!
!

IMPACTS OF LIGHTING SCHEDULES ON BUILDING ENERGY SIMULATIONS
By
Federico Steinvorth

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER IN SCIENCE
Construction Management
2012

!
!

!

ABSTRACT
IMPACTS OF LIGHTING SCHEDULES ON BUILDING ENERGY SIMULATIONS
By
Federico Steinvorth
Common errors in building energy simulation include incorrect lighting and plug load
power densities as well as peak occupancies in spaces (Rosenbaum 2003). Inaccurate results due
to these uncertain parameters support industry opinions that energy models present a pattern of
behavior rather than predicting true energy performance (Clevenger 2010). This research
analyzed the impact of lighting schedule definition in energy use prediction and included a
limited survey addressing methods used by professionals. Three case buildings were analyzed:
the Chemistry Building and the Geography Building at the Michigan State University campus,
and a church project in Lansing, Michigan. Actual lighting schedules were collected and
redefined through field observations and data from portable light loggers. These schedules were
then used in energy models created for each building, and predicted energy use was compared to
results from models previously developed during the design phase to default schedules within the
program, and to available building utility data. Results showed an improvement in energy
prediction accuracy when defining schedules based on averaged use patterns. As a result of the
research, recommendations were developed for approaches to improve schedule definition and a
data gathering tool was developed to collect required information for energy analysis during
design.

!
!

!

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................................... v!
LIST OF FIGURES ..................................................................................................................... viii!
INTRODUCTION .......................................................................................................................... 2!
1.1.! Overview!..........................................................................................................................! !
2
1.2.! Need!for!the!Study!...........................................................................................................! !
5
1.3.! Research!Goal!and!Objectives! .........................................................................................! !
.
7
1.4.! Research!Scope!................................................................................................................! !
8
1.5.! Expected!Contributions! ...................................................................................................! !
.
9
1.6.! Limitations!.....................................................................................................................! 0!
1
1.7.! Chapter!Summary!..........................................................................................................! 1!
1
LITERATURE REVIEW ............................................................................................................. 13!
2.1.! History!............................................................................................................................! 3!
1
2.2.! Energy!Estimation!and!Modeling!Methods! ...................................................................! 5!
.
1
2.3.! Simulation!Tools! ............................................................................................................! 9!
.
1
2.3.1.! DOEP2!......................................................................................................................! 9!
1
2.3.2.! HAP!.........................................................................................................................! 0!
2
2.3.3.! eQuest!....................................................................................................................! 2!
2
2.3.4.! EnergyPlus!..............................................................................................................! 2!
2
2.3.5.! BEOpt!......................................................................................................................! 4!
2
2.3.6! COMcheck!...............................................................................................................! 5!
2
2.4.! Validation!and!Calibration!of!Energy!Simulations!..........................................................! 6!
2
2.5.! Program!Comparison!.....................................................................................................! 4!
3
2.6.! Occupancy!Schedule!Testing!..........................................................................................! 7!
3
2.7.! Sources!of!Information!..................................................................................................! 3!
4
2.8.! Summary!........................................................................................................................! 6!
4
METHODOLOGY ....................................................................................................................... 49!
3.1! Literature!Review!...........................................................................................................! 9!
4
3.2! Survey!............................................................................................................................! 9!
4
3.3! Identification!and!Selection!of!Case!Studies!..................................................................! 0!
5
3.4! Extreme!and!Intermediate!Boundaries!..........................................................................! 3!
5
3.5! Methods!for!Each!Case!Study!........................................................................................! 5!
5
3.5.1! Chemistry!Addition!Building!...................................................................................! 5!
5
3.5.2! Geography!Building!................................................................................................! 8!
5
3.5.3! Church!Project!........................................................................................................! 1!
6
3.6! New!Schedules!and!Analysis!..........................................................................................! 2!
6
iii!
!

!

3.7!
3.8!

Guidelines!Development! ...............................................................................................! 2!
.
6
Chapter!Summary!..........................................................................................................! 3!
6

DATA ANALYSIS ....................................................................................................................... 65!
4.1! Interviews!......................................................................................................................! 5!
6
4.2! Lighting!Use!Range!Testing!............................................................................................! 7!
6
4.3! Building!Case!Studies!.....................................................................................................! 2!
8
4.3.1! Chemistry!Building!..................................................................................................! 3!
8
4.3.2! Geography!Building!..............................................................................................! 15!
1
4.3.3! Church!Project!......................................................................................................! 27!
1
4.4! Chapter!Summary!........................................................................................................! 47!
1
GUIDELINES, TOOLS AND RECOMMENDATIONS........................................................... 151!
5.1! Guidelines!and!Recommendations!..............................................................................! 51!
1
5.1.1! Breakdown!of!Spaces!According!to!Use!...............................................................! 51!
1
5.1.2! Grouping!of!Spaces!with!Similar!Use!Characteristics!...........................................! 52!
1
5.1.3! Owner!and/or!Users!Specify!Schedules!and!Room!Use! .......................................! 55!
.
1
5.1.4! Request!Space!Use!Tendency!Information!Available!...........................................! 56!
1
5.1.5! Automatic!Lighting!Control!...................................................................................! 56!
1
5.1.6! Daylight!and!Occupancy!Sensors!..........................................................................! 56!
1
5.1.7! Use!of!Default!Schedules!......................................................................................! 57!
1
5.1.8! Submetering!and!Calibration!................................................................................! 57!
1
5.1.9! Integrated!Design!.................................................................................................! 58!
1
5.2! Data!Gathering!Tool!.....................................................................................................! 58!
1
5.3! Chapter!Summary!........................................................................................................! 68!
1
CONCLUSIONS......................................................................................................................... 170!
6.1! Summary!......................................................................................................................! 70!
1
6.2! Conclusions!..................................................................................................................! 71!
1
6.3! Limitations!...................................................................................................................! 72!
1
6.4! Future!Research!...........................................................................................................! 73!
1
6.5! Expected!Contributions! ...............................................................................................! 74!
.
1
APPENDIX A ............................................................................................................................. 176!
APPENDIX B ............................................................................................................................. 180!
APPENDIX C ............................................................................................................................. 184!
REFERENCES ........................................................................................................................... 193!
iv!
!

!

LIST OF TABLES!

Table 1-1: Sample Lighting Power Densities Table Using the Building Area Method (ASHRAE
2007) ............................................................................................................................................... 4!
Table 1-2: Sample Lighting Power Densities Table Using the Space by Space Method (ASHRAE
2007) ............................................................................................................................................... 5!
Table 2-1: Summary of Some Simulation Programs .................................................................... 26!
Table 2-2: Contrasting Capabilities of Programs (Crawley et al. 2005)....................................... 36!
Table 2-3: Space Utilization Study (Dober, Lidsky, Craig and Associates, 2003) ...................... 44!
Table 2-4: Space Utilization with Target Objectives (University of Oregon 2007) ..................... 45!
Table 2-5: Sample Schedule for Assembly Occupancy (ASHRAE User's Manual 2004) ........... 46!
Table 3-1: Scheduled Visits for Chemistry Building .................................................................... 56!
Table 3-2: Chemistry Observation Chart ...................................................................................... 57!
Table 3-3: Scheduled Visits for Geography Building................................................................... 59!
Table 3-4: Geography Observation Chart ..................................................................................... 60!
Table 4-1: ASHRAE Requirements for Climate Zone 5 (ASHRAE, 2007)................................. 69!
Table 4-2: ASHRAE Definition of HVAC Systems (ASHRAE, 2007) ....................................... 70!
Table 4-3: eQuest Default Schedule for Weekends, Holidays, and Break Periods ...................... 73!
Table 4-4: eQuest Default Schedule for Weekdays ...................................................................... 74!
Table 4-5: Schedules for Test Cases ............................................................................................. 74!
Table 4-6: Monthly Energy Consumption (kWh) Resulting from Lighting Schedule Variations in
a Hypothetical Building ................................................................................................................ 78!
Table 4-7: Room Distribution in Chemistry Addition Building ................................................... 84!
v!
!

!

Table 4-8: Schedule Used in Chemistry Addition Model during Design Phase........................... 90!
Table 4-9: Default Software Schedule for Unoccupied Periods in Academic Buildings ............. 92!
Table 4-10: Default Software Schedule for Occupied Periods in Academic Buildings ............... 94!
Table 4-11: Percentage of Lighting Fixture Use Schedule from Observations by Zone in the
Chemistry Building ....................................................................................................................... 97!
Table 4-12: Annual Electric Consumption Predicted Measured Utility by Original Model in
Chemistry Building ..................................................................................................................... 100!
Table 4-13: Annual Electric Consumption Predicted by Research Model without Break Periods
and Software’s Default Schedules .............................................................................................. 100!
Table 4-14: Modified Original and Research Models to Normalized Conditions ...................... 101!
Table 4-15: Annual Electric Consumption in Chemistry Building by Different Lighting
Schedules and no Break Periods ................................................................................................. 102!
Table 4-16: Annual Electric Consumption in Chemistry Building by Different Lighting
Schedules and Break Periods ...................................................................................................... 103!
Table 4-17: Information from Campus Monitoring by Capacity................................................ 106!
Table 4-18: Load Analysis from Observed Schedules in Chemistry Building ........................... 108!
Table 4-19: Weekly Lighting Use Estimation from Registrar's Classroom Schedule ................ 112!
Table 4-20: Room Distribution in Geography Building ............................................................. 115!
Table 4-21: Lighting Fixture Use Schedule by Zone in Geography Building ............................ 122!
Table 4-22: Annual Electric Consumption Predicted v Measured Utility Data for Geography
Building....................................................................................................................................... 124!
Table 4-23: Annual Electric Consumption Predicted with 100% OA Requirements v Measured
Utility Data for Geography Building .......................................................................................... 125!
Table 4-24: Software Default Schedule for Church Projects...................................................... 133!
vi!
!

!

Table 4-25: Average Values Recorded by Logger in Main Meeting Room at Church Project .. 134!
Table 4-26: Average Values Recorded by Logger in Meeting Room at Church Project ........... 135!
Table 4-27: Average Values Recorded by Logger in Library at Church Project ....................... 136!
Table 4-28: Average Values Recorded by Logger in Fellowship at Church Project.................. 137!
Table 4-29: Building Average Schedule from Measurements for Church Building .................. 138!
Table 4-30: Normalized Comparison of Models ........................................................................ 139!
Table 4-31: Predicted Electric Load by Research Model with Different Lighting Schedules ... 140!
Table 4-32: Predesign Occupancy Schedule Provided by Owner .............................................. 143!
Table 4-33: Schedule Defined from Owner's Predesign Occupancy Statement ......................... 144!
Table 5-1: Estimated Hours of Use and Percentages for Spaces Observed ................................ 155!
Table B-1: Main Façade Wall Layer Chemistry Building .......................................................... 180!
Table B-2: Facebrick Wall Chemistry Building ......................................................................... 180!
Table B-3: Utility Duct Wall Chemistry Building ...................................................................... 181!
Table B-4: Roof Layers Chemistry Building .............................................................................. 181!
Table B-5: Main Wall Assembly Geography Building .............................................................. 182!
Table B-6: Roof Assembly Geography Building........................................................................ 182!
Table B-7: Wall Assembly Church Project ................................................................................ 183!

vii!
!

!

LIST OF FIGURES
Figure 1-1: 2010 Buildings Energy End-Use Splits (U.S. Department of Energy, 2010) .............. 2!
Figure 2-1: DOE-2 Simulation Engine Structure (Hirsch 1994) .................................................. 20!
Figure 2-2: EnergyPlus Internal Elements (University of Illinois 2009) “For interpretation of the
references to color in this and all other figures, the reader is referred to the electronic version of
this thesis” ..................................................................................................................................... 23!
Figure 2-3: BEOpt Internal Structure (Christensen et all 2006) ................................................... 25!
Figure 2-4: Simulation Error Sources (Judkoff et al. 2008) ......................................................... 29!
Figure 2-5: Measured Energy v. Simulated Energy (Wasilowsky and Reinhardt 2009).............. 39!
Figure 2-6: Sample Graph for Diversity Measurements in Buildings (Abushakra et al. 2001) ... 40!
Figure 2-7: Measured Impacts of Different Performance Values of Building Characteristics
(Heller, Heater, and Frankel, 2011) .............................................................................................. 42!
Figure 3-1: Methodology of Research .......................................................................................... 51!
Figure 3-2: Change in Heating and Cooling Loads Caused by a 1kWh decline in lighting loads in
new construction buildings (Sezgen and Koomey 1998) ............................................................. 54!
Figure 4-1: Schematic VAV with Reheat Coil System (NREL, 2006) ........................................ 72!
Figure 4-2: Monthly Electrical Consumption (kWh) of Test Building ........................................ 76!
Figure 4-3: Annual Energy Consumption of Test Building.......................................................... 80!
4-4: Peak Energy Consumption Consumption of Test Building .................................................. 81!
Figure 4-5: Chemistry Building .................................................................................................... 85!
Figure 4-6: Typical Floor Plan of Chemistry Addition Building ................................................. 85!
Figure 4-7: Chemistry Building Model Screenshot ...................................................................... 86!
viii!
!

!

Figure 4-8: Schematic VAV with Reheat Coil System (NREL, 2006) ........................................ 88!
Figure 4-9: Building Average Lighting Profile v Default Lighting Schedule Profiles in Chemistry
Building......................................................................................................................................... 98!
Figure 4-10: Comparison of Annual Electric Consumption in Chemistry Building .................. 104!
Figure 4-11: Monthly Electric Consumption by Different Models and Lighting Schedules in
Chemistry Building ..................................................................................................................... 104!
Figure 4-12: Classroom Schedule in April from Registrar’s Office ........................................... 110!
Figure 4-13: Hourly Lighting Use Percentages Obtained Through Several Data Collection
Methods in Basement Classroom at Chemistry Building ........................................................... 114!
Figure 4-14: Geography Building ............................................................................................... 116!
Figure 4-15: Typical Floor Plan of Geography Building............................................................ 116!
Figure 4-16: Geography Building Model Screenshot ................................................................. 117!
Figure 4-17: Schematic Design of Constant Volume AHU with Reheat (NREL, 2006) ........... 118!
Figure 4-18: Geography Average Lighting Profile from Observations v Software Default
Schedule ...................................................................................................................................... 123!
Figure 4-19: Comparison of Annual Electric Consumption in Geography Building ................. 126!
Figure 4-20: Monthly Electric Consumption by Lighting Schedules in Geography Building ... 127!
Figure 4-21: Church Project Building......................................................................................... 128!
Figure 4-22: Floor Plan for Church Project ................................................................................ 128!
Figure 4-23: Church Project Building Model Screenshot........................................................... 129!
Figure 4-24: Schematic Design for Furnace System (NREL, 2006) .......................................... 130!
Figure 4-25: Monthly Consumption by Different Simulation Runs ........................................... 141!
Figure 4-26: Annual Electric Consumption Predicted by Each Model ...................................... 142!
ix!
!

!

Figure 4-27: Average Daily Lighting Use Profiles for Various Methods in Main Meeting Room
..................................................................................................................................................... 146!
Figure 5-1: Architectural Information Summary Chart .............................................................. 160!
Figure 5-2: Space Summary and Design Values Summary Chart .............................................. 164!
Figure 5-3: Schedule Summary Chart......................................................................................... 165!
Figure 5-4: Systems Description Summary ................................................................................ 166!
Figure 5-5: Miscellaneous Equipment Description Summary Chart .......................................... 167!

x!
!

!

CHAPTER 1
INTRODUCTION

1!
!

!

INTRODUCTION
1.1. Overview
Buildings account for 39% of U.S. energy use (U.S. Department of Energy 2010) which is
more than any other industry. Local energy codes and sustainable certification processes seek to
reduce energy consumption through design specification of materials and equipment efficiencies.
As shown in Figure 1.1, HVAC equipment and lighting account for almost 50% of energy use in
buildings. Focusing on improving equipment efficiency and controllability of lighting fixtures is
essential in minimizing energy consumption.

Figure 1-1: 2010 Buildings Energy End-Use Splits (U.S. Department of Energy, 2010)
In order to find optimal solutions for energy performance, many decisions are made during
the design phase of a project, and simulation tools can help in evaluating energy impacts of
design choices. Analysis of temperature, envelope losses, system performance, and electric loads
2!
!

!

establishes energy flow within a structure and its behavior under certain climatic conditions
(Judkoff 2008). Computer simulation software is helpful in calculating the impact of all these
parameters in overall energy consumption.
Predicting energy use can become challenging when lacking appropriate information. Energy
modeling accuracy can range from +/- 10% to +/- 40% in residential models. However, industry
experts indicate relative prediction of performance is more useful than absolute values
themselves (Clevenger 2010). Results from modeling programs may not be able to accurately
predict actual building energy consumption, but they can show important patterns of behavior of
design options.
The need for validation and verification of software is essential. Inputs involve many
assumptions and require the modeler to define human behavior in facilities. Decisions made
during simulation impact results, and distance them from actual conditions. Understanding the
influence of parameters that cause inaccuracies can be helpful in matching outputs closer to
actual performance. This research focuses on analyzing the impact of lighting schedules.
Schedule definition relies on the simulator’s own judgment and information available at the time
of design or, use of program default values, and is therefore an important source of inaccuracy in
final results.
Codes and standards can help standardize assumptions in energy use prediction. For example
ASHRAE Standard 90.1 Chapter 9 lighting requirements define the paths for compliance and
specify minimum requirements, mandatory provisions and either of two prescriptive options: the
building area method or space by space method. Mandatory provisions state that all interior
2

lighting shall be controlled by an automatic system for any building larger than 5000 ft . This
3!
!

!

can be accomplished by installing a system that can be programmed to turn on and off lights at
desired times of day or use of occupancy sensors to control lighting.
Both alternative compliance methods specify maximum densities of lighting. The building
area method establishes a maximum lighting density per square foot for a specific type of
building, and installed lighting power shouldn’t exceed this density. The space by space method
specifies a lighting density as well, but for particular spaces in a building, so the sum of all
spaces establishes the maximum lighting power. Figures 1.2 and 1.3 show examples of allowable
lighting power densities:

Table 1-1: Sample Lighting Power Densities Table Using the Building Area Method
(ASHRAE 2007)

4!
!

!

Table 1-2: Sample Lighting Power Densities Table Using the Space by Space Method
(ASHRAE 2007)
Mandatory provisions can impact lighting scheduling definition, since central lighting control
systems can be implemented during the design phase and the programmed schedule can be
defined as the lighting schedule in energy simulation. Default use percentages in simulation
programs consider that lights won’t be used 24 hours per day. Code prescriptive paths can help
define lighting wattage if an electric design is not yet complete for preliminary energy
simulation.
As newer editions of codes are published, energy requirements become more stringent.
Higher standards may translate into higher initial investment. Understanding lighting schedules
becomes important in fine tuning energy simulations and can help to establish better approaches
for predicting consumption and in life cycle cost analysis.
1.2. Need for the Study
According to Judkoff and Neymark (2006), “Extensively reducing the energy intensity of
buildings through better design is possible with the use of simulation tools. However, widespread
5!
!

!

use of building energy simulation software will not occur unless the design and engineering
communities have confidence in these programs.” Judkoff and Neymark indicate the need for
evaluation and improvement of energy simulation tools, especially when sustainable design and
green certification rely on computer simulations to estimate energy savings.
There are three areas where research must be focused to improve simulation tools: 1) testing
and comparing predicted results with measurements from full scale buildings and data from
rooms with controlled parameters, 2) verification of program results through mathematical
calculations and direct comparison of available software programming codes, and 3) the
development of validation methodologies. Previous studies have not reached reasonable
conclusions on errors induced by different processes or on a range of errors where energy
simulations can be considered valid or real. Most of the work completed so far has focused on
the impact of building envelope elements, but internal loads need to be evaluated and tested as
well (Judkoff et al. 2008).
Simulation program validation can be oriented toward two goals. The first is to establish
whether the software is reliable for predicting true energy when compared to field measurements
and the second is to find sources of error which lead to inaccuracies in the results. Most efforts
have been invested in the first goal, while the path towards program prediction improvement
relies on expanding research and has less attention (Bloomfield et al. 1995). Identifying error
sources in thermal analysis can contribute to the accuracy of energy predictions.
Assumptions and shortcuts in calculation methods of simulation programs are the main
sources of error in energy prediction. Other contributing errors are variations in weather data,
building envelope component performance and schedule definition. To improve internal load
6!
!

!

approximations in simulation, influences of external sources such as environmental and building
envelope should be considered and reduced. Therefore, measurements and experiments on
specific parameters must be done carefully to obtain accurate values for these inputs (Judkoff et
al. 2008).
Analyzing all parameters involved in an energy simulation is a difficult task because much
time, equipment, and effort are needed. Consequently this thesis research focuses on lighting,
which has a significant impact on energy use. Figure 1.1 showed the impact on overall energy
use in buildings. A report by the IEA (2006) recognized lighting as an opportunity to improve
energy efficiency in buildings indicating up to a 70% savings could be achieved with new design
strategies. Energy simulation can help to find optimal solutions. Having a better understanding
of the impact of each parameter can lead to improved accuracy. This study is targeted to
establishing a better understanding of lighting use and its impact on energy use prediction.
The need for further research is recommended in a University of Massachusetts study that
compares energy use predicted by modeling with measured post-occupancy consumption in
LEED certified schools (Lowell 2007). Although it was found that energy codes were
outperformed, predicted energy use was as high as 40% more than measured consumption. The
research team concluded that one of the main reasons for the difference was the energy
modeler’s inability to predict the building’s plug loads, occupant behavior, occupancy levels, and
operating hours.
1.3. Research Goal and Objectives
The primary goal of this exploratory research is to obtain a deeper understanding of lighting
use schedules in energy use simulation and their impact on overall predicted results. The
7!
!

!

researcher established three activities to carry out the study. Professional energy modelers were
interviewed to obtain an insight into simulation techniques, their views on software ability to
predict energy use and how to handle uncertainty in schedule definition. Then, a sample building
was modeled and different percentages of lighting use were assigned to establish boundary
ranges of predicted energy values. The cases modeled ranged from 0% light use annually to
100% light use annually at 24 hours per day. Finally, three case buildings were selected for
modeling in order to predict their energy use, and were compared with actual utility
measurements. Buildings selected included the Chemistry and Geography buildings at Michigan
State University and a Church Project in Lansing, Michigan.
The following objectives were established for the research:
1. Develop methods to gather lighting use data from actual buildings
2. Inquire about methods used in energy simulation addressing schedule accuracy
through telephone interviews of energy modelers
3. Explore the influence of lighting schedules in energy simulations and determine
methods to improve the accuracy of schedule assumptions during modeling
4. Develop recommendations for future schedule definition processes
1.4. Research Scope
The aim of this study is to better understand the impact of lighting schedule on energy
simulation. Understanding errors that may occur will help the energy modeler to determine if it is
worth investing time in this task or allocating efforts to others. In the future, if all efforts are

8!
!

!

combined, a code or standard can be published which defines and outlines modeling procedures
for parameters that can reduce uncertainty.
Occupants impact a number of systems including lighting use, fan cycles, water heating
requirements, thermostat settings, plug loads, and miscellaneous equipment use. Although a
deeper understanding of energy simulation requires consideration of all these parameters, the
best approach is to isolate one parameter at a time. Plug loads, fan cycles, lighting, and water
heating in total represent more than 50% of energy consumed (Figure 1.1), but it is beyond the
scope of this study to determine the effects of each individual parameter. Consequently, this
study focuses on lighting.
This research was exploratory and conclusions were not expected to have a valid statistical
significance, but it is expected to contribute to a growing effort to analyze distinct energy
parameters and leads to a better understanding of the subtleties of energy simulation. It is also
beyond the scope of this study to test internal calculation methods of software or discover “bugs”
to be fixed within software. The intention is to explore how to approach models realistically and
not limitations in software.
1.5. Expected Contributions
The fundamental purpose of this research is to determine if errors induced by occupancy
schedule parameters are significant enough to recommend that energy simulators invest time in
developing the information to define this parameter, or if the difference between one defined
schedule and another won’t substantially influence the final result. This parameter is usually
defined without exact knowledge of occupant behavior and different schedules must be defined
for different rooms. Knowing that predefined schedules by software companies would be
9!
!

!

accurate enough would allow the modeler to concentrate more on other parameters such as
HVAC equipment definition. This research may also encourage further research on the subject
and lead to a proposal for a standardization procedure for schedule definition in codes addressing
energy simulation.
Results may help clarify any general misunderstandings about energy simulations among
professionals. If most people in the simulation field define occupancy based on their own
judgment, they might feel comfortable knowing that their colleagues also carry the same
uncertainty of defining schedules without a standardized method.
1.6. Limitations
Certain limitations are embedded in this research. The first limitation is that field
observations are limited to short durations. Energy simulations are based on a 365-day and 24hour analysis. For a completely realistic schedule definition, 24-hour observations during every
day of the year (perhaps over several years) should be conducted, which is not feasible unless
electronic sensors and metering devices are installed to keep track of energy use and occupancy.
Also, some types of building such as educational facilities will have seasonal variations in
schedule operation, which makes it even harder to determine occupancy schedule.
Since human behavior is unpredictable and the number of people in a structure is dynamic, it
is hard to predict what actual consumption and energy conservation habits will be. The
importance of this study lies in our awareness of how these inaccuracies affect energy
simulations, and in determining how to account for these inaccuracies.

10!
!

!

1.7. Chapter Summary
This chapter establishes the main energy uses in buildings and describes the need to improve
energy simulations as a design tool. It outlines the objectives, research scope and limitations of
the research. Chapter 2 presents the literature review including a historical background of
simulation programs along with an overview of modeling methods, simulation programs, and
efforts to date in energy modeling research.

11!
!

!

CHAPTER 2
LITERATURE REVIEW

12!
!

!

LITERATURE REVIEW
This chapter outlines the literature review conducted to develop a background understanding
on energy simulation and software. First, historical events that led to energy conservation
policies are reviewed, followed by an explanation of mathematical methods used in simulation
program calculation as well as the detailing of several available programs. Finally, several
studies regarding energy simulation are presented in three categories: overall calibration and
validation of programs, testing between various available software, and studies regarding
occupancy testing.
2.1. History
The use of computer programs for energy modeling purposes is not new for designers.
Although programs in the late 1960s were not flexible in terms of input information, they still
allowed the calculation of heating, ventilation and air conditioning loads. Initially, envelope
loads were not considered and no alternative window or wall types could be considered.
Calculations were simplified enough to improve computational efficiency but were not adequate
for optimizing building design (Judkoff 2008).
In October 1973, members of the Organization of Arab Petroleum Exporting Countries
(OAPEC) proclaimed an oil embargo on the United States in response to the decision to resupply
the Israeli military during the Yom Kippur war. Since the industrial economy relied mostly on
crude oil and the OAPEC was the main supplier, the U.S. government responded with a wide
variety of new initiatives to contain further dependency. The Department of Energy was also
created in 1973, and the National Energy Act was established in 1978. Along with government

13!
!

!

solutions, the crisis led to greater interest in renewable energy which accelerated research,
especially in solar and wind power (Fehner and Hall 1994).
As a consequence, after 1973 the approach for energy use in buildings changed as well.
The general public became aware of energy saving in homes and buildings, searching for new
ways to achieve lower consumption. Designers felt the need to develop tools for weighing
different solutions in order to better understand building behavior and improve energy savings.
The initial trend was toward active solar systems since they could be added as a supplemental
HVAC system. In 1976 and 1977, architectural options and modifications gained interest as
innovative design concepts were developed. However, building energy analysis systems (BEAS)
were no longer appropriate under these new conditions, since subroutines to existing programs
did not ensure accurate energy analysis. Changes to analysis approaches were often necessary to
handle innovative design strategies, fostering an entirely new generation of building energy
analysis simulations (Judkoff 2008).
A third generation program began to emerge in the mid-1980s. With the advent of more
powerful personal computing, parameters could be modeled as dependent, such that energy
transfer processes were not solved in isolation, with only space and time considered as
independent variables. Thermal, visual, and acoustic aspects of performance had to be integrated
into modeling (Clarke 2001).
By the mid-1990s, program interoperability and integration were added. A growing
understanding by practitioners of energy analysis tools stimulated developers to create
knowledge-based user interfaces, application quality control, calibration, validation, and user
training (Clarke 2001).
14!
!

!

Between 2000 and 2010, programs such as EnergyPlus from the Department of Energy
were developed, taking capabilities to a new level by combining BLAST and DOE-2, the most
common simulation engines. EnergyPlus uses computer code written in Fortran 90, a modern
language with good compilers on many platforms that allows C-like data structures, mixed
language modules, and backward compatibility during the development process. In these modern
programs, third parties are invited to create user interfaces (Crawley et al. 1999).
2.2. Energy Estimation and Modeling Methods
A mathematical model is defined by ASHRAE (2009) as a description of behavior of a
system made up of three basic components: input, system structure and parameters or properties,
and output variables. Modeling can be approached from two different perspectives: forward or
classical, and data driven or inverse (ASHRAE 2009). In a forward approach, output variables of
a model are predicted with specified input variables in a defined structure or calculation method.
The data-driven approach uses known and measured inputs and outputs, and the objective is to
determine a mathematical model of the system. The main advantage of the forward approach is
that systems don’t need to be physically built to predict behavior, making them ideal for design
and analysis stages of a project. Government sponsored simulation tools such as BLAST, DOE-2
and EnergyPlus are based on forward simulation models. A data-driven approach contains a
smaller number of parameters, and thus allows identification of simpler and easier models and
more accurate prediction of future system performance under determined conditions, since
parameters come from actual building performance.
In choosing an analysis method, capabilities and project requirements should be matched.
The method must be capable of evaluating design options with sufficient accuracy to make
15!
!

!

correct choices. ASHRAE (2009) lists several important factors that should apply for every
method selected:
•

Accuracy, the method should be sufficiently accurate to allow correct choices even
though absolute energy prediction is not possible

•

Sensitivity, where the differences in energy use between two choices can be adequately
reflected

•

Versatility to allow analysis of options under consideration

•

Speed and cost is important as the total time (gathering data, preparing input,
calculations, and analysis of output) to make an analysis should be appropriate to the
potential benefits gained

•

Reproducibility, meaning that the method should not allow so many vaguely defined
choices that different analysts would get completely different results

•

Ease of use considering the importance of speed in obtaining results and making results
easy to reproduce

Clarke (2001) elaborates a series of considerations and processes that should be handled by
any advanced system:
•

Lag and thermal storage effect of heat transfer through a building

•

Radiant and convective gains from occupant, light, and processing equipment that varies
with time

16!
!

!

•

Infiltration, natural and controlled ventilation, and inter-zone air

•

The effects of shortwave solar radiation impacting exposed external and internal surfaces

•

Longwave radiation exchange between exposed external surfaces and the sky vault and
its surroundings, as well as between internal surfaces

•

Shading of external opaque and transparent surfaces caused by surrounding buildings, as
well as a variety of facade shading elements

•

Mapping of moving insolation patches from windows to internal receiving surfaces

•

Effects of moisture

A widely used analogy is to view the building as a complex network of thermal resistances
and capacitances, linking regions and representing conductive, convective, radiative, and heat
storage processes. The manner in which this network can be treated mathematically is by
neglecting some portions, assigning fixed values, or simplifying boundary condition
assumptions. This will determine the flexibility of the techniques to emerge for the program
(Clarke 2001).
Methods have changed over the years and it is important to acknowledge former models that
led to current methods. J.A. Clarke (1985) defines former methods: “Steady state approximation,
these methods had no mechanism for the accurate inclusion of solar gains, casual gains,
longwave radiation exchanges, plant operational strategies etc., and so such models only
addressed fabric heat flow (under very special boundaries conditions) and not building energy.
Typical inadequacies include omission of any consideration of the dynamic response of the
building, inability to deal realistically with many of the energy flows occurring within the
17!
!

!

building, and an inability to effect the correct relationship between building fabric and installed
plant operation … Simple dynamic, where the simplified methods of energy assessment have
been produced which address dynamic performance. These methods are mostly based on
regression techniques applied to the results of multiple parametric runs of more powerful
modeling systems. The results to emerge can often be reduced to simple relationships or
presented in tabular or graphical form.”
J.A. Clarke (2001) also describes modern contemporary programs: “Most contemporary
simulation programs are based either on response function methods or on numerical methods in
finite difference or, equivalently, finite volume form. The former method is appropriate for the
solution of systems of linear differential equations possessing time invariant parameters. In use,
it is usual to assume a high degree of equation decoupling. Numerical methods, on the other
hand, can be used to solve time varying, non linear equations systems without the need to assume
equation decoupling as a computational convenience. Numerical methods are favored for a
number of reasons. First, to ensure accuracy it is essential to preserve the spatial and temporary
integrity of real energy systems by arranging that whole system partial differential equation sets
be solved simultaneously at each computational time step. Second, numerical methods, unlike
their response function counterpart, can handle complex flow path interactions. Third, time
varying system parameters can be accommodated. Fourth, processing frequencies can be adapted
to handle so called ‘stiff’ systems in which time constants vary significantly between the
different parts of the problem (building fabric, HVAC component, control system elements, fluid
flow domains, control system elements etc.).”
One last method described by Clarke (1985) is the electrical analogue, which uses the
analogy of electrical flow and heat flow: “The analogy that exists between electrical flow and
18!
!

!

heat flow has led to construction of electrical analogue devices useful in the study of complex
heat flow phenomena. The technique is extremely useful as a research tool, allowing long term
simulations to be completed in a short elapsed time, but has little application in a design
context.”
2.3. Simulation Tools
2.3.1. DOE-2
The engine on which much current simulation software relies is DOE-2. The basic
mechanism of the engine is to process the information through five different subprograms, one
after the other, using the previous subprogram’s output as the next subprogram’s input. DOE-2
runs on five different subprograms that interact with each other. The first one is the Building
Description Language (BDL) subprogram that translates user inputs into computer language. The
BDL subprogram also calculates thermal factors in walls and envelope components. The second
module is the loads module, which calculates the heating and cooling gain loads inside the
building. It considers internal heat gains such as lighting, occupancy, equipment, and plug loads.
The third subprogram addresses systems, and it calculates heating and air conditioning system
requirements. It considers requirements for outside air, temperature controls, and regulating
accessories for equipment. The fourth module is the plant subprogram, which sizes chillers,
boilers, cooling towers, and other equipment necessary to run a number of systems together. The
last module is the economics subprogram, which translates calculated energy into costs
according to the type of fuel specified. This subprogram is also used to carry out life cycle cost
analysis as well (Hirsch et al. 1994). Figure 2.1 shows the structure of DOE-2.

19!
!

!

USER INPUT
FILE

WEATHER
DATA

CONSTRUCTION
LIBRARY

MATERIALS
LIBRARY

BDL PROCESSOR

LOADS

SYSTEMS

PLANTS

ECONOMICS

LOADS
REPORTS

SYSTEMS
REPORTS

PLANTS
REPORTS

ECONOMICS
REPORTS

Figure 2-1: DOE-2 Simulation Engine Structure (Hirsch 1994)
2.3.2. HAP
The Hourly Analysis Program (HAP) from Carrier uses the Transfer Function
Methodology (TFM) for calculating heat transfer phenomena in the software. This methodology
is a derivative of the Heat Balance Method, so shortcuts and assumptions are used to reduce the
volume and detail of required input to speed calculations. The methodology is best described in
the Transfer Function Methodology (TFM) paper (Carrier Corporation 2005) and it is explained
as follows:

20!
!

!

“Methodology is based under the principle that conduction, convection, and radiation are
the main drivers of heat transfer to or from the air in the room. The radiation heat gains from
sources such as solar, lights, and even people take time to become a load; the radiant heat must
first heat up the building and contents and then be conducted and released over time to the room
by air convection processes causing a delay between the time a heat gain occurs and the time its
full effects as cooling load appear. The convection process is governed by the temperature
difference between the mass and the room air. Convection decreases as the room air temperature
rises and increases as the room air temperature decreases. Due to throttling ranges in the control
systems that vary the room air temperature calculations can’t be made assuming constant air
temperature at all times. Using night set up or not cooling during unoccupied times may cause an
increase in room temperature and decrease in convection storing heat for release. Later on
system start up, the room air temperature rapidly decreases and a convective rush of heat can
occur. The TFM calculates the effect of the changing room air temperature on the heating and
cooling requirements. This is done through the Space Air Transfer functions known as Heat
Extraction.
The transfer function with heat extraction is implemented in three steps:
1. Conduction equations are used to analyze the heat flow through walls and roofs.
2. The room transfer functions are used to analyze radiative, convective, and heat storage
processes of all components. Convective components are instantaneous and radiative
components are stored and released over time.

21!
!

!

3. The space air temperature transfer functions (heat extraction equations) are used to
analyze the effects of the changing room air temperature on convective heat flow from
mass to room air that includes behavior of room thermostat.
Results of calculations can yield important benefits such as the ability to analyze realistic
transient heat transfer that occurs in all buildings. Loads can also be accurately computed for any
heat gain sequence and wall or roof construction. Resulting loads are specific and customized for
each application analyzed, accounting for local weather conditions, building constructions and
operating schedules.”
2.3.3. eQuest
A user’s manual (Hirsch 2009) details the basic operation of eQuest: “The program uses a
simulation engine derived from the principles of DOE-2, however it extends and expands its
capabilities in several ways, including: interactive operation, dynamic and intelligent defaults,
and improvements to numerous flaws that have limited its use by designers and building
professionals. The program allows avoiding all the programming process while guiding the user
through the creation of a detailed building model, performing automatic parametric simulations
of the design alternatives.”
2.3.4.

EnergyPlus

EnergyPlus is best explained in its engineering reference manual (University of Illinois
and LBNL 2009): “The EnergyPlus program is a collection of many program modules that work
together to calculate the energy required for heating and cooling a building using a variety of
systems and energy sources. It does this by simulating the building and associated energy
systems when they are exposed to different environmental and operating conditions. The core of
22!
!

!

the simulation is a model of the building that is based on fundamental heat balance principles.”
Figure 2.2 shows the organization of EnergyPlus

Figure 2-2: EnergyPlus Internal Elements (University of Illinois 2009) “For interpretation
of the references to color in this and all other figures, the reader is referred to the
electronic version of this thesis”

The main difference from the DOE engine is (University of Illinois and LBNL 2009): “In
programs with sequential simulation, such as BLAST and DOE-2, the building zones, air
handling systems, and central plant equipment are simulated sequentially with no feedback from
one to the other … This simulation technique works well when the system response is a welldefined function of the air temperature of the conditioned space … lack of feedback from the
system can lead to non-physical results … In EnergyPlus all elements are integrated and
controlled by an Integrated Solution Manager. The loops are divided into supply and demand
sides, and the solution scheme generally relies on successive substitution iteration to reconcile
supply and demand using the Gauss-Seidell philosophy of continuous updating.”
23!
!

!

2.3.5. BEOpt
BEOpt is a recent tool developed by the National Renewable Energy Laboratory (NREL)
to aid in the Building America project development by the US Department of Energy (DOE).
The scope of the program is mostly residential, and it is a useful tool for research towards zero
energy homes. In a report by Christensen et al. (2006), the basic explanation of the software is
given: “Beopt software calls the DOE2 and TRNSYS simulation engines and uses a sequential
search technique to automate the process of identifying optimal building designs along the path
to Zero Net Energy (ZNE) … The DOE-2 simulation program is used to calculate energy use as
a function of building envelope options and HVAC equipment options. Appliance and lighting
option energy savings are calculated based on energy use intensity factors and schedules input
into DOE-2. TRNSYS simulation program is used to calculate water heating loads and energy
savings for solar water heating. TRNSYS is also used to calculate annual electrical production
from a grid tied Photo Voltaic system … A prototype version of BEOpt using EnergyPlus has
demonstrated technical feasibility for future use of a single simulation engine with extended
capabilities.” Figure 2.4 shows the internal structure of Beopt.

24!
!

!

Figure 2-3: BEOpt Internal Structure (Christensen et all 2006)
It is important to emphasize that BEOpt is a tool to obtain optimal solutions for a Net
Zero Structure, unlike the other tools that will report a more detailed solution of the energy
behavior of the structure.
2.3.6

COMcheck

Some energy simulation programs are developed to demonstrate compliance with energy
codes instead of predicting energy performance. Such is the case of COMcheck developed by
Pacific Northwest National Laboratories (PNNL) for the U.S. Department of Energy (DOE).
This has been a helpful tool for designers and builders to understand the requirements of building
energy codes and for local authorities to enforce these codes. COMcheck is based on the
ASHRAE 90.1 standard and it addresses minimum requirements for building envelope, HVAC,
water heating and lighting systems (Bartlett et al. 2011). Bartlett et al. (2011) explain in their
report: “Each major system must comply on its own; trade-offs between major systems are not
permitted, although they are permitted under the Standard 90.1 Building Energy Cost Budget
Method or under the Total Building Performance section in the IECC editions … Compliance
25!
!

!

forms similar to those used in printed guidelines are generated by the software for submission
with plans and specifications …. Program enables Standard 90.1 users to only make trade-offs
between above-grade wall and window components. Trade-offs allow the use of components
exceeding minimum criteria to be used to offset components that fall below minimum criteria.”
This kind of software is limited to calculations for code compliance and it is not intended for
energy analysis or approved for LEED analysis. Table 2.1 summarizes advantages and
disadvantages of the modeling programs discussed above.
Description

Advantages
Disadvantages
Software
Parameters are
Result from one
DOE
Linear differential
assumed so shortcuts
system is input for the
• eQuest
equations with time
are used to reduce the
next one, but there is
• BEOpt
invariant parameters
volume of required
no feedback from
HAP
input
system to system
Systems can be solved
simultaneously,
Non linear equations
allowing feedback
The data input process
with time varying
EnergyPlus
among them; can
can be more complex
parameters
handle complex flow
path interactions
Table 2-1: Summary of Some Simulation Programs
2.4. Validation and Calibration of Energy Simulations
In an effort to increase accuracy of simulation tools, two concepts are important to
define: calibration and validation. Strachan (2008) indicates that these two processes are
differentiated: “Calibration involves using experimental data to ensure that the model predictions
align with the measured data over a realistic range of operating conditions … Validation is aimed
at ensuring that this (the predictions) assumption is true.” Calibration is the first step toward
achieving validation of a program. In simpler words, calibration ensures that computer

26!
!

!

calculations are developed according to theoretical mathematical functions, and validation
ensures that assumptions made to predict energy consumption are as close to reality as possible.
Judkoff (2008) not only defines the main goal of simulation but also identifies the need to
validate results and identify where inaccuracies may be present. Building energy simulation is
described as a tool to predict energy flows (including temperature, envelope losses, system
performance and electric loads) either in an existing structure, modification of an existing
structure or new design. The accuracy of simulation outputs depends on three factors: accuracy
of input data, applicability of the tool to the building and the climate being analyzed, and the
ability of the tool to predict real building performance when given perfect input data. Judkoff
(2008) states: “It is rarely feasible to collect sufficient experimental data or to apply a given
analysis tool to a sufficient number and range of test cases to achieve complete confidence for all
situations. Therefore engineering judgment is commonly used to select test cases that represent
typical applications of the tool.” Many studies have been developed in search of better
approximations and understanding of variables involved in energy simulations.
In a validation process, three different approaches can be taken: first, empirical validation,
where calculated results from a program are compared to monitored data from an actual building,
test cell, or laboratory experiment; second, an analytical approach, where results from a program
are compared to results known from analytical solutions or accepted numerical method
calculations; and finally comparative testing where the program is compared to itself or to other
programs (Judkoff et al. 2008).
Empirical validation implies gathering data from functional buildings or test cells where
parameters may be controlled. The aim of this method is to quantify inaccuracies in calculation
27!
!

!

methods by the program, but not without considering the uncertainties from both field
measurement and data analysis. Recommendations for field measurement include testing in
unoccupied buildings, since it is difficult to separate building performance itself from uncertain
occupant behavior such as doors opening or use of hot water (Judkoff et al. 2008). Figure 2.5
illustrates sources of simulation error identified by Judkoff.

28!
!

!

Figure 2-4: Simulation Error Sources (Judkoff et al. 2008)
29!
!

!

The International Energy Agency (IEA) (Judkoff 2008) developed a comprehensive and
integrated report of energy simulation tools which involve analytical comparative, and empirical
methods. The scope included the investigation of availability and accuracy of building energy
analysis tools and engineering models to evaluate the performance of innovative low energy
buildings. The report was developed for use by simulation software developers, codes and
standards organizations, architects, engineers, energy consultants, building owners, and
researchers. There are several definitions regarding the kind of tests carried out, including
comparative and empirical tests. Comparative tests determine and compare methods used to
calculate parameters in different software tools in order to identify the differences between their
results. Empirical validation consists of comparing predicted values with measured data in
existing buildings or test facilities, and these tests result in higher costs and time-consuming
activities. According to Judkoff, over eighty simulation programs have been amended through
empirical validation tests, improving accuracy in the evaluation of energy efficiency measures.
Recommendations of the IEA report include the support of continued development of models
and more validation studies, as the few buildings that are truly useful for empirical validation
studies were designed primarily as test facilities. Emphasis was placed on the growing body of
literature and activity demonstrating the importance of simulation for reducing energy
consumption through energy design. The continued development and validation of wholebuilding energy simulation programs is one of the most important activities meriting the support
of national energy research programs (Judkoff 2008).
Research and experimentation on calibration was done in another study (Lee and
Claridge 2002) where HVAC system performance (both mechanical and electrical) was
30!
!

!

measured in a prototype building, and results were used to optimize the simulation software
through the Solver package, which is a tool for optimization. As a conclusion the authors
recommend future study and development of a plug-in for an auto-calibration feature of
simulation programs.
An important study was also developed using a calibration method that incorporates short
term monitoring instead of long term monitoring. This study confirmed that short term methods
can be as accurate as long term monitoring, although it is only possible if simulation input is
derived from the procedure prescribed in the study. This finding opens the door for short term
studies, potentially reducing data collection time (Soebarto 1997).
Another research effort (Zmeureanu et al. 1995) used three different analysis programs
(BESA-Design, PC-BLAST, and MIRCRO-DOE2.1D) to evaluate energy consumption and cost
savings in a large existing office building in Montreal. The aim of the research was to test
program capabilities for predicting energy performance of buildings using information usually
available to a design professional or energy consultant, as well to evaluate the impact of several
energy efficiency measures or parameters with various software. The details of the research
were:
•

Area of building: 10,410 square meters (112,052 square feet)

•

Variable air volume HVAC system

•

The energy saving measures used were:
− Insulation of floor between garage floor and ground while decreasing garage
temperature
31!

!

!

− Night setback of indoor temperature to 16 C (60 F) in the winter using
programmable thermostats
− Increasing of set point to 25-26 C (77-79 F)
− Decrease of lighting power density on first floor from 25.8 to 22 W/m

2

2

− Decrease of lighting density from 25.8 to 11.8 W/m on the first floor

− Use of task lighting of 1.4 W/m

2

− Turning off of lights in washrooms, stairwells, and garages during unoccupied
hours
− Installation of timers and motion detectors as well as power management devices
to turn off office equipment when not in use
There was not a significant trend in the results, although overestimates or underestimates
were expected depending on the program or the user. As a recommendation the authors suggest a
need for a detailed methodology for verification of the quality of inputs and calibration with
utility bills, since models can be easily calibrated with these bills and better definition of
unknown parameters.
A similar study (Guyon and Rahni 1996) was performed using CLIM2000 software; this
program had already been tested once in small test cells where each parameter was controlled or
well known. The authors carried out their research in a 100 square meter (1,076 square foot)
house in an eight month time span, and close attention was paid to energy related measurements
32!
!

!

on the heating system. First, the house was modeled with no knowledge of experimental data,
and this model was established as a baseline to which results were compared with final data
collected. Some important results included observed decreases in error percentages during colder
months, probably due to lower solar gains during these periods. Overall consumption of the
building was overestimated by 5.3% over a long period. It was also emphasized that experiments
carried out in laboratories gave less uncertainty than experiments in actual buildings.
Karlsson et al. (2007) conducted a study to observe the variation in predicted energy use
by different programs. Values obtained through field measurements were compared to initial
predictions obtained by simulating the buildings using three different software tools which
remain unnamed and are addressed as programs 1, 2 and 3. The data was collected in twenty
2

terraced low energy houses in Sweden, well insulated (U value = 0.17 W/m K) with a 900 W
2

heat exchanger, with a ventilation rate of 0.35 l/m and an average window U value of 0.85
2

W/m K. The monitoring program continued for approximately three years after the buildings
were inhabited. Results showed that some measured values differed by 50% of the designed
values due to additional heating; higher indoor temperatures were measured in the actual houses
than those assumed in the initial simulation. Karlsson et al. also concluded that it is very
important to study which components affect energy the most and focus on these throughout the
design and construction process. The range of values measured through the twenty houses varied
significantly, between 50% higher than the average and 20% below average, showing large
differences depending on whether the walls were affected by sunlight, whether they were
between two other houses, and on which direction the windows were facing. An important
finding was that among the programs (all using as similar inputs as possible) the total energy
33!
!

!

consumption differed by 2% between programs. The study concluded that tenants are the most
difficult component to model in a simulation. Recommendations were made for the need for
future research on the impact of human behavior on energy use.
Research in full scale buildings usually includes many parameters and sources of error
that are difficult to control. Some studies have been developed in laboratories and test cells
under controlled circumstances and known conditions. Bloomfield et al. (1995) conducted a
study of this kind concluding that there is still considerable doubt about the absolute accuracy of
simulation program predictions, and defining three main techniques recognized for validating
software: analytical, inter-model, and empirical. The research focused on comparing four
programs with simple test rooms, and posed two questions: can the program be deemed valid due
to the level of agreement of prediction and measurements, and what might be the reasons for
discrepancies between programs and measurements. The data was collected in test rooms
developed at the Energy Monitoring Company (EMC), with controlled variables such as
temperatures and on-off sequence of radiators. The programs used were APACHE, CLIM2000,
ESP-r, and SERI-RES. The authors defined air heating units as the biggest source of errors,
followed by prediction of solar radiation entering a zone and its distribution to the enclosing
surfaces.
2.5. Program Comparison
As important as it is to know how accurate results are, it is also important to know
capabilities, strengths, and weaknesses of available software. No comprehensive comparative
survey of tools had been conducted since 1995, until Crawley et al. (2005) conducted an
important study in 2005. The results are presented in a matrix (a sample of which is included in
34!
!

!

Figure 2.6) that places software in one column, and their features in rows such as method of
calculations, capability of including renewable energy, ability to change schedules for every
component, and others. Software in the matrix includes eQuest, HAP, and EnergyPlus among
others. Crawley concluded that there is increased interest in validation by energy analysts. As a
consequence, several program developers indicated their plans to make simulation inputs
available for users to download in the near future. Crawley mentioned the issue of trust, and
asked whether or not tools really perform the capabilities indicated.

35!
!

Building!Envelope,!Daylighting!and!Solar!(9!
of!the!52!rows!from!Table!3!in!the!report)!

BLAST!

B!SIM!

DeST!

DOEH2.1E!

ECOTECT!

EnerHWin!

Energy!
Express!

EnergyH10!

EnergyPlus!

eQuest!

!

Outside!Surface!Convection!Algorithm!
•!BLAST!/!TARP!
•!DOEH2!
•!MoWiTT!
•!ASHRAE!Simple!
•!Ito,!Kimura,!and!Oka!correlation!
•!User!selectable!
Inside!Radiation!View!Factors!
RadiationHtoHair!component!separate!from!
detailed!convection!(exterior)!
Solar!gain!and!daylight!calculations!
account!for!interHreflections!from!external!
building!components!and!other!buildings!

!!
X!
!!
!!
X!
!!
!!
!!

!!
!!
!!
!!
!!
!!
!!
X!

!!
!!
!!
!!
!!
!!
X!
X!

!!
!!
X!
!!
!!
!!
!!
!!

!!
!!
!!
!!
!!
!!
!!
!!

!!
!!
!!
X!
!!
!!
!!
!!

!!
!!
!!
!!
!!
!!
!!
!!

!!
!!
!!
X!
!!
!!
!!
!!

!!
X!
X!
X!
X!
!!
X!
X!

!!
!!
X!
!!
!!
!!
!!
!!

!!

X!

X!

!!

!!

!!

!!

!!

X!

X!

!!

P!

!!

!!

X!

!!

!!

!!

X!

!!

Table 2-2: Contrasting Capabilities of Programs (Crawley et al. 2005)

36!
!

!

Under the standards developed by the IEA for tool validation, the Building Research
Establishment (BRE) undertook work in the UK during a span of five years, from November
1989 until September 1994. A report compiling these results (Lomas et al. 1997) describes the
tests carried out in test rooms with controlled conditions. This study is important because it was
the largest validation work to date and compared a number of simulation programs. An important
conclusion drawn is the necessity of a “blind” phase, that is, an experiment conducted without
knowledge of actual measurements, providing a clear and clean snapshot of a program’s
capabilities when used as it would be used in the design phase of an actual construction project.
2.6. Occupancy Schedule Testing
Research has also focused on determining the impact of a single parameter in overall
simulation results. This approach helps in finding sources of errors for a particular variable
instead of considering the sum of errors from various parameters. This can improve modeling,
since the impact of that particular parameter can be determined and evaluated. Also, uncertainty
about values chosen as inputs can be reduced. If enough data can be evaluated, it may be
possible to standardize variable values and develop norms or codes.
Occupancy schedule is a parameter that carries uncertainty in simulation. Improving
definition of occupancy schedules, lighting schedules, schedules for plugged loads, and
schedules for thermostat settings allows the modeling program to apply factors to electric
consumption and HVAC systems operation, leading to more accurate approximations of energy
use. However, as unpredictable as human behavior is, inaccurate assumptions will be made about
these schedules. Additional research is necessary regarding this variable, because even though
real models of exact occupancy schedules cannot be obtained, standardized profiles could be
37!
!

!

determined to provide a narrower range of differences among models. Standardization can be
reached within buildings of the same occupancy type, and therefore, profiles for office buildings,
school buildings, hospital buildings etc. could follow certain norms for schedule definition.
Efforts have been made to estimate occupancy through agent-based models such as work
by Liao and Barooah (2010). This model involved simulating behavior of occupants and then
graphing the results so statistical information could be used for real time estimations. Although
the experimental data collected by Liao and Barooah was made in a simplified environment of
only one room and one occupant, the authors concluded that simulation can produce accurate
estimates for a whole building. This presents the possibility of integrating agent-based simulation
into energy simulation programs so an approximation of occupancy can be defined.
A study carried out at Cambridge, Massachusetts aimed to survey a building to assess
internal loads and schedules (Wasilowsky and Reinhart 2009). Researchers conducted twenty
walk-through observations of a building, and also conducted online questionnaires regarding
occupant schedule and appliance use. The model was run with default values in the program,
which were replaced with observed values. Conclusions included the importance of carefully
surveying a building during retrofit and emphasizing the importance of adequately modeling
internal loads. The authors calculated the mean bias error (MBE) for both cases: 0.2% for
customized loads and -17% for the default inputs. Likewise, mean square errors showed
improvement from 64% for default loads to 23% for customized loads. Figure 2.7 shows
measured energy against simulations with either custom (measured) or default (by the program)
inputs:

38!
!

!

Figure 2-5: Measured Energy v. Simulated Energy (Wasilowsky and Reinhardt 2009)
Another study attempted to test the sensitivity of occupancy related parameters
(Clevenger and Haymaker 2010). A range of values were defined assuming states of low and
high occupancy, that is, how lighting, plug loads, hot water and people affect the structure during
high and low occupancy. The authors concluded that comparing results obtained from lower
occupancy values to results obtained from higher occupancy values showed differences of up to
150%, demonstrating the range of inaccuracies and the need for the study of occupancy
schedules. The authors envisioned the creation of occupant files, similar to weather files which
may serve as benchmarks for future models. Recommendations for future research included
extended work towards a better understanding of the impact of uncertainty introduced by
variable occupant behavior on accuracy of simulation results, as well as identification of the
overall shortcomings of current occupant models.
A significant project developed by ASHRAE (Abushakra et al. 2001) was the 1093-RP,
which focused on developing a set of library schedules and diversity factors for electricity
39!
!

!

components such as lighting and plug loads. This was developed by measuring data from thirtytwo office buildings throughout one year, differentiating between weekdays and weekends. Files
released were compatible with DOE-2, BLAST, and EnergyPlus, and allowed for modification
as well if the profile didn’t suit a proposed schedule. Figure 2.8, shows measured patterns of
various projects and an average line is shown.

Figure 2-6: Sample Graph for Diversity Measurements in Buildings (Abushakra et al.
2001)
Even though defined profiles might be of great use, separate research (Bourgeois et al.
2005) highlights shortcomings of these results in the 1093-RP. The authors claimed the
technique was valid, although it was derived independently of weather and meteorological
patterns, and lighting use will show different patterns in buildings of large envelope-to-floor area
ratio with daylighting design approaches. This research focused on enabling manual control, or
individual instance modeling of people and equipment instead of groups. It also used energy

40!
!

!

simulations instead of predefined diversity profiles. This technique gave results as much as 62%
better on energy prediction than with the defined schedules.
A recent study conducted for the New Building Institute (NBI) by Heller, Heater, and
Frankel (2011) determined envelope components and operational systems that have an impact on
building energy use. They defined a range of performance values for various components and
systems as poor, baseline, and good. Performance values were tested using energy simulation for
each characteristic and incorporated weather data from sixteen cities. The characteristics
included envelope, occupancy, HVAC, lighting, operations (daylight controls, economizer and
thermostats) and other direct loads. The authors found operations and occupant behavior heavily
impact building energy use. Figure 2.9 shows an example chart included in their study:

41!
!

!

Figure 2-7: Measured Impacts of Different Performance Values of Building Characteristics (Heller, Heater, and Frankel,
2011)

42!
!

!

The authors main conclusion was identifying responsible parties for on-going building
energy performance since it was found how operational and occupant behavior heavily impacts
building energy use. Study also suggests how many design options are not taken advantage of
according to the climatic zone and how these should be recognized and regulated in codes.
2.7. Sources of Information
The energy modeler may use default schedules in software but there are a number of sources
of information which can be available. Published sources or the owner may have information
available which can assist the modeler when defining schedules.
In school classrooms, scheduling can be supported by existing and ongoing space utilization
studies. In a report from Marshall University (Dober, Lidsky, Craig and Associates 2003), the
purpose of these studies was stated as identifying space use percentage to make the most out of
every available room and optimize scheduling. As a consequence, space utilization patterns of
use were also examined. Figure 2.9 shows a sample chart from Marshall University.

43!
!

!

Table 2-3: Space Utilization Study (Dober, Lidsky, Craig and Associates, 2003)

44!
!

!

Similar studies have been conducted on other college campuses, where not only actual utilization
was determined, but also target occupancy hours used by administration and planning
departments. These types of reports may be available from an owner during design and modeling
of a specific project and should be considered by the modeler. A sample chart of this information
from the University of Oregon (2007) is shown in Figure 2.10:

Table 2-4: Space Utilization with Target Objectives (University of Oregon 2007)
Another available source is the ASHRAE 90.1 User’s Manual (2004) which include
suggested schedules by building type being modeled. According to the standard (2004): “Where
the schedules are not known, you may consult values in Table G-E through G-N for guidance …
Schedules can vary significantly from building to building and tenant to tenant. Over the years a
number of schedules have been defined using a variety of sources … Schedules are reproduced
as an example of typical input data.” A sample schedule from ASHRAE 90.1 user’s manual is
shown in figure 2.12.

45!
!

!

Schedule0for0Occupancy
Hour0of0Day0
(Time)

1"(12%1"am)
2"(1%2"am)
3"(2%3"am)
4"(3%4"am)
5"(4%5"am)
6"(5%6"am)
7"(6%7"am)
8"(7%8"am)
9"(8%9"am)
10"(9%10"am)
11"(10%11"am)
12"(11%12"pm)
13"(12%1"pm)
14"(1%2"pm)
15"(2%3"pm)
16"(3%4"pm)
17"(4%5"pm)
18"(5%6"pm)
19"(6%7"pm)
20"(7%8"pm)
21"(8%9"pm)
22"(9%10"pm)
23"(10%11"pm)
24"(11%12"am)

Schedule0for0Lighting0
Receptacles

Percent0of0Maximum0
Load
Wk
Sat
Sun
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
20
20
10
20
20
10
20
20
10
80
60
10
80
60
10
80
60
70
80
60
70
80
60
70
80
60
70
80
60
70
20
60
70
20
60
70
20
60
70
20
80
70
10
10
20
0
0
0

Percent0of0Maximum0
Load
Wk
Mon
Tue
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
40
5
5
40
30
30
40
30
30
75
50
30
75
50
30
75
50
30
75
50
65
75
50
65
75
50
65
75
50
65
75
50
65
75
50
65
75
50
65
75
50
65
75
50
65
75
50
65
25
50
5
5
5
5

Table 2-5: Sample Schedule for Assembly Occupancy (ASHRAE User's Manual 2004)
2.8. Summary
This chapter summarizes the literature that was used as background for this research. The
intention of the literature review was to gain an understanding of the history and development of
energy simulation software, as well as to investigate existing research related to energy
46!
!

!

simulation and prediction. The literature review helped in identifying gaps where this thesis
research could be focused in order to avoid repeating existing studies and to determine the areas
of interest for this study.

47!
!

!

CHAPTER 3
METHODOLOGY

48!
!

!

METHODOLOGY
This chapter outlines the methodology used for this thesis research; the goals were to
determine the impact of schedule definition on energy simulation and to develop
recommendations on procedures helpful in achieving more accurate occupancy schedules. The
research used three case studies. Additionally a survey of energy modelers was conducted to
obtain information on their practices for establishing occupancy schedules during simulation.
Figure 3.1 shows activities carried out for the research and includes problem definition, literature
review, boundary testing for the software, selection of case studies, methods for gathering
information for each case study, development and administration of a survey of energy modelers,
and development of recommendations, guidelines, and tools for energy modeling processes.
3.1 Literature Review
The literature review covered three primary areas. First, background research on energy
modeling software development along with modeling methods and techniques was conducted.
Secondly, review of existing programs was conducted with a focus on understanding the types of
tests carried out in the software. Finally, articles and documents were reviewed to explore
existing research on occupancy related schedules for energy simulation.
3.2 Survey
The first step in the study was to develop a survey of energy modeling professionals on how
occupancy schedules are defined within their projects. General perceptions about uncertainty in
simulation results and scheduled parameters were obtained as well as insight on existing
49!
!

!

procedures for defining user profiles. Upon development of the survey, proposed procedures
were submitted for Human Subject review by the Institutional Review Board (IRB) at Michigan
State University. Surveys were carried out by telephone. The questions used in the survey and
consent forms are shown in Appendix A. The four surveyed professionals were selected based on
their known involvement with LEED and energy simulation projects and also past working
interactions with the research team.
3.3 Identification and Selection of Case Studies
Three buildings were selected as case studies for carrying out measurement and analysis.
These buildings were selected in collaboration with Michigan State University Engineering and
Architectural Services, Physical Plant Division, as well as a local architect. After meeting with
these organizations and presenting the study, they consented to help and offered the selected case
studies. Architectural and engineering drawings for the Chemistry Building addition and the
Geography building at Michigan State University and a Lansing, Michigan church were
provided. Specific data collection methods for each case are discussed below.

50!
!

!
Definition of the project
Need for research, goals and objectives of the research
Literature Review

Need for research, goals, and
Objectives of the research

Existing Tests on Software:
Validation and Calibration

Similar Research:
Testing Occupancy Schedules

Selection of specific simulation parameter to
study: Lighting Schedules

Selection of
case studies

Chemistry
Building

Geography
Building

Test extreme boundaries and specified
Intermediate conditions

Church
Project

Gather available data: drawings,
existing simulation, CAD files

Building walk around, observations, and
portable light logger placement

Gather data from
monitoring room

Figure 3-1: Methodology of Research
51!
!

Development of survey for
professional energy simulators

!

Figure 3-1 (cont’d)

Define new occupancy schedule
and run simulation
Compare results with actual
consumption data

Determine simulation accuracy
Improvement through schedule
Defined by data gathered and observed

Does lighting schedules have a high or low
impact on an overall energy simulation?

Decrease uncertainty on lighting
schedule definition

Recommendations on definition of occupancy
Schedules for future projects

Help standardize and create guidelines
for energy modeling process

Recommendations and tools for owner
and engineering parameters

Create a tool for model calibration and
future building analysis reference

Conclusions

52!
!

!

3.4 Extreme and Intermediate Boundaries
Before measurements in the buildings were undertaken, boundary conditions in the software
were tested. The program used throughout the research was eQuest. A simple building model
(not any of case base buildings) was created in the program, guided by ASHRAE minimum
envelope and HVAC system requirements and a simple rectangular shape. This simple structure
helped in determining the range of predicted energy consumption when utilizing 0% lighting and
100% lighting throughout the 24 hour day in the simulations. Intermediate ranges such as 8 hour
periods and 12 hour periods at 100% lighting use were also tested to understand the impact of
increasing lighting use in energy predictions. These boundary tests were useful in determining
the relative impact of lighting under minimum, maximum and intermediate values.
Varying occupancy percentages affect several parameters including: lighting, HVAC fans,
water heating, thermostat settings, and plug loads; Siva (2010) establishes a rule of thumb where
for every 3 Watts of reduced lighting, 1 Watt of HVAC is reduced. A more detailed study was
carried by Sezgen and Koomey (1998), where through simulation of varying lighting loads and
HVAC loads impacts were determined. Figure 3.2 shows changes in heating and cooling loads
from Sezgen and Koomey’s work.

53!
!

!

Figure 3-2: Change in Heating and Cooling Loads Caused by a 1kWh decline in lighting loads in new construction buildings
(Sezgen and Koomey 1998)
54!
!

!

3.5 Methods for Each Case Study
Due to the type of data available from each of the buildings, the data gathering process varied
from case to case. The following section describes each case study and shows a breakdown of
activities and data available.
3.5.1

Chemistry Addition Building

The steps for gathering data for the Chemistry Addition building included the following:
•

The Engineering and Architectural Services Division from the Physical Plant at Michigan
State University provided construction drawings which were used to obtain information
for model development, including architectural details of the envelope, HVAC design,
and lighting fixture layouts

•

An existing energy simulation conducted during the design phase of this building was
provided and reviewed to obtain schedules used during building design

•

Visual observations of lighting use were made for spaces within the building. All 24
hours were recorded during a one week period. Observations were scheduled as shown
in Table 3.1:

55!
!

!

1st

2nd

3rd

4th

5th

Monday

12:00 AM

5:00 AM

10:00 AM

3:00 PM

8:00 PM

Tuesday

1:00 AM

6:00 AM

11:00 AM

4:00 PM

9:00 PM

Wednesday

2:00 AM

7:00 AM

12:00 PM

5:00 PM

10:00 PM

Thursday
Friday

3:00 AM
4:00 AM

8:00 AM
9:00 AM

1:00 PM
2:00 PM

6:00 PM
7:00 PM

11:00 PM

Table 3-1: Scheduled Visits for Chemistry Building
•

Observations were recorded in a chart where hourly data defined the lighting use profile
for each room, as well as an hourly average profile for the whole building as shown in
Table 3.2

•

Data loggers were installed for verification purposes in the basement spaces and the
computer office in the third room

56!
!

!
HOUR OF THE DAY
Rooms

Floor

Study Room

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16 17 18 19 20 21 22 23

Basement

Office

1

Basement

Classroom

0

Basement

General Office

1st

Classroom

1st

Copy Room

1st

Undergrad Office

1st

Classroom

2nd

Computer Office

3rd

Bussiness Office

3rd

Graduate Office

3rd

M ain Office

4th

Conf./ Sem. Room

4th

Grad. Study Room

5th

Instrument Lab

5th

Conf./ Sem. Room

5th

Hallways

B-5

Bathrooms

B-5

Stairway

B-5
%

Table 3-2: Chemistry Observation Chart
57!
!

!

•

Hourly lighting schedules were defined from these field observations and then used for
simulation

•

The Engineering and Architectural Services from the Physical Plant at Michigan State
University provided electric and steam utility metering data from the building, which was
used to compare predictions from the model and to determine accuracy of the simulation

3.5.2

Geography Building

The steps for gathering data for the Geography building were the following:
•

The Engineering and Architectural Services from the Physical Plant at Michigan State
University provided construction drawings necessary to obtain information required to
develop the model, including architectural details as well as available HVAC and
miscellaneous equipment information

•

Lighting fixture drawings were only available for some spaces where prior remodeling
work had been conducted; consequently the researcher conducted a physical inspection to
determine fixture number and type in spaces where no information was available from
drawings. This approach was necessary to account for the number and type of fixtures in
determining electric loads associated with lighting and to define a baseline number for
100% lighting use

•

Since no previous model was developed for this building, the first simulation run used
default lighting schedules within the eQuest software

58!
!

!

•

Visual observation was carried out for spaces by conducting a walk-through of the
building. Access to the building after normal business hours was requested from building
authorities and approved. Table 3.3 shows the schedule for the visits

1st

2nd

3rd

4th

5th

Monday

12:00 AM

5:00 AM

10:00 AM

3:00 PM

8:00 PM

Tuesday

1:00 AM

6:00 AM

11:00 AM

4:00 PM

9:00 PM

Wednesday

2:00 AM

7:00 AM

12:00 PM

5:00 PM

10:00 PM

Thursday
Friday

3:00 AM
4:00 AM

8:00 AM
9:00 AM

1:00 PM
2:00 PM

6:00 PM
7:00 PM

11:00 PM

Table 3-3: Scheduled Visits for Geography Building
•

Observations were recorded in a chart to create a profile for each room, as shown in
Table 3.4. A whole building average lighting use profile was defined from observations
as well

59!
!

!

Rooms
Study
Room
Classroom
Office
Classroom
Copy Room
Classroom
Computer
Office
Bussiness
Office
Main
Office
Conf./ Sem.
Room
Grad. Study
Room
Instrument
Lab
Bathrooms
Stairway

Floor

0

1

2

3

4

5

6

7

8

9

10

HOUR OF THE DAY
11 12 13 14 15

Base.
Base.
Base.
1st
1st
2nd
3rd
3rd
4th
4th
5th
5th
B-5
B-5
%
Table 3-4: Geography Observation Chart

60!
!

16

17

18

19

20

21

22

23

!

•

The Engineering and Architectural Services Department from the Physical Plant at
Michigan State University provided electric and steam utility metering data of the
building, which was used to compare predicted results from the model to actual
consumption

3.5.3

Church Project

The steps for gathering necessary information for the Church project included the following:
•

Drawings were provided by a local architect, where architectural details of envelope
assemblies were obtained, as well as engineering details for HVAC and lighting systems

•

An energy simulation conducted during the design phase was provided for reference;
assumed schedules for occupancy used in the simulation were extracted from this existing
model and predicted results obtained

•

As the project is farther away than the other two buildings, LEVITON portable light
loggers were used for data collection rather than physical observation. These were
installed in the main rooms (Meeting and Fellowship), library, and one of the small
meeting rooms. The loggers record when lights are on or off, lighting levels and if the
room is occupied or not. Lighting status (being on or off) was used to define lighting use
schedules for the research model in this building

•

Monthly metered data was provided by the project owner to compare actual consumption
to predicted results

61!
!

!

•

Additional information requested from the owner for analysis was provided including a
predesign occupancy schedule and a calendar showing actual use of spaces

3.6 New Schedules and Analysis
After field observations were collected, percentages of lighting use were defined from the
observation charts for all 24 hours, allowing definition of new lighting schedules for each case
study according to recorded profiles. These observed schedules were used to run the simulation
model developed during the research to compare and analyze differences in energy prediction
with the original simulation created during the design phase. Analysis also compared metering
data with the simulations of the buildings. Metered data for Chemistry and Geography buildings
was provided by the Architectural and Engineering Services department from Physical Plant at
Michigan State University, and monthly utility billing history was provided by the owner for the
church project.
3.7 Guidelines Development
Recommendations and guidelines for lighting schedule definition during design phase were
developed following the core study analysis. The recommendations and guidelines identify
sources of information that can better enable the modeler to create realistic profiles of occupancy
rather than using program default schedules. To facilitate collecting relevant information a
collection tool was developed. The tool can help the modeler to collect information for
occupancy, lighting, HVAC, for input into an energy simulation. This tool is used in conjunction
with the owner, architect, and engineers so all relevant information can be collected and included
in a single document.
62!
!

!

3.8 Chapter Summary
This chapter describes the methodology used during the research. The research has six main
components: literature review, a survey of energy modeling professionals, a boundary test to
determine the impact of lighting variations within the software, case study observations and
measurements carried out in three selected buildings, case study simulations and development of
recommendations and tool for collecting information when modeling. The selected buildings
were the Chemistry building and Geography building at Michigan State University, and a Church
project in Lansing, Michigan. Procedures for data collection were described for each building
and addressed how lighting schedules were developed from observations and measurements.
Data collection and its analysis are discussed in Chapter Four. Finally, guidelines, and tools were
identified as deliverables of the research and are presented in Chapter Five.

63!
!

!

CHAPTER 4
DATA ANALYSIS

64!
!

!

DATA ANALYSIS
The following chapter presents the data and analysis for the research. First a survey of
professional energy modelers is presented and provides background on modeling processes.
Then, impacts of lighting schedule variation on a sample building modeled under ASHRAE
requirements are presented. Finally, lighting use patterns were recorded in three case study
buildings, which were used to define lighting schedules for energy models developed for each of
these buildings. Electric consumption from simulation results were compared to actual utility
data which led to conclusions about the impact of schedules. As a result recommendations and a
checklist for collecting information during initial simulations was developed and is presented in
Chapter Five.
4.1 Interviews
Telephone interviews of five energy professionals were conducted in order to obtain their
insights into energy simulation processes. A summary of the interview questions and verbatim
responses are included in Appendix C.
The first question asked of the interviewees was about the software they used for
simulation models, eQuest was used by three of the interviewees and TRANE Trace was used by
two interviewees. Other programs also used by the interviewees included Energy 10, Ecotech,
HAP from Carrier, and EnergyPlus. When asked what the purpose of the models was, all
answered that they used models for energy analysis in buildings and were not limited to
compliance with LEED credits. One of the participants indicated: “We are trying to move away
65!
!

!

from LEED certifications, we find more value in helping teams through schematic and design
phases on a project to help evaluate design decisions”.
When asked about model validation, three interviewees agreed that validation is not
carried out because of associated costs and that the main objective of modeling was for
comparison of design options rather than predicting energy use. One interviewee indicated
however that models are used to measure and verify actual consumption.
There was unanimous agreement that one of the main sources of inaccuracy was
modeling occupancy behavior and scheduling system operation hours (HVAC, lighting). Air
leakage of the building was also cited by one of the professionals, as well as sloppy or careless
assumptions. When asked if they spent enough time on their schedule definition process,
responses varied as two interviewees indicated they do spend time on their schedules while two
others indicated they use default schedules within the program. One interviewee indicated that
for some projects default schedules might be used at the beginning of the project and fine-tuned
after additional information can be gathered. One interviewee indicated that some consultants
might overestimate occupancy to be on a safe side, since it may be better to overestimate energy
use than underestimate it.
When asked about the owner’s information available for creating occupancy schedules,
all professionals agreed that owners generally don’t have a clear idea of how a new building is
going to operate; it was clarified that better information might be obtained from owners that are
just moving between buildings and already have an understanding of building behavior and
operation schedules. Regarding methodologies or codes followed by modelers, most of them
66!
!

!

indicated that schedule definition was developed based on each modeler’s criteria, although one
participant indicated they use ASHRAE developed schedules when working on LEED projects.
There was common agreement among interviewees that standard and more accurate
schedules would help better predict energy use, but they indicated it would be difficult since
every project is different. Interviewees commented that some universities were establishing
certain requirements and guidelines for external consultants to follow, but parameters such as
plug loads and equipment would be hard to normalize until submetering is carried out once the
building is in operation. Again ASHRAE profile schedules were mentioned as a possible
standard to follow.
From the interviews, it was concluded that there was agreement that schedules were
important sources of uncertainty in energy use prediction. There is interest in reducing the
uncertainty of this parameter, but energy modelers were unsure on where to get the right
information, or how to use it in preparing occupancy schedules. Understanding the impact of a
variable parameter such as lighting use can help in making better decisions and can shed some
light on uncertainty during simulations.
4.2 Lighting Use Range Testing
In order to test the impact of lighting schedules on simulation, a model was developed of
a one story rectangular sample building in compliance with ASHRAE 90.1-2007 Energy
Standard minimum baseline construction and lighting requirements for climate zone 5. All
lighting fixtures were specified as fluorescent. The building was modeled as having 32,000 gross
square feet, approximately the size of Chemistry Building Addition, one of the case study
67!
!

!

buildings presented in Section 4.3. Window areas were taken as 40% of the wall area. Thermal
transmission values of the envelope components were: U= 0.064 for exterior walls, U= 0.048 for
the roof, F slab factor = 0.73 for the floor, U= 0.55 for windows, and U= 0.7 for swinging doors.
Table 4.1 and 4.2 show the ASHRAE 90.1 (2007) requirements:

68!
!

!

Nonresidential!
Opaque!Elements!

Assembly!
Maximum!

Insulation!Min.!R=Value!

Insulation!Entirely!above!Deck!

U=0.048!

R=20.0!c.i.!

Metal!Building!

U=0.065!

R=19.0!

Attic!and!Other!

U=0.027!

R=38.0!

Mass!

U=0.090!

R=11.4!c.i.!

Metal!Building!

U=0.113!

R=13.0!

Steel=Framed!

U=0.064!

R=13.0!+!R=7.5!c.i.!

Wood=Framed!and!Other!

U=0.064!

R=13.0!+!R=3.8!c.i.!

Roofs%

Walls,%Above.Grade%

!

!

Fenestration!

!

! Assembly!!!!Max.!
U!

!

Assembly!Max.!SHGC!

Vertical%Glazing,%.%.40$%of%Wall%
Nonmetal!framing!(all)!

U=0.35!

Metal!framing!(curtainwall/!storefront)!

U=0.45!

Metal!framing!(entrance!door)!

U=0.80!

Metal!framing!(all!other)!

U=0.55!

SHGC=0.40!all!

Table 4-1: ASHRAE Requirements for Climate Zone 5 (ASHRAE, 2007)

69!
!

!

Building(Type(

Fossil(Fuel,(Fossil/(Electric(Hybrid,(
and(Purchased(Heat(

Electric(and(Other(

Residential!

System!1!=!PTAC!

System!2!=!PTHP!

System!3!=!PSZ=AC!

System!4!=!PSZ=HP!

System!5!=!Packaged!VAV!with!
Reheat!

System!5!=!Packaged!VAV!with!
PFP!Boxes!

System!7!=!VAV!with!Reheat!

System!8!=!VAV!with!PFP!Boxes!

2

Nonresidential!and!3!Floors!or!Less!and!<!25,000!ft !
2

Nonresidential!and!4!or!5!Floors!and!<!25,000!ft !or!!5!
2

Floors!or!Less!and!25,000!ft2!to!150,000!ft !!
2

Nonresidential!and!More!than!5!Floors!or!>!150,000!ft !

Table 4-2: ASHRAE Definition of HVAC Systems (ASHRAE, 2007)

70!
!

!

The HVAC system was modeled as a packaged variable air volume (VAV) with reheat
coils, in accordance with ASHRAE 90.1 as shown in Table 4.2. A schematic diagram is shown in
Figure 4.1. Thermostat set points were fixed at 72 F for heating and 75 F for cooling. Functional
areas of the structure were allocated as: 40% classrooms, 40% office space, 15% hallways, and
5% restroom space. The spaces were not physically modeled in the program, but functional
percentages entered allowed the program to calculate lighting, people, and equipment loads
according to assigned ratios of areas in the building.

71!
!

!

VAV
BOX

Reheat
Coil

VAV
BOX

Reheat
Coil

Exhaust
Air

Figure 4-1: Schematic VAV with Reheat Coil System (NREL, 2006)
72!
!

VAV
BOX

Reheat
Coil

!

Several lighting schedules were used for this test including schedules using 100%
lighting use during none, some, or all hours of the day, as well as default schedules defined
within eQuest. Lighting schedules percentages were entered into the model reflecting the lighting
which is “on” either as a total wattage load, or as the total of fixtures switched on. For example,
if 100 watts are installed, 1.0 would be interpreted as 100 watts consumed and 0.0 as 0 watts
consumed. If the system had a dimming system, 0.5 would simulate 50 watts being consumed.
On the other hand, if three fixtures were installed, 1.0 would mean three fixtures on, 0.0 no
fixtures on, and 0.33 only one fixture switched on. Tables 4.3 and 4.4 show the default schedules
in eQuest and Table 4.5 show the lighting schedules for the test case.

Time
0*1
1*2
2*3
3*4
4*5
5*6
6*7
7*8

%&use
0.05
0.05
0.05
0.05
0.05
0.05
0.0512
0.0535

Time
8*9
9*10
10*11
11*12
12*13
13*14
14*15
15*16

%&use
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.054

Time
16*17
17*18
18*19
19*20
20*21
21*22
22*23
23*0

%&use
0.0535
0.0521
0.0512
0.0512
0.0502
0.0502
0.05
0.05

Table 4-3: eQuest Default Schedule for Weekends, Holidays, and Break Periods

73!
!

!

Time
0*1
1*2
2*3
3*4
4*5
5*6
6*7
7*8

%&use
0.05
0.05
0.05
0.05
0.05
0.05
0.1605
0.424

Time
8*9
9*10
10*11
11*12
12*13
13*14
14*15
15*16

%&use
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.713

Time
16*17
17*18
18*19
19*20
20*21
21*22
22*23
23*0

%&use
0.611
0.5005
0.679
0.679
0.6705
0.288
0.101
0.05

Table 4-4: eQuest Default Schedule for Weekdays

Time
0,1
1,2
2,3
3,4
4,5
5,6
6,7
7,8

0% 8hr 12hr 100%
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
0
0
1
1

Time 0% 8hr 12hr 100%
8,9
0
1
1
1
9,10
0
1
1
1
10,11 0
1
1
1
11,12 0
1
1
1
12,13 0
1
1
1
13,14 0
1
1
1
14,15 0
1
1
1
15,16 0
1
1
1

Time 0% 8hr 12hr 100%
16,17 0
0
1
1
17,18 0
0
1
1
18,19 0
0
1
1
19,20 0
0
0
1
20,21 0
0
0
1
21,22 0
0
0
1
22,23 0
0
0
1
23,0
0
0
0
1

Table 4-5: Schedules for Test Cases
The intent of this test was to determine the boundary conditions of lighting impact on
simulation examining five different scenarios, all with the same input information except with
varying lighting schedules. The scenarios were: using the software default schedules, 0% lighting
use, 8 hours per day at 100% lighting use, 12 hours per day at 100% lighting use, and 100%
lighting use throughout 24 hours of the day. Vacation periods were registered as: summer break
th

st

th

th

from May 6 to August 31 , winter break from December 16 to January 9 , and spring break
from March 5

th

th

to March 13 . During the break periods, schedules were not modified; the
74!

!

!

schedules for lighting and miscellaneous loads are shown in Table 4.3. Cooling was assumed to
be electric driven, while heating was provided by gas fired hot water coils. Monthly energy
consumption is shown on Figure 4.2:

75!
!

!

Figure 4-2: Monthly Electrical Consumption (kWh) of Test Building

76!
!

!

Figure 4.2 shows that less energy is used during May through August when compared to
other months. Although cooling load demand is higher during this period, lighting and
miscellaneous equipment loads are less because of decreased use schedules defined in Table 4.3.
Since lighting and equipment are defined with the same schedules during these periods, little
variation in consumption occurs from case to case. Consequently variations in energy
consumption are due mostly to weather dependent variables, mainly cooling loads. The monthly
analysis is shown in Table 4.6:

77!
!

!

%9Increase9
Against90%

%9Increase9
Against9
Default

6.93 26.77 19.99 17.41 15.39 177.8

58.4%

,,,,

7.64

6.52 15.41 10.24 9.31 11.23 112.3

,,,,

,36.9%

4.27

7.52

6.84 23.92 17.57 15.39 14.39 161.7

44.0%

,9.1%

4.27

7.51

6.96 26.74 19.95 17.46 15.51 178.3

58.8%

0.3%

22.72 26.3 24.81 28.54 10.74 4.26

7.49

7.51 38.31 30.05 25.65 19.82 246.2 119.3%

38.4%

Jan

Default9Schedules
0%9Lighting9
Schedule
89hr9Lighting9
Schedule
129hr9Lighting9
Schedule
100%9Lighting9
Schedule
Range9Between90%9
and9100%

Feb

Mar

Apr

May

Jun

Jul

16.96 18.39 17.29 18.98

7.9

4.27

7.53

11.29 10.94 10.36 9.68

5.36

4.32

15.47 16.7 15.64 16.71

7.3

16.94 18.52 17.47 19.06 7.95

Aug

11.43 15.36 14.45 18.86 5.38 ,0.06 ,0.15 0.99

Sep

Oct

Nov

Dec

Total

22.9 19.81 16.34 8.59 133.9

Table 4-6: Monthly Energy Consumption (kWh) Resulting from Lighting Schedule Variations in a Hypothetical Building

78!
!

!

The range of impact from no use of lighting and having all lighting on 24 hours is 120%.
It is unclear what percentage is attributable solely to lighting since total building consumption is
reported, and some of the variations are due to space heating, cooling, plug loads, fans, and
equipment. Changes result from occupancy in addition to lighting loads. It is worth mentioning
the highest and lowest monthly differences between the 0% case and 100% case within the year,
occur in December, with 8.7 kW and September with 22.96 kWh. September had the highest
energy use for every case due to cooling loads of occupied periods. June, July, and August have
similar consumption patterns since they are driven by cooling loads and lighting does not have as
much influence. For heating season months, electric consumption decreases since unlike cooling
which is electric driven, heating is gas fired hot water based. Energy use for the eight hour and
twelve hour assumptions are similar to results obtained with default schedules. Default schedules
and extreme schedules vary by 38%, which is considerable if predicting actual energy use and
utility cost. Figure 4.3 shows the annual peak consumption of the test building by lighting
schedule and system.

79!
!

!

Figure 4-3: Annual Energy Consumption of Test Building

80!
!

!

4-4: Peak Energy Consumption Consumption of Test Building

81!
!

!

Figure 4.3 shows the impact of lighting schedule for the test building. A small amount of
energy use is solely attributable to lighting in the 0% case due to use of default schedules (Table
4.3) during June, July, and August. As lighting use increases other parameters decrease, on a
percent basis. The impact of lighting use ranges from 6% to 57% of total annual electrical energy
use for the 0% and 100% cases, and varies 6% for intermediate situations. Default schedules
show a similar behavior for intermediate testing ranges. Consequently using default schedules
may be a safe approach for staying at intermediates energy prediction values.
The annual peak demand by end-use was similar for all the scenarios except for the 0%
case. This makes sense if the concept of annual peak demand is considered, where even though
total electric consumption from lighting is not the same for all the cases, at certain times of day,
every fixture will be switched on, causing a peak demand for electricity. At this same time of
day, the 0% case doesn’t show a peak demand since lighting was specified as turned off.
The analysis shows the ranges of variations in prediction, but doesn’t answer the
question of how useful a detailed schedule will be at approximating actual energy use. As a first
conclusion, lighting has a significant impact on energy and electricity use in simulation. The
analysis also shows that at assumed intermediate lighting intensities, the energy use is similar to
that from default schedules defined by the software.
4.3 Building Case Studies
After boundary testing of the impact of lighting on simulations three case studies were
selected. Through measurements and observations in the buildings, lighting use patterns were
recorded and then defined in simulation models. Along with observed and measured schedules,
82!
!

!

the simulations were run with default schedules within the program and research schedules
constructed from observations and measured data for each building. Energy predicted from
simulations was compared to measured utility data, and also compared among the schedules.
Architectural drawings were used to determine information necessary for describing wall types
and mechanical and electrical systems for each case study building. Additionally, the researcher
met with the design engineer or the building maintenance manager. Each building is described
separately below.
4.3.1

Chemistry Building
2

The Chemistry addition building is a 32,034 ft structure built to expand the original
Chemistry building built in 1963 at Michigan State University, East Lansing, Michigan. This
eighteen million dollar project was developed in 2008 and it achieved LEED Silver certification
under LEED NC 2.2. It consists of five above grade floors and two below grade floors. Functions
of rooms by floor in the building are described in Table 4.7.

83!
!

!

Fifth&Floor

1"Graduate"Study"Room
1"Instrument"Lab
1"Conference"/"Seminar"Room
1"Administrative"Office
Hallway
Restrooms
Stairways
Fourth&Floor 1"Main"Office
1"Conference"/"Seminar"Room
1"Mail"/"Copy"Room
Hallway
Restrooms
Stairways
Fourth&Floor 3"Computer"Offices
1"Business"Office
1"Graduate"Office
1"Teaching"Lab
Hallway
Restrooms
Stairways
Second&Floor 3"Classrooms
Hallway
Restrooms
Stairways
First&Floor
General"Chemistry"Office
1"Classroom
1"Copy"Room
Hallway
Restrooms
Stairways
Basement
3"Teaching"Labs
1"Classroom
Hallway
Restrooms
Stairways
Sub&Basement Mechanical"Room
Stairways
Table 4-7: Room Distribution in Chemistry Addition Building
84!
!

!

Figure 4-5: Chemistry Building

Figure 4-6: Typical Floor Plan of Chemistry Addition Building

85!
!

!

The Chemistry building has three different types of wall construction: a main curtain wall
façade, masonry walls at the East and West stair cases, and a steel louver system at a utility
duct wall. The roof uses pavers over polystyrene insulation and roof membrane. Detailed
descriptions of construction assemblies are provided in Appendix B.

Figure 4-7: Chemistry Building Model Screenshot
The chemistry building HVAC system uses a chilled water air handler unit with variable
air volume (VAV) boxes and hot water reheat coils. A schematic generic design of a VAV
system is shown in Figure 4.7. The cooling coil is fed by an absorption chiller that distributes

86!
!

!

water to other buildings in addition to the Chemistry Building. Stair cases are not air conditioned
but have unit ventilator fans with hot water coils to provide heat.

87!
!

!

VAV
BOX

Reheat
Coil

VAV
BOX

Reheat
Coil

Exhaust
Air

Figure 4-8: Schematic VAV with Reheat Coil System (NREL, 2006)
88!
!

VAV
BOX

Reheat
Coil

!

Chemistry Building Data and Results
As stated before, the Chemistry Addition underwent the LEED certification process and
an energy model was developed for LEED documentation. This model was reviewed and results
were compared to predictions obtained from models developed for the research.
The first schedule considered was the lighting profile used for the original model
developed during the design phase and for LEED certification of the project. Schedules showed
throughout the research represent either wattage percentages or percentage of fixtures turned on.
This schedule was used for all spaces regardless of their use: classrooms, offices, bathrooms,
hallways and conference rooms. The same schedule was used throughout the year without break
occupancy variations considered during break periods. Lighting use is indicated as a percentage
of light use, where 0 represents all lights switched off and 1 represents 100% of lights turned on.
The schedule used for the original design model is shown in Table 4.8:

89!
!

!

Hour
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10
10-11
11-12
12-13
13-14
14-15
15-16
16-17
17-18
18-19
19-20
20-21
21-22
22-23
23-24

Monday
0.05
0.05
0.05
0.05
0.05
0.05
0.2965
0.7980
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.798
0.5005
0.2965
0.2965
0.1010
0.1010
0.05
0.05

Tue sday
0.05
0.05
0.05
0.05
0.05
0.05
0.2965
0.7980
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.798
0.5005
0.2965
0.2965
0.1010
0.1010
0.05
0.05

We dne sday Thursday
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.2965
0.2965
0.7980
0.7980
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.798
0.798
0.5005
0.5005
0.2965
0.2965
0.2965
0.2965
0.1010
0.1010
0.1010
0.1010
0.05
0.05
0.05
0.05

Friday
0.05
0.05
0.05
0.05
0.05
0.05
0.2965
0.7980
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.798
0.5005
0.2965
0.2965
0.1010
0.1010
0.05
0.05

Saturday
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015

Sunday
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015

Table 4-8: Schedule Used in Chemistry Addition Model during Design Phase

90!
!

Holiday
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015
0.015

!

The software can define a set of default schedules based on what the user defines as the
main function of the building. Tables 4.9 and 4.10 show the default lighting schedules within the
software for academic buildings. For the energy model developed during this research, break
periods were defined for the academic year to simulate occupancy variation. Two default
schedules were created, one which considered break periods and one which didn’t. These default
schedules were used for the first run of the model to obtain initial results and subsequent
simulations were run with the research schedules.

91!
!

!

Hour
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10
10-11
11-12
12-13
13-14
14-15
15-16
16-17
17-18
18-19
19-20
20-21
21-22
22-23
23-24

Monday

Tuesday

0.05
0.05
0.05
0.05
0.05
0.05
0.0512
0.0535
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.0535
0.0521
0.0512
0.0502
0.0502
0.0502
0.05
0.05

0.05
0.05
0.05
0.05
0.05
0.05
0.0512
0.0535
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.0535
0.0521
0.0512
0.0502
0.0502
0.0502
0.05
0.05

Wednesday Thursday
0.05
0.05
0.05
0.05
0.05
0.05
0.0512
0.0535
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.0535
0.0521
0.0512
0.0502
0.0502
0.0502
0.05
0.05

0.05
0.05
0.05
0.05
0.05
0.05
0.0512
0.0535
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.0535
0.0521
0.0512
0.0502
0.0502
0.0502
0.05
0.05

Friday

Saturday

Sunday

Holiday

0.05
0.05
0.05
0.05
0.05
0.05
0.0512
0.0535
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.0535
0.0521
0.0512
0.0502
0.0502
0.0502
0.05
0.05

0.05
0.05
0.05
0.05
0.05
0.05
0.0512
0.0535
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.0535
0.0521
0.0512
0.0502
0.0502
0.0502
0.05
0.05

0.05
0.05
0.05
0.05
0.05
0.05
0.0512
0.0535
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.0535
0.0521
0.0512
0.0502
0.0502
0.0502
0.05
0.05

0.05
0.05
0.05
0.05
0.05
0.05
0.0512
0.0535
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.054
0.0535
0.0521
0.0512
0.0502
0.0502
0.0502
0.05
0.05

Table 4-9: Default Software Schedule for Unoccupied Periods in Academic Buildings
92!
!

!

Table 4.9 shows a maximum 5.4% of total installed lighting wattage used during peak
periods. The chemistry building has many office spaces that continue to be occupied during
academic break periods. Consequently, defining office spaces with the default schedule shown in
Table 4.9 results in an underestimation of lighting use.

93!
!

!

Hour
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10
10-11
11-12
12-13
13-14
14-15
15-16
16-17
17-18
18-19
19-20
20-21
21-22
22-23
23-24

Monday
0.05
0.05
0.05
0.05
0.05
0.05
0.1605
0.424
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.713
0.611
0.5005
0.679
0.679
0.6705
0.288
0.1010
0.05

Tuesday
0.05
0.05
0.05
0.05
0.05
0.05
0.1605
0.424
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.713
0.611
0.5005
0.679
0.679
0.6705
0.288
0.1010
0.05

Wednesday Thursday
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.1605
0.1605
0.424
0.424
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.713
0.713
0.611
0.611
0.5005
0.5005
0.679
0.679
0.679
0.679
0.6705
0.6705
0.288
0.288
0.1010
0.1010
0.05
0.05

Friday
0.05
0.05
0.05
0.05
0.05
0.05
0.1605
0.424
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.713
0.611
0.5005
0.679
0.679
0.6705
0.288
0.1010
0.05

Saturday
0.05
0.05
0.05
0.05
0.05
0.05
0.1605
0.424
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.713
0.611
0.5005
0.679
0.679
0.6705
0.288
0.1010
0.05

Sunday
0.05
0.05
0.05
0.05
0.05
0.05
0.1605
0.424
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.713
0.611
0.5005
0.679
0.679
0.6705
0.288
0.1010
0.05

Table 4-10: Default Software Schedule for Occupied Periods in Academic Buildings
94!
!

Holiday
0.05
0.05
0.05
0.05
0.05
0.05
0.1605
0.424
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.713
0.611
0.5005
0.679
0.679
0.6705
0.288
0.1010
0.05

!

A research schedule, shown in Table 4.11, was defined from observations conducted in
the building. The purpose was to see if schedules developed from observations of occupancy
could improve the accuracy of simulations from those developed during the design stage. The
process for gathering field information included the following:
•

Access to the building was requested from the building manager for after-hours operation
observation

•

A preliminary walkthrough of the building was conducted to determine if all rooms were
observable from the interior of the structure

•

From engineering drawings, lighting fixtures were counted. No specific fixture model was
specified on the drawings so the following assumptions were made according to lamp type:
o 2’x4’ troffer – 65 Watts with electronic ballast (2 tubes assumed)
o 4’x4’ troffer - 130 Watts with electronic ballast (4 tubes assumed)
o Downlight – 50 Watts
o Wall Mounted – 50 Watts

•

Observations inside the building were carried out during a span of one week, for several
hours per day until all 24 hours were observed

•

During walkthroughs, fixtures which were switched on were counted, and percent lighting
use determined. If three out of four lamps were on, the lighting use for that space was taken
to be 75%

95!
!

!

•

Spaces that had closed or locked doors typically still had a window which allowed the space
to be observed

•

The instrument lab on the fifth floor was the only inaccessible room due to safety concerns.
However, this room has an exterior window which allowed observations of lighting use from
the outside of the building

•

Hallways, bathrooms, and stairs were considered together throughout all floors and were
assumed to use the same schedule

•

Lighting loggers were installed in three rooms: a classroom in the basement, a study room in
the basement, and a computer office. These were intended to provide additional information
for these rooms and were considered when defining schedules for simulation but because
there were not enough sensors for installation in every room they could only provide
supplementary information to the observations

96!
!

!
HOUR OF THE DAY
Rooms

Floor

0

1

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16 17 18 19 20 21 22 23

Study Room

Basement

0

0

0

0

0

0

0

0

0 100 100 100 100 100 100 100 100 100 100 0

0

0

0

0

Classroom

Basement

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

Office

Basement

0

0

0

0

0

0

0

0

0 100 100 100 100 100 100 100 100 100 0

0

0

0

0

0

General Office

1st

0

0

0

0

0

0

0

0

90 100 100 100 100 75 100 100 100 0

0

0

0

0

0

Classroom

1st

0

0

0

0

0

0

0

0

0

0 100 100 100 0

0

0

Copy Room

1st

0

0

0

0

0

0

0

50 50 50 50 50 50 50 50 50 50 50 50 50 50

0

0

0

Undergrad Office

1st

0

0

0

0

0

0

0

0

0

0

0

0

Classroom

2nd

0

0

0

0

0

0

0

0

0 100 100 100 100 33

0 100 100 100 33 33 66 33 33

0

Computer Office

3rd

0

0

0

0

0

0

0

90 90 90 90 90 90 90 90 80 80

0

0

0

Bussiness Office

3rd

0

0

0

0

0

0

0 100 100 100 100 100 100 100 100 100 100 100 100 100 100 0

0

0

Graduate Office

3rd

0

0

0

0

0

0

0

0

0

M ain Office

4th

10 10 10 10 10 10 10 60 80 90 100 100 90 100 100 95 95 90 65 80 50 25 50 25

Conf./ Sem. Room

4th

50

0

0

0

0

0

0

Grad. Study Room

5th

0

0

0

0

0

0

0 100 100 0

Instrument Lab

5th

100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100

Conf./ Sem. Room

5th

50 50 50 50 50 50 50 50 50 25 25 25 25 100 100 0

Hallways

B-5

50 50 50 50 50 50 50 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100

Bathrooms

B-5

100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100

Stairway

B-5

100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100

0 100 100 0

0

0

0

0

0 100 100 100 100 0

0

0

0

0 100 100 100 100 100 100 100 100 0

0 100 100 100 100 100 100 100 50 50
0

0

0
0

0
0

0
0
0

0
0
0

0

0 100 100 50 100 75 75 100 50 50 100 50 50 50 50 50
0

0

50 100 100 0

0

0

0

0

0

0

0

0

50 100 100 100 50 100 75 50

Average % 24 22 22 22 22 22 22 45 56 71 82 85 84 85 85 72 72 63 55 48 46 32 32 28

Table 4-11: Percentage of Lighting Fixture Use Schedule from Observations by Zone in the Chemistry Building
97!
!

!

Figure 4-9: Building Average Lighting Profile v Default Lighting Schedule Profiles in
Chemistry Building
Table 4.11 shows the number of fixtures turned on at the moment of observation. All
rooms were observable enabling fixture counts for each room. For example, in the hallways,
there were fourteen fixtures installed which were all on from 7:00 a.m. to 11:00 p.m., and seven
of them were off from 12:00 p.m. to 6:00 a.m., yielding 100% use and 50% use respectively.
Average lighting use was calculated for each hour and a profile was developed as shown in Table
4.11 and Figure 4.8 labeled as average. The model developed during the research, called the
“Research Model”, used three types of schedules: one using individual schedules for each room
as recorded in Table 4.11 during observations, a second one using a single schedule for the
whole building defined by the average use percentages calculated from hourly observations in
Table 4.11 and a third one using default schedules within the program (Tables 4.9 and 4.10).
Total predicted annual electric consumption from these three cases and from the original model
developed during the design phase were compared to actual measured electric consumption from

98!
!

!

the control room at Physical Plant during the 2010 year. MSU Physical Plant monitors
approximately fourteen million square feet of space at this central control room.
The models developed during the design phase and the research models were developed
for different reasons and used different approaches, so direct comparison of these models
involved many variables besides lighting variations that makes comparisons difficult. The
original model substitutes metered chilled water for the chiller and cooling tower loads,
consequently cooling loads do not show in the electric consumption predictions. Miscellaneous
loads were lower compared to those considered in the research model. The original model did
not consider seasonal variations, and used the same schedules throughout the year, while the
research model used different schedules for the academic year and break periods. Table 4.12
shows annual electric consumption predicted by the original design phase model. Table 4.13
show annual electric consumption using program default schedules and again without
consideration of break periods.

99!
!

!

Original(Model

Electricity)kWh)
(x000)
0
0
0
0
18.61
9.9
29.07
102.46
160.04
214
25.2%

)Space)Cool
)Heat)Reject.
)Space)Heat
)Hot)Water
)Vent.)Fans
)Pumps)&)Aux.
)Misc.)Equip.
)Area)Lights
)Total
Measured)Utilities
%)Difference

Table 4-12: Annual Electric Consumption Predicted Measured Utility by Original Model in
Chemistry Building
Research(Model(/(
Default(Schedules(
Without(Break(Periods

Electricity)kWh)
(x000)
)Space)Cool
2.07
)Heat)Reject.
7.88
)Space)Heat
0
)Hot)Water
0
)Vent.)Fans
23.57
)Pumps)&)Aux.
21.71
)Misc.)Equip.
184.32
)Area)Lights
64.03
)Total
303.58
Measured)Utilities
214
%)Difference
41.9%
Table 4-13: Annual Electric Consumption Predicted by Research Model without Break
Periods and Software’s Default Schedules
100!
!

!

To compare the original and research models, modifications were made to normalize
them and remove variables that cause differences between the models. Miscellaneous loads
included in the research model were added to the original model to balance this load. Tables 4.12
and 4.13 shows the difference is large. The chiller was substituted by a chilled water meter in the
research model to remove this load and to match the system to the original model. Also,
occupancy schedules were kept constant throughout the year to test results without lower
occupancy due to break periods. Results from these normalized schedules are shown in Table
4.14.
Normalized+Original+
Model+without+Break+
Periods+

Normalized+Research+
Model++with+Break+Periods+
and+Default+Schedules

Normalized+Research+
Model++without+Break+
Periods+and+Default+
Schedules

Electricity)kWh)(x000)

Electricity)kWh)(x000)

Electricity)kWh)(x000)

)Space)Cool
0
0
0
)Heat)Reject.
0
0
0
)Space)Heat
0
0
0
)Hot)Water
0
0
0
)Vent.)Fans
25.71
14.39
23.57
)Pumps)&)Aux.
12.47
5.91
6.55
)Misc.)Equip.
129.35
141.24
184.32
)Area)Lights
102.46
47.86
64.03
)Total
269.99
209.4
278.47
214
214
214
Measured)Utilities
%)Difference
26.2%
2.1%
30.1%
Table 4-14: Modified Original and Research Models to Normalized Conditions

The three models were developed in eQuest, although the original model was developed
under version 3-63 and the research model under version 3-64. The research model that
101!
!

!

considers unoccupied periods shows decreased electric use due to lower consumption during
break periods. The original model and research model using default schedules without
consideration of break periods predict electric consumption less accurately, but similarly with
269 MWh for the original model and 278 MWh for the research model.
To compare impacts of lighting variations in simulations, schedules were varied in the
research model. Simulation results were obtained considering break periods and also for constant
occupancy throughout the year. For each of these scenarios, three simulations were run: using
default schedules within the program (Tables 4.9 and 4.10), one with observed schedules (Table
4.11) and another with an average schedule obtained from observations (Table 4.11). Results are
displayed in Tables 4.15 and 4.16.
Research(Model(/(
Research(Model(/(
Research(Model(/(
Default(Schedules(
Observed(Schedules(
Average(Schedule(
Without(Break(Periods Without(Break(Periods Without(Break(Periods

Electricity)kWh)
Electricity)kWh)
Electricity)kWh)
(x000)
(x000)
(x000)
)Space)Cool
2.07
2.07
2.07
)Heat)Reject.
7.88
7.9
7.9
)Space)Heat
0
0
0
)Hot)Water
0
0
0
)Vent.)Fans
23.57
24.27
23.7
)Pumps)&)Aux.
21.71
21.72
21.71
)Misc.)Equip.
184.32
184.32
184.32
)Area)Lights
64.03
75.86
66.66
)Total
303.58
316.14
306.36
Measured)Utilities
214
214
214
%)Difference
41.9%
47.7%
43.2%
Table 4-15: Annual Electric Consumption in Chemistry Building by Different Lighting
Schedules and no Break Periods

102!
!

!
Research(Model(/(
Default(Schedules(With(
Break(Periods

Research(Model(/(
Research(Model(/(
Observed(Schedules( Average(Schedule(With(
With(Break(Periods
Break(Periods

Electricity)kWh)
Electricity)kWh)
Electricity)kWh)
(x000)
(x000)
(x000)
)Space)Cool
1.72
1.72
1.72
)Heat)Reject.
6.02
6.09
6.03
)Space)Heat
0
0
0
)Hot)Water
0
0
0
)Vent.)Fans
14.39
14.69
14.43
)Pumps)&)Aux.
18.47
18.47
18.47
)Misc.)Equip.
141.24
141.24
141.24
)Area)Lights
47.68
55.08
49.46
)Total
229.52
237.29
231.35
Measured)Utilities
214
214
214
%)Difference
7.3%
10.9%
8.1%
Table 4-16: Annual Electric Consumption in Chemistry Building by Different Lighting
Schedules and Break Periods

Tables 4.15 and 4.16 show that not including break periods in the model has an impact on
electric use prediction. Energy predictions differ over 40% in all cases where constant yearly
occupancy is considered, while cases with break periods considered were within 11% of
measured utilities. Comparing the different schedules used in Table 4.16, use of the default
schedule predicted electric use slightly more accurately with a 7.3% difference compared to
10.9% with observed schedules. A comparison of annual electric consumption prediction and
monthly electric consumption prediction is shown in Figures 4.9 and 4.10.

103!
!

!

Figure 4-10: Comparison of Annual Electric Consumption in Chemistry Building

Figure 4-11: Monthly Electric Consumption by Different Models and Lighting Schedules in
Chemistry Building

104!
!

!

Additional Information for Chemistry Building
Additional information for classroom spaces was also available for the Chemistry
building from Michigan State University (MSU). MSU similar to most universities tracks
classroom space utilization. This information was considered along with the observations in
forming the research model. Table 4.17 shows an example of reports from MSU that may be
available from other owners.

105!
!

!

Capacity(
Group(
30&39!
!
!
!
!
!
`!
!
!
!
!
!
!
!
!
!
!

Room(

BCC!N124!
BCH!111!
BH!106A!
BH!106B!
BH!120!
BH!12C!
CEM!085!
CEM!109!
CEM!110!
CEM!183!
CEM!281!
CEM!283!
CEM!287!
COM!165!
COM!171!
COM!173!
COM!175!

Total(
Capacity(

Enroll(
Limit(

Actual(
Enroll(

Actual(
Enroll(%(
of(
Capacity(

Actual(
Enroll(
%(of(
Limit(

Weekly(
Hours(

Use(
Hours(

Use(Hours(
%(of(
Weekly(

468!
320!
240!
330!
455!
240!
720!
870!
780!
570!
930!
870!
690!
342!
340!
340!
238!

374!
272!
204!
276!
360!
208!
681!
792!
726!
534!
851!
838!
568!
235!
260!
206!
167!

316!
222!
135!
233!
283!
188!
657!
852!
732!
496!
909!
786!
539!
217!
214!
165!
155!

68%!
69%!
56%!
71%!
62%!
78%!
91%!
98%!
94%!
87%!
98%!
90%!
78%!
63%!
63%!
49%!
65%!

84%!
82%!
66%!
84%!
79%!
90%!
96%!
108%!
101%!
93%!
107%!
94%!
95%!
92%!
82%!
80%!
93%!

50!
50!
50!
50!
50!
50!
50!
50!
50!
50!
50!
50!
50!
50!
50!
50!
50!

40!
14!
24!
22!
40!
28!
28!
30!
28!
26!
31!
31!
31!
26!
29!
30!
24!

80%!
28%!
48%!
44%!
80%!
56%!
56%!
60%!
56%!
52%!
62%!
62%!
62%!
52%!
58%!
60%!
48%!

Table 4-17: Information from Campus Monitoring by Capacity
106!
!

!

Figure 4.17 shows weekly hours of use for classrooms as well as expected occupancy.
Twenty eight hours are recorded for classroom use in CEM 085, which corresponds to the
basement classroom observed in table 4.11. Assuming a classroom with a 500 W lighting load
installed would yield: 28 hours x 500 W @ 100% use = 14 kWh per week. The same approach
was used with the observed classroom information from table 4.9 and results are summarized
below. A 500 W load per class was also estimated:

107!
!

!
0
Class%2nd%Floor
Class%1st%Floor
Class%Basement
Average%Use%(%)
Consumption%(kWh)

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

0

0

0

0

0

0

0

0

0

100

100

100

100

33

0

100

100

100

33

33

66

33

33

0

0

0

0

0

0

0

0

0

0

0

0

100

100

100

100

0

0

0

100

100

100

0

0

0

0

0

0

0

0

0

0

0

0

0

100

100

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

11

11

0

0

0

0

0

0

0

0

0

0 0.17 0.33 0.50 0.33 0.22 0.17 0.17 0.17 0.17 0.22 0.22 0.28 0.06 0.06 0

33.3 66.7 100 66.7 44.3 33.3 33.3 33.3 33.3 44.3 44.3 55.3

Table 4-18: Load Analysis from Observed Schedules in Chemistry Building

108!
!

Total0(Watts)
3.052

!

Table 4.18 shows an average lighting use for classrooms observed in the Chemistry
Building, and this overall average is then multiplied by an assumed 500 Watt installed load. This
results in 3.052 kWh per day and 15.2 kWh per week. This is an 8% increase from the use
estimated from the 14 kWh obtained from campus information. As stated before, additional
factors should be considered in lighting use estimation, for example rooms are often used before
or after scheduled class periods or for meetings. Figure 4.11 shows information collected from
the Registrar’s Office for regularly scheduled events and unplanned events including study
sessions and meetings.

109!
!

!

Figure 4-12: Classroom Schedule in April from Registrar’s Office

110!
!

!

From Figure 4.11 a schedule can be defined assigning to each hour the percentage of time
a classroom will be used. If the class is scheduled to be used from 9:10 a.m. to 10 a.m., then the
room will be scheduled to be used fifty minutes out of sixty, therefore 83% will be assigned.
Following this procedure for each day, and obtaining a weekly average use for each hour, the
following schedule was defined:

111!
!

!

Hour
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10
10-11
11-12
12-13
13-14
14-15
15-16
16-17
17-18
18-19
19-20
20-21
21-22
22-23
23-24

Monday
0
0
0
0
0
0
0
0
0
0.83
0.67
0.67
0.67
1.00
0
0.83
0
0
0
0
0
0
0
0

Tuesday Wednesday Thursday
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.83
0
0.83
0.83
0.83
0.83
0.67
0.67
0
0.67
0.67
0.5
0.67
0.67
0.67
0.67
1
0.67
0.67
0
0.67
0.83
0.83
0.83
0.83
0
0
0
0
0.5
0
0
0.33
0.5
0
0
0.5
0
0
0
0
0
0
0
0
0
0
0

Friday
0
0
0
0
0
0
0
0
0
0
0.67
0.67
0.67
0.5
0
0.83
0.5
0.5
0
0
0
0
0
0

Average
0
0
0
0
0
0
0
0
0.33
0.66
0.54
0.64
0.67
0.77
0.27
0.83
0.27
0.20
0.07
0.10
0.10
0
0
0

Table 4-19: Weekly Lighting Use Estimation from Registrar's Classroom Schedule
112!
!

!

Different profiles obtained using different approaches for classroom CEM 085 (basement
classroom) are shown in Figure 4.12. Comparing the profiles with the default schedule and
average building profiles, which predicted electric consumption more accurately, it was expected
that using Registrar’s profiles and the sensor log profiles would give more accurate predictions.
However the data seem to show that for overall building energy prediction it may be more
accurate to define an average occupancy schedule for all spaces with similar use than defining an
individual schedule for each individual room.

113!
!

!

Figure 4-13: Hourly Lighting Use Percentages Obtained Through Several Data Collection Methods in Basement Classroom at
Chemistry Building

114!
!

!

4.3.2

Geography Building
2

The Geography building was built in 1965 and its total area is 31,221 ft . It is located on the
Michigan State University Campus in East Lansing, Michigan. It consists of three floors, two
above grade, and one below grade. Function of rooms by floor in the building is shown in Table
4.20.

Second'Floor

First'Floor

Basement

27#Offices#
3#Computer#Labs
2#Work#Rooms
2#Server#Rooms
20#Offices
3#Conference#Rooms
2#Storage
1#Kitchen
1#Lounge
27#Offices
6#Labs
3#Media#Rooms
1#Computer#Lab
1#Storage#Room

Table 4-20: Room Distribution in Geography Building

115!
!

!

Figure 4-14: Geography Building

Figure 4-15: Typical Floor Plan of Geography Building
116!
!

!

Figure 4-16: Geography Building Model Screenshot
This building has a concrete masonry unit (CMU) construction for all walls and an insulated
concrete slab roof. Descriptions of construction assemblies are shown in Appendix B.
The HVAC system in the Geography building consists of two air handling units (AHU) fed
by one chiller. One AHU supplies cooling for the first and second floor and the other one
provides for the basement only. The AHU system provides only cooling and ventilation. Hot
water reheat coils provide heating during the winter.

117!
!

!

Subzone
Reheat
Coil
To other
zones
Cooling
Coil
Outside
Air

Supply
Air Fan

Return
Air

T

Subzone
Reheat
Coil

Supply Air Controller

T
Return
Air Fan
Relief
Air

Figure 4-17: Schematic Design of Constant Volume AHU with Reheat (NREL, 2006)
118!
!

Conditioned
Space
(Control
Zone)

!

Geography Data
There was no previous energy model of the Geography building, making this case
slightly different from Chemistry. Only a research model was developed and no original model
was available for comparison. Since it is a campus building, it was defined as an academic
building with the same default schedules represented in Tables 4.9 and 4.10. Procedures for
obtaining required information for the Geography building included:
•

An initial walkthrough was carried out to become familiar with the building and to
understand space distribution, and how measurements would be collected

•

Drawings provided included lighting fixture layouts for renovated areas, but not for nonrenovated areas. Consequently fixtures were counted and lamp types were recorded
during the first walkthrough
o Lighting fixtures were recessed fluorescent troffers, at 32 Watts per tube. Some
lamps had four tubes, other had two tubes, and some just one tube

•

Walkthroughs were conducted during a span of one week, and covered all 24 hours of the
day. Access to the building was requested from and granted so that the researcher had
access to the building at all times

•

Most of the spaces were closed and locked during the visits. Access to some office
spaces, conference rooms, or computer labs access was not granted. These spaces were
counted as “on or off” and the researcher could not determine exactly how many lamps
were switched on, but a small window allowed observations of the interior. Office spaces
were counted and the percentages of use were defined as the number of spaces using
119!

!

!

light. It is worth mentioning that sometimes offices were occupied but with the lights off,
due to good daylighting within the space
•

In spaces where access was allowed, lamps were counted to determine the percentage of
lights on and to define use profiles

•

Basement spaces were not as easily observable, as some of the rooms were not only
locked but behind another door. Also, some doors did not have a window through which
observations could be made. These spaces were considered as “on” if they were open,
and “off” if unoccupied and closed

•

Lighting loggers were not used in this building

•

Results in Table 4.21 are shown as percentages. Use percentages were calculated as
follows:
o Hallways (1

st

Floor, 2

nd

Floor, and Basement) and bathrooms: fixtures were

counted, and the percentage represents number of fixtures on to the total of
fixtures installed. For example, out of twenty six lamps in the basement hallway,
twenty two were on so 85% use is defined
nd

o Perimeter offices, Internal Offices (1st Floor, 2

Floor, and Basement): use

percentage was defined as the total of spaces observed on to the total of existing
spaces; for the first floor, there are fourteen offices in the perimeter, if four were
observed to be on, then 28% was defined as the lighting use. Spaces that were
unobservable were considered as lights off. Some spaces were closed but it was
possible to determine if the lights were on or off
120!
!

!

o Perimeter conference room, lounge, kitchen, server room, and internal work
rooms: these spaces were observable all times so the same procedure as for
hallways was applied
o Computer labs and internal conference rooms: these were spaces with restricted
access, so only when a door was open could lights be observed. One of three
computer labs was opened at all times, so only when the other labs were opened
could a full count could be made, otherwise, these closed labs were considered as
off
•

Bathrooms were considered as a single space throughout the building; hallways were
grouped for the first and second floor and considered individually for the basement due to
asymmetry in design and differences in fixture quantities

121!
!

!

HOUR OF THE DAY
Rooms

Floor Qty

0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Perim. Offices

1st

14

0

0

0

0

0

0

0

7

29

43

36

43

50

50

57

29

21

21

0

7

0

0

7

7

Intern Offices

1st

5

0

0

0

0

0

0

0

0

0

0

20

20

20

20

20

20

20

20

0

0

0

0

0

0

Perim. Conf. Room

1st

1

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

Intern. Conf. Room

1st

2

0

0

0

0

0

0

0

0

0

100 100 75

75

45

0

0

0

0

40

25

25

75

0

0

Intern Work Rooms

1st

1

0

0

0

0

0

0

0

50

0

50

50

50 100 100 100 100 100

0

0

0

0

0

0

0

Lounge/ Kitchen

1st

1

0

0

0

0

0

0

0

0

0

0

0

0

100

0

0

0

0

0

0

0

0

0

0

0

Perim. Offices

2nd

28

0

0

0

0

0

0

0

18

32

32

39

32

32

36

29

32

25

18

18

7

4

7

9

9

Computer Labs

2nd

3

50

50

50

50

50

50

50

50

0

50

75

50

40

50

50

50

50

50

50 100 50

45

45

20

Intern. Work Rooms

2nd

2

0

0

0

0

0

0

0

0

100 100 50 100 50

50

0

0

0

0

0

0

0

0

0

0

Server Rooms

2nd

2

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

Bathrooms

Gen

4

100 50

50

50

50

50

50

50

0

100 100 100 100 50

0

100 100 100 100

0

50

50

25

25

B

35

0

0

0

0

0

3

9

17

29

14

9

6

6

3

3

Offices
Hallways

1+2

Hallways

B

0

0

0

0
14

0
23

0
26

40

17

11

11

100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100
10

10

10

10

10

10

10 100 100 10

10

10

10

85

85

85

85

85

85

85

85

85

85

85

AVG 19

15

15

15

15

15

15

42

43

50

45

34

38

37

29

29

24

23

26

20

18

27

26

43

Table 4-21: Lighting Fixture Use Schedule by Zone in Geography Building

122!
!

!

Figure 4-18: Geography Average Lighting Profile from Observations v Software Default
Schedule
The average lighting profile shows much lower lighting use compared to the Chemistry
building, due to the nature of the building. The Geography building has a high percentage of
offices, as the majority of the spaces are intended for faculty and administrative use, while the
Chemistry has a broader space distribution including laboratories, offices, classrooms,
conference rooms, and study rooms.
In addition to the default schedules (Tables 4.9 and 4.10), two schedules were defined for
the research model from Table 4.18. Individual schedules for each room as well as a schedule of
average hour by hour use profile for the entire building were created. Figure 4.17 compares
hourly profiles from default schedules with averaged observed schedule. Results from
simulations are presented in Table 4.22 and it shows a comparison against actual metered
electrical use from the Physical Plant throughout the 2010 year. No original or previously
developed model exists so this analysis was not carried out in the Geography building.
123!
!

!

Research(Model(/(
Default(Schedules(

Research(Model(/(
Observed(Schedules(

Research(Model(/(
Average(Schedule(

Electricity)kWh)
(x000)

Electricity)kWh)
(x000)

Electricity)kWh)
(x000)

)Space)Cool

46.6

89.46

45.33

)Heat)Reject.

2.54

4.47

2.51

)Space)Heat

0

0

0

)Hot)Water

0

0

0

)Vent.)Fans

69.48

124.7

68.53

)Pumps)&)Aux.

6.97

11.7

6.76

)Misc.)Equip.

12.87

12.87

12.87

)Area)Lights

64.4

103.08

40.12

Measured)Utilities

202.86
741

346.28
741

176.12
741

%)Difference

72.6%

53.3%

76.2%

)Total

Table 4-22: Annual Electric Consumption Predicted v Measured Utility Data for
Geography Building
Table 4.22 shows a large difference of simulation energy use against measured utilities.
This difference resides in the air handling unit in the basement working on 100% outside air,
consequently large heating and cooling loads result from conditioning intake air flow during
winter and summer. This was not considered in the research model at first. A second research
model was developed to include these loads and the results obtained are presented in Table 4.23.
The software program does not model these types of systems directly, so an alternative method
had to be developed to be able to reflect these loads. The system was sized to move 12,000 CFM
of outside air according to guidance from the department of Engineering and Architectural
Services at Michigan State University.

124!
!

!

Research(Model(/(
Research(Model(/(
Research(Model(/(
Default(Schedules(w/( Observed(Schedules(w/( Average(Schedule(w/(
100%(OA(
100%(OA(
100%(OA

Electricity)kWh)
Electricity)kWh)
Electricity)kWh)
(x000)
(x000)
(x000)
)Space)Cool
186.37
246.89
182.29
)Heat)Reject.
9.47
10.75
9.38
)Space)Heat
0
0
0
)Hot)Water
0
0
0
)Vent.)Fans
227.64
313.25
228.89
)Pumps)&)Aux.
35.18
48.47
34.74
)Misc.)Equip.
12.87
12.87
12.87
)Area)Lights
64.4
103.08
40.12
)Total
535.93
735.31
508.29
Measured)Utilities
741
741
741
%)Difference
27.7%
0.8%
31.4%
Table 4-23: Annual Electric Consumption Predicted with 100% OA Requirements v
Measured Utility Data for Geography Building

Table 4.23 shows increased cooling loads due to the 100% outside air which must be
cooled during summer or heated during the winter. Miscellaneous loads and lighting loads
remained the same and the only variation was the quantity of air conditioned. Since the only
parameter varied was lighting schedule, impacts of lighting were able to be determined. Lighting
consumption based on observed schedules was calculated as 103 MWh annually, while for the
average and the default schedules were 64 MWh and 40 MWh annually, yielding a 200 MWh
difference in electric annual loads for the building among cases. A graphic comparison of these
predictions is shown in Figure 4.18. Observed schedules gave the most accurate prediction in this
case with a 0.8% difference from metered consumption. Default schedules yielded a 27%
difference and averaged schedules yielded a 31% difference from metered consumption.
125!
!

!

Figure 4-19: Comparison of Annual Electric Consumption in Geography Building
Figure 4.19 compares profiles obtained from the schedules used in the building model.
Peak consumption is shown during summer months, June to September which is influenced by
cooling loads. Heating is steam based so the load during winter months will mostly be reflected
in steam measurements, not in the electric components considered in this research. The average
schedules show a steady behavior throughout the year and follows a similar pattern to the default
schedules within the program.

126!
!

!

Figure 4-20: Monthly Electric Consumption by Lighting Schedules in Geography Building
4.3.3

Church Project
2

This building is a 4,050 ft new building and is intended for weekly worship meetings. The
building is located in Lansing, Michigan. The building was completed in 2010 and is pursuing
LEED certification. It is a one story building, and has the following rooms:
•

One main meeting room

•

One main Fellowship room

•

A kitchen

•

A library

•

2 smaller meeting rooms

•

A small attic
127!

!

!

Figure 4-21: Church Project Building

Figure 4-22: Floor Plan for Church Project

128!
!

!

The church project has 2x6 wood frame construction with R-19 insulation, and a
facebrick exterior layer. The roof uses a wood truss structure with R38 insulation. Roof and
wall construction is shown in Appendix B.

Figure 4-23: Church Project Building Model Screenshot
This project was not air conditioned but incorporates operable windows and natural wind
turbines for natural cooling. Its design includes two separate gas furnaces supplying different
zones. One furnace supplies the meeting zone and the other the fellowship zone. A schematic
design of the system is shown in Figure 4.23. The building also utilizes an energy recovery
unit which captures heat from exhaust air during cold months. Furnace fans provide fresh air
for occupancy by drawing in outside air.
129!
!

!

Figure 4-24: Schematic Design for Furnace System (NREL, 2006)
Church Project Data and Results
Data gathering for this project varied from the other case studies. Continuous observation of
spaces was not possible since there was limited access to the building so the researchers used
occupancy sensors to collect field data. The data collection for this project is described below:
•

Drawings were provided by the architecture and engineering team from the project. A
previously developed energy model was provided with this information package

•

Calendar records of scheduled meetings in the building were provided and were used as a
documentation of room use

•

The owner’s original predesign program statement indicating hours of operation and
occupancy level by room was reviewed
130!

!

!

•

Lighting sensors were installed in four rooms where occupancy was most significant, and
sensors measured the presence of people in the room, if the lights were on or off, and
lighting levels in the rooms

•

Sensors were used to record two weeks of data and then results were averaged for the
period

•

Electric utility records were provided by the owner for comparison to energy simulations
from the design phase and from measured schedules. These records included
measurements from July to December in 2010 and January to July 2011; predicted
electric consumption from research models was compared against available data for 2011
while the remaining months were assumed the same as 2010 data

The analysis for this building includes a previously developed model referred to as the
original model, and a model created for the research referred to as the research model. The
original model was created for the LEED certification process which requires use of ASHRAE
90.1 Appendix G, which requires buildings be modeled as fully air conditioned, even though no
air conditioning system was going to be installed. During the research two simulations were run
to determine electric consumption with the air conditioning system modeled and electric
consumption without the air conditioning. The intention of removing the air conditioning system
was to compare the original model to the research model, developed without an air conditioning
system.
Tables 4.24 to 4.29 present the schedules used for the research model. Table 4.24 shows
the default schedule within the software, used for all spaces in the building. Measurements
recorded with the light loggers for the Main Meeting room, a smaller meeting room, the library,
131!
!

!

and the Fellowship room are shown in Tables 4.25, 4.26, 4.27, and 4.28 respectively. Each of
these schedules was used for its corresponding room for the simulation while the other rooms
where simulated with the default schedule. The light loggers recorded data every twenty minutes,
with additional recordings every time movement was detected, so more than three recordings per
hour could be obtained from each logger. The loggers collected information on occupancy, light
level, and if lights were on or off. A sample logged session is shown in Appendix D. Once the
recordings were collected, schedules were defined as followed:
•

External light (sunlight, street lights) influenced readings on the sensor so lighting levels
were difficult to set as baseline to determine use percentage. Instead, schedules were
defined according to the “on” or “off” readings. If an event was recorded as “on”, 100%
lighting use was assumed

•

The number of “on” readings for every hour were counted, and a percentage of “on” to
the total number of readings during that hour defined the light use percentage. For
example, if the Fellowship room logged six recordings from 12:00 p.m. to 1:00 p.m., and
two were recorded as “on”, then 33% was taken as the light use for that hour (two out of
six measurements within an hour the lights were on)

•

Loggers recorded periods where lights were “off” but showed the space as occupied. The
defined schedule would show 0% light use for those periods but this doesn’t imply the
rooms were not in use
The third schedule defined was an average of lighting use from all four sensors to define

a single profile for building lighting use. Each hour of every day was averaged, and finally
132!
!

!

values for every day were averaged into a single value. A composite schedule is shown in Table
4.29. This schedule was used for every room of the building during the third run of the
simulation.

Hour% Monday Tuesday Wednesday Thursday
051 0.035
0.035
0.035
0.035
152 0.035
0.035
0.035
0.035
253 0.035
0.035
0.035
0.035
354 0.035
0.035
0.035
0.035
455 0.035
0.035
0.035
0.035
556 0.035
0.035
0.035
0.035
657 0.035
0.035
0.035
0.035
758 0.128
0.128
0.128
0.128
859 0.204
0.204
0.204
0.204
9510 0.204
0.204
0.204
0.204
10511 0.204
0.204
0.204
0.204
11512 0.204
0.204
0.204
0.204
12513 0.204
0.204
0.204
0.204
13514 0.204
0.204
0.204
0.204
14515 0.427
0.427
0.427
0.427
15516 0.427
0.427
0.427
0.427
16517 0.427
0.427
0.427
0.427
17518 0.580
0.580
0.580
0.580
18519 0.569
0.569
0.569
0.569
19520 0.547
0.547
0.547
0.547
20521 0.460
0.460
0.460
0.460
21522 0.460
0.460
0.460
0.460
22523 0.460
0.460
0.460
0.460
23524 0.259
0.259
0.259
0.259

Friday
0.035
0.035
0.035
0.035
0.035
0.035
0.035
0.128
0.204
0.204
0.204
0.204
0.204
0.204
0.427
0.427
0.427
0.580
0.569
0.547
0.460
0.460
0.460
0.259

Saturday Sunday
0.040
0.050
0.040
0.050
0.040
0.050
0.040
0.050
0.040
0.050
0.040
0.050
0.040
0.050
0.128
0.407
0.210
0.730
0.482
0.824
0.482
0.900
0.482
0.900
0.659
0.577
0.720
0.220
0.720
0.220
0.720
0.220
0.659
0.220
0.543
0.161
0.598
0.050
0.638
0.050
0.550
0.050
0.550
0.050
0.550
0.050
0.312
0.050

Table 4-24: Software Default Schedule for Church Projects

133!
!

!

Hour% Monday Tuesday Wednesday Thursday
051
0
0
0
0
152
0
0
0
0
253
0
0
0
0
354
0
0
0
0
455
0
0
0
0
556
0
0
0
0
657
0
0
0
0
758
0
0
0.7
0.6
859
0.83
0.6
0
0
9510
1
1
0
0
10511
1
1
0
0
11512
1
1
0
0
12513
1
1
0
0
13514
1
1
0
0
14515
0.6
1
0
0
15516
0
1
0
0
16517
0
1
0
0
17518
0
1
0
0
18519
0.25
0.25
0
0
19520
0.2
0
0
0
20521
0
0
0
0
21522
0
0
0
0
22523
0
0
0
0
2350
0
0
0
0

Friday
0
0
0
0
0
0
0
0.8
0.167
0
0
0
0
0
0
0
0
0
0.25
0
0
0
0
0

Saturday
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.143
0
0
0
0
0
0

Sunday
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.25
0.4
0
0
0
0
0

Table 4-25: Average Values Recorded by Logger in Main Meeting Room at Church Project

134!
!

!

Hour% Monday
051
0
152
0
253
0
354
0
455
0
556
0
657
0
758
0
859
0
9510
0
10511
0
11512
0
12513
0
13514
0
14515
0
15516
0
16517
0
17518
0
18519
0
19520
0
20521
0
21522
0
22523
0
2350
0

Tuesday Wednesday Thursday
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.25
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

Friday
0
0
0
0
0
0
0
0
0.43
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

Saturday
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.6
0
0.43
0.43
0
0
0
0
0

Sunday
0
0
0
0
0
0
0
0
0
0.57
0.5
0
0
0
0
0
0
0
0
0
0
0
0
0

Table 4-26: Average Values Recorded by Logger in Meeting Room at Church Project

135!
!

!

Hour% Monday
051
0
152
0
253
0
354
0
455
0
556
0
657
0
758
0
859
0
9510
0
10511
0
11512
0
12513
0
13514
0
14515
0
15516
0
16517
0
17518
0.25
18519
0.5
19520
0.25
20521
0
21522
0
22523
0
2350
0

Tuesday Wednesday Thursday
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.25
0
0.25
0.66
0
0.66
0.5
0
0.66
0.5
0
0
0.25
0
0
0
0
0
0

Friday
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

Saturday
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.18
0.5
0.28
0
0
0
0

Sunday
0
0
0
0
0
0
0
0
0.33
0.63
0.5
0
0
0
0
0
0
0
0
0
0
0
0
0

Table 4-27: Average Values Recorded by Logger in Library at Church Project

136!
!

!

Hour% Monday Tuesday Wednesday Thursday
051
0
0
0
0
152
0
0
0
0
253
0
0
0
0
354
0
0
0
0
455
0
0
0
0
556
0
0
0
0
657
0
0
0
0
758
0.43
0
0
0.25
859
0.13
0
0
0.5
9510
0
0
0
0.66
10511
0
0
0
0.5
11512
0
0
0
0.5
12513
0
0
0
0.5
13514
0
0
0
0.66
14515
0
0
0
0
15516
0
0
0
0
16517
0
0
0
0
17518
0
0
0
0
18519
0
0
0
0
19520
0
0
0
0
20521
0
0
0
0
21522
0
0
0
0
22523
0
0
0
0
2350
0
0
0
0

Friday
0
0
0
0
0
0
0
0.54
0.5
0.66
0.66
0.5
0.5
0.57
0.66
0.63
0.22
0
0
0
0
0
0
0

Saturday Sunday
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.43
0.22
0.5
0.5
0.5
0.5
0.66
0.66
0.63
0.66
0.13
0.66
0
0.22
0
0.22
0
0
0
0
0
0
0
0
0

Table 4-28: Average Values Recorded by Logger in Fellowship at Church Project

137!
!

!

Hour% Monday Tuesday Wednesday Thursday

Friday

Saturday

Sunday

0:1
1:2
2:3
3:4
4:5
5:6
6:7
7:8
8:9
9:10
10:11
11:12
12:13
13:14
14:15
15:16
16:17
17:18
18:19
19:20
20:21
21:22
22:23
23:0

0
0
0
0
0
0
0
0.335
0.274
0.165
0.165
0.125
0.125
0.143
0.165
0.158
0.055
0.000
0.063
0
0
0
0
0

0
0
0
0
0
0
0
0.000
0.000
0.000
0.000
0.000
0.055
0.125
0.125
0.315
0.165
0.353
0.288
0.125
0
0
0
0

0
0
0
0
0
0
0
0.000
0.083
0.300
0.250
0.108
0.125
0.125
0.165
0.158
0.033
0.063
0.100
0.000
0
0
0
0

0
0
0
0
0
0
0
0.108
0.240
0.250
0.250
0.250
0.250
0.250
0.150
0
0
0.063
0.188
0.113
0
0
0
0

0
0
0
0
0
0
0
0
0.150
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.250
0.063
0
0
0
0
0

0
0
0
0
0
0
0
0.175
0
0
0
0
0
0
0
0
0
0
0.063
0.165
0.165
0
0
0

0
0
0
0
0
0
0
0.213
0.125
0.165
0.125
0.125
0.125
0.165
0.000
0.000
0.000
0.063
0.228
0.125
0.125
0.063
0
0

Weekly%
Average
0
0
0
0
0
0
0
0.119
0.125
0.161
0.149
0.123
0.133
0.151
0.122
0.126
0.072
0.113
0.141
0.075
0.041
0.009
0
0

Table 4-29: Building Average Schedule from Measurements for Church Building
Because the original model and the research model were developed under different
assumptions and different software, both models had to be modified so that parameters other
than lighting that could influence results were normalized, minimizing their impact. The software
used for the original model was HAP from Carrier Corporation and the research model was
developed with eQuest. The original model was developed under ASHRAE Appendix G
requirements due to LEED certification process. ASHRAE Appendix G requires the model to
138!
!

!

include air conditioning throughout the building even though it was not part of the project. The
research model did not include air conditioning, so the first step toward normalization was
removing the air conditioning system from the original model. Miscellaneous loads assigned in
the research model were also set as the same in the original model. Table 4.30 shows the original
model and both the modified original model and research model. The original model predicted
68 kWh consumed annually against 38.3 kWh and 38.98 kWh from the modified models,
showing the impact of including the uninstalled air conditioning systems in the original model.

Original(Model

Original(Model(
without(A/C

Normalized(
Research(Model(

Electricity)kWh)
(x000)

Electricity)kWh)
(x000)

Electricity)kWh)
(x000)

)Space)Cool

2.817

0

0

)Heat)Reject.

0

0

0

)Space)Heat

0

0

0

)Hot)Water

0

0

0

)Vent.)Fans

29.953

2.193

2.79

Ext.)Usage

0

0

1.09

)Misc.)Equip.

26.637

26.637

22.52

)Area)Lights

9.454

9.454

12.58

)Total
Measured)Utilities

68.861
16.71

38.284
16.71

38.98
16.71

%)Difference

312.1%

129.1%

133.3%

Table 4-30: Normalized Comparison of Models

139!
!

!

Research(Model(/( Research(Model(/( Research(Model(/(
Default(Schedules Measured(Schedules( Average(Schedule(

)Space)Cool
)Heat)Reject.
)Space)Heat
)Hot)Water
)Vent.)Fans
Ext.)Usage
)Misc.)Equip.
)Area)Lights
)Total
Measured)Utilities
%)Difference

Electricity)kWh)
(x000)
0
0
0
0
2.8
1.09
9.08
12.58
25.55
16.71
52.9%

Electricity)kWh)
(x000)
0
0
0
0
2.81
1.09
9.08
6.64
19.62
16.71
17.4%

Electricity)kWh)
(x000)
0
0
0
0
2.81
1.09
9.08
3.21
16.19
16.71
3.1%

Research(Model(/(
With(Original(
Model's(Schedules

Electricity)kWh)
(x000)
0
0
0
0
2.81
1.09
9.08
11.57
24.55
17.71
38.6%

Table 4-31: Predicted Electric Load by Research Model with Different Lighting Schedules
The research model was modeled without an air conditioning system, and with heating
and hot water as gas fired, no electric loads were estimated for cooling loads. Yet, ventilation fan
electric loads were obtained because fans for the heating system, ventilation system, and energy
recovery system still run on electricity. Miscellaneous loads were reduced from those in the
original model, according to conversations with the building administration about actual use
patterns within the kitchen, which is the room with the highest installed electric load. Lighting
schedule was the only parameter which varied in the research model. Predicted electrical energy
use using the average building schedule (Table 4.29) predicted electric consumption most
accurately. The default schedule (Table 4.24) estimates a much higher use than the measured
utilities because it assumes more hours of use in the building. Although observed schedules
(Tables 4.25 to 4.28) define low occupancy hours for the building, rooms that were not logged
were defined with the default schedules (Table 4.24) resulting in higher predicted electric
140!
!

!

consumption, yet lower than using solely default schedules in every room. Monthly consumption
for the models is shown in Figure 4.24.

Figure 4-25: Monthly Consumption by Different Simulation Runs
Table 4.31 and Figure 4.25 allow comparison of different schedules used for the
simulations. The average schedule used in the research model yielded consumption predictions
which most closely matched measured data. It is also noticeable that the data used by the
modeler based on ASHRAE Appendix G requirements overestimates energy use in the building.

141!
!

!

Figure 4-26: Annual Electric Consumption Predicted by Each Model
Default schedules in the church project (Table 4.24) show a greater use of lighting than
the estimated use by the building administration. Default schedules assume lighting use
throughout all 24 hours of the day, while the building is used only a limited number of hours per
week. This is why default schedule electric use prediction is 52.9% off against 3% when using
the average building schedule. It appears that average use of several rooms is a better approach
than individual schedules for each room.
Additional Information for Church Building
During the research, a predesign occupancy schedule was provided by the owner and
shown in Table 4.32. Although no simulation was run with this schedule, a comparison of
schedule profiles was conducted in order to draw conclusions about the accuracy of predesign
occupancy estimations to measured schedule occupancy. This is shown in Figure 4.26.
142!
!

!

The schedule from the owner’s predesign occupancy statement was defined under the
following assumptions: scheduled meetings are assumed to use 100% lighting; what is scheduled
as occasional use was assumed as three hour meetings, two times per week (out of five weekday
evenings) meetings, resulting in 40% lighting use; finally, very occasional use was defined as a
three hour meetings, one time per week, every two weeks resulting in one time out of ten days,
or 10%. The schedule is presented in Table 4.33.

LOCATION

SUNDAYS

SATURDAYS

WEEKDAY
EVENINGS

12:30 - 1:45 pm Meeting for Worship Occasional use Very occasional use Meeting Room
25-30 adukts jooined Weddings, Memorials,
50 people
by 20-25 children at Retreats - 75 people
1:15 pm
12:30 - 1:15 pm - 20
Occasional use - 75 Very occasional use Fellowship Room children for First Day
people
50 people
School
12:30 - 2:30 pm - 5
Kitchen
Occasional use
Occasional use
people
5-10 People 2hr.
12:30 - 1:30 pm - 5 2 Small Rooms
?
Meetings, 2-3 days a
10 children, adults
week
5-10 People 2hr.
9:30 - 11:00 pm Library
?
Meetings, 2-3 days a
Small Worship Group
week

WEEKDAY
BUSINESS HOURS

Very occasional use

Very occasional use
Very occasional use
Very occasional use

Very occasional use

Table 4-32: Predesign Occupancy Schedule Provided by Owner

143!
!

!

Hour Monday Tuesday Wednesday Thursday
0:1
1:2
2:3
3:4
4:5
5:6
6:7
7:8
8:9
9:10
10:11
11:12
12:13
13:14
14:15
15:16
16:17
17:18
18:19
19:20
20:21
21:22
22:23
23:24

0
0
0
0
0
0
0
0
0
0.5
0.5
0.5
0
0
0
0
0
0
0
0.5
0.5
0.5
0
0

0
0
0
0
0
0
0
0
0
0.5
0.5
0.5
0
0
0
0
0
0
0
0.5
0.5
0.5
0
0

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

Friday
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

Saturday Sunday
0
0
0
0
0
0
0
0
0
0
0
0
0.25
0.25
0.25
0
0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0
0
0
0
0
0.5
0.75
0
0
0
0
0
0
0
0
0
0

Weekly5
Average
0
0
0
0
0
0
0
0
0
0.143
0.143
0.143
0.107
0.143
0.036
0
0
0
0
0.143
0.143
0.143
0
0

Table 4-33: Schedule Defined from Owner's Predesign Occupancy Statement
From Table 4.33, Mondays and Tuesdays were planned as very occasional use, resulting
in 50% weekly use (it was assumed to be used once every two weeks) for three hours. Very
occasional use on Saturdays was assumed as one event per month resulting in 25% average per
week. For Sundays, the percentage was based on the minutes within an hour the room was
scheduled to be used. The meeting room was scheduled from 12:30 p.m. to 1:45 p.m., so 50%
(30 minutes) from 12:00 p.m. to 1:00 p.m. is defined and 75% (45 minutes) from 1:00 p.m. to
144!
!

!

2:00 p.m. A weekly average was then calculated from these assumptions and compared to
weekly averages from the sensor collected data. Differences from Table 4.25 (Main Meeting
Room Schedule from Loggers) can be seen, since during the time frames services are scheduled,
the loggers recorded the room as occupied but lights as “off”. During the services, lighting may
not have been necessary due to sufficient daylighting in the room.

145!
!

!

Figure 4-27: Average Daily Lighting Use Profiles for Various Methods in Main Meeting Room

146!
!

!

From Figure 4.26, the profile defined by the owner’s information is closer to the average
building schedule defined from logger measurements; data recorded with the logger present a
higher use than what the owner estimated, but the logger recorded lighting use which doesn’t
necessarily reflect actual occupancy of the room due to lights that are left switched on after the
room is used.
4.4

Chapter Summary
This chapter shows the information collected during the research and its analysis. Insights

from four energy modeling professionals were gathered through interviews and opinions from
their experience were collected. Then, a boundary test range was carried out to test impact of
lighting in a building modeled in compliance with ASHRAE 90.1 Energy Standard requirements.
Finally, it describes the process of observation through the Chemistry and Geography buildings
at Michigan State University and the Church project in Lansing, and how schedules were defined
from data gathered. Also, the process of recording occupancy and lighting use with data loggers
and how the schedules were defined according to this gathered information was described. The
main conclusions from interviews, boundary analysis and case studies include:
•

Energy modeling professionals agree that main sources of inaccuracy in energy
simulation are occupancy assumptions and system scheduling

•

The lighting use range test showed that boundary values for electric consumption from
lighting can vary by as much as 120%. The difference between default schedules and
intermediate schedules (8 and 12 hour) is within 10%, showing that use of default
schedules may be appropriate for average structures
147!

!

!

•

The Chemistry building case study showed fairly accurate predictions with 7.3% over
metered readings when using default schedules and 10.9% and 8.1% with observed and
averaged schedule definitions. For academic buildings defining occupancy break periods
is important and has an impact on energy prediction. Without this differentiation, the
research model predicted 40% more electric use than metered data

•

The Geography building case study showed electric consumption predictions differed
from metered data by 27.7% for default schedules, 0.8% for observed schedules and
31.4% for average schedules (after including the outside air requirements of the system).
Large differences in lighting loads seem to result in large differences in HVAC loads
which indicates the importance of educated assumptions when defining lighting
schedules

•

The Church case study project showed 52.9% increase over metered data when using
default schedules, 17.4% for measured schedules and 3.1% for averaged schedules. From
this project, it was concluded that use of ASHRAE Appendix G for energy analysis on
LEED buildings may introduce inaccuracies

•

Daylighting may influence actual lighting use in rooms, and should be considered when
analyzing the owner’s predesign occupancy schedule

•

From the case base buildings, even though the averaged schedules were not always the
option that predicted electrical consumption most accurately, they were generally more
accurate than default schedules and the levels obtained from observed use. Average
148!

!

!

schedules differed 10% in Chemistry and the Church project, and 31% in Geography.
Considering all parameters that vary in a project, averaging hourly use by space type in a
building may be the best approach for estimating lighting use
•

Even existing buildings are difficult to model and predict energy consumption, and there
are many interactions among parameters to isolate only one. Equipment such as pumps,
snow melting devices and elevators can yield large differences in miscellaneous loads.
Depending on assumptions made for various parameters, energy predicted can be
accurate for the wrong reasons and the model can mislead further energy analysis

•

LEED requirements for certification require following ASHRAE 90.1 Appendix G, may
require the modeler to make assumptions that don’t necessarily reflect the actual
building. ASHRAE Appendix G based models are intended to compare to a proposed
building to a minimum standard, but not necessarily to model actual performance

149!
!

!

CHAPTER 5
GUIDELINES, TOOLS, AND RECOMMENDATIONS

150!
!

!

GUIDELINES, TOOLS AND RECOMMENDATIONS
This chapter presents recommendations for steps relating to occupancy that can improve
accuracy when creating energy models. These recommendations and tools flow from the data
and analyses presented in Chapter 4 and are based on the interviews, observations, measurements
and simulations. The recommendations address using available information from owners,
breaking down spaces into groups with similar use, defining an averaged schedule for each of
these groups, and considering daylighting, occupancy sensors and automatic lighting controls
when defining lighting use schedules.
5.1 Guidelines and Recommendations
5.1.1

Breakdown of Spaces According to Use
Loads and energy modeling programs require the user to specify the main building use

when starting a new project. This specification for the building generates a list of default values
for several parameters such as outside ventilation requirements, occupancy schedules, and
lighting schedules. These default values are applied to the building as a whole regardless of
space distribution within the building. This means for example that office spaces in a school
building are assigned the same schedule as classrooms. Spaces could be broken down into
specific use classifications to improve accuracy. The following are detailed space use
descriptions used in the case studies of academic buildings, but could be expanded to reflect the
types of spaces that can be found within any structure:

151!
!

!

•

Classroom

•

Hallway

•

Bathroom

•

Faculty Office

•

Administration Office

•

Copy / Office Equipment Room

•

Study Room

•

Computer Room

•

Meeting / Conference Room

•

Laboratory

•

Kitchen / Break Room

•

Mechanical/ Electrical/ Storage

5.1.2

Grouping of Spaces with Similar Use Characteristics

From the Chemistry Building and the Church Project, schedules defined from averaging the
observed patterns predicted energy consumption which most closely matched metered data.
Recordings from sensors in individual rooms showed that scheduling based on an individual
space can result in underestimating weekly use, because there are unforeseen situations such as
lights being left on, people using the rooms after hours or people that occupy the room and don’t
use the full lighting load or no load at all. The following characteristics may make averaging
spaces better use approximations:

152!
!

!

a) Offices: depending on a space user, offices will be constantly occupied or occupancy will
be very variable. From observations, there was a noticeable difference between faculty
offices (single occupancy) and administrative offices (single or multi occupancy). Faculty
tends to occupy their spaces for shorter periods than administration offices, so decreasing
lighting use in these spaces is adequate. Secretarial offices and lobby spaces can be
considered to use 100% lighting during office hours, while single occupancy
administrative offices were observed to be occupied most of the time with short periods
of absence.
b) Classrooms and Study Rooms: these are very unpredictable spaces within a building.
Even though some information can be obtained for these rooms by requesting an owners
predesign occupancy schedule or from administrative offices as shown in chapter four,
events such as people staying after class, extra class sessions, study groups, and people
that don’t turn off the lights should be accounted for, but are difficult.
c) Laboratories: laboratories can include computer laboratories where access is granted to
any occupant, or laboratories where access is limited to specific occupants. Regardless of
the final use, these spaces are occupied whenever people need to use the labs, and an
average use should be a better approach than fixing weekly hours for lighting in these
spaces.
d) Common Spaces: we can include in this category restrooms, hallways, and stairways, or
spaces that are not controlled by the occupants of the building, and are controlled by
custodians or building administration. These spaces will have fairly constant lighting use
since safety of building users should be ensured. It was observed during night hours and
153!
!

!

times of low occupancy, light levels maybe reduced at times, so a 50% load reduction can
be assumed for some hours. Bathrooms tend to be spaces that are kept with lights on even
afterhours, consequently load reduction shouldn’t be considered unless the owner’s
requirements and design documents support this decision, for example when occupancy
sensing controls are used.
e) Conference and Meeting Rooms: these spaces usually need to be reserved in order to
avoid conflicting activities from groups within a building. People leaving the lights on
after using the room or having lights off for screen projections during meetings should
also be considered. Averaging the hours of use of these spaces will be a better approach
than assigning 100% use for expected use hours.
f) Religious or Mass Meeting Rooms: these rooms will show a low use percentage because
people occupy these rooms for specific weekly or monthly activities. This is not a place
of work that holds occupants for several hours per day, but people instead gather for
specific activities and times, although there might be general illumination. These spaces
will usually have well defined use hours and expected occasional events, so an analysis
similar to that shown in Table 4.28 can be helpful.
Table 5.1 shows a summary of spaces observed throughout the three base cases, and
considers an average of observed lighting use and hours of use from the Chemistry building,
Geography Building, and the Church project. For example, general office data averages
observations from Chemistry and Geography.

154!
!

!

Avg Hours Use Avg Lighting
Daily
% Used Daily
General Office
Personal Office

12
14
7
7
12
24

65%

24
24
4

Classroom
Study Room
Conference Room
Common Spaces
(Restrooms, Hallways)
Specialty Laboratory
Computer Laboratory
Mass Gathering Rooms

40%
15%
25%
30%
30%

100%
50%
10%

Table 5-1: Estimated Hours of Use and Percentages for Spaces Observed
5.1.3

Owner and/or Users Specify Schedules and Room Use

The owner should be the first source of guidance for definition of building use patterns. It is
expected that the owner will have an idea of how many hours rooms will be used. The energy
modeler should seek this information along with building policies regarding working schedules
with questions such as:
•
•

Are people expected to work after hours?

•

Are classes given on particular time schedules, or does this vary between semesters?

•

Are there regularly scheduled meetings every week in certain rooms?

•

Are there fixed meetings in the meeting rooms or are events planned weekly?

•

Can laboratories be accessed at any time or is entrance restricted?

•

!

What are the working hours for occupants?

Are there occasional and non-typical events?
155!

!

5.1.4

Request Space Use Tendency Information Available
University campuses are usually under continuous monitoring to control how spaces are

being used within buildings. These studies show occupancy numbers and hours of use for a
variety of spaces that could be used to help define a schedule. Information such as that presented
in section 4.3.1 and in Tables 4.17 and Figure 4.11 for the Chemistry Building can be useful.
5.1.5

Automatic Lighting Control
Automatic lighting control systems can help program desired use patterns. If automatic

controls are being planned as part of the design process, the model simulation should be
coordinated with the system to match its schedule. Not only timely control of lights can be
achieved, but intensities as well dimming features can be considered. This will define both
hourly control and wattage. Programmed schedules can be used in the simulation program, or the
modeler can recommend to the lighting control programmer the schedule to use in order to
achieve the owner’s desired requirements.
5.1.6

Daylight and Occupancy Sensors
When sensors are included as part of the design, uncertainty in schedules can be reduced.

Daylight sensors are set to attenuate lighting fixtures to a desired level of luminance within a
space according to incidence of natural light. Although the percentage of light used will be
dependent on how much external light enters through windows, percentages of light level used
throughout day lit hours can be determined with a daylight analysis. Better control can be
obtained from occupancy sensors assuring spaces will be turned off once people have left the
room. This will decrease uncertainty for unforeseen activities in rooms such as offices and
156!
!

!

classrooms, where once working hours are over, percentages can be defined as 0%. Some rooms
will still carry uncertainty due to unscheduled use, such as in study rooms and conference rooms.
However, including sensors in these rooms will support assumptions of lower percentage of use
per hour, since there is some certainty that if people leave the room, lights will be turned off.
5.1.7

Use of Default Schedules
Using standard default schedules in energy simulation programs may not be the best

approach occupancy behavior for a specific project. Using default schedules becomes convenient
when all spaces in a building share the same function and have similar occupancy patterns.
Example of these cases can be office buildings where all spaces are designed as working spaces
and have similar working hours. Using a single default schedule for mixed buildings where
spaces vary may not be accurate. Analysis of default schedules and their validity for a building
design should be carried out to avoid representing varying spaces with the same schedule, such
as office spaces in academic buildings.
5.1.8

Submetering and Calibration
If energy monitoring and analysis is part of the ongoing activities once the building is

completed, it’s important to monitor main end-use electric loads and consumption patterns. This
information will help in comparing results to the energy model and to calibrate the model. As the
model is more finely calibrated, more accurate results can be obtained and future predictions can
be more realistic. Lighting load circuits should be designed to be installed from the same electric
center so submetering can be easily installed.

157!
!

!

5.1.9

Integrated Design
Developing an energy model involves summing assumptions from many different fields:

from understanding optimal architecture and materials used to optimization of engineering
systems. Gaining team input on use patterns will be the best approach to an accurate model, and
will provide a good baseline that can be improved as utility measurements are obtained. The
following team members can help to provide the following inputs:
• Owner: occupancy schedules, lighting schedules, building policies, and design
requirements. These should set the energy efficiency goals, and the baseline energy
model created is a first step towards reaching this goal.
• Architect/ Interior Design: materials, windows, finishes, and structural characteristics that
can help influence thermal performance and aesthetics.
• Electrical Engineer: lighting design densities, plug load estimations, and information
regarding installed equipment and important loads in the building.
• Mechanical Engineer: HVAC system optimization, design features that help to reduce
thermal loads and utility costs. Design values used for some parameters become an
important input for the model, and it’s important to keep track of the design
assumptions made through the design phase.
5.2 Data Gathering Tool
The research showed some obstacles to gathering information for developing an accurate
model. One of the reasons is that building maintenance crews are not fully involved during the
design process. They typically have knowledge of how equipment actually works, but they lack
158!
!

!

knowledge of design decisions and parameters during this phase. Also, equipment models
specified in construction documents were different from the models installed, generally with
similar capabilities but probably with better pricing when the purchase was done. On the
Geography building, the HVAC system was shown in the drawings without specification of
models, design or sizing parameters.
It is important during the design phase, modeling to implement a data gathering tool
where decisions can be documented, so people that build the model or people conducting future
energy analysis and verification can have access to this information. Also, schedules that were
defined during the design process can be recorded and verified if necessary during the occupancy
period.
The proposed tool lists three main types of data to be collected. First, a summary of
available construction data and specific architectural details should be compiled. If desired,
individual details of construction assemblies, wall, roof, and floor construction can be attached.
Contact information for professionals responsible for design should also be included.

159!
!

!
Project!
Owner!
Architect!
Civil!Engineering!
Electrical!Engineering!
Mechanical!Engineering!
Energy!Modeler!
!
!
ARCHITECTURE)
!
!
FEATURES)
))
!
EXTERIOR)WALLS)
)
#)

!
Wall!Type!

1!
2!
3!
4!
5!

Wood!Studs!
Concrete!
Wood!Studs!
Metal!Studs!
Wood!Studs!

!
!
ROOF)
)
#!
!!
!!

!

!!
!!

Type!of!
Construction!

!
!

!
!

!
!

!
!
!
!
!
!
!
!
!

!
!
!
! Insulation!
type!/!R!
Value!
!!
!!
!!
!!
!!

!
!
!
!
Cont.!Ins.!R!Value!

!
!
!
!

!
!
!
!

!
!
!

!
!
!

Insulation!R!
Value!

!!
!!

!
!
!
!
!
!
!
!
!

!
!
!
!
!
!
!
!
!

!
!
!
!
!
!
!
!
!

!
!
!
!

!
!
!
!

!
!
!
!

!
!
!
!

!
!
!

!
!
!
!
!
!
!
!
!

!
!
!
!
!
!
!
!
!

!
!
!

!
!
!

!
!
!

Manufacturer!

Brand!

Detail!on!
Sheet!#!

!!
!!
!!
!!
!!

!!
!!
!!
!!
!!

!!
!!
!!
!!
!!

Brand!

!!
!!
!!
!!
!!

!
!
! Detail!
on!
Sheet!#!
!!
!!

!
!
!
Manufacturer!

!!
!!

!
!
!
!
!
!
!
!
!

!!
!!

!
!
!

!

!!
!!
!!
!!
!!

Comments!
!!
!!

Figure 5-1: Architectural Information Summary Chart
160!

Comments!

!

Figure 5-1 (cont’d)
WINDOWS)
!
)
!
#)
1)
2)
3)

Type!of!
Window!
Fixed!!
Fixed!!
Operable!

4)
5)
6)

!
!
Type!of!Door!!
#)
1)
Swinging!
2)
Non!Swing!
3)
Swinging!
4)
Overhead!
!
!
GROUND)FLOOR)

1)

!

Exposure!
Earth!Contact!

!

!
!

!

!

!

!!
!!
!!

!!
!!
!!

!!
!!
!!
!
!
!

!!
!!
!!

Frame!
Material!
!!
!!
!!

!!
!!
!!

!!
!!
!!

!!
!!
!!

!!
!!
!!

Manufacturer!

Manufacturer!
!!
!!
!!
!!

!
!
! Construction!
Material!
2!in!Concrete!

!
!
!

Model!

Model!

!
!
!

!!
!!
!!
!!
!
!
! Ext/!Cav!
Insulation!!
!!

U!Value!

U!Value!
!!
!!
!!
!!

!
!
!

Interior!
Insulation!
!!

SHGC!

!

! Detail!on!
Sheet!#!
!!
!!
!!

!

!!
!!
!!
!!
!!
!!

!

!

!!
!!
!!

!!
!!
!!
!
!
!
Comments!

Location!

!
!
!
!
! Glass!U!
!
Location!
Value!(if!any)!
!!
!!
!!
!!
!!
!!
!!
!!

!
!
! Detail!on!
Sheet!#!
!!
!!
!!
!!

!
!
!

!
!
! Slab!Edge!
Insulation!

!
!
!
!
!
!
Comments!
!!

Finish!

!!

!
!
! Detail!on!
Sheet!#!
!!

!!

Comments!
!!
!!
!!

!
!
! Material!
Type!
!!
!!
!!
!!

!!
161!

!

!

!

!!
!!
!!

Door!
Fixed!!
Fixed!!

!
!
DOORS)

!
#)

!

!

!!
!!
!!
!!

!!

!

Figure 5-1 (cont’d)
BASEMENT)
!

!

#)
1)

Construction!
6!in!Concrete!

! !
CEILING)
!
#)

!
! Insulation/!R!
Value!/!Depth!

!

Construction!

!!
!
!
! Insulation/!R!
Value!!

1) Lay!in!Acoustic!Tile!
! !
PARTITIONS)
!! !!
#)
1)

!
!
!
!
Manufacturer!
!!

!
!!
!!

!!

!
!
!
!
!
!
Manufacturer!

!!

!!
!
!!
!!

Brand!

!
!

!!

Brand!

!
!
! Detail!on!
Sheet!#!

!!
!
!!
!!

!
!!
!!

Construction!

Insulation/!R!
Value!!

Manufacturer!

Brand!

Detail!on!
Sheet!#!

Frame!

!!

!!

!!

!!

Comments!
!!

!
!
!

!!

162!
!

!
! Detail!on!
Sheet!#!

Comments!
!!

!
!
!

Comments!
!!

!

The second information block includes descriptions and assumptions for spaces made during
the design phase. This information will be useful for energy simulation, and eventually can be
verified through measurements during the occupied periods yielding better model calibration.
Information collected should include lighting fixture specification sheets and information on
special equipment included in any space. If an energy model is not developed during the design
phase of the project, this information will be helpful to record for cooling and heating load
calculations, as well as future reference for development of energy models at later stages of the
project. Figure 5.2 shows the Space Design Value Summary. Figure 5.3 shows an hourly
breakdown for spaces.

163!
!

!

Spaces'
and'Zones'
!
!
Space!
Name!

Description!

Area!
(ft2)!

Occupants!

Lighting!
Load!(W)!

Lighting!
Density!
(W/ft2)!

!!

!!

1!

!!

0!

0!

0!

!
! Task!
Lighting!
Density!
(W/ft2)!
0!

!!

!!

1!

!!

0!

0!

0!

!!

!!

!!

!!

!!

!!

!!
!!
!!
!!
!!
!!

!!
!!
!!
!!
!!
!!

1!
1!
1!
1!
1!
1!

!!
!!
!!
!!
!!
!!

0!
0!
0!
0!
0!
0!

0!
0!
0!
0!
0!
0!

!

!
!

!
!

!
!

!
!

!
!

0!

!
! Plug!
Load!
Density!
(W/ft2)!
0!

0!

0!

0!

!!

!!

!!

!!

0!
0!
0!
0!
0!
0!

0!
0!
0!
0!
0!
0!

0!
0!
0!
0!
0!
0!

0!
0!
0!
0!
0!
0!

Task!
Lighting!
(W)!

!
!

Plug!
Load!
(W)!

Figure 5-2: Space Summary and Design Values Summary Chart

164!
!

!
!

Outside!
Airflow!
(CFM)!

!
!

Design!
Flow!
(CFM)!

!!
!!

!!
!!

!!
!!
!!

!!
!!
!!

!!
!!
!!
!!

!!
!!
!!
!!

!

Space:!
Schedule!for:!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!
!!
!
!!
!
!!
!
!!
!
!!
!
!!
!
!!
!
!!
!
!!
!
!!
!

Hour
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10
10-11
11-12
12-13
13-14
14-15
15-16
16-17
17-18
18-19
19-20
20-21
21-22
22-23
23-24

Mon

Tue

Wed

Thur

Fri

Sat

Sun

Holiday

Figure 5-3: Schedule Summary Chart
Engineering systems and miscellaneous equipment installed in the building is the last set
of required information. This information should be supported with attached specifications and
technical manuals of the brand and model installed. This will provide for future references and
accurate descriptions of the HVAC system and equipment loads in the model. Figure 5.4 shows a
sample System Description Summary chart.
165!
!

!
HVAC%and%Equipment%
!!
General%Description%
Cooling!!

!
!!

!
!!

!
DX!Coils!

!
!

Heating!
System!Type!

Furnace!
Split!System!

!
Other!/!Observations!

!
!!

!
Model:!
Manufacturer:!
Efficiency!/!COP!
Add!Spec!sheet!/!Info!

!
!!
!!
!!
!

!
!
!
!
!
!
!
!
!

!
!

!
!

!
!

!
!
!

!
!
!

!
!
!

!
!

!
!

!
!

!
!
!!
!!
Design%Parameters%
!
Thermostat%
!
!
Setpoint%
!
Cooling:!
Heating:!
!
!
Design%Temperatures%Inside%
Cooling:!
Heating:!
!
!
Design%Temperatures%Supply%
Cooling:!
Heating:!
%
Direct!
!!

!
Return!Air!Path!
!
Minimum!Outside!
Air!%:!

!!

Fans%
!

!
Pressure!Drop!(in!H2O):! !
!!
Pressure!Drop!(in!H2O):! !!
!

Chiller%
Chiller!Type!
Condenser!Type!
Model!
Manufacturer!
!
Boiler%

Economizer%
!
!
!!
!!
!!

!
Capacity!
Efficiency!
Fuel!
Model!
Manufacturer!

!
!
!
!!
!!
!!
!!
!!

!

!

!
!
!
!
!
!
!
!
!
!
!
!
!
!

Type!
Limit:!

Drybulb!
Temperature!
!!

!
!
!
!
!
!
!
!
!
!
!

!
!
!
!
!
!
!
!
!
!
!

Figure 5-4: Systems Description Summary
166!
!

!

Miscellaneous+
Equipment+
!
Equipment!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!

!
!
!
!Description!/! !
!
Observation!/! Model! Manufacturer!
Notes!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!

!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!

!
!

!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!

Location!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!

!
! Amps!/!
Volts!/!
Phases!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!

!
!

Power!
(kW)!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!

Figure 5-5: Miscellaneous Equipment Description Summary Chart

167!
!

!
!

Comments!/!
Observations!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!

!

Information in these charts is intended to be maintained in the building administration or
building maintenance office. All information should be supported with architectural drawings or
technical drawings, so the reviewer or energy modeler can have all important information
available from a single source. The more detail provided, the easier it will be for an external
entity to analyze the building either for energy improvements, envelope retrofitting, ongoing
model verification or load analysis.
5.3 Chapter Summary
This chapter presents recommendations based on the experiences and observations from the
researcher during the investigation. Guidelines presented recommend the breakdown of spaces
according to use and grouping these similar spaces so they are defined under a single schedule
representing average use for these spaces. Also, support of schedule definition can include
additional information gathered from owners and administrative entities. A data collection tool
was presented for recording design assumptions and decisions during the design phase of a
project. Chapter Six provides a summary of the research, presents conclusions, identifies
limitations and suggests future areas of research.

168!
!

!

CHAPTER 6
CONCLUSIONS, LIMITATIONS, AND FUTURE RESEARCH

169!
!

!

CONCLUSIONS
Chapters 4 and 5 presented impacts of lighting schedules from case studies and provided
recommendations for improving schedule definition in simulation. This chapter presents lessons
learned from the analysis, limitations of the research and suggestions for future research.
6.1 Summary
The research focused on determining the impact of lighting use variations on building
energy simulations and how assumptions carried out during the early stages of a project can
influence energy consumption predictions. Three case study buildings were selected where actual
lighting use patterns could be observed and measured to define “use” schedules. Observations
consisted of walkthroughs of the buildings and recording of percentage of fixtures turned on and
spaces with lights on. Measurements included installation of light sensors, which recorded the
status of lighting fixtures as on or off, and were used to determine percentage of time lights were
used.
The selected case studies were the Chemistry Addition and Geography buildings at
Michigan State University and a Church project in Lansing, Michigan. The schedules for the first
two were collected by walkthroughs while for the church the sensor method was used.
Information collected was used to prepare research energy models which were compared to
metered data.

170!
!

!

Results showed that using research schedules averaged among spaces was the best
approach throughout the case studies. The research also showed that assumptions made during
the design phase model heavily impact accuracy.
6.2 Conclusions
Based on the interviews, the boundary test, and the case studies analyzed, the following
conclusions can be drawn regarding lighting use predictions:
•

Well defined lighting schedules can help close the gap between inaccuracies in energy
modeling. Although it was found that improving one parameter improves energy
prediction, it is difficult to ignore relationships among other parameters that will
influence overall energy use

•

Miscellaneous equipment use is an important factor in energy prediction. It is hard to
define how often a pump or a snow melting system is going to function. Schedules for
these loads can impact energy prediction and can mislead project teams when considering
energy efficiency improvements. Total energy consumption can coincidentally be
accurate but for the wrong reasons. The Chemistry building showed similar energy
prediction between the design phase model and the research model even though the
design model had high lighting loads and lower equipment loads while the research
model estimated lower lighting loads and higher miscellaneous loads

•

Differences in lighting load estimations can result in differences of up to 120% in energy
prediction. Default schedules are defined for specific building types with a constant
behavior pattern. Using default schedules may be a good approach for average buildings
171!

!

!

but deeper analysis is necessary before using default schedules in a model for more
complex buildings
•

Analysis of individual spaces and their distribution should be conducted for each energy
model. The academic buildings included in the study showed variations of spaces and
included classrooms, offices and laboratories. Use patterns varied among these spaces
and using one standard schedule for all of them leads to inaccuracies

•

Professionals agreed during interviews that occupancy and systems schedules (HVAC,
lighting, equipment) are important sources of inaccuracy. There is no defined
methodology or standard for establishing schedule definitions, a code or standard would
help to make the modeling process faster and more accurate but it would be difficult
since every building and owner is different

6.3 Limitations
Conclusions were drawn from the observations and models created, but there are several
limitations of the study as indicated below:
•

In research of this nature, three case studies are not sufficient to draw conclusions of
statistical value. This research serves more an educational purpose and hopefully sets a
path for similar studies where energy simulation can be further investigated

•

Observations were demonstrative of procedures that can be used and implemented to
record schedules. Expanding the number of observations could lead to detailed and
accurate schedule definition. This research also considers only observations when classes
are in session and not break periods for the case studies of the buildings at Michigan State
172!

!

!

University. Observing Holidays and unoccupied periods as well as academic periods
would help to profile an entire year
•

Influence of other electric loads was not explored

•

Not all building types were analyzed, results may vary in other buildings types

6.4 Future Research
The research focused only on lighting, other research can be conducted addressing other
parameters. Some possible additional studies which can add to the understanding of energy
simulations can include:
•

Miscellaneous equipment schedules research. Patterns of specific equipment could be
determined so they can be more accurately modeled in future projects. Patterns of
equipment such as elevators, snow melting devices and water pumps can be monitored
with an amperage meter and average use can be determined

•

This research applied methodologies to gather data that fit with the time and budget of
the research. Better technology and constant monitoring can be implemented to develop
other measuring techniques not only for lighting use control, but for other parameters so
use patterns can be better defined. Cameras could be installed throughout a building to
monitor occupancy behavior and patterns. Electric measuring of specific end use loads
can be monitored to test accuracy of specific parameters and to isolate individual
parameters from overall energy consumption

173!
!

!

6.5 Expected Contributions
Lack of knowledge of actual lighting patterns for a building during design makes simulation
difficult. Accurate occupancy and lighting schedule definition can be time consuming and
confusing if the proper information is not at hand. The fundamental purpose of this research was
to evaluate the impact of lighting schedules on energy simulation and to find ways to improve
lighting schedule input. A better understanding of the impact of lighting schedules on simulation
can lead to refined outputs and more accurate simulations. The research may lead to addressing
other parameters or standardize codes for schedule definition and energy simulation parameters.
Improvement of energy prediction tools is also beneficial for life cycle cost analysis of
projects. Increasing accuracy in energy simulation tools increases reliability when comparing
design options, particularly when the simulations can be supported with economic analysis.

174!
!

!

APPENDICES

175!
!

!

APPENDIX A
Survey Questionnaire
Date __________
Name ________________________
Title __________________________
Energy Simulation Experience (Years) ___________
Phase 1 interview: Perceptions and Experiences
1. What simulation(s) tool(s) or software do you use?
2. What is the main use of the simulation tools? LEED EA credits? Design tool during
planning and design phase of project?
3. Do you carry out validation of your energy prediction results?
4. How accurate compared to real energy performance of a building does energy
simulations are?
5. What do you think is the biggest source of inaccuracy in energy simulations?
6. In your experience with projects, what has been the most challenging process of the
simulation?
7. Do you use energy simulations as design tools, or only as a mean to comply with LEED
EA Credits?
8. How accurate do you think occupancy schedules can be defined in energy simulation
compared to real occupancy?
9. Do you consider that inaccuracies in scheduled definition increase or decrease
considerably building energy simulation? Is this of your concern when in design?
176!
!

!

10. What information is usually given by (or requested to) the owner/ architect for you to
have the adequate information for the simulation?
11. How accurate do you think occupancy schedules can be defined with the appropriate
information through the design phase?
12. Is there a technique, model, or code you follow to define occupancy schedules? Do you
use the program’s default schedules?
13. Is there an established methodology throughout your company to define occupancy
schedules?
14. How would you improve occupancy and lighting use schedule definition?
15. Do you verify occupancy and lighting schedule in the post construction phase?
16. How much do you think standardization (or a set of guidelines) schedule definition
according to building type will improve energy simulations?

Phase 2: Validation
1. What do you think of the presented information? Any particular interesting points?
2. What do you think of the guidelines and recommendations presented?
3. What do you think of the data gathering tool?
4. Any of this information is applicable to future projects?
5. Any additional recommendation/ suggestions?

177!
!

!

Michigan State University
School of Planning, Design and Construction
Construction Management Program
PARTICIPANT CONSENT FORM
Construction Professionals
SENSITIVITY OF LIGHTING SCHEDULE PARAMETER IN BUILDING ENERGY
SIMULATION SOFTWARE
Principal Investigator: Tim Mrozowski
Secondary Investigator: Federico Steinvorth
Interview
The School of Planning, Design and Construction at Michigan State University is
conducting research to evaluate impact of assumed lighting use schedules on computer-based
building energy predictions. Energy modeling programs have become important design and
decision making tools. This research aims to improve understanding of energy prediction results
and create recommendations and tools to improve the modeling process.
As a participant in this research, you are being asked to answer interview questions,
relating to your experience in energy modeling processes and high performance building design.!
You must be at least 18 years old to participate in this research. Your participation in this
research project is completely voluntary. You have the right to say no. If you are uncomfortable,
you may change your mind at any time and withdraw from the interview. You may choose not to
answer specific questions or to stop participating at any time. Whether you choose to participate
or not will have no effect on your grade or evaluation. Your privacy will be protected to the
maximum extent allowable by law. Your name and title will not be used in any publication. The
estimated time to complete this interview is approximately 45 minutes. As a participant, you may
request a copy of this consent letter for your records.
This research project is not funded. The researchers are employed by Michigan State
University and the data collected will be used for a graduate Master’s thesis.
If you have concerns or questions about this study, such as scientific issues, how to do
any part of it, or to report an injury, please contact:
Tim Mrozowski, A.I.A., LEED® AP
Professor of Construction Management, School of Planning, Design and Construction, Michigan
State University, 102B H.E. Bldg., East Lansing, MI-48824, USA, Email:
mrozowsk@egr.msu.edu, Phone number : +1 517.353.0781.
Federico Steinvorth
Graduate Student, Construction Management Program
178!
!

!

School of Planning, Design and Construction, Michigan State University, 102 H.E. Bldg., East
Lansing, MI-48824, USA. Email: steinvor@msu.edu, Phone numbers: +1 517.775.3573
If you have any questions or concerns about your role and rights as a research participant,
would like to obtain information or offer input, or would like to register a complaint about this
research study, you may contact, anonymously if you wish, the Michigan State University
Human Research Protection Program at 517-355-2180, FAX 517-432-4503, or e-mail
irb@msu.edu, or regular mail at: 207 Olds Hall, MSU, East Lansing, MI 48824.
You indicate your voluntary agreement to participate by beginning this phone interview.

179!
!

!

APPENDIX B

Chemistry Building Construction Data
Thickness
(ft)

Density
(lb/ft3)

Spec. Heat
(Btu/lb-°F)

R-Value
(h-ft2°F/Btu)

0.59

172

0.2

0.04

0.296
0.167
0.083
n/a
0.052
n/a

1/4in
Spandrel
Glass
MinWool Batt R11
MinBd 2in R-6.9
Wood 1in
Air Lay <4in Vert
GypBd 5/8in
Surf Air Film Vert

Conductivity
(Btu/h-ft-°F)

0.021

Material Name

0.025
0.024
0.07
n/a
0.0926
n/a

0.6
15
37
n/a
50
n/a

0.2
0.17
0.6
n/a
0.2
n/a
Total R
U Value
Table B-1: Main Façade Wall Layer Chemistry Building

Thickness
(ft)

Material Name
Facebrick
Air Lay <3/4in Vert
Cellulose 3.5in R-13
Plastic Film Seal
CMU
MW
12in
Hollow
Surf Air Film Vert

Conductivity
(Btu/h-ft-°F)

Density
(lb/ft3)

Spec. Heat
(Btu/lb-°F)

0.333
n/a
0.292
n/a

0.7576
n/a
0.0225
n/a

130
n/a
3
n/a

0.22
n/a
0.33
n/a

RValue
(h-ft2°F/Btu)
0.44
0.9
12.98
0.01

1

0.4959

58

0.2

2.02

n/a

n/a

n/a

n/a
Total R
U Value

0.68
17.02
0.06

Table B-2: Facebrick Wall Chemistry Building

180!
!

11.84
6.96
1.19
0.89
0.56
0.68
22.15
0.05

!

Thickness
(ft)

Material Name
Facebrick
Air Lay <3/4in
Vert
Cellulose 3.5in R13
Plastic Film Seal
Steel Siding
Surf Air Film Vert

Conductivity
(Btu/h-ft-°F)

Density
(lb/ft3)

Spec. Heat
(Btu/lb-°F)

0.333

0.7576

130

0.22

R-Value
(h-ft2°F/Btu)
0.44

n/a

n/a

n/a

n/a

0.9

0.292

0.0225

3

0.33

12.98

n/a
0.005
n/a

n/a
26
n/a

n/a
480
n/a

n/a
0.1
n/a
Total R
U Value

0.01
0.00
0.68
15.01
0.07

Table B-3: Utility Duct Wall Chemistry Building

Material Name
Asph Roll Roof
Polystyrene 4in
Polystyrene 4in
Chem Roof M4
Felt 3/8in
Conc LW 80lb
8in

Thickness
(ft)
n/a
0.333
0.333
0
0.031

Conductivity
(Btu/h-ft-°F)
n/a
0.02
0.02
n/a
0.11

Density
(lb/ft3)
n/a
1.8
1.8
n/a
70

Spec. Heat
(Btu/lb-°F)
n/a
0.29
0.29
n/a
0.4

R-Value (hft2-°F/Btu)
0.15
16.65
16.65
1.7
0.28

0.503

0.2083

80

0.2

2.41

Total R
U Value
Table B-4: Roof Layers Chemistry Building

181!
!

37.85
0.03

!

Geography Building Construction Data

Thickness
(ft)

Steel Siding
Polystyrene 2in
MinWool Batt R7
Steel Siding
CMU MW 6in Hollow
GypBd 5/8in
Surf Air Film Vert

Density
(lb/ft3)

0.005
0.167
0.188
0.005
0.5
0.052
n/a

Material Name

Conductivity
(Btu/h-ft-°F)
26
0.02
0.025
26
0.3571
0.0926
n/a

Spec. Heat
(Btu/lb-°F)

480
1.8
0.6
480
65
50
n/a

0.1
0.29
0.2
0.1
0.2
0.2
n/a
Total R
U Value
Table B-5: Main Wall Assembly Geography Building

Thickness Conductivity
(ft)
(Btu/h-ft-°F)

Material Name
Felt 3/8in
Insul Bd 1in
CMU MW 4in Hollow
Surf Air Film Horiz

0.6
0.083
0.333
n/a

0.11
0.025
0.3003
n/a

R-Value
Density Spec. Heat
(h-ft2(lb/ft3) (Btu/lb-°F)
°F/Btu)

0.4
0.2
0.2
n/a
Total R
U Value
Table B-6: Roof Assembly Geography Building

182!
!

R-Value
(h-ft2°F/Btu)
0.00
8.35
7.52
0.00
1.40
0.56
0.68
18.51
0.05

70
2
76
n/a

5.45
3.32
1.11
0.76
10.64
0.09

!

Church Project Construction Data

Material Name
Face Brick 4in
GypBd 5/8in
FiberGlass Batt R19
Church Ext Wall M4
PartBd Underlay 5/8"
Surf Air Film Vert

Spec.
Thickness Conductivity Density
Heat
(ft)
(Btu/h-ft-°F)
(lb/ft3)
(Btu/lb°F)
0.333
0.7576
130
0.22
0.052
0.0926
50
0.2
0.511
0.025
0.6
0.2
n/a
n/a
n/a
n/a
0.052
0.1796
75
0.29
n/a
n/a
n/a
n/a
Total R
U Value
Table B-7: Wall Assembly Church Project

183!
!

R-Value
(h-ft2°F/Btu)
0.44
0.56
20.44
0.12
0.29
0.68
22.53
0.04

!

APPENDIX C
Interview #1
What simulations tools do you use?
“For smaller projects I use Energy-10 or eQuest Wizard. For bigger projects Ecotech, Trane
TRACE or eQuest”
What is the main use of simulation tools?
“Use of simulation tools is not only for LEED certification, we also use simulation tools for
energy analysis. Also we use the tools for Life Cycle cost analysis, and to implement energy
conservation measures”
Do you validate results?
“We rarely do validations mainly because there is a cost associated that is usually not paid, and
the owner doesn’t ask for this validation”
What do you think are the main sources of inaccuracies in simulations?
“Schedules of operation such as HVAC, light, and people”
Do you spend some time defining the schedules?
“I don’t spend much time defining the schedules; I usually use the default schedules within the
software; not even owners know how buildings are going to work”
What information is usually requested from owners for better simulations?
184!
!

!

“It is difficult to get information from owners since they are not sure how the building is going to
operate either”
How much do you think standardization or guidelines to schedule definition can help
improve energy simulations?
“Better schedules would probably make simulations more accurate, but I don’t think
standardization is possible since all buildings are different.”
Interview #2
What simulations tools do you use?
“HAP from Carrier, and recently started using eQuest”
What is the main use of simulation tools?
“We use simulations both for LEED credits and energy analysis in buildings”
Do you validate results?
“We collect data from owners for as long as they are willing to provide us with it. Usually
energy simulations turn out predicting less energy than actual use. We don’t verify lighting use
itself because it involves a cost that won’t be paid. This would definitively be a method of
improving the model.”
What do you think are the main sources of inaccuracies in simulations?
“Occupancy schedules, and lately I realized how leaky buildings impact energy predictions”
185!
!

!

What do you think is the most challenging process of the simulation?
“The simulation itself is not challenging, but the USGBC and LEED reviewers make the process
harder.”
Do you spend some time defining the schedules?
“Yes, I like spending time on defining schedules; I try talking to users and owners for
information, and analyze room by room to make better schedules”
What information is usually requested from owners for better simulations?
“It is difficult to get information when it’s a new building; the owner knows more about his/her
project when it is a replacement building than when it is a new one.”
Is there a methodology or code followed in your company to define schedules?
“There is not really a method we all follow; when it’s a LEED project we are working on, we use
the ASHRAE proposed schedules, when it’s an energy analysis project we all define schedules
independently and with our own criteria”
How much do you think standardization or guidelines to schedule definition can help
improve energy simulations?
“There are already some schedules in ASHRAE 90.1 that can be used, and programs such as
HAP have a database of schedules that can be used. Improving schedules makes better
simulation models”.

186!
!

!

Additional Comments
“LEED don’t require modeling some models such as snow melting systems and dehumidification
systems, so in one project we projected $18000 on energy cost annually, and the actual cost from
the building turned out to be $34000. After model calibration and analysis, we predicted annual
energy cost to be $25000”
Interview #3
What simulations tools do you use?
“Mainly eQuest and EnergyPlus. For smaller projects we’ve been using BEOpt and Open
Studio”
What is the main use of simulation tools?
“We don’t do many LEED projects, we are trying to move away from LEED certifications. We
work mostly on projects during schematic and design phase to help evaluate design decisions.”
Do you validate results?
“No, we don’t intend to validate results, our objective is to compare design solutions throughout
schematic and design phase, but it is not our intention to predict energy use.”
What do you think are the main sources of inaccuracies in simulations?

187!
!

!

“Starting with weather files we have inevitable inaccuracies, but within the project occupancy
behavior I think is the biggest source on inaccuracy; it is important to take educated guesses on
this issue.”
Do you spend some time defining the schedules?
“I would say we do like spending time on our schedules, although we might use the default
schedules within the program if we don’t obtain real data from the owners. It is better to spend
time on schedules when data is gathered from existing buildings. We might not spend more time
on the schedules on the first stages of the simulation, but as the design is moving along we try to
calibrate the model with data.”
What information is usually requested from owners for better simulations?
“Usually owners are in the same “dark room” as we are when assuming conditions for the model.
Sometimes they might have the information, but most of the times they have the same
knowledge on the building as we do.”
Is there a methodology or code followed in your company to define schedules?
“I would say it’s done on an individual basis, we all know where to get the required information
but it is up to each modeler to come up with their own assumptions.”
How much do you think standardization or guidelines to schedule definition can help
improve energy simulations?

188!
!

!

“I could divide this answer in two: I think it would make the modeling process itself faster and
more efficient, but in terms of accuracy it would be hard to tell if a standard would improve
energy predictions since every project is different and has its own particularities”.
Interview #4
What simulations tools do you use?
“TRACE 700 for whole building analysis and we have our own excel sheets for system loads and
utility calculations and projections”
What is the main use of simulation tools?
“We do both energy analysis, and LEED credits along with external designers, we review the
designs of external designers so they comply with our requirements. We use the models to carry
out measurement and verification process.”
Do you validate results?
“We do verification process with utility data from measurements.”
What do you think are the main sources of inaccuracies in simulations?
“It’s hard to tell, some predictions go under and some go over. I would say sloppy assumptions
have a big impact on accuracy, and trying to determine the user’s behavior within the building.”
What is the most challenging process of the simulation?

189!
!

!

“People are complex and buildings are complex, so it’s very easy to make mistakes and difficult
to reflect building behavior.”
Do you spend some time defining the schedules?
“I don’t think consultants involve the right people in the schedule discussion, so they overstate
occupancy to be on a safe side.”
What information is usually requested from owners for better simulations?
“They apply general details, usually not much information about details on how buildings will be
occupied are requested”
How accurate can predictions are with the right schedule definition?
“They could be very accurate, it’s a matter of finding the right people. There are so many
parameters during design that designers are concerned about, that they give little importance and
don’t focus on occupancy”
How much do you think standardization or guidelines to schedule definition can help
improve energy simulations?
“There’s a big potential to create standards or guidelines and that’s where I think we are going
to. Even though we can normalize these issues, plug loads and equipment with electric load
impact as big as HVAC systems are still unpredictable during the design phase”.
How can simulations be improved?

190!
!

!

“If a space by space analysis is done, it could be more accurate, and by surveying the occupants
that know how the building behaves. Once again, if the right answers are asked to the right
person can bring good solution, but it’s a challenge when not knowing who the occupants are”

191!
!

!

REFERENCES

192!
!

!

REFERENCES
Abushakra, B., Sreshtraputra, A., Haberl, J., and Claridge, D., (2001) “Compilation of Diversity
Factors and Schedules for Energy and Cooling Load Calculations” College Station, Texas:
Texas A&M University.
ASHRAE (2009) “ASHRAE Fundamentals 2009” Atlanta, Georgia: American Society of
Heating, Refrigerating, and Air Conditioning Engineers
ASHRAE (2004) “90.1 User’s Manual ANSI/ ASHRAE/ IESNA Standard 90.1-2004” Atlanta,
Georgia: American Society of Heating, Refrigerating, and Air Conditioning Engineers
Bartlett, R., Connell, LM., Gowri, K., Halverson, MA., Lucas, RG., Richman, EE., Schulz, RW.,
Winiarski, DW., (2011) “Technical Support Document for Version 3.9.0 of the Comcheck
Software” Richland, Washington, Pacific Northwest National Laboratory
Birdsall, B., Buhl, W.F., Winkelman, K.L., (1994) “DOE-2 Basics” Berkeley, California:
University of California.
Bloomfield, D., Candau, Y., Dalicieux, P., Delille, S., (1995) “New Techniques for Validating
Building Energy Simulation Programs” Madison, Wisconsin: International Building
Performance Simulation Association Conference.
Burgeois, D., Reinhart, C., Macdonald, I.A., (2005) "Assessing the Total Energy Impact of
Occupant Behavioural Response to Manual and Automated Lighting Systems" Ottawa, Ontario:
National Research Council Canada.
The Cadmus Group, Inc., (2009) “Massachusetts Green Schools Post Occupancy Study on
Energy Efficiency”. Massachusetts: Massachusetts Renewable Energy Trust.
Christensen, C., Anderson, R., Horowitz, S., Courtney, A., Spencer, J., (2006) “BEopt Software
for Building Energy Optimization: Features and Capabilities” Golden, Colorado: National
Renewable Energy Laboratories.
Clarke, J. A., (1985) “Energy Simulation in Building Design” Oxford: Butterworth-Heinemann
Clarke, J. A., (2001) “Energy Simulation in Building Design” Oxford: Butterworth-Heinemann
Clevenger, C., and Haymaker, J., (2010) “The Impact of the Building Occupant on Energy
Modeling Simulations” Stanford, California: Stanford University
Crawley, D., Hand, J., Kummert, M., and Griffith, B., (2010) "Contrasting the Capabilities of
Building Energy Performance Simulation Programs." Montreal: Ninth International IBPSA
Conference.
193!
!

!

Crawley, D., Lawrie, L., Pedersen, C., (2001). "EnergyPlus: Creating a New-Generation
Building Energy Simulation Program," Energy & Buildings, pp. 319-331, Volume 33, Issue 4
Dober, Lidsky, Craig, and Associates (2003). “Instructional Space Utilization Summary.”
Huntington, West Virginia: Marshall University
Fournier, D., and Deal, B., (2008) "Towards an Energy Efficient Future: The Illinois Smart
Energy Design Assitance Program." Journal of Green Building 2.3 (2008): 52-58. College
Publishing.
Guyon, Gilles, and Nadia Rahni. (1997) "Validation of a Building Thermal Model in CLIM2000
Simulation Software Using Full-Scale Experimental Data, Sensitivity Analysis and Uncertain
Analysis." Prague, Czech Republic: International Building Performance Simulation Association
Conference 1997
Hall, J., Fehner, T., (1994) “Department of Energy 1977 - 1994: A Summary History”
Washington, DC: Department of Energy
Heller, J., Heater, M., and Frankel, M., (2011) “Sensitivity Analysis: Comparing the Impact of
Design, Operation, and Tenant Behavior on Building Energy Performance.” Vancouver,
Washington: New Buildings Institute
Hirsch, J. (2009) “EQuest Introductorial Tutorial Version 3.63”. Camarillo, California: James J.
Hirsch & Associates, 2009
Judkoff, R., and Neymark, J., (2006) “Model Validation and Testing: The Methodological
Foundation of ASHRAE Standard 140”. Quebec City, Canada: ASHRAE 2006 Annual Meeting
Judkoff, R. (2008) “Testing and Validation of Building Energy Simulation Tools”. Golden,
Colorado: International Energy Agency
Karlsson, F., Rohdin, P., and Persson, M., (2007) "Measured and Predicted Energy Demand of a
Low Energy Building: Important Aspects When Using Energy Building Simulation." Building
Services Engineering Research and Technology 28: 223-35.
Lee, S., Claridge, D., (2002) "Automatic Calibration of a Building Energy Simulation Model
Using a Global Optimization Program." Richardson, Texas: Second International Conference for
Enhanced Building Operations
Liao, C., and Barooah, P., (2010) "An Integrated Approach to Occupancy Modeling and
Estimation in Commercial Buildings." Baltimore, Maryland: 2010 American Control Conference
Lomas, K.J., Eppel, H., Martin, C.J., and Bloomfield, D.P., (1997) "Empirical Validation Of
Building Energy Simulation Tools." Energy and Buildings 26: 253-75.
194!
!

!

Rosenbaum, M., (2003) "Understanding the Energy Modeling Process: Simulation Literacy
101." BuildingGreen.com: http://www.buildinggreen.com/features/mr/sim_lit_101.cfm (27-102010)
Sezgen, O., and Koomey, J., (1998) “Interactions Between Lighting and Space Conditioning
Energy Use in U.S. Commercial Buildings.” Berkeley, California: Lawrence Berkeley National
Laboratories
Soebarto, V., (1997) "Calibration of Hourly Energy Simulation Using Hourly Monitored Data
and Monthly Utility Records for Two Case Study Buildings." Prague, Czech Republic:
International Building Performance Simulation Association Conference 1997
Strachan, P.A. (2008) "Simulation Support for Performance Assessment of Building
Components." Building and Environment 43: 228-36.
Siva K. Haran. (2010, April). Part 1: Basics of sustainable lighting. Consulting - Specifying
Engineer, 47(4), 18-22. Retrieved March 13, 2011, from ABI/INFORM Global. (Document
ID: 2015811491).
Transfer Function Methodolgy. Carrier Corporation, 2005.
Wasilowski, H., and Reinhart, C., (2009) "Modelling an Existing Building in
Designbuilder/EnergyPlus: Custom Versus Default Inputs." Glasgow, Scotland: International
Building Performance Simulation Association Conference 2009
Zmeureanu, R., Pasqualetto, L., and Bilas, F., (1995) "Comparison of Cost and Energy Savings
in an Existing Large Building As Predicted by Three Simulation Projects." Madison, Wisconsin:
Fourth International Conference Building Simulation

195!
!