United States
Environmental Protection
Agency
National Risk Management
Research Laboratory
Cincinnati, OH 45268
Research and Development
EPA/600/SR-98/056 July 1998
Project Summary
Demonstration of the
Environmental and Demand-side
Management Benefits of
Grid-connected Photovoltaic
Power Systems - 1994-1997
Daniel L. Greenberg, Edward C. Kern, Jr., Miles C. Russell, and Priscilla D. Kern
The report gives results of an inves-
tigation into the pollutant emission re-
duction and demand-side management
potential of 12 photovoltaic (PV) sys-
tems that began operation between
June 1994 and July 1996 in various
locations in the U.S. The project was
sponsored by the U.S. EPA and 12 elec-
tric utilities. The report documents the
project and presents analyses of each
system's ability to offset emissions of
sulfur dioxide, nitrogen oxides, and car-
bon dioxide, and to provide power dur-
ing peak load hours for the host
buildings and the participating utilities.
The analyses indicate a broad range
in emission offsets resulting from PV
system operation due to variation in
the solar resource available to each
system and variation in the marginal
emission rates of the participating utili-
ties. Each system's ability to provide
power during peak load periods is in-
vestigated using gross and net (of PV
generation) load duration curves. Dif-
ferences between these curves provide
insight into each PV system's ability to
reduce the highest loads experienced
by the host building or utility. One set of
such curves is presented for each sys-
tem and for each month of a 12-month
performance monitoring period.
This Project Summary was developed
by the National Risk Management Re-
search Laboratory's Air Pollution Pre-
vention and Control Division, Research
Triangle Park, NC, to announce key find-
ings of the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
In May 1993 the U.S. EPA issued the
second of three solicitations for the instal-
lation of grid-tied PV systems with the
goal of measuring their environmental and
demand-side benefits. Following up on
its success with the first solicitation, As-
cension Technology's proposal was again
selected from the proposals submitted.
Ascension's proposal was supported
by 12 electric utilities, all of which ulti-
mately participated in the project by in-
stalling a PV system. The participating
utilities were 1) Boston Edison Company
(BECO), serving the greater Boston area;
2)Public Service of Oklahoma (PSO),
which has three service areas in eastern
and central Oklahoma; 3) Consolidated
Edison (CONED), serving New York City
and Westchester County, New York; 4)
Idaho Power Corporation (IPC), serving
most of Idaho; 5) New York State Electric
and Gas (NYSEG), with service areas
concentrated in western New York state,
6) Nevada Power Company (NPC), serv-
ing Las Vegas and southern Nevada; 7)
the Los Angeles Department of Water and
Power (LADWP), which serves Los An-
geles; 8) Public Service Company of Colo-
rado (PSCO), serving Denver and other
urban areas of Colorado; 9) the Arizona
-------
Electric Power Co-Operative (AEPCO),
which serves rural areas in southern Ari-
zona; 10) Florida Power Corporation
(FPC), serving the panhandle and gulf
coast of Florida; 11) Atlantic Energy (ACE)
(formerly Atlantic City Electric), which
serves southern New Jersey; and 12)
Duke Power (Duke), with a service area
covering central and eastern North Caro-
lina. In addition to the geographic diver-
sity of the service areas represented by
these utilities, their pollutant emission
characteristics also proved to be quite
divergent. Ascension Technology's part-
ners from the PV industry were ASE
Americas, which provided PV modules,
and Omnion Power Engineering Corpo-
ration, which provided the inverters.
EPA awarded the contract for this
project to Ascension Technology in Sep-
tember 1993. The final system design ef-
fort began shortly thereafter, and the first
system was installed and operating in
May 1994. Installations of the remaining
11 systems were completed over the
course of the next 2 years.
Monitoring of each system began con-
currently with initial system operation, al-
though the "official data start date" was
delayed at sites where there were initial
technical problems with either instrumen-
tation or PV system hardware. At each
site, 15-minute average values of solar
irradiance, ambient temperature, PV sys-
tem power output, and building load were
recorded and stored for subsequent re-
trieval by modem. Monitoring of each site
for the purposes of this study continued
for a period of 1 year.
Emission rate and load data provided
by each participating utility were used in
conjunction with the data collected from
each system to conduct analyses of (1)
the ability of each PV system to reduce
the peak power demand of the building
on which it was installed; (2) the chrono-
logical correlation of each PV system's
power output to the respective utility's
peak loads; and (3) the emission offsets
resulting from operation of the PV sys-
tems.
Chapter 1 of this report provides a gen-
eral introduction to the project. Chapter 2
describes the design, installation, and cost
of each system. Chapter 3 describes the
data acquisition system and presents data
collection and review procedures. Chap-
ter 4 describes the operating history and
performance of each system in turn, and
presents the results of the three analyses
described above. Chapter 5 discusses
the common operational problems en-
countered by this set of PV systems, and
reviews the results of the emission offset
and load matching analyses across all
12 systems. The authors' conclusions re-
garding the use of PV systems to offset
pollutant emissions or for purposes of re-
ducing peak building or utility loads are
presented in Chapter 6.
PV Systems
All PV systems used in this project
shared the same design but were installed
in different parts of the country. Their per-
formance, as measured by PV system
outage, was affected by both systemic
and environmental factors.
System Design
All PV systems installed under this
project have a peak rating of 18 kW, and
consist of three independent 6 kW sub-
systems, each with its own inverter. The
inverter in each subsystem is fed by three
parallel source circuits, each consisting
of 10 PV modules in series. There are
thus 90 modules in each system. All 12
installations utilize Ascension Technology
RoofJack PV array supports, which have
been used to install more than 1 MW of
PV systems. PV arrays are held in place
by ballast on flat roofs; this approach re-
quires no roof penetrations for hold-down
of the PV arrays. System design details
were developed in close cooperation with
Mobil Solar (now ASE Americas, Inc.),
the PV module supplier. Omnion Power
Engineering was selected as the supplier
of power conditioners. The PV systems
were designed to accommodate the speci-
fications of the 6 kW-rated Omnion Series
2200 unit.
System Installation
Although the typical installation took
only 3-4 days to complete, the final sys-
tem did not begin operation until nearly 2
years after the first system's installation
was complete. This prolonged period was
due to numerous siting and code compli-
ance difficulties encountered in the pro-
cess of installing the systems.
With one exception, the systems were
installed on the roofs of commercial and
industrial buildings. The exception was
the ground-mounted array on the cam-
pus of the University of Nevada at Las
Vegas (EPA23). Table 1 summarizes the
location of each system.
System Performance History
Of the 12 PV systems installed by this
project, all but one suffered one or more
events during the study period which
temporarily limited system output or pre-
vented generation altogether. Inverter-re-
Table 1.
Participating Utilities and
Installation Locations
Site Name
EPA 18
EPA 19
EPA 20
EPA 21
EPA 22
EPA 23
EPA 24
EPA 25
EPA 26
EPA 27
EPA 28
EPA 29
Utility
BECO
PSO
CONED
I PC
NYSEG
NPC
LADWP
PSCO
AEPCO
FPC
ACE
Duke
Location
Boston, MA
Lawton, OK
Greenburgh, NY
Boise, ID
East Aurora, NY
Las Vegas, NV
Los Angeles, CA
Henderson, CO
Benson, AZ
Clearwater, FL
Pomona, NJ
Huntersville, NC
lated problems were the most vexing of
the generation-limiting events. In all, 22
inverter-related events resulted in a gen-
eration loss of 12,200 kWh, approximately
3.3% of the combined generation of these
systems over the relevant time periods.
Snow cover was also a frequent cause of
PV system outages for those systems lo-
cated in northern locations or at high alti-
tudes. Accurate estimation of generation
losses due to snow cover is not possible
because the sensor used to measure sun-
light (the primary input to simulation of
the systems' performance) was usually
covered by snow when the arrays were.
Only about 600 kWh of lost generation
may be attributed specifically to module
failures. Although such failures are known
to have resulted in greater losses, it is not
possible to separate these losses from
those due to other equipment failures that
occurred simultaneously. About 1,040
kWh (approximately 0.3% of gross gen-
eration) were lost due to faults in array
wiring or problems with source-circuit pro-
tectors. Total generation by the PV sys-
tems was reduced by another 6,650 kWh
(1.8% of gross generation) due to failures
in utility equipment.
Results
The report describes results of this
project in terms of pollutant emission off-
sets and load reductions for both the util-
ity and the host building.
Pollutant Emission Offsets
Models of marginal emission rates (i.e.,
emission rates of load following units)
were developed for each utility based on
utility-provided data. The hourly emission
-------
500
400
"3
^
^ 300
o
g 200
c
<
100
0
n^
Tl
18 20 22 24 26 28
19 21 23 25 27 29
Site Number
Figure 1. Annual Sulfur Dioxide Offsets.
<
150
100
50
n
18 20 22 24 26 28
19 21 23 25 27 29
Site Number
Figure 2. Annual Nitrogen Oxides Offsets.
60
"55
§ 4°
0
ra 20
c
<
10
0
,— I
—
—
—
18 20 22 24 26 28
19 21 23 25 27 29
Site Number
Figure 3. Annual Carbon Dioxide Offsets.
rates of sulfur dioxide (SO2), nitrogen ox-
ides (NOX), and carbon dioxide (CO2) were
then combined with hourly PV system
generation data to determine hourly emis-
sion offsets.
Annual emission offsets are presented
in Figures 1 through 3. Note that insuffi-
cient data were provided to determine
marginal emission rates for sites EPA23
and EPA28, and that generation by sys-
tem EPA21 offset hydroelectric genera-
tion and therefore resulted in zero offsets
of all three pollutants. Aside from these
three systems, annual SO2 offsets ranged
from less than 80 g for a system offsetting
generation by natural-gas-fired combus-
tion turbines to 475 kg for a system that
offset the generation of a coal-fired power
plant. Annual NOX offsets ranged from 8
to 128 kg, and the range in annual CO2
emission offsets was from 16,700 to
50,400 kg.
The extreme variability in these results
is due to four factors: 1) the pollutant
emission rates of each utility's load-fol-
lowing generating units; 2) the installed
capacity of the PV systems; 3) the local
solar resource available to each system;
and 4) the operating performance (i.e.,
reliability) of each system. Of these, utility
emission rates were the most influential
factor in determining offsets, particularly
for SO2. The installed capacities of the 12
systems installed under this project were
identical, and therefore did not play a
role in inter-site differences in emission
offsets, but variations in the solar resource
and reliability of the systems certainly did.
Since there are currently no mitigation
measures in place for CO2, variation in
utility CO2 emission rates is due only to
the relatively small (about 2:1) variation
in the carbon content of fuels used and
variation in the heat rates of the power
plants. The range of the highest to lowest
annual offset is relatively small at 3.0. For
the other pollutants, variations in the pol-
lutant content of the fuel as well as inter-
utility differences in type and efficiency of
the load-following generators and in-
stalled pollution mitigation equipment give
rise to the tremendous differences be-
tween utility emission rates which under-
lie the differences in emission offsets
described above.
Host Building Load Reduction
Each PV system's ability to provide
power during building peak load hours
was analyzed by comparing each
building's monthly net (of PV generation)
and gross load duration curves (LDC).
The LDC is constructed by sorting all load
values for a given period in descending
order, and plotting each value against its
rank in the sort. Differences in a building's
net and gross LDC for the highest load
values indicate the PV system's ability to
provide power during peak load periods.
Figure 4 presents an example of gross
and net LDCs for one of the host build-
ings.
Some of the host buildings in this
project had loads very well suited to peak-
shaving by the PV system, while others
always experienced their peak loads at
times of low irradiance. During the 25
highest load hours in the months in which
each host building experienced its high-
est load (during the monitoring period),
the average reduction in the building LDC
ranged from 5 to 72% of system rating.
Analysis of the data collected through
this project revealed that the total annual
850 1
800
T3
(3
O
2
'5
750
700
Gross Load
Net Load
0 5 10 15 20 25
Highest Building Load Hours
Figure 4. Gross and Net LDCs.
insolation available at a given location
indicates little about the ability of a PV
system to provide power during peak load
hours. Several of the systems that pro-
vided the most power during peak load
hours also had among the lowest annual
insolation. Conversely, some of the sites
with the greatest solar resources proved
to be among the worst load matchers in
the group.
Another conclusion resulting from this
analysis is that a PV system's operation
during one or a small number of peak
building load hours may provide an inac-
curate picture of that system's capacity
-------
value. If a building's highest load(s) oc-
cur during mid-day hours but the building
also has near-peak loads occurring at
times of low or no irradiance, the near-
peak loalds will remain unaffected by the
PV system, and will migrate upward in
the building's net LDC. If the difference
between the daytime peak load(s) and
the nighttime near-peak load(s) is small
relative the PV system's capacity, the dif-
ference between the building's gross and
net LDCs may be substantially smaller
than the PV system's power generation
during the daytime peaks.
Conversely, if a PV system generates
at a low capacity factor during the high-
est building load hours, but provides more
power at times when building load is near
its peak level, it may be important to con-
sider whether the conditions that resulted
in the building's highest loads are repre-
sentative of typical conditions for the build-
ing.
Use of gross and net LDCs allows one
to review the effect of PV system opera-
tion on as many or as few of the highest
load hours as is desired, and is an es-
sential tool in assessing the capacity
value of a PV system.
Utility Peak Load Reduction
Conclusions regarding peak load re-
duction at the utility system level are much
the same as those discussed above for
building peak load reduction. For some
of the utilities participating in this study,
PV systems can be very effective at pro-
viding power during periods of peak load,
while for other utilities whose peaks oc-
cur at times of low or no irradiance, the
technology provides little capacity value.
In the months in which each utility's peak
load occurred, the reductions of the utili-
ties' LDCs over the 25 highest load hours
ranged from zero to 71% of PV system
rating.
As discussed above for building load,
the annual insolation available to a PV
system has little to do with its ability to
provide power when it is most needed.
The three systems that had the highest
levels of annual insolation yielded among
the lowest LDC reductions during the
utility's peak load month, and the system
that provided the most support to its utility
in the peak load month was ranked 10th
in annual insolation.
Finally, the use of chronological data
alone to interpret the capacity value of
PV may be deceiving. If near-peak utility
loads occur at times of low irradiance,
even a very large PV system (or set of
systems) would do little to reduce the
peak of the utility's LDC, even if that sys-
tem operates at a high fraction of its rat-
ing during the highest load hour or several
hours. Again, in assessing the capacity
value of PV systems, it is essential to
determine how the resource would alter
the utility LDC.
-------
D. Greenberg, E. Kern, Jr., M. Russell, and P. Kern are with Ascension Technology,
Inc., Lincoln, MA 01773.
Ronald J. Spiegel is the EPA Project Officer (see below).
The complete report, entitled "Demonstration of the Environmental and Demand-side
Management Benefits of Grid-connected Photovoltaic Power Systems - 1994-
1997," (Order No. PB98-145311; Cost: $44.00, subject to change) will be available
only from
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection Agency
CenterforEnvironmental Research Information
Cincinnati, OH 45268
BULK RATE
POSTAGES FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use
$300
EPA/600/SR-98/056
------- |