing decades and in
the late1990s, some developers in the US were selected by the Department of Energy (DOE) and were then invited to
collaborate with DOE National Laboratories to enhance EC glazing market (Lee et al, 2012). National Renewable
Energy Laboratory conducted a series of tests to evaluate the durability of EC glass. Full scale EC prototypes were
assembled at the Lawrence Berkeley National Laboratory, combined with automatically control and dimmable lighting
control system (Lee et al, 2006). Various studies like the technical potential of reducing energy use, peak load demand
and improved user comfort were assessed (Lee et al, 2004, 2005, 2012). However, about measuring data under full
scale real occupied conditions, a few human subject tests had been done over a short period, thus longer term post-
occupancy studies need to be conducted to better evaluate user acceptance and satisfaction of EC windows (Lee et al,
2012).

2.2 Previous Electrochromic Glass Energy Performance Studies

Even though the main objective of this thesis to study the thermal comfort of EC glass, previous energy performance
studies related to EC glass are summarized because the simulation process and inputs data of thermal comfort and
energy performance are similar.

2.2.1 Energy simulation in BEopt

A potential energy savings of EC windows in residential buildings were undertook by Sullivan et al. A single-family
home in Atlanta was modeled in BEopt software and EC windows were assumed to operate automatically based on
incident solar during the cooling season. It was predicted that EC windows would reduce 9.1 % of the whole house
energy consumption, and save 13.5% in whole house electricity demand and demand savings of 10.3% in 50%-level
Building America Home (Sullivan et al, 1994).

2.2.2 Energy simulation in eQuest

Various parametric modeling simulations were completed for a standard eight-story office building using eQuest v3.63
based on ASHRAE 90.1-2007 in research (Pease et al, 2010). To compare the energy performance of Sage EC
windows and other conventional static glazing, each window type was modeled. The simulation was conducted for
three climates: Minneapolis, Phoenix and Washington. According to ASHRAE 90.1-2007 code, a specified static glass
per climate zone is selected (Figure 8).

Figure 8 Glazing characteristics (Pease et al, 2010)

23
The minimum annual energy savings of double and triple pane EC glazings for all the three climate zones were
compared with static single pane glass, baseline ASRAE 90.1-2007 glazing and commercial triple panes glass
respectively (Table 4) (Pease et al, 2010). This study approved that EC glass can save a certain amount of energy,
however only two tinting states of EC glass were considered. And the reference case parameter is not a fixed variable
because the static glass property is different in each climate zone, which makes the results complicated and
incomparable.

Static Single Pane (no ASHRAE 90.1-2007 Commercial Triple

daylighting controls)
SageGlass Double 45% 20% NA

SageGlass Triple 53% 34% 14%

Table 4 Minimum Annual Energy Savings for SAGE Dynamic Glazing (Pease et al, 2010)

2.2.3 Energy simulation in EnergyPlus

A further HVAC and lighting systems energy use were determined using EnergyPlus simulation in a study by Lee et
al (Lee et al, 2013). A west facing conference room in Washington DC with automatic EC windows was modeled.
The EC windows were switched using the same control algorithm as the real product except for that the window was
triggered by exterior vertical irradiance, while the physical EC window is triggered by exterior vertical illuminance
level which is not available in EnergyPlus. Equivalent switching thresholds were calibrated using measured data under
clear sky conditions. Total annual energy use was estimated to be reduced by 39-48 %, and peak electric demand
reduced by 22-35% compared to the existing condition through solar heat gain reduction in summer and harvest in
winter, and thermal conduction reduction through the window and frames, and lighting energy use reduction. Predicted
percentage of people dissatisfied with the thermal environment was also calculated in EnergyPlus simulation. The
number reduced from 18% with the exiting single-pane window to 13% with clear and tinted dual-pane, argon-filled
EC windows (Lee et al, 2013). In this study, the EC windows were automatically controlled with fully clear or fully
tinted mode, again not all four tinting states were considered.

2.2.4 Energy Performance Evaluation through Monitored Data

Monitored results from full-scale large area electrochromic windows were given in a research carried in LBNL (Lee
et al, 2000). Two office rooms, located in Oakland, California, were simultaneously exposed to almost the same
interior and exterior environments (except for the window types) so as to compare the two rooms. The window to wall
ratio was 0.40 and the electrochromic windows had a visible transmittance range of Tv=0.11-0.38. Dimmable electric
lighting system was integrated to provide constant work plane illuminance. The EC window control strategy used in
the simulation was based on solar control. Daily energy usage from the automated EC window system decreased by
6% to 24% compared with static, low transmission (Tv=0.11) window. However, daily lighting energy use increase
13% compare with the reference case with static windows that had Tv=0.38 (Lee et al, 2000). Unfortunately, no
thermal comfort measurements were made in this study.

Another 20-month field study by Lee et al. was carried out to monitor the energy performance of south-facing large-
area absorptive EC windows in a private office setting (Lee et al, 2005). The EC windows were controlled by a variety
means, and some cases applied venetian blinds and some cases involved dividing the EC windows wall into zones.
When visual comfort requirements were in priority, compared to the reference case with fully lowered Venetian blinds,
a 2-zone EC window configuration provided average daily lighting energy savings of 10±15% and cooling load
reductions of 0±3%. Lighting energy savings would be 44±11% if the reference case has no daylighting controls.
When a demand-response mode was given, peak demand due to window cooling loads reductions were 19-26%
maximum on clear sunny days and the peak lighting energy use reductions were 72-100% compared with a reference
case with lighting controls (Lee et al, 2005). From this study, it is concluded that how glare and sun is controlled have
a great significance on lighting energy use and further study need to be done on occupant comfort.

24
 2.2.5 Summary
There are many results to characterize the energy impacts of EC windows. Overall, the simulations and monitored
data indicate a modest level of energy savings achieved by electrochromic glazing. However, those results cannot be
generalized for the current EC products, because the software used in the simulations have limitations on the glass
properties which deflect from the most recent EC glass product. Monitored data is not reliable in the same way and
might have random errors, for example occasional mistake in collecting the data. Performance simulations for EC
glazing is not rare, while thermal comfort studies with human subjects seem to be very rare in the field.

2.3 Previous Thermal Comfort Study

Occupant thermal comfort is different when exposed to sunlight near windows. Post occupancy study is an alternative
way to get occupants comfort satisfaction reactions. Studying about what softwares people use in the practice is helpful
in finding existing tool limitations and selecting more effective software in future.

2.3.1 Thermal Comfort Near Glass Facades

Glass offers great design opportunities and the transparent area may provide abundant daylight and views. However,
the glazed areas often result in excessive energy consumption, and occupants being exposed to sun light are likely to
have thermal discomfort. When a person is sitting or standing in the perimeter zone of the glass area, the effects of the
facade on air temperature, long wave radiation, solar load, and air movement have to be taken into consideration
(Hoffmann, 2012). The glazed surface influence the occupants thermal comfort in two ways: the transparent property
let solar radiation into the space and the glazing’s inner surface temperature may be different from other interior
surface temperature, causing long-wave radiant heat exchange, and convective heat flows. Overall the glazed area
affects the body heat balance through convection, long-wave radiation and short-wave radiation (Hoffmann, 2012).

2.3.2 Post Occupancy Study

A monitored evaluation was done in the Times Company building located in Manhattan, New York by Lawrence
Berkeley National Laboratory (LBNL) together with the Center for the Built Environment (CBE) at the University
of California (Lee et al 2013). The building is a new, 52-story,1.5 million gross square feet commercial office
building. An online survey was issued to all Times employees and 665 responses were received, accounting for 35%
of the total number of occupant among which several questions with corresponding responses were related to the
thermal and quality of the thermal environment (Table 5) (Lee et al, 2013). Responses were on 7-point scale. 1 is
“very dissatisfied”, 4 is “neutral” and 7 is “very satisfied”.

Questions Greater than neutral satisfaction Average Rating

Temperature in your workspace 46% 4.06
Humidity Level in your workspace 68% 5.26
Does thermal comfort in your workspace enhance 39% 4.14
or interfere with your ability to get your job done?
Table 5 Questions Related to Thermal Comfort (Lee et al, 2013)

Statistical analysis concluded that how satisfied the occupants to the building overall were strongly correlated with
humidity satisfaction, not related to thermal comfort. While responses to “overall, does the new office building
enhance or interfere with your ability to get your job done?” were strongly related to thermal comfort, not humidity
(Lee et al, 2013).

In terms of temperature, 206 individuals thought the thermal environment being too cold and 31 individuals felt the
temperature too warm. Based on the occupant concerns, the building operations had been making adjustment and fine-
tuning the temperature set points and energy savings were achieved at the same time (Lee et al, 2013). Even though
EC glazing is not the window system used in this building, the post occupancy study done in this building suggests
one possible way of evaluating occupants thermal comfort, and the results can be very helpful for building system
adjustment.

25
 2.3.3 Thermal Comfort Simulation Software in Practice
There are some commonly used thermal comfort simulation tools used in practice. EnergyPlus is maintained by the
U.S. Department of Energy and is used by architects, engineers and researchers to model energy use and thermal
comfort output as part of its people object (Webb 2006). EnergyPlus can provide analysis output based on Fanger
Model or the ASHRAE Standard 55 adaptive model. The software is free for download and more information on
thermal comfort simulation can be found in the program input-output reference guide.

ROOM is a thermal simulation model developed at Arup over the last 30 years which provides energy analysis,
radiation and shading analysis, and thermal comfort analysis and the thermal comfort results can be displayed in a 2-
D spatial mapping at a set vertical distance from the floor (White et al, 2009). However, the results are for an average
day for each month of the year, not being able to produce results for every hour of the year (White et al, 2009).

A comprehensive web application was developed by the Center for the Built Environment (CBE) for thermal comfort
calculation, visualization, design and compliance according to the latest ASHRAE Standard 55 and more importantly,
the Comfort Model has features beyond the current standard requirement (Arens et al, 2015). The tool does not provide
building energy use or thermal conditions and is only able to provide information about thermal conditions at a single
point in space and a single point in time, which is freely available at http://comfort.cbe.berkeley.edu/ (Schiavon, et al,
2014). A more advanced version of CBE Thermal Comfort Model has detailed human body with the surrounding
thermal environment. This tool can predict thermal comfort for human body as a whole as well as for specific body
parts. Like the web application, this tool cannot predict energy use and only provides results based on specific point
in time (Webb, 2012).

Despite that the Fanger PMV model and the relevant indoor environment standards ASHRAE Standard “Thermal
environmental conditions for human occupancy” did not mention solar radiation impacts on thermal comfort, the
influence of solar radiation directly incident on human body cannot be ignored (Arens, et al, 2015). The subject of
how the shortwave gain affects thermal comfort had been almost absent until recently and the method of calculating
solar gain to the human body can be found on the paper by Arens et al.

2.4 Limitations of Thermal Comfort Study on EC Glazing

Important parameters neglected using current software or simulation metrics, limitations on human subject test, timing
consuming and high cost issues are the main problems for thermal comfort study of EC glazing.

2.4.1 Parameters limitation

As for thermal comfort software simulation for EC glazing, no papers can be found. However, through studying related
software and the energy performance of EC glass, it is not hard to find that several important parameters are missing
when doing any kind of simulation for EC glass. For example, glass libraries are not sufficient to provide all the tint
states, and glass tinting schedule, as one of the most important variables that influence the results, cannot be selected
or manually input in the software.

On the other hand, PMV and PPD, as the most frequently used thermal comfort metrics, do not consider the solar load
impacts on occupants. Besides, the traditional comfort models do not take into account the occupant’s position and do
not consider the physical presence of the occupant. Therefore, some of the main influences of how occupants perceive
the environment are neglected within existing assessment (Hoffmann et at, 2012).

2.4.2 Human Subject Test Limitation

In a research done by Clear, forty-three subject responses and data were collected in a private office with switchable
electrochromic windows, manually operated Venetian blinds, and dimmable fluorescent lights in the testbed facility
at the Lawrence Berkeley National Laboratory in Berkeley, California (Clear, 2006). The collected data supported
that a dimmable window could reduce blind use. However, the direct correlations suggest that electrochromic
operation increased fluorescent light and energy use were less satisfactory. Part of the unexpected result was because
of the control algorithms used in the experiment and turning lights fully off from a significant fraction of the sample

26
is an unexpected behavior which resulted in the surprising results (Clear, 2006). Separate study need to be done to
confirm the anomalous results. Unexpected factors may happen during the human subject tests.

2.4.3 Time Consuming

Some tools might be time consuming because of their computational methods are too complicated or because they are
an entirely separate tool which cannot fit in a company’s existing analysis workflow. On the other hand, thermal
comfort feedbacks from post occupancy evaluation also involves too much time and human efforts. However, as time
is limited and extremely valuable, it won’t work if the analysis takes too much time.

2.4.4 High Cost

In most of the cases, the tools require advanced input and user knowledge, which would require expert users to ensure
quality results. Especially for design firms, it is more difficult for them to find prospective employees with such
knowledge. On the other hand, many of the tools are not publicly available. Training the employees with related skills
or buying a particular software are a great cost and many firms may not want to bear. Those problems and the low
occurrence of thermal comfort analysis in practice indicate the need for a more cohesive process which happen at
multiple stages during the design process. However, there are no existing professional guides or workflow suggesting
such a process (Webb, 2012).

2.5 Physical Model Fabrication

The fabrication of a full-size model is a separate part from all the above software simulation. It is a great opportunity
to learn from the physical model fabrication process. Each part of the material is from different outside company
(Figure 9). Getting materials from different manufactures is also a good chance to get involved with the industry.
Giroux Glass kindly donated four pieces of Sage Glass (34 7/16” x 35 1/8” for each) and delivered the glass to the
shop of Arcadia Inc, an aluminum window and façade system manufacturer and re-seller. Arcadia Inc. provided the
aluminum frame, gaskets for the unit. Frame cables, grommets, and control systems were donated by Sage Glass. Note
that EC coating degrade quickly if water penetrate into the air gap, consequently the glass edge need to be sealed
properly. Due to the sealed nature, the glass cannot be cut at the building site (Munshi, 2012), the glass size and
shapes should be pre-fabricated before shipment and installation.

Figure 9: Glass Unit Materials from Different Company

Unlike traditional glass fabrication, EC glass need more attention during the whole process. Under the guidance of
SageGlass, following are some basic instructions for the fabrication process (Figure 10):

 Have the aluminum frame assembled

 Measure the frame wire diameter so as to choose the right drill size.
27
  Drill a hole in the mullion for the frame wire and place a grommet into the hole which can prevent the frame
wire from damaged.
 Make sure an appropriate length of wire is left in the glazing pocket to connect to the pigtails.
 Clean the glass edges and surface before place the glass in the frame.
 Shift the glass to one side and connect the IGU pigtail to frame wire.
 Insert gasket on the four sizes of the glass to prevent infiltration.
 Attach the wires at the frame so that they are protected during shipment.
 Seal the hole so that no air or water will infiltrate in the framing system.

Figure 10 EC Glass Unit Fabrication Process

Wiring network is critical at the building installation site. The wiring is connected from the pigtail of EC IGU to the
controller through the hollow framework and is fed into the wall box (Figure 11) (Lee et al, 2006). On the other side,
the controller is connected to the power supply and the control system need to pre-programmed based on project needs,
such as triggered by daylighting level, incident radiance, air temperature or space load. The full size EC glass unit
fabricated was transport to the third floor of School of Architecture, USC for exhibition. Each piece of the EC glass
was connected to the basic manual four states controller. The one on the left was switched to fully tinted state and the
one to the right was in clear state (Figure 12).
28
 Figure 11 Small Scale EC Window Installation (Lee et al, 2006)

Figure 12 EC Glass Exhibition

2.6 Summary
Most of the previous research about electrochromic glass examined the glass using only two control states (fully
bleached and fully tinted states). As technologies has been developed, the intermediate-state control only become
commercially available in recent years. More research need to be done to keep up to date with the newest products
and to get a better understanding of the EC glass which have four tinting states. As for evaluating thermal comfort
near windows, the impact of solar radiation is nonnegligible. The physical model fabrication process is considered as
background research to learn more about the product, but no further research will be done with it.

29
 Chapter 3: Methodology
This chapter illustrates the overall workflow and methodologies for each workflow. The case study rooms input
perimeters are also explained. In order to work out effective methods to evaluate electrochromic glass thermal comfort,
four workflows are developed in different software. Each workflow is evaluated to find out the strengths and
limitations of the process.

3.1 Workflow Overview

COMFEN, DesignBuilder, Grasshopper (with Ladybug and Honey plug-ins), CBE Thermal Comfort Model are used
to evaluate the effectiveness of the tool in terms of thermal comfort for EC glazing. Among existing software that
have the ability of conducting thermal comfort simulation, COMFEN, DesignBuilder, and Grasshopper (with Ladybug
and Honey plugins) were chosen because these are the ones that have the access to choose EC glass from the software
glazing library, or users can create specific glass properties inside the software. CBE Thermal Comfort Model was
chosen because this software takes solar load in to account when doing simulations for thermal comfort. Some other
software might also have the capability of doing thermal comfort simulations, but were not considered.

The overall workflow can be divided in to two parts: workflow development and workflow evaluation (Figure 12).
Workflow involves studying the software’s existing capabilities of doing thermal comfort simulations and improving
the process if possible. The workflows are evaluated based on three main criteria: if all four tinting states of the current
electrochromic glazing are considered, if solar load impacts are taken into account, and if the tool could apply glass
tinting schedule in the simulation process. Case studies including the test room with EC glass and reference room with
low-e glass are done in each software to compare the results of the two case rooms.

Figure 13 Workflow Overview

3.2 Case Study Room

Research about electrochromic glazing had been done at the 71T test facility of the National Lawrence Berkeley Lab,
San Francisco, California. The studies were very detailed and accessible to public, so this lab was chosen to serve as
the simulation prototype. The facility was built to imitate a commercial office environment which consists of three
identical side-by-side test rooms. Two rooms are modeled with the same settings and exposed to the same exterior
and interior environment simultaneously except for the window type so that the results can be comparable. The rooms
are 3.05m (10ft) wide by 4.57m (15ft) deep and 3.35m (11ft) high (Figure 14). Fifteen 0.85x0.46m electrochromic
insulating glass units facing due south are installed, forming a 3-wide by 5-tall array of windows. The test room used

30
EC glass with four tinting states, while the reference room windows are double pane low-e glass fabricated with the
same dimensions and spacers as the test EC IGUs. The reference low-e glass properties are Tv=0.42, SHGC=0.219,
U-Value=1.408 W/m2-ºC (0.25 Btu/h-ft2-ºF) (Lee et al., 2006).

Figure 14 Floor Plan and Section of Test Model

3.3 Workflow Development

Detailed software interface and specific steps related to thermal comfort are illustrated. Workflows in COMFEN and
DesignBuilder follow the basic functions of what the software could do. However, workflows in Grasshopper (with
Ladybug and Honeybee plugins) and CBE Thermal Comfort Model are developed based on new methods and the
limitations of the glass libraries are solved through these two workflows.

3.3.1 COMFEN
COMFEN is a tool that gives the decision maker early feedback on the design variables and helps in integrated design
decisions (Selkowitz et al, 2011). COMFEN provides outputs such as energy consumption, peak energy, façade and
window loads, thermal comfort, daylight illuminance which are presented in graphical and tabular format within the
simplified user interface and the analysis engine for those outputs are EnergyPlus and Radiance (Selkowitz et al, 2011).

When start a new project in COMFEN, project name, location, building type, orientation, cost and HVAC can be
selected. Because weather file cannot be specified, building location is also relevant to the weather input used for
simulation. Building type is also the determinant factor for occupancy, lighting and equipment schedules. The building
geometry that COMFEN could model is limited to regular rectangular space. The software interface is divided into 3
sections: the project browser on the left, scenario graphic in the upper right and results tabs in the lower right (Figure
15). A project is composed of a collection of scenarios which can be added in the scenario edit tab. Geometry,
orientation, glazing systems, shading systems, framing systems, lighting controls, and occupancy and plug load values
can be specified in each scenario.

31
 Figure 15 COMFEN User Interface

After creating a set of scenarios and clicking the calculate button, COMFEN automatically generates and exports input
data file for each scenario into EnergyPlus. After performing the required calculations in EnergyPlus, the output data
file for each scenario can be displayed in COMFEN with different graphs. Each graph can be saved as PNG images
or CSV files for further study.
More detailed steps are demonstrated as follows. After giving the basic project information, then create room geometry
by defining height, width and depth which will generate a graphic image of the exterior façade. Orientation is defined
in terms of geometry coordinates and project north feature definition which can be found on the COMFEN knowledge
base website. COMFEN then generates lighting controls, lighting and equipment loads, and number of people
automatically. Then move on to the window system where glazing system, frame, and shading can be defined
(Selkowitz et al, 2011). In COMFEN, there are two kinds of electrochromic available in the glass library (Table 6).
After all the settings are done, users can drag the scenarios to the comparison tab and annual average thermal comfort
graphs will be shown in summary. The graphs can also be exported to .csv file for further analysis.

Name Tvis SHGC U-factor (Btu/h-ft2-F) Thickness Cost ($/ft2)

Sage 4mm 0.630 0.488 0.331 0.78 10.84
Sage 6mm 0.633 0.477 0,329 0.84 10.84
Table 6 EC glass in COMFEN

3.3.2 DesignBuilder
DesignBuiler is an established user interface to EnergyPlus and provides access to commonly required simulation
capabilities, such as building fabric, thermal mass, glazing, HVAC systems, financial analysis, etc. A longer
description can be found on the DesignBuilder website https://www.designbuilder.co.uk/.

As for thermal comfort simulation for electrochromic glazing, users can use the window shading data and defining
the properties of the outer window pane to be used when the shading is active (Figure 16). According to the user
manual, the procedure is the following:
● Specify the glassing type on the opening tab under external windows, under shading.
● Select the type as “'Electrochromic absorptive 6mm” or “'Electrochromic absorptive 6mm”.
● Select the Position as 4-Switchable.
● Set the control type with corresponding set point.

32
Control type defines what the electrochromic glass control algorithm would be. For example, if choose the control
type to be “4-solar”, a solar set point can be set to a given number. The glass would enter its dark state whenever the
solar radiation incident on the outside of the window is greater than that number (DesignBuilder Program Help, 2017).

Figure 16 DeisgnBuilder User Interface

The default simulation results in DesignBuilder don’t have the thermal comfort outputs; one has to go to the output
interface and check the Fanger option under comfort and environment tab (left side of Figure 17). Once the simulation
is done, users alsoneed to go to the display options on the left side of the screen and check the Fanger PMV option
again (right side of Figure 17).

33
 Figure 17 Output Fanger PMV in DesignBuilder

3.3.3 Ladybug and Honeybee

Grasshopper was chosen to be one of the softwares tested, because it has many powerful plugins, making simulation
tools available in a parametric way. For example, Ladybug as an open source environmental plugin for Grasshopper
can import standard EnergyPlus Weather files (the weather files are available for download at the website:
http://www.ladybug.tools/epwmap/) into Grasshopper, while Honeybee can connect Grasshopper to EnergyPlus,
Radiance, Daysim and OpenStudio for building energy and daylighting simulation (Ladybug Tools, 2017). Another
important advantage of Ladybug and Honeybee is that the comfort models are derived from the Java script of the CBE
Thermal Comfort Model (Mackey, 2010). So even though the comfort results generated from Ladybug and Honeybee
are PMV data, the PMV model here is based on solar adjusted mean radiant temperature, making it different from the
Fanger PMV model. Solar adjusted MRT is a term used in the method of calculating solar gain to the human body
which takes solar radiation into account in comfort. The component in Grasshopper uses radiance functions to
determine the amount of direct and diffuse solar radiation falling on a comfort mannequin. The formulas to translate
the radiation into an effective radiant field and into a solar-adjusted mean radiant temperature and detailed process
can be found in the study done by Arens et al (Aren, et al, 2015)

Importing a climate file and running an energy simulation with this climate file and building is the pre-step to thermal
comfort analysis. From the energy model and the corresponding weather file, four main outputs: surface temperature,
air temperature, solar radiation incidents on occupants and air temperature stratification can be obtained (Figure 18).
These four output are required to produce an indoor comfort map. Before the comfort map calculation, a set of test
points and mesh is needed to represent where temperature and comfort analysis will be performed. Note that the
distance unit in EnergyPlus is the meter, and the bigger the grid size, the higher resolution of the results and the more
time will need for the simulation (Mackey, 2010).

34
 Figure 18 PMV Model in Ladybug and Honeybee (Mackey, 2012)

Additional shading input is needed to accept shading geometry which will be used to block sunlight that could fall
directly onto occupants (Mackey, 2010). With the Honeybee window shades generator component (Figure 19) shade
type can be selected to switchable glazing (input number 2 under Shade Type) that represents electrochromic glazing
that can be switched on to reflect the material state of the shade material. However, only one shade material can be
plugged in under shade material which limited the EC glass with only two states. The shade control types have various
options, for example, shading is on if outside air temperature exceeds set point or shading on if total horizontal solar
irradiance exceeds set point. A specific shade set point input can be given based on the control type chosen from the
previous step. This is very useful when assuming the EC glass only have two tinting states. However, the objective
for this thesis is to study the performance of EC glass that have four tinting states, so a new workflow using
Grasshopper was developed.

Figure 19 Honeybee Window Shade Generator

35
Because of the dynamic nature of the EC glazing and most exiting software lack the intermediate states in their library,
it is very challenging to simulate the systems that are constantly changing the glass parameters (e.g., VLT and SHGC)
(Ardakan, 2015). The workflow in Grasshopper can solve this problem by modeling the glass as four different glass
types (Figure 19).

Figure 20 Model Four States EC Glass in Grasshopper

Components for both adaptive comfort and PMV analysis also need epw weather file to provide key inputs, for
example, direct and diffuse solar radiation data are used as inputs to calculate solar adjusted MRT (Figure 21). There
are also several other optional inputs (e.g. window transmissivity, eight percent satisfied occupants) if the users don’t
want to use the default values. The more detailed descriptions can be found on Chris Mackey’s thesis (Mackey, 2010).

Figure 21 PMV and Adaptive Comfort Recipe

36
The overall workflow in Grasshopper is displayed in Figure 22 and based on the methodology used in Ardakan’s
research for EC glass energy performance simulation (Ardakan, 2015).
 Build the room model in Grasshopper and define input parameters that needed for energy and thermal comfort
simulation.
 To determine the EC tint schedule, a daylighting simulation is performed based on weather data to find the
illuminance level incident on the window. (As the algorithm that control the glass tinting is already known,
namely using the upper and lower threshold to establishing the relation between the outside vertical
illuminance and the EC tinting state. See more information in Chapter 3.5.4.
 Perform thermal comfort simulations for each state of EC window (63%, 20%, 6%, 1%).Visible light
transmittance 63%, 20%, 6%, 1% are described as State 1, State 2, State 3, and State 4 accordingly in the
workflow diagram.
 Extract the thermal comfort simulation results based on the predicted EC tint states at each time step.
 Combine the extracted thermal comfort results for each day so as to come up with daily results.

Figure 22 Workflow in Grasshopper

3.3.4 CBE Thermal Comfort Tool

Because the PMV model was developed using test in uniform thermal environments and did not consider the solar
radiation impacts, UC Berkeley has developed a more sophisticated thermal comfort model that can predict thermal
comfort in non-uniform conditions. This model is capable of accessing local comfort for different body parts which is
influenced by radiant heat exchange through the window, solar gain, air motion, and non-uniform air temperatures,
and integrates local comfort levels to yield whole body thermal comfort (Zhang et al. 2004). The program is based on
large number of human subject tests within asymmetric and transient thermal environments (Zhang, 2003).

Figure 23 is the interface of the UCB Thermal Comfort Model. The simulation output including local thermal comfort,
local thermal sensation, temperature, heat flux, solar load, etc. Those results can be displayed in image or exported to
csv file for further analysis. Note that the simulation results table on the right part of Figure 23 are a series data set
for every minute and image is linked to the table. When clicking on different points of the simulation, the image
displays that data set.

37
Users can define their own room and window geometry through the “room editor.” As the person can be placed at any
place inside the space, the model can be used to predict thermal comfort in a perimeter zone or center zone. Non-
uniform temperature can be defined because the surface temperatures, such as the window or wall temperature, are all
determined individually. The glazing products database is also incorporated in the model (Lyons et al., 2000).

Figure 23 CBE Thermal Comfort Model User Interface

As there is no EC glass or alternative glass with similar glass properties in the glazing products database, no
simulations could be done for EC glass. CBE Thermal Comfort Model is no doubt a good tool for thermal comfort
simulations, but the existing glass library with only a small number of glass types available in the database limits the
use of this tool. In order to solve this problem, a more in-depth investigation was done to improve the glazing database.

The first step is to find the software installation location in the computer. There are three sub-folders for the UCB
Comfort Model in local Disk (C:): Client, Dispatch Server and Simulation Server (Figure 24). Glass library named
“DelphiGlassType” in .txt format can be found both in the Client and Simulation Server Folder. Modifications are
made in both .txt file. It has been proven that only changing the “DelphiGlassType” file under the Client folder would
work. In this case, if one add a glass type in this file, the glass type would appear in the software interface. Following
the folder path C:\UCBComfortModel\Client\System\GlassLib, one can find the “DelphiGlassType” file and edit it.

38
 Figure 24 CBE Thermal Comfort Model Software Installation Folder

When adding a new glazing product in the “DelphiGlassType” file, many properties of the glass is needed (e.g. Tsol,
Rsol, Tvis, Rvis, Emis, etc.) To find the exact values for those properties, software WINDOW 7.5 which has a much
bigger glazing products database is used. Sage glass with five states can be found in the WINDOW 7.5 glass library.
For simplicity, the four states are selected (highlighted in Figure 25) and the corresponding values are extracted as
inputs for simulations by adding these data to the “DelphiGlassType” file under Client folder (Figure 26).

Figure 25 EC Glass Properties in WINDOW 7.5

39
 Figure 26 Editing the “DelphiGlassType”

Within CBE Thermal Comfort Model, only thermal conditions at a single point in space and a single point in time,
can be achieved. In order to get a one day thermal comfort result, simulations are done for each hour from 6am to 6pm
based on the glass tinting schedule generated from Grasshopper. In order to take solar load into account, under the
solar load tab, automatic solar load calculation was selected (Figure 27). Solar altitude and solar azimuth was
determined by the sun path code in Grasshopper (Figure 28). The influencing factors are latitude and longitude based
on site location, along with specific day and time.

Figure 27 Solar Load Setting in CBE Thermal Comfort Model

Figure 28 Sun Path Code

40
3.4 Inputs
Different software use different inputs as simulation prerequisite for thermal comfort simulation. Those needed inputs
are summarized including weather data, thermal comfort parameters, occupant locations, tinting schedule which are
required simulation parameters for the software described in chapter 3.3.

3.4.1 Weather Data

Different software provides different weather data input format. In COMFEN and DesignBuilder, specific location
can be selected that corresponds to the local weather data. In CBE Thermal Comfort Model, no location or time can
be specified; the only way to connect to weather data is by inputting the solar altitude, solar azimuth and direct solar
intensity. For Ladybug and Honeybee, epw weather file can be loaded directly from the weather file website.

3.4.2 Thermal Comfort Parameters

The determining thermal comfort factors have two categories: environmental factors including air temperature, mean
radiant temperature, air speed and relative humidity, and personal factors that includes metabolic rate, clothing
insulation. The environmental parameters are influenced indirectly in COMFEN, DesignBuilder and Grasshopper by
defining a specific HVAC system, while these parameters can be specified directly in CBE Thermal Comfort Model.
As for the human factors, users are not allowed to change metabolic rate and clothing level in COMFEN, but in
DesignBuilder, Honeybee and CBE Thermal Comfort Model those two factors can be modified according to the
human activity and season.

3.4.3 Occupant Location

Normally, occupant’s location in a room is based on furniture arrangement or different tasks. Different location varies
differently on solar radiation and contact surface temperature. Obviously, the closer a person is to a window, the
greater the impacts on comfort. COMFEN, DesignBuilder are focus on the primary zone, while in CBE Thermal
Comfort Model users can adjust occupant position at any point in the room, and in Grasshopper the zones can be
separated into based on grid size, thus multiple data set can be achieved.

3.4.4 Tinting Schedule

In the following simulation, the EC windows are assumed to operate automatically, tinting in response to light sensors
to maintain a comfortable level of daylight in the space. Daylighting control strategy is used more logical than other
control strategies because EC glass is applied in the building mainly for the purpose of maintaining good lighting
condition and reducing glare problems. For simplicity, other kind of control strategies (e.g. temperature control, glare
control are not considered). The amount of light entering through the window is defined by giving an upper and lower
illuminance threshold. When the upper set point is crossed for a certain amount of time, the glass will tint to the next
state. Similarly, when the lower set point is crossed, the glass will change to the previous state (Ardakan, 2015). The
thresholds that trigger the state transitioning are from the research done by Ardakan and are showed below (Table 7).

EC State Lower Threshold Upper Threshold

Clear State (63% VLT) 0 Lux 8772 Lux
Intermediate Sate 2 (20% VLT) 8772 Lux 27237 Lux
Intermediate Sate 1 (6% VLT) 27237 Lux 101449 Lux
Fully Tinted Sate (1% VLT) 101449 Lux ----
Table 7 EC Glass Tinting State Threshold (Ardakan, 2015)

After establishing the relationship between the vertical illuminance and the EC tinting state, a daylighting simulation
is performed in Grasshopper with the exact same system configurations. EC glass tinting schedule is determined based
on the exact illuminance level at each hour.

41
3.5 Workflow Evaluation
There are three main factors that will influence the thermal comfort results for EC glass: if solar load impacts are taken
into account, if all four tinting states of the current electrochromic glazing are considered, and if the tool could apply
glass tinting schedule in the simulation process. Some other categories are also considered including the results metrics,
analysis period generated in the workflow, complexity of the overall workflow and if the results can be visualized.

3.5.1 Solar Load Impacts

Sunlight hitting the human body is always the source of thermal discomfort near windows. Even when only the human
head or arm is exposed into direct sunlight, the most uncomfortable part determines the overall thermal conditions.
When calculating the thermal comfort, the existing mean radiant temperature has to be adjusted based on solar
radiation. Different workflow use different thermal comfort metrics. It is critical to find out the specific metrics for
each workflow.

3.5.2 EC Glass Tinting States

As illustrated in Chapter 1.2.4, the EC glass on market now have at least four tinting states, and each tinting state has
different performance specifications including Tv, SHGC, %Tvis., %Rb Int., %Tsol, %Tdw-K. The amount of solar
light or heat transmitted through the glass is quite different under different glass states. Thus the thermal comfort
results vary in different tinting states. Only one or two states cannot fully represent how EC glass influence the indoor
environment if the glass is using more than two states.

3.5.3 EC glass Tinting Schedule

The EC glass tested has four tinting states; depending on different control algorithms, the glass won't stay in the same
state in a normal sunny work day. The results can be quite different if the glass stays fully clear or fully tinted
throughout a day. In reality the EC glass tinting state schedule is a combination of different tinting states followed by
different time periods. Only when the simulations follow the glass tinting states base on the tinting schedule, the results
can be convincing enough represent the impacts of EC glass.

3.6 Summary
This chapter introduced the four software that would be used in each workflow. Detailed workflow development in
Grasshopper and CBE Thermal Comfort Model were discussed. Input parameters needed for the simulation process
were summarized. Workflow evaluation criterial was explained and the actual data inputs and results will be further
discussed in the next chapter.

42
 Chapter 4: Results
A series of tests were done to evaluate the capabilities of the proposed workflow in thermal comfort simulations for
EC glass. Following the workflow of each tool, the test room and reference room were modeled, and simulations were
run based on the simulation time step. This chapter describes the detailed inputs needed and shows the results for each
major step of each software. As the only changing parameter, the glass type used in each software is specified that
best matches the EC glass. If no EC glass is available in the software library, the method of choosing the glass used
in each software was to find the glass properties that best match (or model the exactly same glass) with EC glass used
in the test room and low-e glass used in the reference room.

4.1 Results from Each Workflow

Results from different workflow have different thermal comfort metrics to represent occupant thermal comfort level.
In COMFEN, the percent of people satisfied is used as evaluation metric. DesignBuilder uses Fanger PMV, while
PMV base on solar-adjusted mean radiant temperature can be achieved in Grasshopper. CBE Thermal Comfort Model
has its own unique overall comfort and overall sensation metric as outputs.

4.1.1 COMFEN
In COMFEN, users can build models simultaneously with different glazing systems. Because the room window was
facing south, project north degree was set to 0 degrees. The building type was an office, and its location was selected
to San Francisco. The HVAC system type used was packaged single zone (COMFEN currently allows only packaged
single zone systems). Outdoor air control flow rate was based on flow/person, and the input number was 21.19
cfm/person (which is the default number based on building type). After giving the room geometry and materials,
Scenario 1 chose to use PPG—SB R100 (2) clear/clear (air) for the glazing system because this type of glass properties
matches most with the reference low-e glass properties described in chapter 3.4, while scenario 2 used Sage glass 6
mm (Table 8).

Glass Type Tvis SHGC U-factor Thickness

Sage Glass 6mm 0.644 0.477 0.329 0.84
PPG—SB R100 (2) clear/clear (air) 0.415 0.233 0.291 0.95
Table 8 Glass Used in COMFEN

Thermal comfort data were calculated from 8 am to 6 pm throughout the year. The Y axis represents the percentage
of people satisfied in regards to thermal comfort. The room with conventional glass is slightly more comfortable than
the room with EC glass in most of the time, which are not as expected (Figure 29). However, the results in both rooms
are within 2% difference which are almost identical. To figure out the reason, a breakdown results for each season are
shown in Figure 30-33 are further discussed on chapter 5.1.1. Note that Figure 29 to 33 are all graphs based on the
CSV file exported from COMFEN.

43
 91.5
91
90.5

Percent People Satisfied

90
89.5
89
88.5
88
87.5
87
86.5
86
8:00 9:00 10:00 11:00 12:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00
AM AM AM AM PM PM PM PM PM PM PM PM PM

Reference Glass EC Glass

Figure 29 Annual Average Thermal Comfort

91
90.5
Percent People Satisfied

90
89.5
89
88.5
88
87.5
87
86.5
8:00 9:00 10:00 11:00 12:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00
AM AM AM AM PM PM PM PM PM PM PM PM PM

Double Low-e Glass EC Glass

Figure 30 Spring Average Thermal Comfort

44
 94
92

Percent People Satisfied

90
88
86
84
82
80
78
76
74
8:00 9:00 10:00 11:00 12:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00
AM AM AM AM PM PM PM PM PM PM PM PM PM

Reference Glass EC Glass

Figure 31 Summer Average Thermal Comfort

93
92
Percent People Satisfied

91
90
89
88
87
86
85
84
83
8:00 9:00 10:00 11:00 12:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00
AM AM AM AM PM PM PM PM PM PM PM PM PM

Double Low-e Glass EC Glass

Figure 32 Autumn Average Thermal Comfort

45
 Winter
96

Percent People Satisfied

94
92
90
88
86
84
82
80
8:00 9:00 10:00 11:00 12:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00
AM AM AM AM PM PM PM PM PM PM PM PM PM
Axis Title

Double Low-e Glass EC Glass

Figure 33 Winter Average Thermal Comfort

Annual energy performance data can also be exported under the comparison tab in COMFEN. The energy results
include heating, cooling, fans, and lighting. However, for simplicity, only heating and cooling energy consumption is
considered (Table 9).

Scenario Name Heating and Cooling (kBtu / ft2)

Reference Glass 8.7
EC Glass 16.75
Table 9 Annual Energy Results from COMFEN

4.1.2 DesignBuilder
The case study room was built in Revit and exported to DesignBuilder through gbXML file (Figure 34). After the
building geometry was input, the activity template was set to generic office area. Occupancy, schedule, and
environment control were all based on this template. The HVAC system selected was “Fan Coil Unit” and the
construction materials used were generic construction template. As for the glass settings, there are two types of EC
glass available in DesignBuilder: Double Electrochromic Absorptive and Double Electrochromic Reflective. Both of
the two types have two states: bleached and colored. “Dbl LoE Elec Abs Bleached 6mm/13mm Air” was selected for
the following simulation because its glass properties better match with the properties of EC glass (fully clear sate) in
the case study. Likewise, “Dbl LoE Spec Sel Tint 6mm/13mm Air”, having the similar glass properties with the low-
e glass in the case study, was selected for further simulation (Table 10).

Figure 34 Models in Revit and DesignBuilder

46
 Glass Type SHGC Direct Solar Light Transmission U-Value
Transmission
Dbl LoE Spec Sel Tint 0.219 0.208 0.408 0.264
6mm/13mm Air
Dbl LoE Elec Abs Bleached 0.476 0.364 0.657 0.261
6mm/13mm Air
Table 10 Glass Properties Used in DesignBuilder

The simulation period was one day, and the output was generated hourly. Figure 35-38 show the PMV results of both
test room and reference room on winter solstice day and winter solstice day. Zero is comfortable; negative values are
cold; positive values are hot.

Figure 35 Test Room PMV Results on Winter Solstice Day in DesignBuilder

Figure 36 Reference Room PMV Results on Winter Solstice Day in DesignBuilder

Figure 37 Test Room PMV Results on Summer Solstice Day in DesignBuilder

Figure 38 Reference Room PMV Results on Summer Solstice Day in DesignBuilder

4.1.3 Ladybug and Honeybee

The workflow developed in Grasshopper need to get the EC glass tinting schedule first, then the schedule will be used
as reference guide to extract data from the four simulations data set.

 Tinting Schedule
To do a lighting illuminance simulation in Grasshopper, the first step was to convert the room geometry to Honeybee
zones with all the properties needed. The interior daylight analysis was done assuming that the glass stays in EC clear
state all the time, so that the light sensors know when to change the EC glass from clear state to the next state. As the
47
control strategy used is illuminance on work plane, so the light sensors were placed on the work plane height 0.762 m
and were distributed evenly (space between each sensor was 0.5m) in the entirely space. The weather file and specific
date and time was given to determine the sun’s position. All the inputs recipes (inputs) are put together and imported
to “Honeybee Daylight Analysis” component (Figure 39). Appendix X-1 has all the Grasshopper code. Illuminance
data for each test point were generated using the code, and only the maximum illuminance data were extracted from
all the sensors for each hour, and the data were used to decide what state the glass should be based on the tinting state
thresholds described in chapter 3.5.4 (Table 11, 12). Each color in the table correspond to one particular state.

Figure 39 Illuminance Calculation Code in Grasshopper

48
 Table 11 Winter Solstice Day Illuminance and EC Glass States

Table 12 Summer Solstice Day Illuminance and EC Glass States

 PMV
The entire process is shown in Figure 40. To get the PMV and energy consumption data, the office building program
was selected from the template. Wall, roof, and floor materials was chosen based on ASHRAE climate zone. As the
room prototype is located at LBNL, Berkeley, California (latitude 37°4'N, longitude 122°1'W), the climate zone was
set to 3. The room was conditioned with fan coil units, which is consistent with the real HVAC system for the lab at
LBNL. The analysis period was for the entire days of June 21th and December 21th, and only the typical office hour
time period was considered (8AM-6PM). Glass input followed the exactly same properties as the EC glass four tinting
states. All these inputs were imported to Open Studio, and the simulation run through EnergyPlus. The output from
the EnergyPus includes the energy consumption. Output such as surface temperature, air temperature, air flow volume
and air heat gain rate can also be used as input for further PMV calculation. Along with test point generated from
“Indoor View Factor” component and occupant activities and clothing level parameters, PMV results were generated
under the same analysis period. Apendix X-2 has all the Grasshopper code.

49
 Figure 40 Thermal Comfort Simulation Code in Grasshopper

As it is still not possible to add a dynamically tinting EC glass with four states in Grasshopper, simulations in
Grasshopper have to be done separately for each state. The simulations were run four times, and data were extracted
based on the EC glass tinting states. Unlike the dynamic character of EC glass, the static window room only need to
have the simulation run once for each day. The thermal comfort results generated from the simulation are shown below
(Table 17 and Table 18). -0.5<PMV<+0.5 is considered as acceptable thermal environments and zero is the ideal value
(ASHRAE Standard 55, 2004).

Table 13 Test Room PMV Results in Ladybug and Honeybee

50
 Reference Room Reference Room
Hour PMV on Winter PMV on Summer
Solstice Day Solstice Day
8:00 AM -0.08 0.37
9:00 AM 0.18 0.42
10:00 AM 0.39 0.46
11:00 AM 0.90 0.44
12:00 PM 1.26 0.44
1:00 PM 1.19 0.45
2:00 PM 0.65 0.45
3:00 PM 0.57 0.48
4:00 PM 0.51 0.50
5:00 PM 0.42 0.50
6:00 PM 0.38 0.49
Table 14 Reference Room PMV Results in Ladybug and Honeybee

4.1.4 CBE Thermal Comfort Model

In CBE Thermal Comfort Model software, the first step in setting up a simulation is to define various parameters for
the environment (room geometry, air temperature, air velocities, radiant temperatures, and solar conditions) and
occupant (position, posture, clothing, body basic information, gender, skin color, body fat). The room geometry is
limited to a rectangular shape in this tool, and simulation results can be shown with specific color scale on the building
geometry and human body (Figure 41). The body segment temperature was set to 26°C, and air velocity was set to 0.1
m/s. The occupant inside the space was a male sitting near the window and the clothing is based on simulation season.
Window and other opaque panel temperatures were all set to 26°C, which is the default temperature in the system.
The occupant is assuming doing office work and seated near the window.

Figure 41 Models in CBE Thermal Comfort Model

Solar azimuth and altitude at San Francisco are derived from the Grasshopper code, which can be used as solar load
input in UCB Thermal Comfort Model (Table 15).

51
 Winter Solstice Day Summer Solstice Day
Time Solar Azimuth Solar Altitude Solar Azimuth Solar Altitude
8 AM 128 9 88 39
9 AM 139 17 98 50
10 AM 152 24 111 62
11 AM 167 28 136 72
12 AM 183 30 186 76
1 PM 198 27 231 70
2 PM 212 22 252 60
3 PM 224 14 265 48
4 PM 235 6 274 36
5 PM 244 -4.6 282 24
6 PM 252 -16 290 13
Table 15 Solar Azimuth and Altitude

Simulations were done for each hour with EC glass and low-e glass separately. The simulation time was set to 60 min
and simulation results output time interval was 1 min. The last-minute thermal comfort results of each hour are shown
below (Table 16, 17). For simulations of each hour, EC glass was set to different tinting states based on the tinting
states (Table 11, Table 12) generated from Grasshopper. The body segments initial temperature were updated based
on previous time step result.

Test Room Reference Room

Time
Overall Overall Overall
Overall Comfort
Sensation Comfort Sensation
8:00 AM -0.31 0.49 -3.12 3.53
9:00 AM 0.04 0.42 -3.24 3.5
10:00 AM 0.05 0.45 -3.33 3.55
11:00 AM 0.00 0.51 -3.35 3.55
12:00 AM -0.01 0.54 -3.39 3.57
1:00 PM 0.06 0.5 -3.41 3.61
2:00 PM 0.09 0.49 -3.41 3.64
3:00 PM 0.16 0.45 -3.32 3.65
4:00 PM -0.24 0.52 -3.19 3.63
5:00 PM 1.16 0.17 -1.09 0.72
6:00 PM 1.15 0.17 -1.08 0.72
Table 16 Thermal Comfort Results at Winter Solstice Day in CBE Thermal Comfort Model

Test Room Reference Room

Time
Overall Overall
Overall Comfort Overall Sensation
Sensation Comfort
8:00 AM 1.21 0.17 1.1 0.18
9:00 AM 1.18 0.16 0.94 0.23
10:00 AM 1.09 0.19 0.33 0.49

52
 11:00 AM 1.02 0.22 -0.41 0.66
12:00 AM 0.96 0.24 -0.71 0.75
1:00 PM 0.99 0.23 -0.54 0.69
2:00 PM 1.16 0.17 0.99 0.22
3:00 PM 1.19 0.16 1.00 0.21
4:00 PM 1.16 0.17 0.88 0.25
5:00 PM 1.15 0.17 0.89 0.24
6:00 PM 1.15 0.17 0.89 0.25
Table 17 Thermal Comfort Results at Summer Solstice Day in CBE Thermal Comfort Model

CBE Thermal Comfort Model can predict thermal comfort for human body as a whole as well as for specific body
parts. The image is an example of the simulation outcome at noon showing the detailed human body comfort level
with the surrounding thermal environment (Figure 42). The simulation results, overall comfort and overall sensation,
use a scale of -4.0 to +4.0 (Figure 43). For the overall comfort, the darker shades of green indicating greater comfort,
while discomfort is indicated through shades ranging from yellow to dark (UCB Thermal Comfort Model Users
Tutorial). Meanwhile, thermal sensation is indicated by image on the right of Figure 43. Red means very hot, and blue
represents very cold. The outputs are given for each of the 16 body segments and the whole body. The most
uncomfortable part determines the overall thermal conditions. For example, if only the head shows a significant
increase in temperature, thus high values for the head sensation and the person overall will feel uncomfortable
(Hoffmann et al, 2012).

Figure 42 Thermal Comfort Results in CBE Thermal Comfort Model

53
 Figure 43 Comfort and Sensation Scale

4.2 Conclusion
Detailed results are collected for each workflow and some are not as expected. For example, in
COMFEN, most of the time the room with EC glass is not as comfortable as the room with conventional
glass. The EC glass tinting schedule in summer solstice day generated in Grasshopper only includes
two tinting states. These unexpected issues will be further analyzed in the next chapter.

54
 Chapter 5: Discussion
In this chapter, the room with EC glass window is referred to as the test room, and the room with low-e glass window
is referred to as reference room. Simulation results are analyzed and evaluated through comparing data of the test
room and reference room. The unexpected results are identified, and potential reasons are discussed. Each workflow
is evaluated by giving the advantages and limitations. Some potential study limitations also need to be clarified. Note
that all the analysis or conclusions are based on the case studies that may not be suitable for other projects. However,
the proposed methodology can be applied to other buildings with EC glass installed.

5.1 Result analysis of Each Workflow

Thermal comfort outcomes from COMFEN and DesignBuilder are summarized and analyzed. The results getting from
each workflow are reorganized as two dimensional charts to compare the test room and reference room performance
difference and are used for further discussion.

5.1.1 COMFEN
Surprisingly, the annual average thermal comfort chart (Figure 29) shows that percentage of people satisfied
differences in the test room and reference room are very small (around 2-3%), which are essentially identical. The
outcomes in the test room is more often slightly lower than the reference room. The test room only performed better
from 8AM-10AM, and from 10 AM to 8PM more people are satisfied with the thermal environment in the reference
room than the test room. However, the break down the annual thermal comfort results of each season (Figure 30-33)
indicate that the test room performs better throughout the day during the summer on thermal comfort, while the test
room loses against the reference room during spring, autumn and winter. In addition, the test room consumes more
energy than the reference room. The results are not as expected because normally EC glass is supposed to be more
comfortable and save more energy in most of the time (Pease et al, 2010). So these results not convincing to show the
advantages of EC glass over static glazing.

When comparing the glass properties, it is found that” PPG—SB R100 (2) clear/clear (air)” has a lower visible light
transmittance (Tvis) and solar heat gain coefficient (SHGC) than “Sage Glass 6mm.” In this case, the test room is
likely to allow more solar radiation coming into the space that may cause human discomfort and increase the energy
consumption. Meanwhile, as for the Sage glass in COMFEN, there is no control strategies or set point settings
available in the software system. As a dynamic glazing, when running simulations in COMFEN users cannot adjust
the Sage Glass states as there is only one tinting state available, which is the main limitation of the software.

The thermal comfort results are yearly or seasonal average data, so it is not possible to predict thermal comfort on a
given day or at a certain time of one day. The only way to evaluate thermal comfort is to see the percentage of people
satisfied. Users cannot tell if the occupants are feeling cold or warm, so the results are not sufficient for further trouble
shooting of the system. On the other hand, as the user interface is quite simple, COMFEN is useful if one wants to get
a roughly estimation of energy or thermal performance of the space and make comparison models to see which one
works better.

5.1.2 DesignBuilder
The thermal comfort results of DesignBuilder show that both the test room and reference room have negative PMV
value at summer solstice day, which means the occupants feel slightly cold. Occupants feeling slightly cold on a hot
summer day is unexpected. While on the winter solstice day, the test room have a positive PMV value 0.45 and the
reference room have a negative PMV value -0.022 (Table 18). According to ASHRAE standard 55, -0.5<PMV<+0.5
is considered as acceptable environments, so the results indicate that the test room are comfortable both at summer
and winter solstice day, while the reference room is slightly uncomfortable at summer solstice day.

PMV at Summer Solstice Day PMV at Winter Solstice Day

Test Room -0.45 0.45
Reference Room -0.65 -0.022
Table 18 PMV Results in DesignBuilder
55
Even though the PMV results generated from DesignBuilder is for a particular day, the data is a consistent number for
the whole day, which is also unexpected. Besides, the results of the two rooms are so close that makes it not convincing
enough to tell if EC glass performs better than static low-e glass. Note that the thermal comfort metrics used in
DesignBuilder is Fanger PMV that didn’t take solar radiation into account, which might be a reason for the unexpected
results. Apart from the simulation results, among the tools that tested in chapter 4, DesignBuilder is the only software
that allows users to export models from other modeling software and do simulations for more than one zones, so it
would be very helpful if the building geometry is complicated or there are multiple zones need to study at the same
time.

5.1.3 Ladybug and Honeybee

During the winter solstice day, the tinting schedule generated based on the illuminance calculation and control
algorithm seems reasonable. The glass turns to relatively dark states from 10AM to 4PM and stays in clear state in the
early morning and late afternoon. However, the tinting schedule on summer solstice day only has two tinting states.
The question arises that the glass never goes to the intermediate state 2 or fully tinted state on a hot summer day. In
order to find out the reason, the interior illuminance distribution was exported from Grasshopper; it is based on the
same code (Figure 39) described in chapter 4.1.3. And the sun path was also studied to find the relationship between
solar position, interior illuminance level, and EC glass tinting schedule (Figure 44).

Figure 44 Sun Path

The interior illuminance level on winter solstice day and summer solstice day from 7AM to 6PM were studied for the
test room assuming the EC glass stayed in clear state so as to get the actual sunlight distribution in the space. As the
solar altitude is low in winter, the sun light goes far into the interior space (Figure 45). However, as the solar altitude
is high in summer, the direct solar radiation only falls in the areas close to the window, while there is not as much
solar light hitting into the inner zone of the room (Figure 46). From the comparison of Figure 45 and Figure 46, it is
obvious to see that due to the sun position difference, the overall illuminance distribution is wider and stronger on
winter solstice day than summer solstice day, which is the reason why the EC glass doesn’t turn to dark state on the
hot summer day.

56
Figure 45 Illuminance Distribution on Winter Solstice Day

Figure 46 Illuminance Distribution on Summer Solstice Day

57
Another thing need to consider is the sensors locations in the study is different from the real life. In the illuminance
calculation process of the case study, in order to catch the highest illuminance data of the entire room, the light sensors
were placed on the work plane height 0.762 m and were distributed evenly with a 0.5m spacing in the entirely space.
However, in the room at LBNL, there are only four sensors in the room (Figure 47). Another illuminance simulation
was done having the same sensors as the real life and the maximum results were compared with the maximum data of
the case study with many sensors (Table 19). It was found that illuminance data collected from the four sensors were
not as high as many sensors. Therefore, the four sensors were not capturing the highest illuminance level in the room.
So it is better to have more sensors to get the highest illuminance data of the room because the occupant might be in
any location of the room at a given time.

Figure 47 Sensor Locations in Case Study VS in Real-life

Results from Case Study Sensor locations Results from Real-life sensor locations

winter summer winter summer

Hour (Illuminance (Illuminance Hour (Illuminance (Illuminance
in Lux） in Lux） in Lux） in Lux）
8:00 AM 2655 2425 8:00 AM 1069 862
9:00 AM 13237 5129 9:00 AM 11430 1866
10:00 AM 24546 12464 10:00 AM 22726 2883
11:00 AM 29964 23434 11:00 AM 28536 2191
12:00 PM 31510 26215 12:00 PM 29663 1685
1:00 PM 26467 23084 1:00 PM 24769 1820
2:00 PM 20015 14912 2:00 PM 18388 1976
3:00 PM 11599 4462 3:00 PM 9630 1697
4:00 PM 5527 3882 4:00 PM 2543 1469
5:00 PM 0 3164 5:00 PM 0 1225
6:00 PM 0 2342 6:00 PM 0 905
Table 19 Illuminance Results in Case Study VS in Real-life

After having the tinting schedule data, the next step is to re-organize the results using two dimensional charts to
compare the thermal comfort results in the test room and the reference room (Figure 48). The grey area in the figure
having PMV values within -0.5 and 0.5 are considered comfortable. On Winter Solstice day, overall EC glass performs
better than Low-e glass throughout the day. For the test room, most of the PMV numbers are within or around 0.5 and
58
0.5 which indicate that occupants are comfortable in most of the time. From 12AM to 2PM, EC glass is in intermediate
state1 and the occupant is slightly warm. This result suggests that if the EC glass is tinted to even darker state, the
uncomfortable problem will be solved. For the reference room, from 10AM to 2PM, is not comfortable as the PMV
values have exceeded 0.5.

Figure 48 PMV Comparison on Winter Solstice Day in Grasshopper

As for the thermal comfort on summer solstice day, from 8AM to 10 AM, the results in both rooms are very close,
and as the EC glass turns to the intermediate state 1 from 10AM to 2PM, the PMV of test room with EC glass is
slightly lower than the reference room. However, as the glass goes back to clear state from 3PM to 6PM, the PMV
in the reference room with low-e glass is lower than the test room, which also indicates that if the glass is set to
darker state during this time period, the thermal comfort level in the test room might be improved. Overall, despite
the slight difference between the two rooms, the PMV data in the test room and reference room are almost identical
and are within a relatively acceptable comfort range (Figure 49).

Figure 49 PMV Comparison on Summer Solstice Day in Grasshopper

59
In conclusion, as the workflow in Grasshopper has considered solar load impacts by using the solar adjusted
temperature when calculating the PMV data and all four tinting states are included during the simulation process, the
results generated from Grasshopper on winter solstice day and summer solstice day are reasonable and convincing.
Most of the results are as expected and the unexpected hourly data are helpful to find problems for current tinting
schedule. Even though it might take some time to develop the code, it can be used on any building type or system.
Changing the parameters in the code is also straightforward, which makes the workflow very efficient for future use.

5.1.4 CBE Thermal Comfort Model

To take solar load into consideration during the thermal comfort simulation process, solar setting inputs for each hour
was calculated using the Grasshopper sun path code as is is an effective method and can be used in any other location.
Note that the thermal comfort results achieved from CBE Thermal Model using overall comfort and overall sensation
as the thermal comfort evaluation metric which is different from PMV or PPD. The overall comfort and overall
sensation data achieved from CBE Thermal Comfort Model for each hour were reorganized as Figure 48 to Figure 51
and the comfortable areas are highlighted in grey.

The overall comfort of the test room on winter solstice day is higher than the reference room which is as expected
(Figure 50). However, both rooms have comfort problems to some extent. According to the comfort scale, numbers
above zero is considered comfortable, and numbers below zero are uncomfortable, the test room is having slightly
uncomfortable issue at 8AM and 4PM, while the reference room is in very uncomfortable condition throughout the
day. As for overall sensation, numbers above zero means hot and below zero indicates cold. The overall sensation
results show that the occupant in test room is slightly warm at 8AM and 4PM, while the occupant in reference room
is hot in most of the time (Figure 51).

Figure 50 Overall Comfort on Winter Solstice Day

60
 Figure 51 Overall Sensation on Winter Solstice Day

For the thermal comfort on summer solstice day, the results in test room have almost constant comfort values around
1 and overall sensation values around 0.1 which means that the occupant is comfortable throughout the day (Figure
52, Figure 53). Meanwhile, for the reference room, from 11AM to 2PM the overall comfort values are below zero
indicting occupant uncomfortable and from 9AM to 2PM the occupant feels warm because the overall sensation
number went higher than zero. From the comparison of the result, it is concluded that EC glass in the test room
enhanced the occupant thermal comfort by admitting less heat gain from the outside.

Figure 52 Overall Comfort on Summer Solstice Day

61
 Figure 53 Overall Sensation on Summer Solstice Day

In summary, the results came from CBE Thermal Comfort Model on summer solstice day and winter solstice day are
as expected and reasonable. The test room performs better than the reference room; this means that EC glass has
helped to improve occupant thermal comfort of the space. CBE Thermal Comfort Model is very useful in getting short
time period thermal comfort results. However, as the simulation process requires continuous inputs for each hour, if
one needs to find the monthly or yearly outcome, it would be very time consuming.

5.2 Workflow Evaluation

After a detailed result analysis of each workflow, the differences of each workflow are summarized in three main
areas including EC glass type, if solar load is considered, and tinting schedule. Other possible evaluation factors
including analysis period, result metrics, workflow complexity, and data visualization are also discussed. These
evaluation factors could be used as a guidance for people who want to study thermal comfort of EC glazing (Table
19).

5.2.1 Solar Load Impacts

The thermal comfort metric that used in COMFEN is percent of people satisfied (similar to PPD), which is another
expression of PMV, while the thermal comfort outcome from DesignBuilder is Fanger PMV data. For both COMFEN
and DesignBuilder, the simulation engine behind the results is EnergyPlus, which provides analysis based on several
comfort models including Fanger model and the ASHRAE Standard 55 adaptive model (Webb, 2012). However, both
these two models used in EnergyPlus did not consider the impacts of solar radiation. Thus, the thermal comfort outputs
in COMFEN and DesignBuilder didn’t take solar radiation impacts into account. On the other hand, EnergyPlus use
the control volume analysis method, which draws a boundary around a volume and solves for the output (Webb, 2012).
As a result, for COMFEN and DesignBuilder users cannot change the analysis location and the scale of focus is the
whole room. Those limitations are responsible for the unexpected results in COMFEN and DesignBuilder.

As for CBE Thermal Comfort Model, the software is developed using a more sophisticated thermal comfort model
that is capable of accessing solar gain impacts. The comfort results generated from Ladybug and Honeybee are PMV
data, and the comfort models are derived from the Java script of the CBE Thermal Comfort Model, besides the PMV
model here use solar-adjusted mean radiant temperature as simulation inputs.

5.2.2 EC Glass Tinting States

The glass library for COMFEN and DesignBuilder is limited for EC glazing, and there might be some ways to add
new glass types into the existing library. This idea was not further developed because the limitation on solar load
62
impacts makes the results not reliable enough to begin with. The existing glass library in Grasshopper and CBE
Thermal Comfort Model also have limitations on EC glass but the developed workflow solved this problem. In
Grasshopper, EC glass with four tinting states was modeled as four independent glass types and then combined
together as inputs. The four tinting states of EC glass were added in the CBE Thermal Comfort Model by editing the
glass library txt file in the software system.

5.2.3 EC Glass Tinting Schedule

No tinting schedule can be applied in COMFEN, while in DesignBuilder tinting schedule can be achieved by giving
a specific set point. The developed workflows in Grasshopper and CBE Thermal Comfort Model follow the tinting
schedule calculated based on interior illuminance level using the lighting control strategy.

5.2.4 Other Evaluation Factors

Other evaluation factors including analysis period, thermal comfort metrics, complexity, and data visualization are
also compared. The thermal comfort analysis period in COMFEN is for the entire year and can be further break down
to each season. The analysis time period in DesinBuilder is flexible and can be set to any time period. With the
developed workflow in Grasshopper, simulations have to be done hour by hour and from CBE Thermal Comfort
Model results are calculated minutely.

Workflow complexity is divided into three aspects: if the tool can import model form other software, the allowable
building geometry, and the time needed to be familiar with the workflow. Among the four software programs, only
DesignBuilder allows users to import model from other software and the others do not. The building geometries are
limited to rectangular shape in COMFEN and CBE Thermal Comfort Model, while DesignBuilder and Grasshopper
give users more freedom in the building form. The time to learn the software is just an approximately time needed to
be very familiar with the workflow, but it might vary depends on users’ ability. Overall the workflows in Grasshopper
and CBE Thermal Comfort Model take more time the workflows in COMFEN and DesignBuilder. The data generated
from COMFEN and DesignBuilder are just a series of numbers, while through Grasshopper and CBE Thermal
Comfort Model the data can be visualized on the human body or the room which makes the results more intuitive and
readable (Table 20).

Software EC Consider Tinting Analysis Result Complexity Visualize

Glass Solar Schedule Period Metrics Data
Type Load Import Geometry Time to Learn
Model
COMFEN One No No Annually, Percent No Rectangle 7 Days No
State Quarterly People
Satisfied
DesignBuilder Two No Yes Any time PMV Yes Any 10 Days No
States period Shape
Grasshopper Four Yes Yes Hourly Solar No Any 20 Days (basic) Yes
(Ladybug and States Adjusted Shape + 5 Days
Honeybee) PMV (development)
CBE Thermal Four Yes Yes Minutely Overall No Rectangle 15 Days (basic) Yes
Comfort States Sensation, + 5 Days
Model Overall (development)
Comfort
Table 20 Workflow Comparison

As COMFEN and DesignBuilder have limitations described above, these two workflows are considered not as
efficient as CBE Thermal Comfort Model and Grasshopper (with Ladybug and Honeybee plugins). Workflow in
Grasshopper solved the three main problems, however it involves more data processing, for example extracting and
reorganizing data based on tinting schedule. The workflow in CBE Thermal Comfort Model has successfully added
the four tinting sates of EC glass in the glazing library, but this workflow includes lots of manual data entry and needs
outside help from Grasshopper to get the solar load values and has to follow the glass tinting schedule obtained from
Grasshopper. Without those inputs from Grasshopper, user can only get momentary results.

63
5.3 Study Limitations
As the results analysis for each software is based on the comparison of the test room and reference room, and the only
changing parameter for the two rooms is glass type, the results comparison is reliable for each software. But because
of some potential issues, for example, input inconsistency between each software, the results comparison between
different software is not applicable.

 Input inconsistency
The inputs that influencing the thermal comfort results in each software is not exactly same but similar, because the
parameter settings and simulation engines are very different in each tool. For example, not all software (COMFEN,
DesignBuilder) have the exact same EC glass or low-e glass performance properties as given in the case studies, so
the only way to simulate the impacts of EC glazing in those software is to use whatever the software can offer for the
EC glass or use alternative glass that have similar glass properties.

 Insufficient Simulation
The simulations in Grasshopper and CBE Thermal Comfort Model were processed for only two days, which is not
sufficient enough to get accurate results. Better results predicting the thermal comfort impacts of EC glazing can be
achieved and random errors can be reduced if more simulations are conducted. In addition, more simulation means
longer period results are available, which would be more convincing for thermal comfort study.

 Neglected parameters
Some parameters including the electric lighting, construction materials were not entirely taken into account. The
neglected parameters were taken care of by using default settings or based on climate zone which might affect the
results indirectly.

5.4 Summary
Results from the test room with EC glass and reference room using low-e glass were analyzed, and the unexpected
data was explained. The four workflows were compared following the evaluation methodology described in Chapter
3.5. Following the methodology, EC glass tinting states, solar load impacts and EC glass tinting schedule along with
some other considering factors were discussed. Study limitations including inconsistent inputs, insufficient
simulations, and neglected parameters were also clarified.

64
 Chapter 6: Conclusions and Future Work
Previous research had been done about energy savings and visual comfort in buildings with EC glass installed.
However, thermal comfort studies for EC glazing is often neglected from the research. The thermal comfort models
are complicated especially when windows allow a certain amount of light into space in which case solar load impacts
need to be taken into account. The dynamically changing features of EC glass along with the sophisticated tinting
schedule make the software simulations even more challenging. Workflows have been developed using different tools
including COMFEN, DesignBuilder, CBE Thermal Comfort Model and Grasshopper. Limitations for each workflow
are clarified and Workflows in CBE Thermal Comfort Model and Grasshopper seems to indicating improved
reliabilities to some extent which can be very helpful for manufacturers, designers, builder owners who want to see
how EC glass can improve the indoor thermal environment, and which are good for EC glass market expansion in
future.

6.1 Result Reliabilities

The result analysis of DesignBuilder, Grasshopper and CBE Thermal Comfort Model are focused on winter solstice
day and summer solstice day, so the conclusions and evaluations are only applicable for these two days. The limited
data sets make it difficult to draw conclusions for the developed workflows. However, studying the most extreme two
days in a year is representative and can be helpful in suggesting trends.

It is easier to get monthly or yearly results in DesignBuilder only by changing the analysis period before run the
simulation, while the limitations of this workflow make it not worthwhile to do more simulations. The workflows in
Grasshopper and CBE Thermal Comfort Model are more reliable but more complicated. Approximately, it would take
one day to calculate results for another day in Grasshopper or CBE Thermal Comfort Model because both the glass
tinting schedule and thermal comfort are run hour by hour which involves lots of data entry and is time consuming.
Getting results for more days is still enough to be able to draw better conclusions and they would likely reveal no
significant difference from the results of two days. If one can spend more time and run simulations for a month or one
year, it would be more convincing but will involve a lot of data processing.
As for accuracy

6.2 Workflow Developed

Figure 54 Workflow Overview

The workflows in COMFEN and DesignBuilder followed what the existing software could do; this turned out to be
limited for EC glazing thermal comfort study and are less likely for users to improve the software capabilities. Thermal
65
comfort results from these two workflows did not consider solar load into the MRT, and the EC glass types are not
sufficient, so the workflows based on these two software are not recommended. Using the current software
Grasshopper and CBE Thermal Comfort Model also have limitations, but after overcoming the difficulties by
improving the simulation system and process, the results getting from these two workflows are more reliable (Figure
54).

At first, CBE Thermal Comfort Model was considered not able to do thermal comfort simulations for EC glass because
there is no EC glass available in the glazing library. However, this problem was solved by adding the EC glass
properties for all four tinting states in the software system. Grasshopper has switchable glazing as window shade, but
the existing shade generator can only accept one shading material and one shade set point over the window.
Simulations could have been done for any time period if only two EC glass tinting states need to be considered. In
order to take four tinting states into account, the EC glass was modeled as four glass types in Grasshopper. The
shortcoming of doing this is that the users have to simulate the thermal comfort hour by hour which increases the
simulation workload. The advantage of CBE Thermal Comfort Model and Grasshopper is that the comfort model
metrics inside have considered solar load impacts.

Overall, among the developed workflows, COMFEN is relatively easier, however the result accuracy is not guaranteed.
DesignBuilder is the only software that can accept a model from other software and it can be very useful for thermal
comfort studies if no solar load needs to be considered. CBE Thermal Comfort Model and Grasshopper did a relatively
good job and results getting from these two are relatively more reliable. From comparing the characteristics of the
software, one can select the most appropriate tool based on project needs.

6.3 Future Work

The workflows developed in Grasshopper and CBE Thermal Comfort Model have provided some methods to evaluate
the thermal comfort for EC glass, however, it is hard to draw conclusions due to the limited data. The assumptions for
the research such as using lighting control strategy and sensor locations need to be verified. More simulations need to
be done to validate the efficiency and accuracy of the proposed workflow. Control algorithms can be evaluated through
the uncomfortable hours’ data. Physical model tests and human subject tests can be carried on to further test the EC
product and validate the efficiency of proposed workflows by comparing the ultimate actual data with the simulated
data. More workflows using other software and dynamically simulation method could also be done in future work.

 Checking assumptions
In this study, EC glass was assumed to be controlled by daylighting, tinting in response to the light sensors to maintain
a comfortable level of daylight, however, this control strategy is not necessarily related to thermal comfort. To get the
maximum thermal comfort, more study need to be done to find the what control strategy would be better for thermal
comfort and find better set points and thresholds that determines what state the glass should be. Another issue is that
assuming using the daylight control strategy, instead of putting the sensors in the interior space, it is better that the
lighting seniors are mounted on the exterior surface of the glass, because the solar light might be filtered out through
the window even if the glass is in clear state. In future work, the glass tinting schedule should be based on exterior
sensor data.

 More simulations
For the proposed workflow in CBE Thermal Comfort Model and Grasshopper, results are simulated only on winter
solstice day and summer solstice day. The conclusions draw on chapters 5 are all based these two days and might not
be applicable at a different day or with different weather conditions. Next step could start to perform thermal comfort
simulation on equinox days, and then get monthly or even yearly data. More simulations would be very helpful to
have a greater data set and conclusions will be more convincing.

 Control algorithms evaluation

As the results from CBE Thermal Comfort Model and Grasshopper discussed in chapter 5, the uncomfortable hours
in the test room with EC glass can be used to improve how the glass should be tinted at different time of the day. On
the other hand, the EC glass was controlled based on interior illuminance level in the proposed workflow of CBE
Thermal Comfort Model and Grasshopper. However, this control method makes more sense from lighting perspective,
66
not from thermal comfort perspective. Using light sensors with fixed thresholds throughout the year to control thermal
comfort is problematic. A control strategy will work better to maximize thermal comfort if it is sensitive to whether it
is heating or cooling season, or whether the building is in heating or cooling mode or whether the occupant in the
perimeter zone is hot, etc. Future work could be done to work out more effective control algorithms using the proposed
workflows in CBE Thermal Comfort Model and Grasshopper.

 Physical model tests and human subject tests

A full scale EC glass unit has already been assembled with two pieces of EC glass. There are some interesting tests
can be done in future. For example, test the illuminance level both outside and inside of the glass so that one can see
how much light is reflected and absorbed by the glass in different state. Another example is to test how long it will
take for the EC glass transit from one state to another and see if the window tint is uniform during the switching
process. On the other hand, the glass unit can be installed in a chamber facing due south for further human subject
test. For Example, validating the numbers used in the software program using the built chamber. The accuracy of the
proposed workflows can also be evaluated by comparing the software predicted results with the human subject test
results.

 More possible workflow development

In the research, only four software are tested and developed, however, there are other software available for simulating
EC glazing. For example, IES-VE, a software developed for architects and engineers can be used for create and capture
performance data during design, commissioning and operation (IES-VE, 2017). In this software, it has electrochromic
glazing construction option included in the latest version and it is possible to modulate gradually between the fully
clear and dark states. More in-depth study could be done to work out a new workflow for EC glazing thermal comfort
study.

 Dynamically simulation method

In CBE Thermal Comfort Model, one has to generate thermal comfort data hour by hour, and in Grasshopper involves
extracting data for each day based on tinting schedule. The data set getting from the proposed workflow in CBE
Thermal Comfort Model and Grasshopper are still not efficient enough, especially if one wants to get a monthly or
yearly thermal comfort result. If user customized tinting schedule can be embedded in the software, it could solve this
problem. On the other hand, the method in Grasshopper is problematic because some parameters including surface
temperature, humidity are considered as consistent inputs for each simulation, but those parameters might change for
the next time step especially when the sun hit into the space and change the surface or air temperature, so a more
dynamically simulation method is needed to solve this problem.

6.4 Summary
Four workflows using software COMFEN, DesignBuilder, Grasshopper, CBE Thermal Comfort Model were studied.
And only workflows based on Grasshopper and CBE Thermal Comfort Model were improved and developed for
further use. Some surprising results emerged in the process of evaluating the workflows because of the limitations of
the software. To test the effectiveness of the improved workflow, case studies were conducted for both the test room
with EC glass and reference room with typical low-e glass. Results have shown that the thermal comfort performance
of EC glass is better than low-e glass in almost all the time. Because the process of improved workflows in
Grasshopper and CBE Thermal Comfort Model are time-consuming, only two days’ data (winter solstice day and
summer solstice day) were collected, so the results reliabilities are clarified. Despite the deficiency in the amount of
data sets, the case studies support the hypothesis well and proved that simulation workflows in CBE Thermal Comfort
Model and Grasshopper (with Ladybug and Honeybee plugins) which take multiple main factors (e.g. four tinting
states, solar load impacts, glass tinting schedule) into consideration seems to indicating improved reliabilities. The
proposed workflows make up the vacancies in EC glass thermal comfort study and are helpful for manufactures and
designers to better evaluate the most recent EC glass product thermal impacts. Future work can be done to further
testify and improve the potential application of the proposed workflow.

67
 Reference

Arens, Edward, Tyler Hoyt, Xin Zhou, Li Huang, Hui Zhang, and Stefano Schiavon. "Modeling the comfort effects
of short-wave solar radiation indoors." Building and Environment 88 (2015): 3-9.

Arens, Edward A., Richard Gonzalez, and Larry Berglund. "Thermal comfort under an extended range of
environmental conditions." ASHRAE Transactions 92 (1986).

Ardakan, Ahoo Malekafzali. A Multi-Zone Electrochromic Window System Integrated with Light Shelf. North
Carolina State University, 2015.

ASHRAE Standard 55-2004, Accessed Dec 23 2016, https://www.ashrae.org/standards-research--

technology/standards--guidelines

Burdis, Mark, and Neil Sbar. Electrochromic Windows: Process and Fabrication Improvements for Lower Total Costs.
Sage Electrochromics, Incorporated, 2007.

CBE Thermal Comfort Tool for ASHRAE-55. "CBE Thermal Comfort Tool for ASHRAE-55." Accessed September
01, 2016. http://comfort.cbe.berkeley.edu/.

Chris Woodford. "How Do Electrochromic (smart Glass) Windows Work?" Accessed February 01, 2017.
http://www.explainthatstuff.com/electrochromic-windows.html.

Clear, R. D., Vorapat Inkarojrit, and E. S. Lee. "Subject responses to electrochromic windows." Energy and
Buildings 38, no. 7 (2006): 758-779.

Deb, S. K. "A novel electrophotographic system." Applied Optics 8, no. 101 (1969): 192-195.

Department of Energy (DOE), Building Performance Database, Accessed Sep 18, 2016
https://energy.gov/eere/buildings/building-performance-database

Fanger, Poul O. "Thermal comfort. Analysis and applications in environmental engineering." Thermal comfort.
Analysis and applications in environmental engineering. (1970).

Haglund, Kerry L. "Decision-making Methodology & Selection Tools for High-performance Window Systems in US
Climates." BEST2, Strategic Issues in Building Design, WB13e4 (2010).

Hoffmann, Sabine, Christoph Jedek, and Edward Arens. "Assessing thermal comfort near glass facades with new
tools." In BEST 3 Building Enclosure Science and Technology Conference. 2012.

Markets and markets. “Electrochromic Glass Market worth $2.59 Billion by 2020." Accessed September 01, 2016.
http://www.marketsandmarkets.com/PressReleases/electrochromic-glass.asp.

Explain That Stuff. "How Do Electrochromic (smart Glass) Windows Work?" Accessed September 01, 2016.
http://www.explainthatstuff.com/electrochromic-windows.html

"Ladybug Tools." Food4Rhino. March 19, 2017. Accessed April 30, 2017. http://www.food4rhino.com/app/ladybug-
tools.

Lampert, Carl M. "Large-area smart glass and integrated photovoltaics." Solar Energy Materials and Solar
Cells 76, no. 4 (2003): 489-499.

LBNL. “Advancement of Electrochromic Windows” Accessed March 20, 2017. http://windows.lbl.gov

68
Lee, Eleanor S, Dennis L. DiBartolomeo, and Stephen E. Selkowitz. "Electrochromic windows for commercial
buildings: Monitored results from a full-scale testbed." Lawrence Berkeley National Laboratory (2000).

Lee, Eleanor, Mehry Yazdanian, and Stephen Selkowitz. "The energy-savings potential of electrochromic windows
in the US commercial buildings sector." Lawrence Berkeley National Laboratory (2004).

Lee, Eleanor S., Dennis L. DiBartolomeo, Joseph Klems, Mehry Yazdanian, and Stephen E. Selkowitz. "Monitored
energy performance of electrochromic windows controlled for daylight and visual comfort." Lawrence Berkeley
National Laboratory (2005).

Lee, Eleanor S., Stephen E. Selkowitz, Robert D. Clear, Dennis L. DiBartolomeo, Joseph H. Klems, Luis L. Fernandes,
Greg J. Ward, Vorkapat Inkarojrit, and Mehry Yazdanian. "Advancement of electrochromic windows." Lawrence
Berkeley National Laboratory (2006).

Lee, Eleanor S., Stephen E. Selkowitz, Robert D. Clear, Dennis L. DiBartolomeo, Joseph H. Klems, Luis L. Fernandes,
Greg J. Ward, Vorapat Inkarojrit, and Mehry Yazdanian. "A design guide for early-market electrochromic
windows." Lawrence Berkeley National Laboratory (2006).

Lee, Eleanor S., Erin S. Claybaugh, and Marc LaFrance. "End user impacts of automated electrochromic windows in
a pilot retrofit application." Energy and Buildings 47 (2012): 267-284.

Lee, Eleanor S., Luis L. Fernandes, Brian Coffey, Andrew McNeil, Robert Clear, Tom Webster, Fred Bauman, D.
Dickeroff, David Heinzerling, and Tyler Hoyt. "A post-occupancy monitored evaluation of the dimmable lighting,
automated shading, and underfloor air distribution system in The New York Times Building." Berkeley National
Laboratory (2013).

LBNL Window & Daylighting Software. "LBNL Window & Daylighting Software -- COMFEN." Accessed
November 10, 2016. https://windows.lbl.gov/software/comfen/comfen.html

Lighting Research Center. “Lighting Answers” Accessed April 15, 2017.

http://www.lrc.rpi.edu/programs/nlpip/lightinganswers/lightpollution/glare.asp

Lyons, P. R., Dariush Arasteh, and Charlie Huizenga. "Window performance for human thermal
comfort." Transactions-American Society of Heating Refrigerating and Air Conditioning Engineers 106, no. 1 (2000):
594-604.

Mackey, Chris Christopher William. "Pan climatic humans: shaping thermal habits in an unconditioned society." PhD
diss., Massachusetts Institute of Technology, 2015.

Munshi, Kavish Prakash. Analysis of Life Cycle Costs and Energy Savings of Electrochromic Glazing for an Office
Building. Arizona State University, 2012.

National Optical Astronomy Observatory (NOAO). “Recommended Light Levels” Accessed March 14, 2017
https://www.noao.edu/education/QLTkit/ACTIVITY_Documents/Safety/LightLevels_outdoor+indoor.pdf

Pease Brad, Paladino, Narayan Thulasi.. Performance Assessment of Sage Glass Electrochromic Coatings and
Control Scenadrios. PDF. Seattle: Paladino and Company, June 2010.

Renewable Energy and Energy Efficient Partnership, “Worldwatch Report 182,”2010. "PMV-PPD." PMV-PPD.
Accessed September 01, 2016. http://www.eat.lth.se/fileadmin/eat/Termisk_miljoe/PMV-PPD.html

Sage Glass. “Frame Cable Cut Sheet” Accessed February 10, 2017
https://www.sageglass.com/sites/default/files/scs-004_sageglass_frame_cable_cut_sheet.pdf
69
Sage Glass. “Gain LEED v4 points with Sage Glass” Accessed March 14, 2017
https://www.sageglass.com/sites/default/files/leed_brochure_us.pdf

Sage Glass. “Improving the Human Experience in the Build Environment” Accessed September 15, 2016
https://www.sageglass.com/sites/default/files/architectbrochure_mkt_23_heatlight.pdf

Sage Glass. “Performance and Acoustic Data” Accessed October 21, 2017
https://www.sageglass.com/sites/default/files/mkt-043.4_performance_and_acoustical_data_flyer.pdf

Sage Glass. "SageGlass Control System." Accessed November 10, 2016.

https://www.sageglass.com/sites/default/files/mkt-061_sageglass_controls.pdf.

Sage Glass. “Sage Glass Product Guide” Accessed April 1, 2017

https://www.sageglass.com/sites/default/files/productguide_mkt_48.pdf

Sage Glass. “Occupant Comfort” Accessed January 24, 2017. https://www.sageglass.com/en/benefits/occupant-

comfort

Selkowitz, Stephen E. "Thermal performance of insulating window systems."Lawrence Berkeley National

Laboratory (2011).

Selkowitz, Stephen. "COMFEN 3.0-Evolution of an Early Design Tool for Commercial Facades and Fenestration
Systems." In Building Enclosure Sustainability Symposium (BESS) 2011, Pomona, CA, 4/29/11-4/30/11. 2012.

Standard 55-2004: Thermal Environmental Conditions for Human Occupancy. The American Society of Heating,
Refrigeration and Air-Conditioning Engineers (ASHRAE).

Sullivan, R., E. S. Lee, K. Papamichael, Michael Rubin, and Steven E. Selkowitz. "Effect of switching control
strategies on the energy performance of electrochromic windows." In Optical Materials Technology for Energy
Efficiency and Solar Energy Conversion XIII, pp. 443-455. International Society for Optics and Photonics, 1994.

Sustainability Workshop. "Human Thermal Comfort” Accessed November 10, 2016.

http://sustainabilityworkshop.autodesk.com/buildings/human-thermal-comfort.

Webb, Amanda Laurel. "Mapping comfort: an analysis method for understanding diversity in the thermal
environment." PhD diss., Massachusetts Institute of Technology, 2012.

White, Andrew, and Michael Holmes. "Advanced simulation applications using room." In Eleventh international
IBPSA conference, pp. 27-30. 2009.

Schiavon, Stefano, Tyler Hoyt, and Alberto Piccioli. "Web application for thermal comfort visualization and
calculation according to ASHRAE Standard 55." In Building Simulation, vol. 7, no. 4. 2014.

U.S. Department of Energy, “Energy Efficiency Trends in Residential and Commercial Buildings,” 2008

View Dynamic Glass. “Energy benefits of View Dynamic Glass in workplaces” Accessed November 1, 2016
https://viewglass.com/assets/pdfs/workplace-white-paper.pdf

Zhang, Hui. "Human thermal sensation and comfort in transient and non-uniform thermal environments." Center for
the built environment (2003).

Zhang, Hui, Charlie Huizenga, Edward Arens, and Danni Wang. "Thermal sensation and comfort in transient non-
uniform thermal environments." European journal of applied physiology 92, no. 6 (2004): 728-733.
70
 Appendix X
X-1: Illuminance Calculation Code in Grasshopper

71
X-2: Thermal Comfort Simulation Code in Grasshopper

72