VC3.5
DEGRADATION ANALYSIS OF SILICON PHOTOVOLTAIC MODULES
E. L. Meyer and E. E. van Dyk
Department of Physics, University of Port Elizabeth, PO Box 1600, Port Elizabeth, 6000, SOUTH AFRICA,
Tel: +27 41 504 2579, Fax: +27 41 504 2573, E-mail: phbelm@upe.ac.za
ABSTRACT
The continuing quest to reduce photovoltaic module degradation and improve performance requires a complete set of techniques to evaluate the module’s entire cells’ parameters.  The foremost technique to observe module degradation is by monitoring the module performance under actual operating conditions.  In this study the performance of silicon photovoltaic modules was monitored over an extended period of time.  Modules comprising cells of three different technologies were used in the investigation.  The cell technologies are Edge-defined Film-fed Growth silicon, mono-crystalline silicon and multi-crystalline silicon.  During the monitoring period degradation in the performance of one of the
modules was observed.  An analysis of this degradation showed that low cell shunt resistance caused the observed degradation.  It is also shown that these shunt paths reduce a module’s efficiency when operating under low light levels.  Hence, verifying the fact that shunt paths caused the observed degradation.
Keywords:  Monitoring – 1: Degradation – 2:  Shunts – 3
Presented at the 16th  EPVSEC.  1–5 May 2000.  Glasgow, UK.
1.INTRODUCTION
Photovoltaic (PV) modules are renowned for their reliability.  However, some modules do degrade and can fail when operating outdoors for extended periods of time.Since there is an ever-increasing deployment of PV as a form of energy in South Africa [1], it is desirable to have a failure reporting and analysis program to support the local industry as illustrated in [2,3].  A failure analysis program would include evaluation of the different degradation and failure modes on a regular basis as well as monitoring module performance under operating conditions.  In this study the performance of three silicon-based PV modules was monitored over an extended period of time.  During this time, performance degradation of one of the modules was observed.  This module was then subjected to a
test procedure to analyse the observed degradation.  Results obtained indicate that the degraded module had cells with low shunt resistances.  The effect of these low shunt resistance cells on module performance at low light levels was investigated.  Results showed that modules with low shunt resistance cells perform worse than “normal”modules at low light levels.2.
MODULE PERFORMANCE
The three modules used in this study comprise Edge-defined Film-fed Growth silicon (EFG-Si), mono-crystalline silicon (mono-Si) and multi-crystalline silicon (multi-Si) cells.  The modules’ powers measured at standard test conditions (STC) both before and after the test period, the manufacturer’s specified power and aperture area efficiency, η, are all listed in table 1.  The measured powers were obtained using a SPI-SUN Simulator 240,with an estimated uncertainty of 5% in the measurement.
The efficiency corresponds to the power measured before deployment.
Table I: Specified power ratings, measured STC power before and after outdoor deployment and efficiency of the modules before the outdoor deployment.
Module
Specified (W)
STC befor e
(W)
STC after (W)
ηbefor e (%)
EFG-Si 50.048.6349.8312.5Mono-Si 53.047.0248.8511.7Multi-Si 53.046.0247.2810.5
The modules were deployed outdoors under operating conditions over a 15-month period from August 22, 1997to October 30, 1998 at the Outdoor Research Facility (ORF) of the University of Port Elizabeth (UPE), South Africa [4].  The modules were mounted outdoors on a north-facing rack at an angle of 34º, the latitude of Port Elizabeth.  The operating currents and voltages of the modules were measured every 15 seconds and stored as 15-minute averages.  The modules were subjected to a regulated voltage and identical environmental conditions.This allowed for comparison between the energy production of all the modules.  The effect of temperature,irradiance and wind speed on module performance was also investigated [5,6].
From June 1998 degradation in the EFG-Si module
performance was observed.  This degradation is clearly seen by comparing the EFG-Si module performance to the other modules’ performances in the months of May and June 1998.  The modules’ performances over the 15-month period are illustrated in figure 1 below.  The months in which the degradation was first observed are circled in figure 1.  The energy conversion efficiency of the EFG-Si module decreased by more than 22% from May 1998 to June 1998 while the efficiencies of the other two modules
decreased by less than 13% for the same time.  These percentages were calculated relative to the energy measured during May 1998.  This was due to the lower,total daily irradiance-levels.  The average daily irradiance for June 1998, was approximately 4500 Wh/m 2/day.  From June to July 1998 the efficiencies of all the modules increased by about 13% indicating that the EFG-Si module has degraded in efficiency by approximately 10% from May to July 1998.  Energy production of the EFG-Si module in months after July 1998 indicates that the observed degradation was permanent.
degrade
50A u g '97
S e p '97
O c t '97
N o v '97
D e c '97
J a n '98
F e b '98
M a r '98
A p r '98
M a y '98
J u n '98
J u l '98
A u g '98
S e p '98
O c t '98
A v e r a g e  E n e r g y  (W h /d a y )Figure 1: Average monthly energy per day produced by each module.  Note the performance degradation of the EFG-Si module.3.
DEGRADATION ANALYSIS
A photovoltaic module or cell can degrade for a number of reasons.  To effectively analyse any observed degradation when a module operates outdoors, a complete set of techniques is required.  Such a set of techniques was formulated for this study.  These techniques include measuring light and dark current–voltage characteristics,electrical evaluation, hot-spot investigation and measuring individual cell parameters non-intrusively.
The EFG-Si module was subjected to these testing
techniques in order to analyse the observed performance degradation.  No visual defects on the cells of the module were observed.  No electrical fault or cell mismatch was observed.  Since the module conformed to all the external tests’ requirements, it was suspected that the degradation was caused by inherent cell defects.
These inherent cell defects may be caused by four
factors, namely, an increase in the cell’s series resistance, a decrease in the cell’s shunt resistance, antireflection coating deterioration, and discoloration of the encapsulant.The module’s total series resistance as obtained from light I–V measurements was 0.96 Ω.  No antireflection coating deterioration or yellowing of the e ncapsulant was observed during visual inspection.
A two-terminal diagnostic technique was then used
to measure the individual cell shunt resistances, non-intrusively, without de-encapsulation [7].  Flawed manufacturing processes and material defects cause a shunt path in a PV cell (cell having low shunt resistance).  The
results of measuring the individual cell shunt resistance of the modules under test are presented and
discussed in the next section.4.RESULTS Results obtained for the individual cell shunt resistances are shown in figure 2 for the EFG-Si module and in figure 3 for the mono-Si module as measured at the time of the observed degradation.
20406080100120
1401
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
Cell
S h u n t  R e s i s t a n c e  (o h m )
Figure 2: Shunt resistance of individual cells in the EFG-Si module after the observed degradation.  Cells 6 and 12will detract from the module output.
24681012141618201
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
Cell
S h u n t  R e s i s t a n c e  (o h m )Figure 3: Shunt resistance of individual cells in the mono-Si module.  This module had no cells with low shunt resistances.
The multi-Si module had a similar trend as that of the mono-Si module (fig. 3).  The two crystalline modules had no cells with low shunt resistances.  The EFG-Si module on the other hand, had a few cells with low shunt resistances as can be seen from figure 2.  The dotted line in figure 2 corresponds to 7.5 Ω, the shunt resistance of all the cells when fully illuminated.  This means that cells with shunt resistance close to or less than 7.5 Ω will detract from the module output under low light levels.
Although the EFG-Si module did not decrease in
maximum power at STC (table 1), its total energy production decreased.  To establish whether the low shunt-
resistance cells were responsible for the observed energy degradation, consider the equivalent cell model for three
cells in series (figure 4).
Figure 4: Equivalent cell model for three cells in series.Cell 2 has a low shunt resistance and carries the current away from the intended load.
Cell 2 has a low shunt resistance.  At high light levels, the photocurrent, I ph , will be relatively high.  A small fraction of this current will be carried away from the intended load by the shunt path, which can only carry so much current [8].  At low light levels, however, this shunt path will dissipate a bigger fraction of the now lower photocurrent.  This situation occurs when a module operates outdoors on cloudy days or when the sun is lower in the sky.
To verify that the degraded energy production of
the EFG-Si module in figure 1 can be attributed to the lower irradiance conditions, the three modules’ efficiencies were characterised as a function of light intensity.  This was done by placing the modules in the solar simulator and covering it with masks of different opacities, attenuating the light levels accordingly.  The module aperture area efficiency was then determined for each light level and is shown in figure 5 for the EFG-Si module and the mono-Si module.  The error in the efficiency at each light level is less than 0.2%.  Therefore, the mono-Si module is expected to be more efficient than the EFG-Si module at low light levels.  It is evident from figure 5 that the efficiency of the EFG-Si module is further reduced than that of the mono-Si module at low light levels as was the case during the months of the observed degradation.
It was also noted that for modules with cells having
high shunt resistances, the efficiency first increases as light intensity is reduced.  This can be explained by again considering the equivalent cell model of figure 4.
By taking the direction of I ph  to be positive, the net
current I out , is given by [9]:
sh
D ph out I I I I −−=          (1)
The dark recombination current I D , which is formed by recombination of electron-hole pairs within the semiconductor, detracts from the I ph .  The rate of recombination depends, among others, on the availability of charge carriers and subsequently on carrier lifetime.Using the fact that the I ph  is directly proportional to the incident irradiance, we can define the efficiency of the cell to be:
ph
out
I I =η          (2)
and from equation (1):
ph
sh
ph D I I
I I 1−−=η          (3)
If the cell’s shunt resistance is sufficiently large so that its shunt current, I sh , is negligibly small, equation (3) reduces to:
ph
D
I I 1−=η          (4)
If it is assumed that an irradiation level of 1000 W/m 2contributes 100% of the incoming photons to I ph  and 30%of that is dissipated in the recombination current, I D ,equation (4) suggests a cell efficiency of 70%.  Since I D  i s highly dependent on carrier lifetime [9] it will be drastically reduced at lower values of I ph .  Hence, when I ph constitutes 50% of the incoming photons, less than 15%will be dissipated in I D .  If, for instance, 10% of the incoming photons at 500 W/m 2 are dissipated in I D , the cell’s efficiency will be 80% according to equation (4),hence an increase in the efficiency as light intensity is reduced from 1000 W/m 2 to 500 W/m 2.  This situation will perpetuate until both I D  and I ph  have the same order of magnitude, which will act to reduce the efficiency.The
refore, modules that first show an increase in efficiency as light intensity is reduced, are expected to have no cells with low shunt resistances.  If, on the other hand, the cell’s shunt resistance is small, the shunt current, I sh , in equation (3) cannot be neglected.  The efficiency in this case is only
expected to decrease as light intensity is reduced.
10.0
10.410.811.211.612.012.412.80200400
600800
10001200
Light Intensity (W/m² )
M o d u l e  E f f i c i e n c y  (%)
Figure 5: Peak-power efficiencies as a function of light intensity for the EFG-Si module and the mono-Si module.5.
CONCLUSIONS
The outdoor performances of 3 silicon, PV module were monitored over a 15-month period from August 22, 1997 to October 30, 1998.  Degradation in the performance of the EFG-Si module was observed since June 1998.  This emphasises the importance of monitoring the performance of PV modules deployed outdoors in order to detect degradation modes.
To analyse these degradation modes, a complete set
of techniques is required.  Such a set of techniques must be able to evaluate all cell parameters.
The EFG-Si module had cells with low shunt resistances causing shunt paths.  These shunt paths lead the current away from its intended load.  These shunt paths
cause a module to be less efficient at low light levels than "normal” modules.  This was verified by measuring the module efficiency as a function of light intensity.
Results showed that all modules operating under reduced light intensity have a significant decrease in their efficiency.  It also showed that if cells in a module have low shunt paths that divert current away from the intended load, the module’s efficiency will be even further reduced under these lower l
ight conditions.  The EFG-Si module exhibited this behaviour.  Knowledge of individual cell shunt resistance of an encapsulated module is therefore imperative to identify cells that are degrading module performance.
6.ACKNOWLEDGEMENTS
The authors wish to express their gratitude to Tom Basso of the National Renewable Energy Laboratory, Golden, CO for useful communications in building the shunt resistance measurement system. The financial assistance of the South African National Research Foundation is also acknowledged.
7.REFERENCES
[1]Swanepoel R. (1998).  “An appropriate energy
source for Africa”.  S. Afr. J. Sci. 94.  pp. 46. [2]Dumas L.N., Shumka, A. (1982).  “Photovoltaic
module reliability improvement through
application testing and failure analysis”.  IEEE
transactions on reliability, vol. R-31, No. 3.
August.[3]Forman S.E. (1982).  “Performance of
experimental terrestrial photovoltaic modules”.
IEEE transactions on reliability, vol. R-31, No.
3.  August.
[4]Meyer E.L. (1999).  Investigation of properties
and energy rating of photovoltaic modules.
M.Sc. dissertation.  University of Port Elizabeth,
South Africa.  January.  pp. 74 – 82.
[5]van Dyk E.E, Meyer E.L., Scott B.J., O’Connor
D.A., Wessels J.B. (1997).  “Analysis of
photovoltaic module energy output under
operating conditions in South Africa”.  26th IEEE
PVSC.  pp. 1197 – 1200.
[6]Meyer E.L. and van Dyk    E.E. (2000).
“Development of an energy model based on total
daily irradiation and maximum Ambient
Temperature”, Renewable Energy Journal 21/1.
pp. 37 – 47.  20th April.
[7]McMahon T.J. (1995).  “Cell shunt resistance
and PV module ratings”.  13th NREL
photovoltaics programs review, Lakewood, CO.
May.
[8]McMahon T.J., Basso T.S. (1995).  “Two
terminal diagnostics for cells in series connected
photovoltaic modules”.  ASME International
Solar Energy Conference, Lahaina, HI, March
19–24 .
[9]Mazer, J. A. (1997).  Solar cells: An introduction
to crystalline photovoltaic technology.  Kluwer
Academic Publishers.  pp. 98 – 99.

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。