ANALYZING AND REDUCING ENERGY CONSUMPTION ON THE MICHIGAN STATE
UNIVERSITY CAMPUS
By
Alexander Renny

A THESIS
Submitted to
Michigan State University
In partial fulfilment of the requirements
For the degree of
Chemical Engineering – Master of Science
2017

ABSTRACT
ANALYZING AND REDUCING ENERGY CONSUMPTION ON THE MICHIGAN STATE
UNIVERSITY CAMPUS
By
Alexander Renny
The reliance on fossil fuels to generate most of the world’s energy has led to a significant
increase in greenhouse gas emissions and the associated negative effects on the climate. The application
of energy reducing retrofits allows for significant reductions in total energy use, providing the opportunity
to achieve cost savings as well as emissions reductions. Students Planning Advanced Retrofit Technology
Applications (SPARTA) was founded to pursue implementation of these retrofits at Michigan State
University (MSU). Under this framework, lighting, device, and HVAC retrofits are examined. In
commercial spaces, lighting comprises ~17% of total electricity used. The economic viability of
retrofitting LEDs in place of fluorescent ceiling fixtures as well as various control methods are
determined. The energy use intensity of the average office space on the MSU campus is also analyzed and
equipment changes are recommended to decrease the total energy footprint. Conversion from electricity
driven cooling to steam driven cooling is evaluated as a method for increasing efficiency of the cogeneration power plant and decreasing total fuel consumption. Finally, activities aimed at engaging the
public on renewable energy generation are described, which are based on the development of paintable
luminescent solar concentrators that combine art and energy. A number of techniques for reducing
energy use on the MSU campus have been analyzed which can be applied across a broad spectrum of
spaces. Continuing to aggressively pursue the implementation of this framework as well as tracking key
economic indicators outlined will ultimately make MSU cleaner and more sustainable.

To my parents, Karen Grady and Craig Renny, for their never-ending love and support.

iii

ACKNOWLEDGEMENTS

I would first like to thank the students of SPARTA – this includes Aryka Thomson, Ben Stephens, Philip
Wandor, Bo Li, Lee Padilla, Yuzhu Liu, and Haojun Wang. Without all of you none of this would have
been possible. To everyone in the MOE lab – especially Chenchen Yang – thank you for your help in and
out of the lab. I would also like to thank Ann Erhardt at Sustainability for championing my cause, as well
as Kane Howard, William Lakos, and Jason Vallance at IPF for all of their help. Also, my deepest thanks
to Dr. Richard Lunt, my advisor, for helping guide me through SPARTA and graduate school in general.
Thank you to Dr. Rebecca Anthony for all of your help co-advising SPARTA. I’d also like to thank Dr.
Maddalena Fanelli for serving on my thesis committee.

iv

TABLE OF CONTENTS

LIST OF TABLES ...................................................................................................................................... vii
LIST OF FIGURES ...................................................................................................................................... x
KEY TO ABBREVIATIONS ..................................................................................................................... xii
INTRODUCTION ........................................................................................................................................ 1
CHAPTER 1: Lighting Retrofits and Control Systems ................................................................................ 2
Introduction ............................................................................................................................................... 2
Hallway LED Retrofits ............................................................................................................................. 4
Rationale for Retrofitting ...................................................................................................................... 4
Characteristics of Lighting.................................................................................................................... 6
Fixture testing ....................................................................................................................................... 8
On-Rate Analyses ............................................................................................................................... 11
Financial Analysis............................................................................................................................... 15
Discussion ........................................................................................................................................... 20
Hallway Occupancy Based Lighting Controls ........................................................................................ 22
Rationale ............................................................................................................................................. 22
Available Energy Reductions ............................................................................................................. 22
Financial Analysis............................................................................................................................... 23
Occupancy Based Lighting Controls Only ..................................................................................... 23
Occupancy Based Lighting Controls Coupled with LED Retrofits ................................................ 26
Discussion ........................................................................................................................................... 26
Daylight Harvesting Light Management................................................................................................. 27
Rationale ............................................................................................................................................. 27
Financial Analysis............................................................................................................................... 28
Sensitivity Analysis ........................................................................................................................ 30
Case Study ...................................................................................................................................... 31
Discussion ........................................................................................................................................... 33
Summary ................................................................................................................................................. 34
CHAPTER 2: Energy Use Analyses ........................................................................................................... 35
Introduction ............................................................................................................................................. 35
Evaluation of Steam Cooling Requirements ........................................................................................... 36
Rationale ............................................................................................................................................. 36
Electricity Use Analyses ..................................................................................................................... 37
Steam Use Analyses............................................................................................................................ 38
Discussion ........................................................................................................................................... 41
Energy Use Intensity of MSU Campus Spaces ....................................................................................... 43
Rationale ............................................................................................................................................. 43
v

Materials and Methods........................................................................................................................ 43
Results................................................................................................................................................. 46
Discussion ........................................................................................................................................... 48
Summary ................................................................................................................................................. 50
CHAPTER 3: Luminescent Solar Concentrator Outreach Activity ............................................................ 51
Introduction......................................................................................................................................... 51
Concepts of Materials ......................................................................................................................... 51
Goals of the Demonstration ................................................................................................................ 53
Design of Materials and Demonstration ............................................................................................. 54
Hazards ............................................................................................................................................... 59
Discussion ........................................................................................................................................... 59
What Is the Best Color? .................................................................................................................. 59
How to Produce a Usable Device? .................................................................................................. 61
Outreach Events .................................................................................................................................. 62
Conclusions......................................................................................................................................... 63
SUMMARY AND FUTURE WORK......................................................................................................... 63
APPENDIX ................................................................................................................................................. 66
BIBLIOGRAPHY ....................................................................................................................................... 78

vi

LIST OF TABLES

Table 1: Values used for sensitivity determination. Baseline values are indicated by bold typeface ........... 5
Table 2: Estimated annual hourly light on-rate, Psychology Building [8].................................................... 5
Table 3: Illumination levels in specified zones, in ft-candles. Value in parentheses indicates # F32T8
bulbs/fixture. ANSI/IES [7] specify 10-40 ft-candle range and 25 ft-candle average .................................. 7
Table 4: Calculated CRI at manufacturer specified CCT and at CCT calculated from spectra. Italics
indicate current standard fluorescent lamps .................................................................................................. 9
Table 5: Illuminance (Lux) and tested power draw compared to manufacturer specified luminous flux
(lumens) and power draw. Italics indicate current standard fluorescent lamps. CREE dimming system set
at ~50% power. *indicates lumens only provided by manufacturer for single bulb ................................... 10
Table 6: Light on-rate for Chemistry Building floors basement and 1-6 .................................................... 14
Table 7: Values used for sensitivity determination in Figure 8. Baseline values are indicated by bold
typeface ....................................................................................................................................................... 18
Table 8: Light on-rates for Anthony Hall and Biomedical & Physical Sciences (BPS) measured with an
Onset HOBO U12 temp/RH/light/ext logger; PH – Penthouse; B – Basement .......................................... 23
Table 9: MSU Standards occupancy sensors; Manufacturer: Leviton; Cost includes OSA20-R00 power
pack required for operation ......................................................................................................................... 24
Table 10: Installed occupancy sensor cost required to achieve financial viability; Manufacturer: Leviton25
Table 11: Installed occupancy sensor cost required to achieve financial viability in conjunction with LED
based fixtures; Manufacturer: Leviton ........................................................................................................ 26
Table 12: Costs required to trigger a 5-year payback period for occupancy sensor installation in four MSU
buildings. OSC10/OSC20 indicate Leviton occupancy sensor models ...................................................... 27
Table 13: Baseline costs used in DLH system analyses ............................................................................. 32
Table 14: Cost of dimmable fluorescent ballast at different electricity reduction rates by DLH system
required to trigger a 5-year payback (at $0.07492/kWh) ............................................................................ 32
Table 15: Quantity of window mount air conditioning units by building, organized by intensity (units/ft2).
Annual energy use for WAC units estimate from US EPA/DOE window air conditioning unit calculator
[19] .............................................................................................................................................................. 37

vii

Table 16: Electricity draw at different states of representative devices. All units are in watts (W).
Refrigerator energy consumption assumed to US Government Energy Guide value ................................. 44
Table 17: Estimated annual energy use on a per device basis .................................................................... 45
Table 18: Surveyed buildings ..................................................................................................................... 46
Table 19: Comparison of payback period at $0.05755/kWh electricity purchse rate with installed system
lifetime ........................................................................................................................................................ 68
Table 20: Agriculture Hall light on-rate ..................................................................................................... 69
Table 21: Computer Center light on-rate .................................................................................................... 69
Table 22: Giltner Hall light on-rate............................................................................................................. 69
Table 23: Berkey Hall light on-rate ............................................................................................................ 69
Table 24: Trout Food Science light on-rate ................................................................................................ 69
Table 25: Packaging Building light on-rate ................................................................................................ 70
Table 26: Natural Resources light on-rate .................................................................................................. 70
Table 27: Wells Hall (B/C Wings) light on-rate ......................................................................................... 70
Table 28: Engineering Building light on-rate ............................................................................................. 70
Table 29: Bessey Hall light on-rate............................................................................................................. 70
Table 30: Wells Hall (A/D Wings) light on-rate......................................................................................... 71
Table 31: Chemistry Building light on-rate ................................................................................................ 71
Table 32: Biochemistry light on-rate .......................................................................................................... 71
Table 33: North Kedzie Hall light on-rate .................................................................................................. 71
Table 34: Biomedical & Physical Sciences light on-rate ............................................................................ 71
Table 35: Plant and Soil Sciences light on-rate .......................................................................................... 72
Table 36: Baker Hall light on-rate .............................................................................................................. 72
Table 37: Kresge Hall light on-rate............................................................................................................. 72
Table 38: Communication Arts & Sciences light on-rate ........................................................................... 72

viii

Table 39: Music Practice Building light on-rate ......................................................................................... 72
Table 40: Student Services Building light on-rate ...................................................................................... 73
Table 41: Economic metrics at fuel cost of electricity($0.05755/kWh) ..................................................... 74
Table 42: Economic metrics at produced cost of electricity ($0.07492/kWh)............................................ 75
Table 43: Economic metrics at delivered cost of electricity ($0.09206/kWh)............................................ 76

ix

LIST OF FIGURES

Figure 1: Sensitivity analysis of factors affecting payback period (in years) for LED retrofits targeting
three-lamp F32T8 fixtures ............................................................................................................................ 4
Figure 2: Measurement locations using Dr. Meter LX1330B, performed at floor level ............................... 7
Figure 3: Spectral tests for (clockwise from top left) two-lamp retrofit (Precision Paragon TKG24-40XWEU), three-lamp retrofit (CREE UR2-40L-40K), and standard fluorescent bulbs (Phillips Alto II T8)....... 8
Figure 4: From top to bottom: CREE UR2-48-45L-40K backing with magnetic mounting, CREE front of
lamp, PP X-X24-40VW-EU backing with adhesive mounting, PP front of light bar................................. 11
Figure 5: Clockwise from top left: Logger installation example; Onset HOBO U12 data logger and
placard; measurements of luminous flux and on-rates in the MSU Natural Resources building 3rd floor;
basement ..................................................................................................................................................... 13
Figure 6: Annual light on-rates for surveyed buildings on the MSU campus............................................. 15
Figure 7: Net present value with respect to time, assuming installation date of 10/1/2017. Large drops in
NPV indicate end of expected retrofit lifetime and complete replacement ................................................ 17
Figure 8: Sensitivity analysis of factors affecting payback period (in years) for LED retrofits targeting 2
lamp F32T8 fixtures in the Plant and Soil Sciences Building .................................................................... 18
Figure 9: Payback period for LED retrofits at 19 surveyed spaces on the MSU campus. .......................... 19
Figure 10: Return on investment for LED retrofits at 19 surveyed spaces on the MSU campus ............... 20
Figure 11: Survey of office occupants in Psychology Building hallways which had LED retrofits installed
.................................................................................................................................................................... 21
Figure 12: Varying hallway widths of MSU buildings; clockwise from top left: Plant and Soil Sciences
(10ft); Engineering Building (12ft); Student Services Building (10ft); Biomedical and Phsyical Sciences
(6ft) ............................................................................................................................................................. 24
Figure 13: Comparison of payback periods for 6 different lighting configurations to implement DLH
systems. Sensor clusters consist of DLH and occupancy sensor ................................................................ 29
Figure 14: Sensitivity analysis variables affecting DLH installation payback. All variables except
[expected] electricity reduction apply to both LED and fluorescent variants. Fluorescent electricity
reduction refers to both 2/3 lamp and ballast/no ballast addition. LED electricity reduction refers to both
2/3 lamp retrofits. ........................................................................................................................................ 31

x

Figure 15: Case study for DLH system installation in Engineering Building room 2250 (two sensor
clusters) ....................................................................................................................................................... 31
Figure 16: MSU campus map highlighting (in red) buildings with window AC units ............................... 36
Figure 17: Average cooling (left) and heating (right) degree days on a monthly basis. Note differing
scales on each plot. A degree day is defined as the number of degrees the average temperature is
over/under a base temperature (65°F). If the average daily temperature for each day in November (30
days) was 50°F, the total number of (heating) degree days would be 450. ................................................ 39
Figure 18: Comparison of HDD, CDD, and total steam consumption for Berkey Hall (WAC cooling) ... 40
Figure 19: Comparison of Berkey Hall and Communication Arts steam use per total degree day scaled by
total building area (top) and average values for heating months (bottom), indicated in the top graph by
unshaded areas (Oct – Apr)......................................................................................................................... 41
Figure 20: Comparison of Communication Arts (central cooling) and Berkey Hall (distributed/WAC
cooling) temperature profiles. Outdoor temperatures are daily averages from LAN (NOAA) .................. 42
Figure 21: Energy use intensity for personal office spaces (total: 539 kWh/yr)......................................... 47
Figure 22: Comparison of total office device (not including ceiling fixtures) electricity use to two
fluorescent fixtures of different configurations........................................................................................... 48
Figure 23: Luminescent solar concentrator schematic ................................................................................ 53
Figure 24: Normalized absorption (solid) and emission (dots) spectra of paints used in the demonstration
(A). Solar cell module (D). Note that the multiple absorption and emission peaks for blue and green paints
stem from two distinct dyes with one dye emitting in the visible and one emitting in the infrared to
reinforce absorptive coloring and provide greater power. .......................................................................... 56
Figure 25: Sample painted device under direct illumination (left) and ambient conditions (right). Note
luminescence along waveguide edge where solar modules are attached during demonstration ................. 57
Figure 26: Current-Voltage (IV, left) and Power-Voltage (right) of device (center) with two edge mounted
solar modules connected in series and parallel. Multimeter is connected in parallel, reading mV (center) 58
Figure 27: Spectrum of sunlgiht, LED lamp, and fluorescent lamp ........................................................... 61
Figure 28: MSU Grandparent's University outreach activity...................................................................... 62
Figure 29: MSU Science Festival student participation .............................................................................. 63

xi

KEY TO ABBREVIATIONS

ANSI – American National Standards Institute
BPS – Biomedical and Physical Sciences Building
CCT – Correlated Color Temperature
CDD – Cooling Degree Day
CRI – Color Rendering Index
DOE – US Department of Energy
EUI – Energy Use Intensity
EPA – Environmental Protection Agency
FLCH – Full Load Cooling Hours
HDD – Heating Degree Day
HVAC – Heating, Ventilation, and Air Conditioning
𝐼𝑆𝐶 – Short circuit current
IES – Illuminating Engineering Society
IPF – Infrastructure Planning and Facilities
IV – Current – Voltage, used for measuring output of photovoltaics
MPP – Maximum Power Point
MSU – Michigan State University
LED – Light Emitting Diode
LEED – Leadership in Energy and Environmental Design
LSC – Luminescent Solar Concentrator
𝑃𝑓𝑙𝑢𝑜𝑟 – Power draw of fluorescent fixture
𝑃𝐿𝐸𝐷 – Power draw of LED fixture
PMMA – Poly-methyl methacrylate
PSS – Plant and Soil Sciences
xii

PP – Precision Paragon TKG24-40VW-EU or equivalent Columbia GORK24-40VW-EU LED retrofit kit
PV - Photovoltaic
R – Reflection of front face of waveguide
SPARTA – Students Planning Advanced Retrofit Technology Applications
STEM – Science, Technology, Engineering, and Mathematics
TDD – Total Degree Day
𝑉𝑂𝐶 – Open Circuit Voltage
WAC – Window Air Conditioning
ηAbs – Efficiency of absorption of luminophore
ηLSC – Efficiency of luminescent solar concentrator
ηPL – Efficiency of photoluminescence
ηPV – Efficiency of photovoltaic
ΡRA - Efficiency of reabsorption
ηWG – Efficiency of waveguiding

xiii

INTRODUCTION
The purpose of the SPARTA group is to harness the enthusiasm, energy, and vision of students,
faculty and staff to conduct applied energy research while also looking for solutions in novel and
impactful arenas. Student members, known as SPARTA Analysts, remain driven by the “Spartans Will”
mantra to create innovative solutions to advance themselves and their university. SPARTA relies on its
Analysts and Faculty Advisers for much of the work within the different project areas, and we are excited
to bring in fresh talent with many of our previous analysts graduating last spring. Our undergraduate
students, often working on a volunteer basis for long hours, have been highly motivated by and dedicated
to energy transition research which provides valuable experience working with software and other tools
which are used widely in industry or actively being developed here on campus. It is under this framework
that the analyses and outcomes of this thesis have been achieved.

1

CHAPTER 1: Lighting Retrofits and Control Systems

Introduction
Lighting accounts for around 17% of total electricity consumption in US commercial buildings
[1]. Reducing the electricity consumption of lighting through retrofitting is possible with several different
emerging technologies which can have few noticeable effects for users of a space. The three technologies
which are evaluated in this analysis are Light Emitting Diode (LED) replacements, occupancy sensing
controls, and daylight harvesting light management systems.
LED lamps have seen declining costs and increasing efficiencies for the past several years, with
trends expected to continue through at least 2030 [2]. LED retrofit kits, units which contain only lamp(s)
and a driver (convert high voltage alternating current to lower voltage direct current) which are installed
in an existing fixture, are now at a price point which allows for economic viability in many spaces.
Commercially available retrofits for 4-foot fluorescent fixtures with three lamps have previously been
analyzed for application in the Psychology Building hallways on the MSU campus [3]. This study seeks
to expand that previous analysis to the main hallways of other viable buildings as well as determine
suitable retrofit options for two-lamp fluorescent fixtures, which make up most ceiling lighting in
building hallways on the MSU campus. When retrofitting fixtures there are three key lighting
characteristics that must meet or exceed the current fluorescent fixtures which are standards. First is the
color rendering index (CRI), a measure of the accuracy of color rendering under illumination by a light
source. Correlated color temperature (CCT) is a measure of the color appearance of a light source, with
low CCT light sources appearing more orange and high CCT sources appearing more blue. The third
characteristic is luminous flux, or the illumination level provided by a light source. A survey of occupants
in offices off retrofitted hallways is performed to determine whether the installation had a noticeable
impact on occupants. After determining suitable retrofits for two-lamp fluorescent fixtures, economic
evaluations are performed to determine viability of installations for nineteen buildings across the MSU
campus.

2

Automatic lighting controls allow for energy use reduction by turning off lights completely or
reducing light output levels to match a predetermined setpoint. The use of scheduling is the cheapest
method of lighting control, where lights are automatically turned off by central controls based on the time
of day. Due to the nature of academic buildings where occupancy by students can often be irregular [4],
having lights turn on and off in hallways by a schedule is less feasible without a safeguard such as
occupancy sensing systems to override the light off state. Occupancy sensing systems toggle lights on
through detection of motion and off after no motion is detected for a user-defined amount of time [5].
While most classrooms on campus have occupancy sensors installed, there are currently only two
buildings on the MSU campus which have installed hallway occupancy sensors. A light on-rate analysis
is used to determine the energy savings provided by the systems in these two buildings, and this reduction
is then used to estimate the required system costs to achieve fiscal viability in the hallways of four
buildings with hallway light on-rates of 100%.
A second type of commercially available automatic lighting control are daylight harvesting
(DLH) light management systems. These systems are generally used in conjunction with occupancy
sensing systems and use light sensors to automatically dim electric lights to balance available daylight
and provide a consistent illumination level. The analysis of DLH systems focuses on classrooms and
other public areas on campus which have favorable window to wall or footprint ratios. The viability of
these systems is evaluated both generally across a wide scale of room sizes as well as a specific room
(EB2250) in the MSU Engineering Building using estimated energy reduction values. Fixture
modification through reduction of fluorescent bulb count or retrofitting an applicable LED system is
included as some spaces on campus have been deemed over-lit by current standards.
Many buildings on campus were initially constructed in the mid-1960s and were therefore built to
American National Standards Institute specifications current at that time. By 1982, these recommended
lighting levels (ANSI/IES) had decreased by 25-50% across space types of interest [6] and have remained
similar through 2017 [7]. Before installing daylight harvesting systems or any other retrofits which target
3

lighting, the possibility of switching to a lower energy consumption lighting system should be evaluated
to match current ANSI/IES recommendations for the space type.

Hallway LED Retrofits
Rationale for Retrofitting
With the rising efficiency and falling prices of LED lighting, the replacement of fluorescent
fixtures has become a key focus of energy conservation measures. Previous studies by SPARTA [3, 8]
found a favorable payback period of 2.9 years for retrofitting fixtures in the Psychology Building on the
Michigan State University campus. When performing a sensitivity analysis, light on-rate was determined
to be the largest determining factor for the payback period on three-lamp retrofit installations (Figure 1,
Table 1).

Installation Labor
Time
10
8
Time Fixtures are
on

6

Maximum
Mid High
Mid
Mid Low
Low
UR2 Pricing (usd)

4
2
0

Real Discount
Rate

Annual Utility
Rate Escalation

Inflation Rate

Figure 1: Sensitivity analysis of factors affecting payback period (in years) for LED retrofits targeting three-lamp
F32T8 fixtures

4

Table 1: Values used for sensitivity determination. Baseline values are indicated by bold typeface

High
Mid High
Mid
Mid Low
Low

Installation
Retrofit
Utility Rate Inflation Discount Light OnTime (mins) Pricing (USD) Escalation
Rate
Rate
Rate
60
$101.32
2.88%
3.48%
7.00%
100%
45
$96.99
2.16%
2.61%
5.25%
75%
30
$93.00
1.44%
1.74%
3.50%
50%
15
$92.60
0.72%
0.87%
1.75%
25%
10
$90.12
0.29%
0.35%
0.70%
10%

Public hallways were the areas found to be the best targets in these previous studies as they
exhibited the highest light on-rates when compared to other building spaces, such as classrooms and
offices (Table 2). Across all surveyed spaces, hallway lighting consumed from 0.5% (Computer Center)
to 20.8% (Baker Hall) of total building energy consumption. These two factors show the clear benefit of
targeting hallway spaces.
Table 2: Estimated annual hourly light on-rate, Psychology Building [8]

Space Type
Classroom
Hallway
Office

Usage (hours)
Walk-throughs Metered Data
3830
--8650
8425
5840
---

The economic viability of the installation of retrofit kits was required by Infrastructure Planning
and Facilities (IPF), the division at MSU which oversees the installation of any capital projects on campus
among other duties. The payback period IPF typically targets must be less than five years for a project to
proceed, and recommended kits must also have a warranty of at least five years.
The analysis results for retrofitting fixtures with three F32T8 lamps show very favorable
economic returns [3, 8], but these types of fixtures are not common on the Michigan State University
campus. After performing walk-through surveys of campus buildings, it was determined that most F32T8
containing fixtures have only two lamps, with the notable exceptions of the Engineering, Psychology, and
Chemistry Buildings which have three-lamp fixtures. Because of the reduced number of lamps in these
5

fixtures, retrofits suitable for three-lamp fixtures see a relatively low (31% vs 55% wattage reduction in
three-lamp fixtures) reduction in electricity consumption, and therefore cost savings. The increased
luminous flux of fixtures containing three lamps instead of two must also be considered. Some hallway
spaces on campus are already over-lit [7], so increasing luminous flux through the installation of retrofits
suitable for replacing three-lamp installations is not desired. For these reasons, determination of a suitable
retrofit solution for fixtures containing two F32T8 lamps was a key step towards translating these
potential retrofits across a greater fraction of MSU.

Characteristics of Lighting
Key specifications for determining a suitable retrofit kit for lighting include the brightness
(luminous flux), CCT, and color quality (CRI). Evaluation of these measures for selected LED retrofit
kits was performed to ensure quality before moving forward with economic analyses and final installation
recommendations.
Luminous flux is defined as “[a] measure of the rate of flow of light, i.e. the radiant flux in the
wavelength range 380–760 nanometres, corrected for the dependence on wavelength of the sensitivity of
the human eye” and is measured in lumens. While lumen specifications are generally provided by
manufacturers, specifications for lighting levels are instead measured in lux, or lumens per m2 (footcandles in SAE). This measurement is indicative of the total amount of light which reaches a surface. The
acceptable range of illuminance for hallways is 10 – 40 foot-candles (107 – 430 lux) with an average of
25 foot-candles, measured at floor level [7]. Illuminance measurements were taken in several buildings
targeted for retrofits (Figure 2, Table 3). Wells Hall is consistently over-lit, while the Engineering
Building only exhibits levels which exceed recommendations directly under fixtures. This is most likely
due to the fact that the Engineering Building had three-lamp fixtures as surveyed.

6

1. Centered beneath fixture
2. Along wall beneath fixture
3. Centered between two fixtures
4. Along wall between two fixtures
Figure 2: Measurement locations using Dr. Meter LX1330B, performed at floor level
Table 3: Illumination levels in specified zones, in ft-candles. Value in parentheses indicates # F32T8 bulbs/fixture.
ANSI/IES [7] specify 10-40 ft-candle range and 25 ft-candle average

Building
Engineering (3)
Plant & Soil Sciences (2)
Wells (2)
Biomed & Physical Sciences (2)
Student Services (2)

1
54.0
34.8
49.8
29.4
37.8

2
29.4
25.9
45.4
25.1
28.6

3
20.4
15.1
44.3
20.4
20.8

4
14.5
12.8
42.7
19.9
16.8

The CCT of a light source is the apparent color of that source in relation to a black body radiator
and is measured in Kelvins (K). A CCT between 3500K and 4100K, which produces a fairly neutral white
light, is recommended by the IESNA for office buildings and schools for daytime use. For reference,
there is a strong preference for warmer, lower CCT (e.g. 2700K) for lighting in homes that more closely
matches that of incandescent bulbs, particularly at night. The Alto II F32T8 lamps generally installed in
hallways and offices on the MSU campus have a manufacturer specified 4100K CCT. The retrofit kits
evaluated for this study are all at a manufacturer specified 4000K.
The color rendering index is a scale (0-100) which indicates how accurate a light source is at
rendering color when compared to a black body radiator of the same CCT. A higher CRI indicates a better
color rendering ability, allowing for greater differentiation between similar colors. Per previous testing by
the SPARTA group, the values provided by manufacturers are not always accurate. Therefore,
independent testing of selected retrofit kits as well as installed F32T8 lamps are also performed.

7

Fixture testing
Previous investigation [3, 8] determined that for three-lamp retrofits, the CREE UR2-48-45L-40K
retrofit satisfied color temperature and CRI requirements, exceeding the performance of standard fixtures
with three Phillips Alto II F32T8 light bars installed. To determine the optimal retrofit kit for two-lamp
fixtures, retrofit candidates from four different manufacturers were acquired and both CCT and CRI
testing were performed using bare light bars, measured with a calibrated Ocean Optics USB4000
spectrometer. Normalized spectral intensity for the retrofit kit targeting three-lamp fixtures, retrofit kit
targeting two-lamp fixtures, and standard fluorescent bulbs were determined using a MATLAB based
CCT/CRI calculator (Figure 3).

Figure 3: Spectral tests for (clockwise from top left) two-lamp retrofit (Precision Paragon TKG24-40XW-EU),
three-lamp retrofit (CREE UR2-40L-40K), and standard fluorescent bulbs (Phillips Alto II T8)

To determine the CRI of the evaluated retrofit kits and fluorescent bulbs, two methods were used
(Table 4). First, the CCT was calculated from the measured spectra (Figure 3) and the CRI was then
8

determined at that CCT. In the second method, the CRI was determined with CCT set at the manufacturer
specification (4000K for both retrofits, 4100K for fluorescent bulbs). Manufacturer listed CRI values are
78 for Phillips Alto II F32T8 fluorescent bulbs and 80 or above for all tested LED retrofit kits.
Table 4: Calculated CRI at manufacturer specified CCT and at CCT calculated from spectra. Italics indicate
current standard fluorescent lamps

Manufacturer
Phillips
Precision Paragon
Precision Paragon
Phillips
Phillips
CREE
Columbia
Columbia

Model
Alto II F32 T8
TKG24-40XW-EU
TKG24-40VW-EU
12T8/40
15T8/40
UR-48 [Dimming]
GORK24-40VW-EU
GORK24-40XW-EU

Specified CCT
CCT (K)
CRI
4100
76
4000
75
4000
74
4000
76
4000
74
4000
76
4000
75
4000
75

Calculated CCT
CCT
CRI
3600
75
3400
73
3800
75
3200
72
3400
72
3600
75
3800
73
3800
73

Luminous flux testing for retrofit kits of interest was then performed using a benchtop test fixture
which allowed for a streamlined installation procedure for ballast and lamps to compare retrofit kit
illumination levels to a baseline provided by two F32T8 lamps. A Dr. Meter LX1330B Digital Lux Meter
was used to measure the luminous flux from the bench top fixture (Table 5) at a fixed position and
orientation for all tests. A light diffuser was not installed over the fixture as diffusers in final installation
areas vary space to space. Because the retrofit kits will utilize the fixture housing and diffusers which are
currently in place, comparison of bare bulbs was determined to be the best method for measuring lighting
characteristics. Actual power consumption was measured to compare with manufacturer specifications
using a P3 Kill A Watt meter.

9

Table 5: Illuminance (Lux) and tested power draw compared to manufacturer specified luminous flux (lumens) and
power draw. Italics indicate current standard fluorescent lamps. CREE dimming system set at ~50% power.
*indicates lumens only provided by manufacturer for single bulb

Manufacturer
Phillips
Precision Paragon
Precision Paragon

Model
Alto II F32 T8
TKG24-40XW-EU
TKG24-40VW-EU

Specified
Lumens
Power (W)
---*
64
2300-2600
22
2900-3200
27

Phillips
Phillips
CREE
Columbia
Columbia

12T8/40
15T8/40
UR-48 [Dimming]
GORK24-40VW-EU
GORK24-40XW-EU

1600
2100
variable
2900-3200
2300-2600

24
30
variable
27
22

Tested
Lux
Power (W)
1230
63
1350
24
1420
26
1020
1330
1220
1430
1330

25
33
31
26
24

While minimizing the variety of light types was desired for maintenance and inventory reasons,
reliability, potential effects on lifetime due to dimming, and long-term quality of illumination were
concerns for implementation of permanently dimmed CREE systems. Long term quality evaluations
would have to be performed before a recommendation for installation could be provided. However,
pricing at the time of the study for CREE UR2-48-45L-40K retrofit kits was not close to providing
feasible payback periods of less than five years, so long-term testing was not performed.
The Phillips LED lamps which were tested exhibited an unsatisfactory magenta tint, significant
flickering, and very low luminous flux. The Phillips LED lamps closely resembled the F32T8 fluorescent
bulbs, with a tubular casing and bi-pin ends. The use of previously installed electrical connection
hardware caused reliability concerns and could also increase overall installation time when compared to
self-contained units seen on other retrofit kits (CREE, Precision Paragon, Columbia).
LED lamps with internal drivers (driver contained within bulb) were also examined but due to
reliability and lifetime concerns due to driver heat production they were not considered for final
recommendation. Failure of only one of part of these systems (driver or bulb) will require a complete
replacement of the system instead of piecewise replacement available with external drivers.

10

More

importantly, bulbs with integrated drivers are often confused for fluorescent bulbs and/or installed
incorrectly with old driver ballasts that significantly reduce the overall efficiency.
While all evaluated LED retrofits either met or exceeded warranty requirements from IPF, the
Columbia and Precision Paragon models (manufactured by the same parent company) have the best
lighting characteristics as well as the lowest power draw of eligible systems (Figure 3). Both retrofit kits
will generally be referred to as Precision Paragon (PP) for simplicity. PP retrofit kits not only had
adequate CRI at specified CCT (75 at 4000K vs 76 at 4100K for F32T8) but also had the lowest power
consumption and overall cost of evaluated retrofits. Coupled with a 5-year warranty and 60,000hr
expected lifespan, PP retrofit kits were determined to be the most cost effective, highest quality option.
The X24-40VW-EU models of Precision Paragon/Columbia (where X represents TKG/GORK
respectively) have a slightly higher energy consumption (2W, ~8%) than X24-40XW-EU variants and are
recommended to account for any possible degradation that may occur over the lifetime of the system.

Figure 4: From top to bottom: CREE UR2-48-45L-40K backing with magnetic mounting, CREE front of lamp, PP
X-X24-40VW-EU backing with adhesive mounting, PP front of light bar

On-Rate Analyses
The light on-rate is the single largest determining factor for the economic viability of a retrofit
installation. For determining the savings afforded by retrofitting fixtures with lower wattage LEDs, the
amount of kWh saved each year is first determined by Equation 1 below.
11

𝑘𝑊ℎ 𝑠𝑎𝑣𝑒𝑑 𝑎𝑛𝑛𝑎𝑢𝑙𝑙𝑦 =

8,760 ℎ𝑟𝑠
𝑘𝑊
∗ % 𝑂𝑛 𝑅𝑎𝑡𝑒 ∗ (𝑃𝑓𝑙𝑢𝑜𝑟 (𝑊) − 𝑃𝐿𝐸𝐷 (𝑊)) ∗
𝑦𝑒𝑎𝑟
1000 𝑊

(1)

After determining the preferred LED retrofit kit (and accordingly, the power draw), the only
variable in the above Equation is the light on-rate. The amount of energy saved is therefore directly
proportional to the light on-rate, with 24/7 operation yielding the largest energy savings. Increasing the
light on-rate decreases the operational lifetime of installed retrofits, which must be accounted for in return
on investment (ROI) determination. With manufacturer specified lifetimes of 50,000 hours for the threelamp retrofit and 60,000 hours for the two-lamp retrofit, installations at all evaluated electricity costs will
pay back the installed cost of the retrofit a minimum of 1.76 years before the end of the specified lifetime
(Appendix, Table 19).
Hallways have previously been determined to have the highest light on-rates in the Psychology
Building on the MSU campus [3, 8]. Investigating the light on-rates of the hallways of other on campus
buildings was performed to capitalize on these previous findings. To determine the light on-rates in the
hallways of other buildings, Onset HOBO U12 temp/RH/light/ext channel data loggers were installed.
The loggers were installed on hallway walls ~4ft from the floor under an illuminated light fixture, and the
number of strings of lights that were on at the time of installation were noted (Figure 5). This allowed for
determination of a baseline lighting level. If the illumination levels dropped below 20 lux, it was assumed
that all strings of lights in the hallway were turned off. A middle level indicates that half of lamps in
hallway were turned off (a single string). An example of data from loggers is included, including each of
the described conditions (Figure 5).

12

100

200
Basement

80

160

70

140

60
50
40

120
100
80

30

60

20

40

10

20

0
7/5/16

7/7/16

7/9/16

7/11/16

3rd Floor

180

Intensity (Lux)

Intensity (Lux)

90

0
2/25/16

7/13/16

Date

2/27/16

2/29/16

3/2/16

3/4/16

Date

Figure 5: Clockwise from top left: Logger installation example; Onset HOBO U12 data logger and placard;
measurements of luminous flux and on-rates in the MSU Natural Resources building 3rd floor; basement

Loggers were installed for one-week intervals and results were categorized by day. Installation
was performed on all floors where possible, and if a floor was not surveyed the on-rate was assumed to be
the average of similar floors in the same building which were surveyed. On-rates were then assumed to be
the same for each day of the week throughout the year. In spaces where hallway lights are not on 24 hours
per day, variations exist in the total amount of light on time due to irregular occupant behaviors (Table 6).
13

Table 6: Light on-rate for Chemistry Building floors basement and 1-6

Floor Monday Tuesday Wednesday Thursday Friday

Saturday Sunday

6

72%

100%

87%

100%

75%

70%

94%

5
4

65%
100%

69%
100%

68%
67%

72%
97%

77%
61%

100%
100%

77%
97%

3
2

100%
66%

100%
100%

100%
67%

100%
100%

100%
0%

100%
0%

100%
1%

1
B

71%
100%

24%
100%

66%
69%

100%
100%

70%
100%

3%
64%

0%
0%

Because weekday and weekend behavior can vary, lower/upper bound on-rate estimations were
determined by assuming that each weekday on-rate was equal to the lowest/highest measured weekday
on-rate respectively. The same was done for weekends to estimate the variability that is inherent in
systems which can be controlled by users. For example, the sixth floor in the Chemistry Building (Table
6) would assume that the total on-rate for Monday through Friday was 72% and for Saturday and Sunday,
70%. This would be taken as the lower bound for the light on-rate. For the upper bound, Monday through
Friday would be assumed to have 100% overall light on-rate, and Saturday and Sunday would have a 94%
total on rate. The hallways of several academic buildings had 100% on-rates measured (24-hour daily
operation) with an average of 83% on-rate across all the buildings surveyed (Figure 6). The combination
of occupancy by students and faculty, the scheduling of janitorial staff during night time hours, and
concerns for safety are all likely explanations for the high on-rates.

14

Light On Rate

100%
80%
60%
40%
20%
0%

Figure 6: Annual light on-rates for surveyed buildings on the MSU campus

Financial Analysis
Determination of the economic viability and ROI for each building/retrofit option was performed.
The typical metric used by IPF for determining the suitability of installation for an energy reduction
technology is a payback period of less than five years, with a total capital investment (C0), cash flows
(CF), and time (t):

𝑃𝑎𝑦𝑏𝑎𝑐𝑘 𝑃𝑒𝑟𝑖𝑜𝑑 =

𝐶0
𝑛
∑𝑡=1 𝐶𝐹𝑡

(2)

C0 consists of labor cost, retrofit hardware cost, and recycling of fluorescent lamps/ballast. Due to
the ease of retrofit installation, expected time per fixture is 25 minutes, and at this rate the capital cost of a
single installed retrofit for a two-lamp F32T8 fixtures is $100.53. The capital cost of a single installed
retrofit for a three-lamp fixture is $120.12. It is expected that previous fixture hardware including lens,
troffer, and wiring, are in acceptable working condition and retrofits can be installed with little to no
fixture modification. Due to the mounting hardware on chosen retrofit kits (Figure 4) the previous
mounting hardware for attachment of F32T8 lamps can be left within the fixture.
15

Cash flows consist of energy savings due to electricity use reduction, any rebates, assumed failure
rate (1 fixture per 3 months), and replacement of all retrofits at end of usable lifetime. An annual utility
escalation rate of 2.23% [12] was used to account for expected changes in the cost of electricity
production. In all surveyed cases, the manufacturer specified lifetime of the chosen lamps exceeds the
projected payback period. To evaluate the effect of end of life replacement of retrofit kits, the net present
value (NPV) of the investment is used, with a discount rate of 3.5% (i), where:
𝑁

𝑁𝑃𝑉 = ∑
𝑖=1

𝐶𝐹𝑡
− 𝐶0
(1 + 𝑖)𝑡

(3)

The cost at the end of system lifetime is assumed to be the same as the initial C0. The evaluation
of NPV with respect to time is determined for two buildings with similar C0, Plant and Soil Sciences
(98% light on-rate) and Biomedical & Physical Sciences (70% light on-rate) (Figure 7). While the cost of
replacement at end of life is significant, a positive return on investment (ROI) is realized several years
before this end of life replacement point.

16

Net Present Value

$60,000
$50,000

Plant and Soil Sciences

$40,000

Biomedical & Physical Sciences

$30,000
$20,000
$10,000
$0
-$10,000
-$20,000
-$30,000
2017

2020

2023

2025
2028
Date (years)

2031

2034

Figure 7: Net present value with respect to time, assuming installation date of 10/1/2017. Large drops in NPV
indicate end of expected retrofit lifetime and complete replacement

The ROI can be used to normalize profits from an investment C0 to easily compare the
profitability from vastly different C0:
𝑁

𝑅𝑂𝐼 = ((∑ 𝐶𝐹𝑡 ) − 𝐶0 ) /𝐶0

(4)

𝑖=1

The two main factors in determination of CF due to electricity use reduction for a given building
are the total amount of electricity reduced (Equation 1) and the rate at which electricity is purchased. As
discussed earlier, the total amount of electricity reduced is dependent solely on the light on-rate in a given
building. Because MSU produces most the energy consumed on campus locally at the TB Simon Power
Plant [9], pricing of electricity occurs at several rates. These rates are 1) fuel only cost ($0.05755/kWh),
2) produced cost (including operations and maintenance, $0.07492/kWh) and 3) delivered cost (including
delivery system operations and maintenance, $0.09206/kWh). The delivered cost is used for buildings
under the administration of Athletics or the Residential and Hospitality Services. Unless otherwise

17

specified, following results are evaluated at the produced cost of electricity. The Plant and Soil Sciences
Building is used to model the sensitivity of a retrofit installation to five factors (Figure 8, Table 7).

Utility Escalation
Rate

Retrofit Cost
7
6
5
4
3
2
1
0

Light On Rate

High
Mid
Low
Installation Time

Electricity Cost

Figure 8: Sensitivity analysis of factors affecting payback period (in years) for LED retrofits targeting 2 lamp
F32T8 fixtures in the Plant and Soil Sciences Building
Table 7: Values used for sensitivity determination in Figure 8. Baseline values are indicated by bold typeface

Low
Mid
High

Electricity
Retrofit
Installation
Light OnUtility
Cost
Cost
Time
Rate
Escalation Rate
$0.05755
$63.48
25
50%
0.2%
$0.07492
$70.53
35
75%
1.2%
$0.09206
$77.58
45
98%
4.2%

Another important factor in the economic viability of LED retrofit installations comes in the form
of rebates from energy providers. Currently, MSU purchases approximately 8% of its electricity from
Consumers Energy (a utility company). This is because the electricity either cannot be provided by the
TB Simon Power Plant or can be purchased at a lower rate than production costs, which can occur during
off-peak pricing hours. Subsequently, this still allows the university to take advantage of rebates provided
by Consumer’s Energy. As of January 2017 [11], the rebate provided for the recommended LED retrofits
is $0.25/Watt reduced. With this rebate, two-lamp (PP) retrofit fixtures receive a rebate of $9.25/fixture
and three-lamp (CREE) retrofit fixtures receive a $13/fixture rebate. Without rebates applied, the payback
18

period for a given installation can be expected to increase by 11Âą1%. There is a $2,000,000 annual cap on
rebates per facility (...contiguous property for which a single customer is responsible for paying the
Consumers Energy electricity and/or natural gas bill [10]) so it should be verified that this limit has not
been exceeded before applying for rebates.
The payback period at the produced cost of electricity is less than five years for spaces which
have an on-rate greater than 70% (Figure 9). Only four of the surveyed spaces do not achieve the desired
payback period: the Computer Center (54% on-rate), Natural Resources (67%), Wells B&C wings (67%),
and Bessey Hall (54%). It should be noted that when looking at Wells Hall in its entirety, the payback

Payback Period (years)

period is less than five years. This is because of the relative scale of the B&C wings to the A&D wings.

10
9
8
7
6
5
4
3
2
1
0

Figure 9: Payback period for LED retrofits at 19 surveyed spaces on the MSU campus.

A positive ROI indicates that an investment pays itself back within the specified time range. The
ROI is evaluated at the end of the lifetime for all installations (Figure 10). While five-year payback
periods are not available with all evaluated buildings, the positive ROI over the lifetime of the
investments suggests that these retrofits are still economically viable over the lifetime of the installation.

19

80%
70%
60%
50%
40%
30%
20%
10%
0%

Figure 10: Return on investment for LED retrofits at 19 surveyed spaces on the MSU campus

Discussion
Installation of LED retrofits across all surveyed spaces (19 buildings) would yield an annual
reduction of 1.06 GWh of electricity consumption, or ~0.4% of the net annual power delivered to
buildings on the MSU campus. Retrofitting fixtures with LEDs is not only economically viable in most
spaces, but also requires no behavioral modification and does not have user overrides as seen with
lighting controls. A pilot installation occurred in hallways of the Psychology Building using the CREE
UR2-48 retrofit. Occupant responses were positive regarding energy efficiency gains and lighting quality
of installed retrofits (Figure 11).

20

Did you notice thtat the hallway lighting in
this building was converted to LEDs in 2014?

Are you currently satisfied with the
hallway lighting?
5

No

4
3
2

Yes

1
0

5

10

15

0

How important is it to you that LEDs has
created significant energy efficiency gains
in the building ? (Rate on a scale of 1 to 5)
5

4

4

3

3

2

2

1

1
2

4

6

4

6

How interested are you that MSU is
investigating ways to implement energy
efficiency technologies?

5

0

2

8

0

2

4

6

8

Figure 11: Survey of office occupants in Psychology Building hallways which had LED retrofits installed

Some of the important cosiderations for IPF to consider before implementation of LED retrofits
include surveys of current lighting levels to determine which retrofits would be suitable for a space to
comply with ANSI/IES standards. As discussed previously, lighting levels are not consistent across
building hallways on campus and are not always within acceptable illumination level ranges. Stepping
down three-lamp F32T8 fixtures to retrofits specified for two-lamp fixtures can be evaluated dependent
on illumination level. Use of lower output X24-40XW fixture could also be examined in hallways with
two-lamp fixtures which are currently overlit.
The use of different lighting systems adds complication and bulk to the inventory system which
would need to be maintained for periodic maintenance and replacement of lamps/drivers which
prematurely fail. A final consideration is the overall application lifetime of these systems. With expected
system lifetimes ranging from seven (100% light on-rate) to twelve (54% on-rate) years, at the end of life

21

there is no guarantee that these systems will still be available. However, the very positive ROI values seen
over the lifetime of an installation is significant enough that the investments are still very favorable.

Hallway Occupancy Based Lighting Controls
Rationale
Besides reducing the amount of energy consumed by changing to lower wattage fixtures, a
second option for reducing the energy consumed by lighting is reducing the total amount of time that
fixtures are powered on. While past efforts (sticker campaigns, recommendations to janitorial staff) have
targeted behavioral changes, results from light on-rate analyses in the LED retrofit study indicate that
these measures often are less effective. This leads to a significant amount of electricity used at times when
lighting is not required – when spaces are not occupied.
Occupancy based lighting controls are systems which are hardwired into lighting strings to toggle
lights on and off based on input from occupancy sensors which detect motion [5]. If no motion is detected
for a set period (specified by the user on installation) lights are automatically turned off. On sensing
motion, the occupancy sensors are triggered which switches paired lighting strings on. Wide ranges of
energy savings are estimated to be available, from 15 to 75% depending on the source [5].
The integration of hallway occupancy sensors were studied as a case study in the Biomedical and
Physical Sciences (BPS) building and Anthony Hall on the Michigan State University campus. This study
seeks to evaluate the energy savings resulting from these installations. These results are then applied to
other buildings on campus to estimate the economic viability of further retrofits of occupancy based
lighting controls.

Available Energy Reductions
From the light on-rate analyses performed for the LED retrofit study, resultant energy savings
from installed hallway occupancy sensors can be observed in both BPS and Anthony Hall (Table 8). Note
that occupancy sensors appear to have been deactivated on floors 1 and 2 in Anthony Hall and floor 2 in
22

BPS. The first floor in BPS contains student study areas, classrooms, and a small convenience store,
while all other floors with installed and functional occupancy sensors are in hallways which primarily
service office and lab spaces.
Table 8: Light on-rates for Anthony Hall and Biomedical & Physical Sciences (BPS) measured with an Onset
HOBO U12 temp/RH/light/ext logger; PH – Penthouse; B – Basement

Floor
Anthony
BPS

PH
6
5
4
3
2
1 B
40% 100% 100%
19% 47% 50% 61% 61% 100% 81% 37%

With an average on-rate reduction of 52 Âą 19% across activated occupancy sensor spaces,
realized electricity use reductions are within the expected ranges [5]. Removing spaces which are not
regularly occupied (BPS basement, penthouse), regularly occupied floors (BPS 1, 3-6, Anthony 3) saw an
average on-rate reduction of 43 Âą 14% due to installed occupancy sensor control systems.
While these energy use reductions are significant, results are lower than would be seen through
the installation of LED retrofits while maintaining 24 hour on-rates. LED retrofits for two F32T8 lamp
fixtures realize a 57% energy use reduction over operating times while retrofits for three-lamp fixtures
yield a 54% reduction in energy consumption. Combination of the two technologies (LED and occupancy
sensors) would result in an estimated average energy use reduction of 79% with respect to lighting.

Financial Analysis
Occupancy Based Lighting Controls Only
Cost estimations are performed on linear foot basis, as occupancy sensors have a defined
distribution and therefore application area size. The light on-rate reduction of 43% Âą 14% is used to
estimate electricity use reductions from installed systems. The current MSU Standards specified
occupancy sensors are highlighted in Table 9.

23

Table 9: MSU Standards occupancy sensors; Manufacturer: Leviton; Cost includes OSA20-R00 power pack
required for operation

Model

Coverage (ft)

Cost

$/ft

OSC05-M0W

23 x 23

$98.46

$4.28

OSC10-M0W

46 x 23

$99.35

$2.16

OSC20-M0W

62 x 32

$103.52

$1.67

Hallway widths in surveyed buildings vary from 6 – 12ft (Figure 12) so hallway length is the
primary factor for quantity of sensors installed. Because each sensor can cover at most 62 linear feet,
multiple occupancy sensors will need to be installed in any hallway over 62ft in length. Occupancy
sensors must be tied into a centralized lighting string controller, so wiring cost will be a function of total
hallway length as well. Due to the difficulty in estimating these different conditions, a $/linear foot cost
required to reach a five-year payback period is determined instead of estimating the cost on a building by
building basis. The total cost can then be compared to the value for economic viability.

Figure 12: Varying hallway widths of MSU buildings; clockwise from top left: Plant and Soil Sciences (10ft);
Engineering Building (12ft); Student Services Building (10ft); Biomedical and Phsyical Sciences (6ft)

Plant and Soil Sciences (PSS) is used as a case study for these estimations as well as for
comparison to installing LED retrofits previously described. From the MSU Planning & Budgets Spartan
24

Space system, the total area of public corridors for PSS is 56,000 ft 2 (5,600 linear feet). A total of 256
fixtures in public corridors in the PSS building were on 100% of the time surveyed. Applying the average
on-rates seen in BPS and Anthony Hall (Table 8) due to functioning occupancy based lighting controls
results in an annual estimated electricity reduction of 60.6 Âą 20.0 MWh/year for PSS. At the produced
cost of electricity ($0.07492/kWh) this results in annual savings of $4,600 Âą 1,500. Over a five-year
lifespan, the total expected returns are $23,000 Âą 7,500. This equates to $4.1 Âą 1.3 per linear foot of
hallway.
Per distribution specifications for the Leviton OSC05, OSC10, and OSC20 occupancy sensors
(above), it is assumed that spacings between sensors will be 23, 46, and 62 feet respectively. Results are
tabulated in Table 10, where viability is defined as achieving a 5-year payback period. The value of
required occupancy sensors is the minimum expected value, with more thorough surveying required to
determine the exact number of occupancy sensors required.
Table 10: Installed occupancy sensor cost required to achieve financial viability; Manufacturer: Leviton

Model

Required Occupancy
Sensors

Cost for
Viability

OSC05-M0W

244

$94.3 Âą 30.7

OSC10-M0W
OSC20-M0W

122
91

188.5 Âą 61.5
$252.7 Âą 82.4

With these costs for viability and the very small cost difference between occupancy sensors with
vastly different coverage areas, installation of sensors with the largest distribution footprint will provide
the highest likelihood of achieving financial viability. Average costs for installation and ancillary
equipment for the OSC10 and OSC20 sensors are $89.2 and $149.2 respectively. OSC05 equipment cost
exceeds the cost requirements for viability.

25

Occupancy Based Lighting Controls Coupled with LED Retrofits
For maximizing the savings of occupancy based lighting controls, combination with LED retrofits
to reduce electricity use when lights are in the on state is performed. Using the same methodology as
above, electricity requirements are reduced by 109.2 Âą 8.8 MWh/year, yielding cost savings of $8,180 Âą
660/year. With a capital investment of $24,500 for LED retrofits, a total installation cost of $16,400 Âą
3,300 is required to achieve fiscal viability of an occupancy sensor installation. This equates to $2.9 Âą 0.6
per linear foot of hallway. The costs for viability of the different occupancy sensors are outlined below
(Table 11).
Table 11: Installed occupancy sensor cost required to achieve financial viability in conjunction with LED based
fixtures; Manufacturer: Leviton

Model

Required Occupancy Cost for
Sensors
Viability

OSC05-M0W
OSC10-M0W

244
122

$67.4 Âą 13.5
$134.8 Âą 27.0

OSC20-M0W

91

$180.7 Âą 36.3

Average costs allowed for installation and ancillary equipment for the OSC10 and OSC20 sensors
used in conjunction with LED fixtures are $35.4 and $77.2 respectively. OSC05 sensor and power pack
costs exceed the minimum cost for financial viability of an occupancy sensor installation.

Discussion
For the purposes of determining when occupancy sensors may be a valid solution for reducing
overall electricity consumption, a more in-depth case study must be performed to determine the costs of
installation and associated equipment. While the sensors and power pack equipment show favorable
comparison on a cost per foot and cost per installed sensor basis, these are only part of the total capital
cost of installation. Once an estimate for installation and ancillary equipment is determined, the overall
cost of an installed sensor can be compared to on-rate reductions in existing occupancy sensor controlled

26

spaces to determine the viability of an occupancy sensing installation. Cost requirements for viability for
four different high (~100%) on-rate buildings are highlighted below (Table 12).
Table 12: Costs required to trigger a 5-year payback period for occupancy sensor installation in four MSU
buildings. OSC10/OSC20 indicate Leviton occupancy sensor models

$/ft
Engineering

5.4 Âą 1.8

Biochemistry

7.9 Âą 2.6

PSS

4.1 Âą 1.3

Wells

4.4 Âą 1.4

OSC10-M0W
Req. Occ.
Sensor
Cost
156 $246.6 Âą 80.3
33 $353.8 Âą 115.2
122 $186.0 Âą 60.6
162 $204.5 Âą 66.6

OSC20-M0W
Req. Occ.
Sensor
Cost
116 $331.6 Âą 108
24 $486.5 Âą 158.4
91 $249.4 Âą 81.2
121 $273.7 Âą 89.1

Daylight Harvesting Light Management
Rationale
Daylight harvesting light (DLH) management systems are used to exert automated control of
lighting levels to both maximize usability of a space and to minimize energy consumption of installed
lighting elements in areas which contain large windows, especially in south, and to a lesser extent east
and west, facing orientations. These systems use sensors which detect illumination levels in a space and
automatically adjust dimming levels to match the total illumination detected to a previously programmed
set-point. Combining scheduling, occupancy sensors, and daylight dimming controls with an
appropriately tuned set point will yield the largest energy reductions [12]. These three separate systems,
however, lead to increased cost of installation as well as increasing the likelihood of control issues which
arise from a larger number of system inputs. Besides the possibility of significant energy use reductions,
greater reliance on daylight instead of electric lighting can see increases in occupant satisfaction with
lighting conditions [13,14]. With both occupancy satisfaction and possible energy reductions, the fiscal
viability is determined in the following analysis.
While the previous lighting retrofit (LED, occupancy) analyses have been performed for interior
hallways within buildings, the DLH systems being considered in this section focuses on classroom and
27

office spaces. Most hallways on the MSU campus are interior hallways and therefore do not have
windows; the majority only experience daylighting in entryway areas. These spaces are therefore
untenable for daylight harvesting systems. However, there are numerous office spaces with adequate
window area for considering such systems.

Financial Analysis
To determine available savings from these installations, a baseline estimate for energy reduction
must first be determined. The US Department of Energy’s (DOE) Energy Efficiency and Renewable
Energy department provides a Daylighting tool for estimating savings available through the
implementation of daylight harvesting light management systems [15]. Minimum illuminance through
panels are provided by NREL for a southern angle of incidence under mostly cloudy conditions. Three
lamp fixtures output 4500 lumens per previous fixture testing for LED retrofit analyses [3]. With two 20’
x 4’ windows, the calculated proposed reduction is 85%. This same energy use reduction is estimated for
fixture lumen outputs between 3200 and 6200 lumens, so can be applied in determining savings for both
two-lamp and three-lamp installations. As shown in the following analysis, however, even these very
large reduction estimations provided by the DOE do not provide paybacks of less than 5 years because of
the high cost daylight harvesting equipment.
To determine the total cost of installation, an installation time of one hour for user control
hardware (Lutron MRF2-F6AN-DV-WH wall dimmer switch), one hour for each sensor cluster, fifteen
minutes for dimming ballast installation in fluorescent fixtures, and thirty minutes for LED retrofit
installation are assumed. Each sensor cluster (Lutron LRF2-OCR2B-P-WH occupancy sensor, LRF2DCRB-WH daylight sensor) had a baseline cost of $116.13 and the wall dimmer switch (one per room
installation) had a unit cost $105.56. LED retrofits were priced at $70.53 and $90.12 for two-lamp and
three-lamp retrofits respectively. The labor rate was set at $60/hour.

28

The payback period for several different configurations are examined below at different numbers
of fixtures per occupancy/DLH sensor cluster. These values are determined for four sensor clusters, where
the total number of fixtures covered by a single point is the product of fixtures per sensor cluster and
number of sensor clusters.
30
2L LED
2L Fluo
3L LED
3L->2L LED
3L Fluo
3L->2L Fluo

Payback Period (years)

25
20
15
10
5
0
0

2

4

6

Fixtures per sensor cluster

8

Figure 13: Comparison of payback periods for 6 different lighting configurations to implement DLH systems.
Sensor clusters consist of DLH and occupancy sensor

The schemes investigated above in Figure 13 are detailed below:
•

2L LED: Two lamp LED retrofit (Precision Paragon TKG24 as described in Hallway LED
Retrofits) with dimming ballast replaces two lamp F32T8 configuration

•

2L Fluo: Two lamp fixture configurations maintained, existing ballast replaced with dimmable
ballast ($62.46/fixture)

•

3L LED: Three lamp LED retrofit (CREE UR2-48 as described in Hallway LED Retrofits) with
dimming ballast replaces three lamp F32T8 configuration

•

3L→2L LED: Two lamp LED retrofit (as above) with dimming ballast replaces three lamp F32T8
configuration

•

3L Fluo: Three lamp fixture configurations maintained, existing ballast replaced with dimmable
ballast ($62.46/fixture)
29

•

3L→2L Fluo: Two lamp F32T8 replaces three lamp F32T8 configuration

The replacement of three lamp fixtures with two fluorescent lamps or retrofits specified for
replacement of two fluorescent lamp containing fixtures is included to cover areas which may be over-lit.
These conditions generally exist because of changes in lighting standards over time [6, 7] which have
seen a decrease in recommended lighting levels since the construction of several buildings on campus.
Sensitivity Analysis
Scale dependency (quantity of fixtures for each DLH/occupancy sensor cluster) of DLH
installations was described above and highlighted in Figure 13. However, other factors which can have a
significant effect on the projected payback period are quantified in Figure 14. Payback periods are most
sensitive to the cost of electricity and electricity use reduction (DOE proposed reduction). For the
fluorescent fixtures, these two factors show an identical relationship, as both act on a direct kWh
reduction basis. Changes in the proposed reduction for LED retrofits see a change in kWh reduction due
to a changed baseline fixture wattage from originally installed fluorescent fixtures.

30

Chane in Payback Period

100%
Cluster Count
Electricity Cost
Elec Reduction (Fluo)
Elec Reduction (LED)
Labor Cost
Eqpt. Cost

75%
50%
25%
0%
-25%
-50%
-100%

-50%

0%

50%

100%

Change of Variable

Figure 14: Sensitivity analysis variables affecting DLH installation payback. All variables except [expected]
electricity reduction apply to both LED and fluorescent variants. Fluorescent electricity reduction refers to both 2/3
lamp and ballast/no ballast addition. LED electricity reduction refers to both 2/3 lamp retrofits.

Case Study
Room 2250 in the Michigan State University Engineering Building was used as a case study to
examine the payback period for an installed DLH system with several different variations on lighting
installations accounted for (Figure 15). The four lighting schemes outlined below are described in detail
above.

9

Payback Period (yrs)

8
7
6
5
4
3
2
1
0
3L Fluo

3L LED

3L->2L Fluo

3L->2L LED

Figure 15: Case study for DLH system installation in Engineering Building room 2250 (two sensor clusters)

The room is 1152 ft2 (24’ x 24’) and contains 18 three lamp F32T8 fixtures as well as an installed
occupancy sensor system. The room consists of an exterior east oriented wall with 176 ft 2 of window

31

space with user operated blinds. The DOE calculator previously described assumes a south facing
orientation for estimating illumination from daylight entering a space. Therefore, the upper bound of
payback periods indicated by the error bars in Figure 15 above are more likely to represent results from
decreased illumination from daylight due to room orientation. For this case study, it is assumed that the
blinds remain raised at all times. Equipment costs are included in Table 13. A previously described sensor
cluster consists of an occupancy sensor and daylight sensor. These clusters would be utilized in spaces
without previously installed occupancy sensing systems.
Table 13: Baseline costs used in DLH system analyses

Manufacturer
Lutron
Lutron
Lutron
Lutron

Model
MRF2-F6AN-DV-WH
LRF2-OCR2B-P-WH
LRF2-DCRB-WH
PJ2-3BRL-GWH-L01

Description
Cost/Unit
Dimmer Switch
$106
Occ. Sensor
$50
Daylight Sensor
$66
Remote (Wireless)
$13

With the results from the EB 2250 case study (Figure 13) and the generalized results (Figure 15)
it is difficult to recommend the installation of DLH systems in classrooms or offices. All schemes
examined, even with an assumed 85% reduction in electricity use, are unable to achieve 5-year payback
periods. The only way that these fast paybacks could be achieved are in spaces which already have
dimming ballasts installed in each fixture. In Table 14, the fluorescent dimming ballast price required to
trigger a 5-year payback period for the case study area (EB 2250) is highlighted at different electricity
reduction rates (assuming occupancy sensor is already installed).
Table 14: Cost of dimmable fluorescent ballast at different electricity reduction rates by DLH system required to
trigger a 5-year payback (at $0.07492/kWh)
Electricity
Reduction
Ballast
3 Lamp
Cost
2 Lamp

39.2%

40%

$0.00
--

$0.77
--

50%

58.8%

60%

70%

80%

90%

$10.52

$19.11

$20.27

$30.02

$39.76

$49.51

$0.00

$0.77

$7.27

$13.77

$20.27

--

32

Discussion
The high installation costs for daylight harvesting light management systems are the major
obstacle in these retrofit implementations. However, if seeking LEED accreditation through retrofits or
for new installations, daylight harvesting is an efficient way to reduce total energy consumption and get
LEED points towards certification when cost (and payback period) are less important [16].
Two trial office installations were performed, with a wired sensor in one office and wireless
sensor in the second. The wireless system exhibited seemingly random dimming of lights, responding to
perceived changes in lighting levels. This continued through swapping out daylight sensors used as well
as placement modification. The wired system experienced none of these fluctuations. As has been noted
in previous literature [5] the issue of occupancy sensors detecting only major motion sometimes resulted
in ‘false-off’ triggers when the space was occupied but occupants were performing desktop work which
resulted in only minor movements. This would in likelihood be mitigated by the larger quantities of
occupants if scaled up to a classroom sized installation.
Initial purchase price on a per fixture basis will be less for LED retrofits than DLH installations
on fluorescent fixtures at values below five fixtures per sensor cluster (three-lamp retrofit) or twelve
fixtures per sensor cluster (two-lamp retrofit). Electricity reductions from these two LED retrofits are
54% and 58% respectively, much lower than the 85% reduction stipulated by the DOE calculator.
However, it should be noted that T5 fixtures dimmed to 0% lumen output have been shown to still
consume approximately 20% of baseline power [17]. While currently installed fixtures contain T8 lamps,
comparable results are likely and would yield lower electricity reduction levels than determined by the
DOE calculator.
Because of the footprint of most office spaces (generally one to two fixtures) sufficient payback
periods are not achievable with DLH systems presently available. However, installation and monitoring
of several case study classroom spaces is recommended to determine actual electricity use reductions due

33

to DLH installation. Different orientations and window to wall ratios should also be tested. Three
orientations (E, N, S) and two window to wall ratios, or 6 trial classrooms, are recommended for testing
for at least a 6-month monitoring period which can be extrapolated to full year results.

Summary
Retrofitting fluorescent fixtures with LED in areas with high light on-rates yields very significant
reductions in total energy use at a relatively low capital investment. Of the nineteen buildings surveyed,
seventeen can achieve payback periods of five years or less by retrofitting fixtures with LEDs.
Recommendations have been made to IPF and installation is now planned for the hallways of the
Engineering Building, Plant and Soil Sciences, Biomedical and Physical Sciences, and Wells Hall which
will yield lifetime energy savings of 3,006 MWh, an ROI of 64%, and lifetime net savings of $98,000.
Occupancy sensing was also investigated as a tool for reducing energy use on campus, but further
analysis must be done to determine total installation costs which can then be compared to the
requirements determined for economic viability. While DLH systems can provide significant energy
reductions, the cost of installing these systems does not allow for suitable payback periods, with a best
case of 5.3 years without completely replacing the installed lighting systems.

34

CHAPTER 2: Energy Use Analyses

Introduction
Lighting consumes 17% of the total electricity use in US commercial buildings [1], which
corresponds to 14% of total energy use [18]. The remaining 86% of energy used can be split broadly into
heating (34%), ventilation (6%), air conditioning (10%) (HVAC, total 50% of energy use), and other
equipment (36%; computing, refrigeration, office equipment, etc.). In commercial buildings, office
equipment and computing consume 5% of the total energy used in office spaces [18] while space cooling
consumes a similar amount of the total energy consumption (10%) [18].
The energy needs of Michigan State University campus are serviced primarily by the TB Simon
power plant, a co-generation facility. Since the installation of the power plant, electricity generation has
nearly tripled from 120GWh in fiscal year 1965-66 to 347GWh in fiscal year 2012-13 [9]. In this same
time frame, steam demand has increased by less than 50%, and excess steam produced is condensed in
cooling towers which results in a significantly lower overall process efficiency than at the beginning of
the facilities lifetime. Steam is primarily used for heating of campus buildings in the winter and to a lesser
extent, cooling in the summer through the use of adsorption chillers.
While steam use for heating is currently at its maximum, increasing the amount of steam used in
the summer for cooling is possible with a significant portion of campus currently underserved with
respect to cooling. Increasing the summer steam load would increase overall efficiency of the cogeneration process by consuming more steam and requiring less electricity. A 630MWh/yr electricity
reduction is available by removing all installed window air conditioning units.
Another way to increase the efficiency of the co-generation process is to decrease overall
electricity generation at the facility. Purchase of off-peak electricity from an outside provider has driven
down total generation to some extent, but decreasing electricity requirements during more expensive peak
times can only be done by decreasing the total required load. While lighting was previously discussed,
computing and other office equipment uses nearly half of the load required by lighting [1, 18]. By
35

determining the composition of electricity consuming devices in the average office on the Michigan State
University campus, recommendations are made for reducing office equipment electricity intensity.

Evaluation of Steam Cooling Requirements
Rationale
Window air conditioning (WAC) units on the MSU campus are present primarily in the older
sections of the campus north of the Red Cedar River. While most of the cooling requirements are
provided to buildings on campus using steam produced in the on-campus co-generation plant, many of the
buildings north of the Red Cedar are primarily cooled using distributed cooling with window air
conditioners.

Figure 16: MSU campus map highlighting (in red) buildings with window AC units

It is estimated that these buildings consume approximately 630 MWhrs of electricity annually,
which could potentially be offset by excess steam production. Currently no reliable source of hourly
steam or electricity production are available to determine actual availability of steam sources. However,
steam use for HVAC generally peaks on a per building basis in January for heating and all spaces on the
main campus are serviced with steam for heating purposes. With this analysis, the amounts of steam

36

required for cooling these spaces will be determined to be available for reference on verifying quantities
of steam available during cooling months (May – September).

Electricity Use Analyses
The buildings with installed WAC units are most easily compared by determining the density of
units for a given building floor area (ft2/unit) (Table 15). Building areas were retrieved from the MSU
Office of Planning and Budgets [19].
Table 15: Quantity of window mount air conditioning units by building, organized by intensity (units/ft2). Annual
energy use for WAC units estimate from US EPA/DOE window air conditioning unit calculator [19]

Building
Linton
Old Botany
Natural Sciences
Computer Center
Berkey
Giltner
Agriculture Hall
Olds Hall
Museum
Kedzie
Old Horticulture
Kresge Art Center
Farrall Hall
Mason/Abbot
Snyder/Philips

Building
Density
Annual Energy
# Units
Area (ft2)
(ft2/Unit)
Use (kWh)
74
32964
445
43660
19
10216
538
11210
186
155173
834
109740
75
65076
868
44250
122
107273
879
71980
208
209084
1005
122720
108
114997
1065
63720
51
56990
1117
30090
29
44938
1550
17110
87
137600
1582
51330
16
35054
2191
9440
27
79585
2948
15930
18
70104
3895
10620
26
165303
6358
15340
23
241074
10481
13570

For estimating annual energy use, the Room Air Conditioner Calculator for consumer grade
window AC units from the US EPA/DOE was used [20]. An Energy Efficiency Ratio of 10.8 was
assumed as specified within the calculator, as well as 578 full load cooling hours (FLCH) for Lansing,
MI. The number of cooling degree days per year (CDD) was 746 ± 166 for 2010 – 2016 in Lansing, MI.
Using a design temperature of 95°F, the calculated annual FLCH (Equation 5) is 596 ¹ 132, in good
agreement with the value of 578 provided by the US EPA/DOE calculator [21, 22].
37

𝐹𝐿𝐶𝐻 =

𝐶𝐷𝐷 ∗ 24
𝑑𝑒𝑠𝑖𝑔𝑛 𝑡𝑒𝑚𝑝 °𝐹 − 65°𝐹

(5)

In total, there were 1069 window AC units counted on these buildings, accounting for
approximately 630 MWh/year in electricity consumption. Combined with the inadequate cooling seen in
public spaces, determining possible savings from upgrading to steam based cooling will be performed.
The electricity consumption outlined can be offset through using of excess steam produced at the TB
Simon cogeneration power plant on the MSU campus. Currently there are no estimators for steam
requirements for new cooling installation so this analysis proposes a comparative estimating method for
determining steam requirements for retrofitting of buildings, which can then be compared against
available steam when determining viability of retrofits.

Steam Use Analyses
Determining average steam use for a given building on the MSU campus is difficult due to the
high variability of both heating and cooling degree days (HDD and CDD respectively). A heating degree
day is the number of degrees below a base temperature that the average daily temperature is. For example,
if the average daily temperature is 50°F and the base temperature is 65°F, that day had an HDD of 15. If
every day in November had an HDD of 15, then the monthly HDD would be 450. Cooling degree days
are determined in the same manner as HDD, but for average temperatures in excess of 65°F. For the
purposes of these analyses, 65°F is used as the base temperature for both HDD and CDD determination.
Total degree days on a monthly or annual basis have significant variability, seen in comparing CDD for
2016 to both 2015 [22] and the 1981-2010 averages [23] (Figure 17).

38

1600

250

Heating Degree Days

200
150
100
50

1981-2010
2015
2016

1400
1200
1000
800
600
400
200

Dec

Nov

Oct

Sep

Jul

Aug

Jun

May

Apr

Jan

Dec

Nov

Oct

Sep

Aug

Jul

Jun

May

Apr

Mar

Feb

Jan

Mar

0

0

Feb

Cooling Degree Days

300

Figure 17: Average cooling (left) and heating (right) degree days on a monthly basis. Note differing scales on each
plot. A degree day is defined as the number of degrees the average temperature is over/under a base temperature
(65°F). If the average daily temperature for each day in November (30 days) was 50°F, the total number of
(heating) degree days would be 450.

This variability in degree days year to year is very important when estimating the potential
requirements for steam based cooling in spaces previously cooled using distributed WAC units. If degree
day values are not accounted for, rates of steam use can appear to have very large deviations. When
plotted against degree days it becomes clear that the large variations in monthly steam use arise from
variations in degree days (Figure 18). For example, during the ‘heating months’ of October – April the
average amount of steam used in Berkey Hall during 1/2013 – 12/2016 is 1.02 ± 0.14

𝑘𝑙𝑏 𝑠𝑡𝑒𝑎𝑚
,
𝑇𝐷𝐷

where

TDD is the total degree days (sum of CDD and HDD). If the steam use is not related to the total degree
days, the average amount of steam used for this date range is 930 Âą 380

𝑘𝑙𝑏 𝑠𝑡𝑒𝑎𝑚
.
𝑚𝑜𝑛𝑡ℎ

Without accounting

for TDD, the standard deviation is 41% of the average, compared to 14% when accounting for TDD.

39

CDD

HDD

1,800

Berkey

1600

1,600

1400

1,400

1200

1,200

1000

1,000

800

800

600

600

400

400

200

200

7/1/2016

1/1/2016

7/1/2015

1/1/2015

7/1/2014

1/1/2014

7/1/2013

1/1/2013

7/1/2012

1/1/2012

7/1/2011

1/1/2011

7/1/2010

1/1/2010

0

7/1/2009

0

Steam Use (klb/month)

Degree Days

1800

Figure 18: Comparison of HDD, CDD, and total steam consumption for Berkey Hall (WAC cooling)

To estimate the amount of steam which is required for operation of a centralized steam based
cooling unit for a building currently cooled by distributed WAC, a comparison method is used between
buildings of similar use type. Berkey Hall and Communication Arts are both buildings which consist of a
first floor of primarily classrooms, three (Berkey) or four (Communication Arts) upper floors of primarily
office space, and a single basement level. Total building area for Berkey and Communication Arts are
107,000 ft2 and 221,000 ft2 respectively. Due to this large discrepancy in building area, the units of
𝑙𝑏 𝑠𝑡𝑒𝑎𝑚
𝑇𝐷𝐷∗1000 𝑓𝑡 2

are used (Figure 19) instead of the previously used

40

𝑘𝑙𝑏 𝑠𝑡𝑒𝑎𝑚
𝑇𝐷𝐷

units (Figure 18).

lbs. Steam/degree day*1000 ft2

40
Berkey
Comm Arts

35
30
25
20
15
10
5
0

Comm Arts
Berkey
0

2

4

6

8

Steam lb/degree day¡ft2

10

12

Figure 19: Comparison of Berkey Hall and Communication Arts steam use per total degree day scaled by total
building area (top) and average values for heating months (bottom), indicated in the top graph by unshaded areas
(Oct – Apr)

With the average scaled steam use of Berkey Hall heating months within one standard deviation
of the Communication Arts average, a direct comparison was used for estimation of steam requirements
for operation of a steam based cooling unit. Cooling months (May – Sept.) for Communication Arts use
24.1 Âą 8.3

𝑙𝑏 𝑠𝑡𝑒𝑎𝑚
.
𝑇𝐷𝐷∗1000 𝑓𝑡 2

Therefore, the estimated amount of steam required for Berkey Hall for the

cooling months is 2.6 Âą 0.9

𝑘𝑙𝑏 𝑠𝑡𝑒𝑎𝑚
.
𝑇𝐷𝐷

Without steam cooling Berkey Hall uses 6090 Âą 1520 klb

steam/year. With steam cooling 9210 Âą 1870 klb steam/year would be required per the 1981-2010 average
monthly TDD values [23].

Discussion
Compared to steam requirements during heating months, the amount of steam required per total
degree day during cooling months is significantly higher for four of the five buildings with centralized
41

steam cooling which were analyzed: Communication Arts, Chemistry, Hannah Administration, and Wells
Hall. Biomedical and Physical Sciences is the only one of the five where there is no significant difference
in steam use intensity between heating and cooling months. It is expected that this is due to the much later
initial construction date of the building of 2001 [24]. Of the other four buildings, the ratio of steam used
in heating months to cooling months shows a downward trend with later installation dates, from an
average 6.9 for Chemistry (1963) to 2.4 for Communication Arts (1981).
In most buildings with WAC cooling, units are installed in personal spaces. While this produces
acceptable conditions within affected spaces, the use of these units does not extend into public spaces
such as hallways and classrooms (Figure 20). This leads to a high likelihood of occupant discomfort in
the WAC cooled spaces compared with those which are centrally cooled.

95

Temperature (°F)

85
75
65
55

Communication Arts
Berkey

45

Outdoor
35
4/5/2017

5/5/2017

6/4/2017

7/4/2017

Figure 20: Comparison of Communication Arts (central cooling) and Berkey Hall (distributed/WAC cooling)
temperature profiles. Outdoor temperatures are daily averages from LAN (NOAA)

A second point which must be addressed is the long-term availability of steam produced by the
TB Simon power plant. The efficiency of a cogeneration plant such as TB Simon is maximized when both
steam and electricity are consumed at the production rates. However, the demand on campus for
electricity is generally in excess of steam requirements [25]. Construction is currently under way for solar

42

carports over which are expected to produce approximately 15,000 MWh/year of electricity, directly
offsetting production at the TB Simon power plant. Assuming a one to one relation of steam produced to
electricity produced, this will result in a 5% reduction in produced steam annually.
When determining the viability of retrofitting existing spaces with centralized steam driven
cooling units, several considerations must be accounted for. First, the availability of steam to power the
system must be determined. With current use trends, unused excess steam could be used to power new
installations providing ‘free’ utility costs. Therefore, any offset in electricity cost from ceasing WAC unit
operation can be directly counted against the capital cost of an installation. Due to the high up-front cost
of these units and required installation materials, however, a purely economic case cannot be made for
any surveyed spaces. However, the increase in occupant comfort due to adequate cooling as well as the
electricity use reduction and accompanied emissions reductions should be considered when determining
the total impact of a proposed installation.

Energy Use Intensity of MSU Campus Spaces
Rationale
Determining the average energy use intensity of different space types on campus allows for
determination of which types of devices consume the most energy. With this knowledge,
recommendations for behavioral modification or equipment changes can then be provided. Equipment
changes could, for example, be implemented by recommendations Spartan Marketplace stocking of
various appliances that are more energy efficient. These changes will have a direct effect on end users
and are not transparent like previously discussed energy saving measures.

Materials and Methods
To minimize intrusion on space occupants during in person surveys, the energy use of
representative devices were applied to device counts from surveyed spaces. These energy use values were
determined using a Kill A Watt meter (P3 International). Measurements were taken under different use
43

conditions (Table 16). Devices measured were in two personal office spaces, one shared office space, and
one computer lab. The generalized power use values also allowed for collection of data through an online
survey which was utilized in buildings which were not targeted by in person surveying.
Table 16: Electricity draw at different states of representative devices. All units are in watts (W). Refrigerator
energy consumption assumed to US Government Energy Guide value

Device

On

Monitor
Refrigerator
Picture Frame
Lamp
Desktop Computer
Toaster
Phone
Printer
Microwave
Coffee Maker
Laptop
Speakers
Water heater
Fan
Fax
Phone Charger
Cable Box (reference)

40
110
4
29
29
1200
1.5
16
1100
3.9
35
15
2000
25
16
3.2
44

Off Standby
0
-0
0
0.8
0
0
0
--0
0
0
0
0
0
40

0.8
-0
0
2
0
0.8
5.7
1.3
3
18
5
0
0
5.7
0
40

Device use rates were estimated for both weekdays (Mon – Fri, 8 hours/day) and weekend days
(Sat – Sun). These were then extrapolated to give annual values (Table 17). Personal refrigerators were
assumed to be equivalent to the US Government Energy Guide value for the refrigerator stocked by
University Stores (campus distributor) as of 01/2016. Calculation method is described in the Appendix,
Equation 11.

44

Table 17: Estimated annual energy use on a per device basis

Device
Monitor
Refrigerator
Picture Frame
Lamp
Desktop
Computer
Toaster
Phone
Printer
Microwave
Coffee Maker
Laptop
Speakers
Water heater
Fan
Fax
Phone Charger

On
Off
Standby
Wh/day
Wh/day
kWh/yr
(hrs) (hrs) (hrs)
(M-F)
(Sa/Su)
8.0
0.0
16.0
332.8
19.2
88.5
24.0
0.0
0.0
--315.0
24.0
0.0
0.0
96.0
96.0
25.0
8.0
16.0
0.0
232.0
0.0
60.3
8.0

0.0

16.0

264.0

48.0

73.6

0.1
0.3
0.0
0.1
0.2
1.0
4.0
0.1
8.0
0.0
1.0

23.9
0.0
16.0
0.0
0.0
0.0
0.0
23.9
16.0
16.0
22.0

0.0
8.0
8.0
23.9
23.8
5.0
4.0

100.0
6.9
46.1
122.8
72.2
125.0
80.0
166.7
200.0
46.1
3.2

0.0
19.2
0.0
31.2
72.0
0.0
0.0
0.0
0.0
0.0
0.0

26.0
3.8
12.0
35.2
26.2
32.5
20.8
43.3
17.0
12.0
0.8

0.0
8.0
1.0

In-person surveys were performed for office spaces on acquiring permission of occupants. If a
room was unoccupied or occupants did not consent to a survey no results were taken for the space. Space
types surveyed included offices (single & multiple occupant), classrooms, and common spaces (eg break
room). Number of devices in each space were recorded manually.
Online surveys consisted of a list of devices with radio buttons for quantity of devices in a space.
Occupants were asked for their space type (personal or shared office) and building. SurveyMonkey was
used for creating the survey and distribution of the survey was performed with department level emails.
There were significant variations in shared office occupant counts, with 8.2 Âą 5.6 per shared
office in areas surveyed on foot. Size of classrooms and lecture halls varied between 20 and 200+ allowed
occupants and were therefore discounted for determining an average energy use intensity. While personal
office spaces did have some variation indicated in surveys by ceiling light fixture counts, personal office

45

spaces were of roughly comparable size across different buildings on campus (75% had 1 or 2 ceiling
fixtures).
Average energy use intensity was determined only for personal office spaces due to high variation
in shared office, common spaces, and classroom/lecture hall size. Each device type quantity was summed
across all surveyed spaces and expected annual energy use was applied (Table 18). The total number of
surveyed spaces was then divided into these values to determine the average annual usage per office
space.

Results
In total 431 occupants were surveyed, consisting of 228 in personal office spaces and 203 in
shared office spaces (Table 18). 198 occupants were surveyed on foot and 233 were surveyed online.
Table 18: Surveyed buildings

Building
Business College Complex
Wells Hall
Erickson Hall
Engineering Building
Communication Arts and Sciences
Biomedical and Phsycial Sciences
Farral Hall
Natural Resources
Human Ecology
Psychology
IM Circle
Oyer Hall
Trout Food Sciences

Method
In-Person/Online
In-Person/Online
In-Person/Online
Online
Online
In-Person
In-Person/Online
In-Person
Online
In-Person
Online
Online
In-Person9

Responses
116
96
67
44
31
24
17
11
10
8
3
2
2

Total energy consumption for a personal office space on the MSU campus is 539 kWh/year if
ceiling lighting fixtures are not included (Figure 21). Laptops, desktops, and computer monitors consist of
2/3 of the total office energy consumption and refrigerators accounting for 1/4. Offices have an average of
1.5 computer monitors. Per results from online surveys (174 samples) 25% of offices had a desktop only,
46

24% had both a desktop and laptop, 45% used a laptop with secondary monitor(s), and 5% used a laptop
only.

Coffee Maker,
1.0%

Refrigerator,
22.1%
Phone, 0.4%

Monitor, 42.8%
Desktop
Computer,
15.5%

Laptop,
Lamp, 7.7%
6.0%

Microwave,
1.3%

Printer, 3.2%

Figure 21: Energy use intensity for personal office spaces (total: 539 kWh/yr)

Ceiling fixture lighting consumes a sizable portion of electricity in most office spaces. A single
fluorescent ceiling fixture consumes approximately 130 (two-lamp) to 200 (three-lamp) kWh/year. In
surveyed spaces, 15% had a single fixture. Most office spaces (60%) had two ceiling fixtures, consuming
between 260 and 400 kWh/year. Compared to the average energy use intensity due to other office
equipment, the most significant contributor to total office energy use both in occurrence rates (all
surveyed spaces contained at least one ceiling light fixture) and total energy consumption is lighting,
which consumes between 33 and 43% of total office electricity (Figure 22).

47

Annual Electricity Consumption
(kWh/office/year)

600
500
400
300
200
100
0
Office
Equipment

2 Lamp
fixture

3 Lamp
fixture

Figure 22: Comparison of total office device (not including ceiling fixtures) electricity use to two fluorescent
fixtures of different configurations

Discussion
The three largest areas of office energy use determined in this study were in lighting, computing,
and refrigeration. Lighting was previously discussed in Chapter 1 regarding retrofitting of hallway ceiling
fixtures and further validates those studies as key target areas for energy reduction. However, the lower
lighting on-rates for office spaces does not present as compelling of a business case. A simple payback
period of 17.2 Âą 2.2 years for 2 lamp retrofits and 14.8 Âą 1.9 years for 3 lamp retrofits is expected in most
office spaces (see Chapter 1 – Hallway LED Retrofits for more detailed description of retrofit systems and
cost assumptions).

Applying the results from the daylight harvesting light management system

installation (Chapter 1), similar paybacks are expected to the installation of LED retrofits. In an office
with 2 ceiling light fixtures, a payback of 24.9 Âą 3.2 years for 2 lamp fixtures and 16.6 Âą 2.1 years for 3
lamp fixtures is expected. Occupancy sensors for personal office spaces are not recommended unless they
have easy overrides due to the degree of control that occupants have over lighting compared to common
spaces and hallways. Light switches within offices are easily accessible and generally used to turn off
lighting when the occupant is leaving the space for an extended period, such as at the end of their work
day.
48

It is difficult (with current pricing) to make an economic case for lighting in office spaces despite
the large energy impact, but savings are available in computing refrigeration. Due to the relatively short
3-5 year life span of computing systems per Michigan State University policy, replacement schedules can
be utilized to replace high energy consumption equipment with lower intensity options. 25% of offices
had a desktop only, while 24% had both a desktop and laptop. The remaining 51% had no desktop and
instead used a laptop, generally with a second monitor (45% had external monitors). Per our
measurements, a laptop consumes 32.5 kWh annually (on, not charging) compared to a desktop under
normal conditions which consumes 73.6 kWh. A single computer monitor consumed 40W under normal
conditions, so a desktop with at least two monitors (21% of total surveyed spaces) consumes 251 kWh
annually under normal conditions. A laptop with a single monitor attached will consume 121 kWh
annually, 48% of the consumption of a desktop system with a similar total screen area. This 130kWh
reduction in annual electricity use corresponds to a 24% decrease in current average office energy use
each year.
After lighting and computer equipment, the next largest consumer of electricity within personal
office spaces is the refrigerator. While the refrigerators within personal office spaces have a smaller total
volume, electricity use is often on par with much higher volume ‘standard’ refrigerators. For example, a
4.5 ft3 unit is listed at 315 kWh/year, while an 18.1 ft3 unit (Frigidaire) is listed at 404 kWh. With a single
personal refrigerator generally consuming more than half of the electricity as a shared refrigerator,
moving to centralized refrigerators in either break rooms or departmental offices can provide a significant
reduction. If one 404 kWh/year shared refrigerator was substituted for every two personal refrigerators,
energy reductions of 388MWh/year can be achieved. If one 18.1 ft3 shared refrigerator is instead
substituted for four personal refrigerators (18.0 ft3 total volume), an energy reduction of 1.47GWh/year is
available.

49

Summary
Reductions in annual electricity use of up to 631 MWhrs can be achieved by utilizing steam
absorption chillers in place of distributed window air conditioning units in all surveyed buildings. In
addition to consuming a significant amount of electricity annually, these systems are often outdated and
do not provide adequate cooling within the building at large. Based on the analysis of Berkey Hall, it is
expected that to achieve this electricity use reduction 44,200 Âą 15,200 klb of steam would be required
annually. Another target for energy reduction on campus lies within the almost 10,000 office spaces on
campus. Lighting, computing, and personal refrigerators consume the largest share of the total office
energy profile, and recommendations are provided for reducing each of these targets. The total impact of
implementing the computer and refrigerator recommendations would provide a ~0.2% reduction in total
electricity consumption. If LED retrofits were implemented in every office on campus, a ~0.6% would be
available. While this is significant, the relatively low light on-rates limits the economic viability of these
installations.

50

CHAPTER 3: Luminescent Solar Concentrator Outreach Activity

Introduction
With increasing greenhouse gas emissions [26], electricity costs [27], and international
agreements on combatting climate change [28], the need for access to inexpensive renewable energy is
increasing every year. Globally, 21.6 trillion kWh of electricity were generated in 2012 and this value is
projected to increase to 36.5 trillion kWh by 2040 [29]. Increasing the fraction of this energy which
comes from renewable sources is necessary for securing the energy future of the world and limiting the
negative environmental impact in meeting that energy demand. The single largest source of renewable
energy generation is our sun, which provides enough energy to power the planet for an entire year in less
than one hour [30]. A key to effectively harnessing this energy lies in finding ways to creatively and
synergistically deploy various solar technologies across a wide variety of spaces and we aim to promote
this kind of creative thinking.
While solar panels are becoming a more ubiquitous and publicly recognizable form of solar
harvesting, current photovoltaic (PV) technologies are often prohibitively expensive in relation to savings
available [31] and can be constrained in application areas due to bulkiness, flexibility, or aesthetics. Thin
film silicon (Si) cells [32], transparent organic PV [33], and luminescent solar concentrators (LSC) [34,
35] are all technologies which seek to address one or more of these issues. For this demonstration, we
have developed a solar module design kit for making artistic and colorful LSC paintings that are both art
and solar energy generators as a route to engage students in the understanding of optics, waveguides,
energy transport, solar energy, and the necessity of renewable energy generating devices.

Concepts of Materials
The fundamental goal of an LSC device is to concentrate light gathered over a large area
concentrator (waveguide) onto a small area of expensive, edge-mounted solar cells to reduce the overall
solar cell system cost (Figure 23). Concentrators are doped with molecules which absorb certain fractions
of the solar flux and re-emit this energy as luminescence at a longer wavelength (lower energy). This
51

photoluminescence is then waveguided by total internal reflection to edge mounted solar cells for energy
generation [34] or lost from the front/back of the waveguide if emitted photons contact the waveguide
surface at angles inside the escape cone. This principle is similar to that experienced at amusement parks
in fiber-connected flashlights. The efficiency of waveguiding (Equation 6) is dictated by the contrast in
the index of refraction of the waveguide and the surrounding media (air):

1

𝜂𝑊𝐺 = √(1 − 𝑛2 )11

(6)

For typical waveguides such as PMMA with an index of n = 1.5, this translates to a waveguiding
efficiency of 75%. This means that up to 75% of the light which is captured within the waveguide can be
directed to and emerge from the edges. Increasing total LSC efficiency (𝜂𝐿𝑆𝐶 ) is a primary goal of current
research in the field and is defined by:
𝜂𝐿𝑆𝐶 = (1 − 𝑅) ∗ 𝜂𝐴𝑏𝑠 ∗ 𝜂𝑃𝐿 ∗ 𝜂𝑅𝐴 ∗ 𝜂𝑃𝑉 ∗ 𝜂𝑊𝐺

(7)

This equation includes the front face reflection (R), absorptive efficiency of the luminophore (𝜂𝐴𝑏𝑠 ),
photoluminescent efficiency (𝜂𝑃𝐿 ), reabsorption suppression efficiency (𝜂𝑅𝐴 ), and photovoltaic efficiency
(𝜂𝑃𝑉 ) [36]. Current LSC research goals focus on increasing the efficiencies above [37, 38] as well as scale
of devices [39, 40], flexibility [40], and increasing concentrator transparency [41, 42, 43]. Ultimately there
are two goals of developing LSC technologies: 1) reduction in total PV surface area required for
collecting light to reduce module costs and 2) the ability to deploy solar harvesting devices in new areas.

52

Figure 23: Luminescent solar concentrator schematic

LSC devices were originally developed as a simpler solution to mirror based concentrators10
which harvest direct-incidence sunlight. LSCs can collect light across a wide range of incidence angles
(including diffuse light), so complicated tracking equipment is not required for efficient collection of
sunlight. Early concentrator designs generally absorbed and emitted light in the visible range, making
them unsuitable for windows and other areas where the accurate transmission of natural light is desired
[44]. However, these visibly absorbing and/or emitting concentrators often boast much higher efficiencies
[45, 46] than transparent LSC devices [41, 42, 43] which allow for increased levels of visible light
transmission. Application of both types of LSC devices into architectural spaces has become a focus of
research [45, 47]. The light transmission characteristics of different devices provides architectural
designers flexibility in implementation of solar harvesting technologies, from standard office windows to
colorful designs such as at the Palais de CongrĂŠs in Montreal, QuĂŠbec. The creation of visually appealing
designs which can harvest energy and demonstrate the working principles of LSC technology are the
focus of the demonstration described in this work.

Goals of the Demonstration
This This work examines an approach to LSC technology implementation which is different than
uniform LSC applications seen in previous studies. By incorporating absorptive and luminescent dopants
into clear paint formulations, they can be painted onto an optically clear sheet which constitutes the
waveguide of the LSC. We are focused on developing an activity generally associated with creativity
53

(painting) with solving an engineering problem (solar energy generation). Increasing student creativity in
the approach of engineering problems helps to drive innovation which is essential for advances in science
to occur [48]. The demonstration has four primary educational objectives:
1) To engage students in discussion of renewable energy
2) To highlight the energy potential from the sun
3) To develop fun and instructive solar cell activities
4) To associate creativity of design with STEM principles (e.g. linking art and energy)

Design of Materials and Demonstration
The key ingredients for building an LSC include a waveguide, PV cell(s), and luminescent
materials. A 6”x6”x1/4” PMMA sheet was used as the waveguide for its suitable trapping (waveguide)
efficiency as well the near complete transmission of light through the material before painting. The sheet
dimensions allow for straightforward design and attachment of solar modules. While power can be
generated with either flashlights or sunlight (indeed, both are encouraged), local illumination with the use
of flashlights to control light penetration area and angle helps to visualize the principles of wave-guiding
and can make the glowing light more apparent. Different penetration angles can also be used to promote
discussion on possible efficiency loss pathways.
Five different colored paints (red, orange, yellow, green, and blue) were designed. Paint color is
determined by the absorptive, luminescent, and reflective properties of constituent dyes. Absorptive
filtering allows for the transmission of certain wavelengths of light which results in visible paint coloring
(the dominant mechanism in this exercise). Under normal lighting conditions, the paints produce
luminescence from absorbed light which also contributes to the overall perceived color by producing a
noticeable glow (the second dominant mechanism of coloring in this exercise). On application, a small

54

amount of reflection from the surface of the waveguide [44] also produces some coloring but is less
important here.
The light absorbed by luminophore (dye) molecules is emitted at a longer wavelength (lower
energy) than the absorptive color of the paint. This is caused by vibrational losses of the energy of
electrons in the excited state before they return to the ground state. For example, a dye which has a yellow
absorptive tint when in solution emits green light under excitation by flashlight; a dye which has a red
absorptive tint (absorbing green light) emits red light which reinforces the color. For this demonstration,
matching paint absorption color with emission color was desired to amplify particular coloring. This was
achieved by mixing commercially available luminescent dyes so the imparted color from absorption and
luminescence were similarly matched. The complementary blending of both the emission and absorption
spectra of different dyes was performed to produce the five-color spectrum used in the demonstration
(Figure 24). This mixing is described in greater detail in the Discussion section. The dyes are dissolved
into an acetone solution and then combined with a poly(butyl methacrylate-co-methyl methacrylate) resin
based polymer matrix solution (n = 1.51) to create the final paint which can adhere to the PMMA
waveguide and dry within 15 minutes of application.

55

Figure 24: Normalized absorption (solid) and emission (dots) spectra of paints used in the demonstration (A). Solar
cell module (D). Note that the multiple absorption and emission peaks for blue and green paints stem from two
distinct dyes with one dye emitting in the visible and one emitting in the infrared to reinforce absorptive coloring
and provide greater power.

Disposable pipettes were used in place of traditional paintbrushes to increase the thickness (and
associated color density) of applied paints. The technique used is similar to drop casting, a common
method18 used in prototyping new dye formulations. Paint brushes can also be utilized but often require

56

multiple coats due to the limited dye solubility of the paints. We note that even if higher solubilities were
possible, concentrations above those utilized in this work often lead to “concentration quenching” of the
luminophore resulting in a reduction in luminescence efficiency [49] and less glowing. Pipettes are
initially less intuitive for participants to use but provide more control than brushes. It is helpful to have
participants practice controlling them with scrap pieces of acrylic before attempting more complicated
patterns and paintings.
Clips were designed and 3D printed to allow easy mounting and dismounting of fragile PV cell
around the edge of the waveguide. A laser-cut 6” x 1/4” silicon solar cell was inserted into the printed clip
with holes for soldered electrical leads (Figure 24D). The cell was then encapsulated in a clear epoxy for
durability. Producing solar modules which were durable and easy to attach was a necessity for hands-on
demonstration, especially for younger participants.

Figure 25: Sample painted device under direct illumination (left) and ambient conditions (right). Note luminescence
along waveguide edge where solar modules are attached during demonstration

The demonstration is comprised of four segments: 1) initial introduction to solar energy and
LSCs, 2) painting of devices, 3) paint drying time, and 4) device testing. The painting is preceded by an
introduction which briefly describes the principles of LSC technology and a demonstration with a

57

previously painted LSC device (Figure 25). Students were then allowed 5-10 minutes to paint their own
LSC devices. During the following drying time (~15 minutes), a variety of topics were covered dependent
on age/educational level of participating students. A discussion on the difficulties facing the
implementation of solar harvesting technologies, including cost, application areas, and integration into the
electrical grid were presented to students at all educational levels. For more advanced students, additional
topics can include: factors which impact LSC efficiency, efficiency records, use of blocking diodes to
prevent battery discharge, maximum power point tracking, and battery recharging time from
instantaneous or average solar flux [50].
Upon completion of drying, solar modules were attached to one edge of each student’s device and
returned to students. Fans were then connected to the LSC to first get a qualitative idea of power
generation by the devices before moving on to multimeters where current and voltage could be measured.
Students were encouraged to collaborate and experiment with linking devices together in parallel and/or
series to determine changes to device output in current, voltage, and fan speed.

Figure 26: Current-Voltage (IV, left) and Power-Voltage (right) of device (center) with two edge mounted solar
modules connected in series and parallel. Multimeter is connected in parallel, reading mV (center)

An IV curve (Figure 26) from two solar cell modules attached to a painted LSC in either parallel
or series displays possible voltage and current combinations. The short circuit current, ISC (current at zero
voltage), and the open circuit voltage, VOC (voltage at zero current), are used to characterize solar
harvesting devices, but operation at either of these points will not provide usable power since 𝑃 = 𝐼 ∗ 𝑉.

58

Power output is maximized at the MPP (maximum power point), which for the painting in Figure 4 was
11.0 mW (0.87V, 12.6mA) or 13.3 mW (0.38V, 35.0 mA) for series and parallel respectively.

Hazards
The solvent used in this demonstration is acetone (the active ingredient in nail polish remover),
which has a high vapor pressure and produces a noticeable smell. The demonstration should be performed
in a well-ventilated space. If participants handle paints the use of gloves and smocks are recommended to
prevent staining of skin or clothing.

Discussion
Two questions were most frequently brought up by participants in the demonstration: 1) which
color is the best and 2) how can a device which powers a fan or other device be produced? These
questions provide an opportunity to discuss the mechanics behind electricity generation using LSC
devices in more detail and are outlined below.
What Is the Best Color?
The question first raised by most students is which paint will produce the most power. While the
most efficient device is not the goal of this activity, in real world applications the highest device
efficiency is desired. Photon energy (E, eV) is inversely proportional to the wavelength (Îť, Îźm) by
Equation 8:

𝐸=

1.240 𝑒𝑉 ∗ 𝜇𝑚
𝜆

(8)

As a result of this equation, many student’s initial (and intuitive) conclusion is that lower
wavelength (higher energy) fluorescence/solar-absorption would yield higher current generation.
However, the energy of the photon does not directly impact the efficiency of current generation because
the excess energy of all the absorbed photons is thermalized to the bandgap and each absorbed solar
photon produces, at most, one luminescence photon. This re-emitted photon can be reabsorbed, lost

59

through the escape cone, or produce a maximum of one electron in the solar cell (in the absence of
multiexciton generation). That is, the energy of the photon is not informative about the power generation
in single junction LSCs. Again, the amount of power generated depends on the overall amount of
absorption within a particular wavelength range, the amount of photons in the light source at those
wavelengths, the luminescence efficiency, reabsorption losses from the dyes, and the efficiency of
converting particular wavelength photons into electrons in the solar cell (Equation 7). The
photoluminescent efficiencies for the constituent luminescent dyes described above are all similarly high
(typically >85%); the wavelength efficiency (or quantum efficiency) for Si PVs is relatively flat across
the spectrum (500-1000nm) except at UV and blue wavelengths where the efficiency drops off by about
half; the solar photon spectrum is also relatively flat in the range of 500nm-1000nm (visible and infrared,
Figure 27). The “best” color can then be attributed to the tuning of absorption widths (total light source
absorption) combined with a particular pairing of light source. For example, the red paint has an
absorption peak of ~570nm, a wavelength which is emitted at a much higher intensity by flashlights used
in the demonstration. This leads to a high current generation potential. For the blue and green paints, a
luminophore which emits in the near-infrared is used with a visibly luminescent dye to reinforce the blue
or green color by absorbing the red part of the spectrum. This causes blue and green paints to absorb light
at ~650-800nm (infrared) and 300-400nm (ultraviolet). Most commercial fluorescent and LED lamps do
not emit at all in the ultraviolet or infrared, and the combination of these factors results in a low total
current generation potential by these paints under artificial illumination. The blue and green paints have a
higher current generation potential under sunlight, where their light production in both the ultraviolet and
infrared portions of the spectrum. Reabsorption of the luminescence in the same or other dyes also plays a
role in the overall current generation potential of a given paint.

60

Figure 27: Spectrum of sunlgiht, LED lamp, and fluorescent lamp

How to Produce a Usable Device?
Generation of the most power possible from a device is important but is often secondary to
producing the correct voltage and current for a particular application. Fortunately, this can be achieved
primarily through different PV cell arrangement. Connection of multiple devices into an array as seen in
large scale solar installations is carefully considered to satisfy voltage/current requirements. Similarly,
connection of PV cells used in the demonstration in series and/or parallel circuits allows for tuning of
voltage or current respectively to meet application requirements. An application requiring higher current
levels than provided by a single device could utilize several devices with solar cell modules connected in
parallel to increase output current; an application requiring higher voltage than produced by a single PV
module (~0.4V for Si based LSCs) could be configured with several modules wired together in series to
build voltage like batteries in series.
In this demonstration, the DC fan required 0.3V to begin operation. Because the operating point
in both series and parallel exceeds the 0.3V threshold, the fan speed is limited only by the current
provided. The combination which produces the most current will then yield the fastest fan speed. Wiring
the edge mounted solar cells in parallel will produce the best device in this case. However, many
applications (e.g. recharging batteries) require at least 1.8V, significantly more than can be generated by
61

two cells connected in series. Connection of several devices (e.g. four cells from the four edges) in series
is then required to produce a device suitable for these applications if they are to be powered directly.

Outreach Events
This demonstration was initially performed during Grandparent’s University, an annual MSU
program for school alumni to bring their grandchildren (8 – 12 years of age) to campus to participate in
educational activities. During the two LSC painting outreach activities, students created their own LSC
designs and tested them under illumination by the sun as well as with the use of flashlights as previously
described (Figure 28).

Figure 28: MSU Grandparent's University outreach activity

A second outreach event participated in was the MSU Science Festival. Posters were designed
covering the energy use intensity surveys (Chapter 2) and other topics investigated by SPARTA and how
they could be applied at participants homes. Colored panels were included (Figure 29) to provide an
explanation of the mechanisms of LSC devices, and small solar toys were provided for further participant
engagement.

62

Figure 29: MSU Science Festival student participation

Conclusions
The goals of these described outreach activities are to engage students in discussion of renewable
energy, highlight the potential from the sun, develop a fun and instructive activity, and associate creativity
of design with STEM principles. Throughout the activity students are engaged in the discussion of
renewable energy, from initial concepts of total available potential through to devices used in harvesting
solar energy. With the use of different measuring devices, the energy potential available from the sun can
be demonstrated both quantitatively and qualitatively. Parallels can be drawn between the activity
provided and current research in the energy generation field. Materials (dyes used in paints) and
techniques (drop-casting) used in the activity are the same as those used in current LSC research. While
the competitive nature of students will often drive them to want to find the ‘best’ design, discussion of the
efficiency factors follows the painting activity and therefore will not influence the painting of student
devices. This allows students the freedom to produce devices which generate energy while expressing
themselves through artistic paintings.
SUMMARY AND FUTURE WORK
The ultimate goal of the Michigan State University Energy Transition Plan is to source 100% of
the required energy for campus from renewable sources [25]. To make this goal more reachable, there are
63

essentially three pathways which can be taken: increasing the amount of energy produced by renewable
energy sources, carbon capture and sequestration, and reduction of overall energy use so that less energy
must be generated from non-renewable sources. In this thesis, several opportunities for reducing total
energy use are investigated and potential savings are quantified. Engaging students in community
outreach activities is important for driving interest and creativity in renewable energy generation. Because
of the long-term investment which is inherent in the MSU Energy Transition Plan as well as any other
approach to decreasing reliance on non-renewable energy sources, engaging students in these discussions
benefits not only the student but can pay dividends in the future by helping steer them towards careers in
the STEM, renewable energy, or sustainability fields.
In the first chapter, energy savings available from lighting retrofits for LED conversion and with
various control methods are investigated. Retrofitting LED bulbs and drivers in ceiling fixtures is very
economical in many spaces on the MSU campus, dependent on pricing scheme used for electricity
purchase and the targeted areas for deployment (e.g. high on-rate). Because much of the electricity used
on campus is purchased internally from the TB Simon power plant, choosing the right price level when
determining economic indicators is very important. Due to the varying widths and light fixture
concentrations, determining a single generalized value required to trigger a 5-year payback for occupancy
sensor installations is not possible. However, most installations require around a $4 - $6 per linear foot of
hallway to reach economic viability. Daylight harvesting systems are not economical as retrofits due to
the exceedingly high equipment costs.
The second chapter focuses on energy use analyses of buildings and their occupants on the MSU
campus. Cooling is a significant portion of HVAC operating load in the summer, and an estimation of
electricity savings and steam utility requirements to switch from distributed window air conditioning units
to steam absorption driven central cooling is performed. Increasing levels of steam use over the summer
as facilitated by new installations of steam-based cooling units for buildings currently serviced by
electrically powered window air conditioning units would cause a favorable increase in overall power
64

plant efficiency. The average energy use intensity is determined for personal office spaces, with
computing and lighting consuming the majority of total power use. A shift from desktop use to laptop use
can see significant energy use reductions of 130kWh/yr in each instance with minimal reduction of total
screen space, and switching from personal to communal refrigerators can realize energy use reductions of
up to 1.47GWh/year.
The final chapter outlines an outreach activity used to engage students in a discussion of
renewable energy generation, the second pathway for reaching the Energy Transition Plans goals.
Luminescent solar concentrators are an alternative method for harvesting solar energy which can provide
a tangible reminder for students of the power available from the sun, the largest source of renewable
energy.
Moving forward, an aggressive approach to energy reductions on campus can be maintained. The
implementation of LED lighting is an inexpensive way to greatly reduce total campus energy
consumption. While four buildings are currently scheduled for retrofits, these reductions are estimated to
reduce total campus energy consumption by 399 MWh annually (0.14%). If retrofits are implemented in
all surveyed spaces, a reduction of 939 MWh annually (0.37%) is readily available. A complete economic
evaluation of steam-cooling implementation is recommended at larger buildings such as Natural Sciences,
Giltner, and Berkey Hall. To continue pursuing the goals of the MSU Energy Transition Plan, novel
methods of energy reduction must continue to be analyzed and implemented alongside renewable energy
generation to eventually reach 100% renewable campus.

65

APPENDIX

66

The determination of installed LED retrofit lifetimes (Equation 9) takes into account the
manufacturer specified lifetime (L) (50,000 hours for CREE UR2-48-4500L, three-lamp retrofit; 60,000
hours for Precision Paragon TKG24-40VW-EU, two-lamp retrofit) and the expected light on rate after
installation. The light on-rate after installation is assumed to be equal to the measured light on rate.
Assuming 365 days per year, each year has 8,760 hours.

𝐼𝑛𝑠𝑡𝑎𝑙𝑙𝑎𝑡𝑖𝑜𝑛 𝐿𝑖𝑓𝑒𝑡𝑖𝑚𝑒 (𝑦𝑒𝑎𝑟𝑠) =

𝐿 (ℎ𝑟𝑠)
ℎ𝑟𝑠
𝑂𝑛 𝑅𝑎𝑡𝑒 ∗ 8760 𝑦𝑒𝑎𝑟

(9)

Comparing this to the payback period at the lowest electricity pricing rate ($0.05755/kWh), the
Packaging building will see the lowest difference between payback period and expected lifetime, with the
retrofit expected to provide enough savings from electricity reductions 1.76 years before the end of the
installation lifetime (Table 19)

67

Table 19: Comparison of payback period at $0.05755/kWh electricity purchse rate with installed system lifetime

Building
Agriculture Hall
Computer Center
Giltner Hall
Berkey
Trout Food Science
Packaging
Natural Resources
Wells - B+C Wing
Engineering
Bessey
Wells - A+D Wing
Chemistry
Biochemistry
North Kedzie
BPS
Plant and Soil Sciences
Baker
Music Practice
Comm. Arts
Wells Combined
Student Services Building

Expected
Payback
Lifetime Lifetime (yrs)
Period (yrs) Payback (yrs)
8.82
6.25
2.56
12.65
9.50
3.14
6.85
4.58
2.27
9.05
6.09
2.96
8.83
6.09
2.74
6.85
5.09
1.76
10.24
6.84
3.40
10.16
6.84
3.33
5.71
3.92
1.79
10.54
7.26
3.28
8.15
5.34
2.81
6.68
4.58
2.10
6.94
4.67
2.27
8.15
5.50
2.64
9.80
6.42
3.38
6.98
4.58
2.39
9.08
6.09
3.00
7.88
5.34
2.54
6.85
4.58
2.27
8.68
5.67
3.01
6.85
4.84
2.01

The following Tables (19-39) outline the results of light on-rate measurements which were
performed under the methodology described in Chapter 1 – Hallway LED retrofits. The row labeled Low
indicates the lower bound on rate (estimate) and High indicates the upper bound on rate (estimate). The
row labeled Act indicates the actual measured on rate. The row labeled ‘Fixtures’ indicates the number
lighting fixtures present in the surveyed hallways. The total on rate is weighted by the number of fixtures
affected by the on rates by Equation 10, where i is the floor number.

𝑂𝑛 𝑅𝑎𝑡𝑒 =

∑𝑛𝑖=1 𝑂𝑛 𝑅𝑎𝑡𝑒𝑖 ∗ 𝐹𝑖𝑥𝑡𝑢𝑟𝑒𝑠𝑖
∑𝑛𝑖=1 𝐹𝑖𝑥𝑡𝑢𝑟𝑒𝑠𝑖

68

(10)

Table 20: Agriculture Hall light on-rate

Floor

B

1

2

3

4 Total

Fixtures

13

11

14

13

13

64

Low
Act
High

100%
100%
100%

50%
50%
50%

63%
68%
74%

52%
70%
85%

87%
97%
100%

71%
78%
83%

Table 21: Computer Center light on-rate

Floor

1

2

3

4

5 Total

Fixtures

11

11

11

11

11

55

Low
Act
High

52%
60%
71%

52%
59%
70%

46%
54%
64%

46%
54%
64%

33%
44%
51%

46%
54%
64%

Table 22: Giltner Hall light on-rate

Floor

1

2

3

4 Total

Fixtures
Low

58
100%

51
100%

45
100%

27
181
100% 100%

Act
High

100%
100%

100%
100%

100%
100%

100% 100%
100% 100%

Table 23: Berkey Hall light on-rate

Floor

B

1

2

3

4 Total

Fixtures

30

33

35

33

25

156

Low
Act

75%
82%

68%
79%

64%
75%

50%
57%

86%
90%

67%
76%

High

89%

99%

84%

72%

95%

87%

Table 24: Trout Food Science light on-rate

Floor

B

1

2

3 Total

Fixtures
Low

22
50%

22
100%

28
50%

26
100%

98
74%

Act
High

60%
71%

100%
100%

53%
61%

100%
100%

78%
82%

69

Table 25: Packaging Building light on-rate

Floor
Fixtures
Low
Act
High

1
37
100%
100%
100%

Total
37
100%
100%
100%

Table 26: Natural Resources light on-rate

Floor

B

1

2

3 Total

Fixtures

56

61

34

34

185

Low
Act
High

50%
69%
75%

36%
42%
46%

100%
100%
100%

51%
75%
90%

55%
67%
73%

Table 27: Wells Hall (B/C Wings) light on-rate

Floor
Fixtures

1
75

2
30

Low
Act
High

50%
50%
50%

100%
100%
100%

3 Total
40
145
70%
76%
82%

66%
67%
69%

Table 28: Engineering Building light on-rate

Floor
Fixtures

1

2

3

4 Total
284

Low
Act

100%
100%

100%
100%

100%
100%

100% 100%
100% 100%

High

100%

100%

100%

100% 100%

Table 29: Bessey Hall light on-rate

Floor
Fixtures

1

Low
Act
High

50%
51%
57%

70

2 Total
158
51%
57%
64%

51%
54%
60%

Table 30: Wells Hall (A/D Wings) light on-rate

Floor
Fixtures

1
36

2
62

3
62

4
40

5
40

6
40

7 Total
40
320

Low
Act

79%
81%

100%
100%

100%
100%

100%
100%

50%
50%

76% 50%
87% 50%

83%
84%

High

83%

100%

100%

100%

50%

100% 50%

86%

Table 31: Chemistry Building light on-rate

Floor
SB
B
Fixtures
52
30
Low 50%
Act 53%
High 60%

75%
88%
95%

1
30

2
30

3
31

4
31

17%
48%
72%

50%
74%
86%

100%
100%
100%

71%
89%
100%

5 Total
32
184
69%
75%
84%

64%
85%
91%

Table 32: Biochemistry light on-rate

Floor
B
Fixtures
21
Low
94%
Act
99%
High 100%

1
21
100%
100%
100%

2
23
76%
95%
100%

3
22
100%
100%
100%

4
22
100%
100%
100%

5 Total
22
131
94% 94%
99% 99%
100% 100%

Table 33: North Kedzie Hall light on-rate

Floor
Fixtures
Low
Act
High

1

2

36
79%
93%
100%

45
64%
84%
100%

3 Total
48
56%
78%
90%

129
65%
84%
96%

Table 34: Biomedical & Physical Sciences light on-rate

Floor

B

Fixtures
Low
Act
High

28%
37%
44%

1

2

3

4

5

58

47

58

58

26

24

271

76%
81%
84%

100%
100%
100%

58%
61%
65%

57%
61%
66%

47%
50%
53%

44%
47%
50%

66%
70%
73%

71

6 Total

Table 35: Plant and Soil Sciences light on-rate

Floor
B
Fixtures
34

1
46

2
44

3
44

4
44

5 Total
44
256

Low
Act

97%
98%

100%
100%

100%
100%

90%
93%

100%
100%

97%
98%

97%
98%

High

99%

100%

100%

97%

100%

99%

99%

Table 36: Baker Hall light on-rate

Floor
Fixtures

1
33

2
33

3
33

4
33

Low
Act
High

71%
89%
100%

75%
87%
93%

55%
60%
67%

57%
73%
93%

5 Total
33
165
59%
68%
74%

64%
75%
85%

Table 37: Kresge Hall light on-rate

Floor
B
Fixtures
58
Low
Act
High

50%
50%
50%

1
50

2
50

75%
77%
81%

50%
73%
86%

3 Total
50
150
50%
56%
72%

56%
64%
71%

Table 38: Communication Arts & Sciences light on-rate

Floor
Fixtures

2
65

3
27

4
27

5 Total
27
146

Low
Act
High

100%
100%
100%

100%
100%
100%

100%
100%
100%

100% 100%
100% 100%
100% 100%

Table 39: Music Practice Building light on-rate

Floor
B
Fixtures
13
Low
Act
High

95%
95%
96%

1
6

2
13

3
13

4
13

82%
86%
89%

75%
80%
84%

89%
91%
94%

82%
86%
89%

72

5 Stairs Total
13
60
118
82%
86%
89%

82%
86%
89%

83%
87%
89%

Table 40: Student Services Building light on-rate

Floor
B
Fixtures
26

1

2
8

3 Total
26
60

Low
Act

100%
100%

100%
100%

100%
100%

100% 100%
100% 100%

High

100%

100%

100%

100% 100%

The following Tables (41, 42, 43) provide the results from Chapter 1 – Hallway LED retrofits.
The upper and lower bounds of each metric are determined by applying the low and high light on-rates in
Tables 20 – 40, while the baseline value is determined at the act measured on rate. The payback period is
determined by Equation 2, net present value by Equation 3, and return on investment by Equation 4.
Annual savings refer to the amount in dollars saved due to the energy reduction (Equation 1) from
installing lower wattage LED retrofits in place of fluorescent fixtures.

73

Table 41: Economic metrics at fuel cost of electricity($0.05755/kWh)
Low
Payback
Period
(years)

Payback
Period
(years)

High
Payback
Period
(years)

Low
Net
Present
Value

High
Net
Present
Value

Agriculture Hall

5.8

6.3

6.9

$6,815

$7,235

Computer Center

7.9

9.5

11.5

$5,035

Giltner Hall

4.6

4.6

4.6

Berkey

5.3

6.1

Trout Food Science

5.8

Packaging

Low
Return on
Investment

Return on
Investment

High
Return on
Investment

Low
Annual
Savings

Annual
Savings

High
Annual
Savings

$7,614

111%

118%

125%

$900

$955

$1,005

$5,622

$6,135

96%

107%

117%

$512

$572

$624

$21,725

$22,388

$23,086

126%

129%

134%

$3,374

$3,477

$3,585

6.8

$17,600

$18,662

$19,524

118%

125%

131%

$2,138

$2,267

$2,372

6.1

6.3

$10,983

$11,447

$11,969

117%

122%

128%

$1,401

$1,460

$1,526

5.1

5.1

5.1

$3,979

$4,099

$4,226

113%

116%

120%

$690

$711

$733

Natural Resources

6.3

6.8

8.3

$20,307

$21,925

$23,059

115%

124%

130%

$2,202

$2,377

$2,500

Wells - B+C Wing

6.7

6.8

7.0

$16,204

$17,001

$17,820

117%

123%

129%

$1,789

$1,877

$1,967

Engineering

3.9

3.9

3.9

$40,313

$41,325

$42,383

123%

126%

130%

$7,263

$7,445

$7,636

Bessey

6.5

7.3

7.8

$20,469

$21,660

$22,890

113%

119%

126%

$2,121

$2,244

$2,371

Wells - A+D Wing

5.3

5.3

5.4

$37,997

$39,366

$40,875

124%

129%

134%

$4,989

$5,168

$5,366

Chemistry

4.3

4.6

6.2

$24,916

$26,473

$27,098

118%

125%

128%

$3,876

$4,119

$4,216

Biochemistry

4.6

4.7

4.9

$15,375

$16,008

$16,538

123%

128%

132%

$2,385

$2,483

$2,565

North Kedzie

4.8

5.5

7.1

$14,405

$15,453

$16,218

117%

125%

132%

$1,942

$2,083

$2,187

BPS

6.2

6.4

6.8

$31,133

$32,707

$34,402

120%

126%

133%

$3,461

$3,636

$3,825

Plant and Soil Sciences

4.6

4.6

4.7

$30,844

$31,697

$32,643

126%

130%

133%

$4,697

$4,827

$4,971

Baker

5.3

6.1

7.2

$18,544

$19,715

$20,821

118%

125%

132%

$2,248

$2,389

$2,524

Music Practice

5.2

5.3

5.6

$13,559

$14,140

$14,610

120%

125%

130%

$1,888

$1,969

$2,034

Comm. Arts

4.6

4.6

4.6

$17,412

$17,943

$18,502

125%

129%

133%

$2,721

$2,804

$2,892

Wells Combined

5.5

5.7

5.8

$55,190

$57,507

$59,498

124%

129%

134%

$6,761

$7,045

$7,289

Student Services Building

4.8

4.8

4.8

$6,814

$7,021

$7,239

119%

122%

126%

$1,118

$1,152

$1,188

Building Name

Net
Present
Value

74

Table 42: Economic metrics at produced cost of electricity ($0.07492/kWh)
Low
Payback
Period
(years)

Payback
Period
(years)

High
Payback
Period
(years)

Low
Net
Present
Value

High
Net
Present
Value

Agriculture Hall

4.4

4.8

5.3

$8,929

$9,463

Computer Center

6.0

7.1

8.5

$6,728

Giltner Hall

3.5

3.5

3.5

Berkey

4.0

4.7

Trout Food Science

4.3

Packaging

Low
Return on
Investment

Return on
Investment

High
Return on
Investment

Low
Annual
Savings

Annual
Savings

High
Annual
Savings

$9,952

146%

155%

163%

$1,173

$1,243

$1,308

$7,472

$8,112

128%

142%

154%

$671

$745

$809

$27,959

$28,825

$29,739

162%

167%

172%

$4,390

$4,526

$4,669

5.3

$22,725

$24,090

$25,190

152%

162%

169%

$2,784

$2,951

$3,086

4.6

4.8

$14,250

$14,853

$15,529

152%

159%

166%

$1,823

$1,900

$1,987

3.8

3.8

3.8

$5,253

$5,415

$5,586

149%

153%

158%

$898

$925

$954

Natural Resources

4.8

5.3

6.4

$26,210

$28,284

$29,754

148%

160%

168%

$2,868

$3,095

$3,256

Wells - B+C Wing

5.1

5.3

5.3

$20,942

$21,985

$23,057

151%

159%

166%

$2,327

$2,443

$2,562

Engineering

3.0

3.0

3.0

$51,526

$52,843

$54,223

158%

162%

166%

$9,451

$9,692

$9,945

Bessey

5.0

5.6

6.0

$26,296

$27,842

$29,432

145%

153%

162%

$2,759

$2,921

$3,088

Wells - A+D Wing

4.0

4.1

4.2

$48,785

$50,572

$52,536

160%

165%

172%

$6,490

$6,728

$6,989

Chemistry

3.3

3.6

4.8

$31,940

$33,922

$34,727

151%

160%

164%

$5,049

$5,362

$5,489

Biochemistry

3.5

3.6

3.8

$19,835

$20,659

$21,353

158%

165%

171%

$3,104

$3,233

$3,341

North Kedzie

3.7

4.3

5.4

$18,644

$19,970

$20,949

151%

162%

170%

$2,532

$2,712

$2,845

BPS

4.8

5.0

5.3

$40,025

$42,073

$44,282

155%

163%

171%

$4,504

$4,734

$4,983

Plant and Soil Sciences

3.5

3.6

3.6

$39,622

$40,737

$41,972

162%

167%

172%

$6,112

$6,284

$6,475

Baker

4.1

4.7

5.5

$23,939

$25,435

$26,860

152%

161%

170%

$2,928

$3,111

$3,285

Music Practice

4.0

4.1

4.3

$17,529

$18,286

$18,898

155%

162%

168%

$2,457

$2,563

$2,649

Comm. Arts

3.5

3.5

3.5

$22,440

$23,135

$23,869

161%

166%

171%

$3,541

$3,651

$3,766

Wells Combined

4.3

4.3

4.5

$70,783

$73,799

$76,391

159%

166%

172%

$8,797

$9,171

$9,494

Student Services Building

3.7

3.7

3.7

$8,880

$9,154

$9,444

155%

160%

165%

$1,455

$1,500

$1,548

Building Name

Net
Present
Value

75

Table 43: Economic metrics at delivered cost of electricity ($0.09206/kWh)
Low
Payback
Period
(years)

Payback
Period
(years)

High
Payback
Period
(years)

Low
Net
Present
Value

High
Net
Present
Value

Agriculture Hall

5.8

6.3

6.9

$6,815

$7,235

Computer Center

7.9

9.5

11.5

$5,035

Giltner Hall

4.6

4.6

4.6

Berkey

5.3

6.1

Trout Food Science

5.8

Packaging

Low
Return on
Investment

Return on
Investment

High
Return on
Investment

Low
Annual
Savings

Annual
Savings

High
Annual
Savings

$7,614

111%

118%

125%

$900

$955

$1,005

$5,622

$6,135

96%

107%

117%

$512

$572

$624

$21,725

$22,388

$23,086

126%

129%

134%

$3,374

$3,477

$3,585

6.8

$17,600

$18,662

$19,524

118%

125%

131%

$2,138

$2,267

$2,372

6.1

6.3

$10,983

$11,447

$11,969

117%

122%

128%

$1,401

$1,460

$1,526

5.1

5.1

5.1

$3,979

$4,099

$4,226

113%

116%

120%

$690

$711

$733

Natural Resources

6.3

6.8

8.3

$20,307

$21,925

$23,059

115%

124%

130%

$2,202

$2,377

$2,500

Wells - B+C Wing

6.7

6.8

7.0

$16,204

$17,001

$17,820

117%

123%

129%

$1,789

$1,877

$1,967

Engineering

3.9

3.9

3.9

$40,313

$41,325

$42,383

123%

126%

130%

$7,263

$7,445

$7,636

Bessey

6.5

7.3

7.8

$20,469

$21,660

$22,890

113%

119%

126%

$2,121

$2,244

$2,371

Wells - A+D Wing

5.3

5.3

5.4

$37,997

$39,366

$40,875

124%

129%

134%

$4,989

$5,168

$5,366

Chemistry

4.3

4.6

6.2

$24,916

$26,473

$27,098

118%

125%

128%

$3,876

$4,119

$4,216

Biochemistry

4.6

4.7

4.9

$15,375

$16,008

$16,538

123%

128%

132%

$2,385

$2,483

$2,565

North Kedzie

4.8

5.5

7.1

$14,405

$15,453

$16,218

117%

125%

132%

$1,942

$2,083

$2,187

BPS

6.2

6.4

6.8

$31,133

$32,707

$34,402

120%

126%

133%

$3,461

$3,636

$3,825

Plant and Soil Sciences

4.6

4.6

4.7

$30,844

$31,697

$32,643

126%

130%

133%

$4,697

$4,827

$4,971

Baker

5.3

6.1

7.2

$18,544

$19,715

$20,821

118%

125%

132%

$2,248

$2,389

$2,524

Music Practice

5.2

5.3

5.6

$13,559

$14,140

$14,610

120%

125%

130%

$1,888

$1,969

$2,034

Comm. Arts

4.6

4.6

4.6

$17,412

$17,943

$18,502

125%

129%

133%

$2,721

$2,804

$2,892

Wells Combined

5.5

5.7

5.8

$55,190

$57,507

$59,498

124%

129%

134%

$6,761

$7,045

$7,289

Student Services Building

4.8

4.8

4.8

$6,814

$7,021

$7,239

119%

122%

126%

$1,118

$1,152

$1,188

Building Name

Net
Present
Value

76

To determine the electricity reduction available from combining occupancy sensing control
systems alongside LED retrofits, the case the Engineering Building is examined. Three lamp F32T8
fixture (96 W) light on-rates are 100%, and with 284 fixtures the total annual electrical draw is currently
238.8 MWh/year. Retrofitting fixtures with CREE UR2-48-4500L LED systems (44 W) will reduce the
annual electricity consumption rate to 109.4 MWh/year (54.2% reduction). Applying the average
expected light on-rate reduction due to occupancy sensing system installation (43%) to the
109.4MWh/year rate of consumption with installed LED retrofits results in an annual electricity
consumption rate of 62.4 MWh/year, a reduction of 73.8% over fluorescent fixtures with no occupancy
sensing systems installed.
*
To calculate the estimated annual energy use for devices in Table 17 (Chapter 2 – Energy Use
Intensity of MSU Campus Spaces), average daily energy use was determined for both weekdays and
weekends. For weekdays, it is assumed that offices are occupied 8 hours each day. For weekends, it is
assumed that offices remain unoccupied. Using the wattage draw values (Wi) from Table 16 and the
On/Off/Standby hours (Timei) from Table 17, the following Equation is used to calculate device draw in
watt-hours (Wh) used per day.
𝑑𝑒𝑣𝑖𝑐𝑒 𝑑𝑟𝑎𝑤 = 𝑊𝑜𝑛 ∗ 𝑇𝑖𝑚𝑒𝑜𝑛 + 𝑊𝑠𝑡𝑎𝑛𝑑𝑏𝑦 ∗ 𝑇𝑖𝑚𝑒𝑠𝑡𝑎𝑛𝑑𝑏𝑦 + 𝑊𝑜𝑓𝑓 ∗ 𝑊𝑜𝑓𝑓

(11)

The values in Table 17 for hourly usage are for weekdays. Weekend operation assumes spaces are
unoccupied, so on time for all devices (except for refrigerators) is assumed to be zero hours. All other
devices are assumed to be left in standby mode except for laptops. It is assumed that laptops are taken
home with occupants and therefore do not draw power within the office space during the weekend.

77

BIBLIOGRAPHY

78

BIBLIOGRAPHY

1. EIA. (2017). Commercial Buidlings Energy Consumption Survey 2012 - Trends in Lighting in
Commercial Buildings. Retrieved from
https://www.eia.gov/consumption/commercial/reports/2012/lighting/
2. Comstock, O., & Jarzomski, K. (2014, March). LED Bulb Efficiency Expected to Continue Improving
as Cost Declines. US Energy Inveromation Administration. Retrieved from
https://www.eia.gov/todayinenergy/detail.php?id=15471
3. Ding, Y. (2015). Solar Energy Related Applications, Education, and Building Retrofits. ProQuest.
Retrieved from http://search.proquest.com.proxy1.cl.msu.edu/docview/1657428372
4. Davis III, J. A., & Nutter, D. W. (2010). Occupancy diversity factors for common university building
types. Energy and Buildings, 42(9), 1543–1551. https://doi.org/10.1016/j.enbuild.2010.03.025
5. Guo, X., Tiller, D., Henze, G., & Waters, C. (2010). The performance of occupancy-based lighting
control systems: A review. Lighting Research & Technology, 42(4), 415–431.
https://doi.org/10.1177/1477153510376225
6. Osterhaus, Werner K.E. (1993). Office Lighting: A Review of 80 Years of Standards and
Recommendations. In 1993 IEEE Industry Applications Society Annual Meeting.
7. Energy Trust of Oregon. (2013, July). Footcandle Light Guide. Retrieved from
https://www.lightingdesignlab.com/sites/default/files/pdf/Footcandle_Lighting%20Guide_Rev.072013.
pdf
8. Grossman, A. J. (2014). Renewable energy and conservation measures for non-residential buildings
(M.S.). Michigan State University, Ann Arbor. Retrieved from
https://search.proquest.com/docview/1547038417?accountid=12598
9. Bauer, W., Bollman, D., Ellerhorst, R., Latta, W., & Verhanovitz, N. (2017). Integration of Research,
Teaching, and Practice in the Implementation of the Michigan State University Energy Transition Plan.
In Handbook of Theory and Practice of Sustainable Development in Higher Education (Vol. 2, pp.
401–412).
10. Consumers Energy. (2017). Business Energy Efficiency Programs: 2017 Incentive Catalog. Retrieved
from
https://www.consumersenergy.com/~/media/CE/Documents/Energy%20Efficiency/Business/businesscatalog.ashx
11. US Department of Energy. (n.d.). Energy Escalation Rate Calculator. Retrieved 1/6/2016
12. Hastbacka, M., Beeson, T., Cooperman, A., Dieckmann, J., & Bouza, A. (2013). Harvesting daylight.
ASHRAE Journal, 55(4), 60+.

79

13. Dubois, M.-C., & Blomsterberg, Å. (2011). Energy saving potential and strategies for electric lighting
in future North European, low energy office buildings: A literature review. Energy and Buildings,
43(10), 2572–2582. https://doi.org/10.1016/j.enbuild.2011.07.001
14. Galasiu, A. D., & Veitch, J. A. (2006). Occupant preferences and satisfaction with the luminous
environment and control systems in daylit offices: a literature review. Special Issue on Daylighting
Buildings, 38(7), 728–742. doi:10.1016/j.enbuild.2006.03.001
15. AMO Energy Resources Center, U. D. of E. (n.d.). Daylighting Tool.
16. Lutron Electronics Co. (2010). Light Control and LEED. Retrieved from
http://e2s2.ndia.org/pastmeetings/2011/tracks/Documents/12559.pdf
17. Doulos, L. T., Tsangrassoulis, A., Kontaxis, P. A., Kontadakis, A., & Topalis, F. V. (2017). Harvesting
daylight with LED or T5 fluorescent lamps? The role of dimming. Energy and Buildings, 140, 336–
347. doi:10.1016/j.enbuild.2017.02.013
18. D&R International, Ltd. (2012). 2011 Buildings Energy Data Book.
19. Office of Planning and Budgets – Facilities and Planning and Space Management. (2016). Room Use
By Building. Retrieved 1/28/2016
20. US EPA. (2013). Room Air Conditioner Calculator. Retrieved from
www.energystar.gov/ia/business/bulk_purchasing/bpsavings_calc/CalculatorRoomAC.xls
21. Korn, D., & Walczyk, J. (2016). Exactly What Is a Full Load Cooling Hour and Does Size Really
Matter? ACEEE Summer Study on Energy Efficiency in Buildings. Retrieved from
http://aceee.org/files/proceedings/2016/data/papers/1_1168.pdf
22. EnergyCAP. (2017). Weather Data Depot. EnergyCAP.
23. NOAA. (2017, June 15). Summary of Monthly Normals 1981-2010. NOAA.
24. Facilities Information Tool. (2017). Michigan State University
25. Michigan State University. (2012). Energy Transition Plan. Michigan State University. Retrieved from
http://ipf.msu.edu/files/pdfs/energy-transition-plan.pdf
26. “Climate Change 2014: Mitigation of Climate Change.” IPCC, 2014.
27. “Short-Term Energy Outlook.” US Energy Information Administration, 2016.
28. Savaresi, Annalisa. “The Paris Agreement: A New Beginning?” Journal of Energy & Natural
Resources Law 34, no. 1 (January 2, 2016): 16–26. doi:10.1080/02646811.2016.1133983.
29. Energy Information Administration. “International Energy Outlook 2016: With Projections to 2040.”
(May, 2016): 19-44.

80

30. Lewis, Nathan S, and Daniel G Nocera. “Powering the Planet: Chemical Challenges in Solar Energy
Utilization.” Proceedings of the National Academy of Sciences of the United States of America 103, no.
43 (October 24, 2006): 15729–35. doi:10.1073/pnas.0603395103.
31. Breyer, Christian, and Alexander Gerlach. “Global Overview on Grid-Parity.” Progress in
Photovoltaics: Research and Applications 21, no. 1 (2013): 121–36. doi:10.1002/pip.1254.
32. Yang, Hung-Jen, Chien-Hsun Chen, Wei-Cheng Lai, Chien-Liang Wu, Chien-Fu Huang, Yu-Hung
Chen, and Jenn-Chang Hwang. “Adjusted Colorful Amorphous Slicon Thin Film Solar Cells by a
Multilayer Film Design.” Journal of The Electrochemical Society 158, no. 9 (September 1, 2011):
H851–53. doi:10.1149/1.3604949.
33. Lunt, Richard R., and Vladimir Bulovic. “Transparent, near-Infrared Organic Photovoltaic Solar Cells
for Window and Energy-Scavenging Applications.” Applied Physics Letters 98, no. 11 (2011).
doi:http://dx.doi.org/10.1063/1.3567516.
34. Batchelder, J. S., A. H. Zewai, and T. Cole. “Luminescent Solar Concentrators. 1: Theory of Operation
and Techniques for Performance Evaluation.” Applied Optics 18, no. 18 (September 15, 1979): 3090–
3110. doi:10.1364/AO.18.003090.
35. Weber, W. H., and John Lambe. “Luminescent Greenhouse Collector for Solar Radiation.” Applied
Optics 15, no. 10 (October 1, 1976): 2299–2300. doi:10.1364/AO.15.002299.
36. Yang, Chenchen, and Richard R. Lunt. “Limits of Visibly Transparent Luminescent Solar
Concentrators.” Advanced Optical Materials 5, no. 8 (2017). doi:10.1002/adom.201600851.
37. Bradshaw, Liam R., Kathryn E. Knowles, Stephen McDowall, and Daniel R. Gamelin. “Nanocrystals
for Luminescent Solar Concentrators.” Nano Letters 15, no. 2 (February 11, 2015): 1315–23.
doi:10.1021/nl504510t.
38. Coropceanu, Igor, and Moungi G. Bawendi. “Core/Shell Quantum Dot Based Luminescent Solar
Concentrators with Reduced Reabsorption and Enhanced Efficiency.” Nano Letters 14, no. 7 (July 9,
2014): 4097–4101. doi:10.1021/nl501627e.
39. Kanellis, Michalis, Minne M. de Jong, Lenneke Slooff, and Michael G. Debije. “The Solar Noise
Barrier Project: 1. Effect of Incident Light Orientation on the Performance of a Large-Scale
Luminescent Solar Concentrator Noise Barrier.” Renewable Energy 103 (April 2017): 647–52.
doi:10.1016/j.renene.2016.10.078.
40. Correia, Sandra F. H., PatrĂ­cia P. Lima, Edison Pecoraro, Sidney J. L. Ribeiro, Paulo S. AndrĂŠ, Rute A.
S. Ferreira, and Luís D. Carlos. “Scale up the Collection Area of Luminescent Solar Concentrators
towards Metre-Length Flexible Waveguiding Photovoltaics.” Progress in Photovoltaics: Research and
Applications 24, no. 9 (September 1, 2016): 1178–93. doi:10.1002/pip.2772.
41. Zhao, Yimu, and Richard R. Lunt. “Transparent Luminescent Solar Concentrators for Large-Area
Solar Windows Enabled by Massive Stokes-Shift Nanocluster Phosphors.” Advanced Energy Materials
3, no. 9 (September 1, 2013): 1143–48. doi:10.1002/aenm.201300173.

81

42. Meinardi, Francesco, Hunter McDaniel, Francesco Carulli, Annalisa Colombo, Kirill A. Velizhanin,
Nikolay S. Makarov, Roberto Simonutti, Victor I. Klimov, and Sergio Brovelli. “Highly Efficient
Large-Area Colourless Luminescent Solar Concentrators Using Heavy-Metal-Free Colloidal Quantum
Dots.” Nat Nano 10, no. 10 (October 2015): 878–85.
43. Zhao, Y., Meek, G., Levine, B., Lunt, R. “Near-Infrared Harvesting Transparent Luminescent Solar
Concentrators.” Advanced Optical Materials 2, no. 7 (2014): 606-611. doi: 10.1002/adom.201400103
44. Vossen, Finn M., Mariëlle P.J. Aarts, and Michael G. Debije. “Visual Performance of Red
Luminescent Solar Concentrating Windows in an Office Environment.” Energy and Buildings 113
(February 1, 2016): 123–32. doi:10.1016/j.enbuild.2015.12.022.
45. Debije, MG Michael, and PPC Paul Verbunt. “Thirty Years of Luminescent Solar Concentrator
Research : Solar Energy for the Built Environment.” Advanced Energy Materials 2, no. 1 (2012): 12.
46. Slooff, L. H., E. E. Bende, A. R. Burgers, T. Budel, M. Pravettoni, R. P. Kenny, E. D. Dunlop, and A.
Büchtemann. “A Luminescent Solar Concentrator with 7.1% Power Conversion Efficiency.” Physica
Status Solidi (RRL) – Rapid Research Letters 2, no. 6 (December 1, 2008): 257–59.
doi:10.1002/pssr.200802186.
47. Chemisana, Daniel. “Building Integrated Concentrating Photovoltaics: A Review.” Renewable and
Sustainable Energy Reviews 15, no. 1 (January 2011): 603–11. doi:10.1016/j.rser.2010.07.017.
48. Thompson, G, and M Lordan. “A Review of Creativity Principles Applied to Engineering Design.”
Proceedings of the Institution of Mechanical Engineers, Part E: Journal of Process Mechanical
Engineering 213, no. 1 (February 1, 1999): 17–31.

49. Lakowicz, J. R. (1983). Quenching of Fluorescence. In J. R. Lakowicz (Ed.), Principles of
Fluorescence Spectroscopy (pp. 257–301). Boston, MA: Springer US.
https://doi.org/10.1007/978-1-4615-7658-7_9
50. Lunt, R. “Solar Charging Calculator.”
http://www.egr.msu.edu/~rlunt/Solar_Charging_Worksheet_Calculator.xlsx

82