indent: 36pt;
">Limitations of accelerated testing to predict all possible degradation
modes and mechanisms in the photovoltaic PV modules necessitate that
actual outdoor monitoring and testing of PV modules be performed out-doors.
For this reason, thin film PV modules from leading US thin film PV manufacturers
namely, First Solar (Glass/CdTe/Glass), Shell Solar Glass/CIS/Glass),
Shell Solar New (Glass/CIGS/Glass), United Solar (a-Si:H on flexible
substrate), Energy Photovoltaics (Glass/a-Si:H/Glass) and Global Solar
(CIS on flexible substrate) with additional one crystalline silicon
module are being tested. The goal is to assess their performance in
the hot and humid climate of Florida and to correlate the PV performance
with the meteorological parameters namely, solar irradiance, temperature,
relative humidity, wind speed, etc. Statistical data analysis of the
recorded data is carried out on a daily basis and on a monthly basis
with PVUSA type regression analysis. Current-voltage characteristics
(I-V) of module arrays taken on a regular basis complement the results
obtained with continuous data monitoring. Moreover, high voltage bias testing of the modules
is carried out to study behavior of leakage currents and detect any
packaging material and processing flaws and consequently the module
reliability.

1. Objectives

The development of cost effective photovoltaic modules with 30 year
useful lifetime is the goal of the U.S. Department of Energy. It is
therefore essential to understand the failure modes and mechanisms in
PV modules and recommend improvements in the manufacturing technology
so as to avoid premature degradation of field-deployed modules and thus
improve their performance, reliability and durability.

2. Technical Approach

Modules from each company are divided in two sets. Within each set
they are connected in series so as to build maximum open circuit voltage
of less than +600 V and -600 V with respect to the ground. The arrays
are connected across fixed load resistors and current measuring shunts.
Two modules from each company are also biased at +600 V and -600 V individually
in order to study the leakage currents from the module frames to the
ground. The datalogger measures the data every five seconds and averages
over 15 minutes intervals are recorded.

3. Results and Accomplishments

The data is recorded on continuous basis and the power versus solar
irradiance graphs are plotted daily. PVUSA type regression analysis
of the recorded monthly data is carried using a LABVIEW program developed
in-house. The program provides power value for each module array under
performance test conditions (PTC) i.e. at 1000 W/m<sup>2
and 25°C. The variation of PTC power calculated for each month is plotted
and studied over a period of time. The quality of data has certainly
improved which can be seen from the PVUSA regression plots of a modules
array (Figure 1). The plots show very little scatter in the data 
compared to the scatter that used to be seen previously.

Figure 1: Plot of PV USA Regression for United Solar negative module
array for Dec 2006

Central Florida region being a hurricane prone area, hurricane protectors
were designed, fabricated and tested by the FSEC team. These hurricane
protectors have reduced DAS downtime to few days from few months during
2004 which was the year with the maximum hurricane activity in over
100 years.

Figure 1 shows consistency in the performance of EPV positive module
array. The plot is expected to remain horizontal if there are no signs
of degradation. Any downward shift in these curves would indicate an
onset of degradation.

Current-voltage measurements are also carried out on the module arrays
to corroborate the results obtained with continuous data monitoring.
Current-voltage measurements provide characteristic data,

Figure 2: Variation of PTC power for EPV positive module array.

such as open circuit voltage, short circuit                        
current, peak power and fill factor of PV arrays. The peak power is
easier to relate of all the PV characteristics since the values are
normalized at 1000 W/m<sup>2 and 25⁰C using the module
temperature coefficients provided by the manufacturers. Additionally,
this data is plotted over time  for all module array to understand
the performance trend. Taking the analyses based on the current-voltage
measurements and the monthly PTC power variation together has prove
to be considerably more effective than taking either of them alone.
An indicator of this fact is the observed trend for shell solar modules
that pointed toward some significant performance loss. This performance
loss was due to degradation in some older generation modules that were
part of the array, which was confirmed after the modules were tested
at NREL.

The Shell Solar modules were redeployed in May 2006 and since then
no further degradation has been observed in their performance. I-V measurements
are being carried out on a monthly basis since May-2006 in order to
gather more data points to fine tune the observed performance pattern.
No degradation has been observed for the module arrays of other companies.
New large-area CIGS modules received from Shell Solar were installed
at the outdoor site in September 2006.

The study of leakage currents in high-voltage biased thin film PV
modules fabricated by FS, SS, USSC and EPV is also being studied at
FSEC. Data is being collected on a continuous basis.  It is plotted
on a daily basis from midnight to midnight and analyzed, to study the
correlation of leakage currents with the relevant meteorological parameters
such as solar irradiance, module temperature and relative humidity over
the entire diurnal cycle (Figure 3).

4. Conclusions

   The PVUSA regression model is a very
efficient and accurate tool for studying performance degradation in
PV module arrays. Regression analysis complemented with I-V characteristics
verify results. No signs of module degradation have been observed in
modules of First Solar, United Solar, Energy Photovoltaics and

     Figure 3: Leakage Current trend
for Shell Solar modules on a cloudy day

new-generation Shell Solar modules. Even though initially a few old-generation
Shell Solar modules showed some degradation, no further degradation
has been observed in their performance after being redeployed in May
2006. Stable performance of PV module arrays over an extended period
of time can provide basis for providing warrantees. Any loss of performance
can assist in pin-pointing processes used in specific production batches
that need to be improved. Further, in-depth study and analysis of the
modules can also be carried out for understanding failure modes and
mechanisms using detailed analysis<sup>1-3.

Overall, data monitoring carried out over an extended period of time
will provide an insight into module performance, useful life-time and
degradation modes and mechanisms.

ACKNOWLEDGEMENTS

  This work is supported by the NREL and the
Thin-Film Partnership under the contract # ZDJ 33360002. The authors
are thankful to Bolko von Roedern, Carl Osterwald, Harin Ullal, Ken
Zweibel, NREL and colleagues from leading US PV companies for their
useful suggestions and discussions.

REFERENCES

1</sup>N. G. Dhere and N. R. Raravikar, Solar
Energy Materials and Solar Cells, vol. 67, (2001), pp 363-367.

2</sup>N. G. Dhere, Proc. 31<sup>st IEEE
Photovoltaic Specialists Conference, Orlando, FL, (2005), pp. 1570-1576.

3</sup>N. G. Dhere, N. R. Raravikar, A. Mikonowicz
and C. Cording, Proc. 29<sup>th IEEE Photovoltaic Specialists
Conference, New Orleans, Louisiana, (2002), pp. 231-234.

MAJOR FY 2006/2007 PUBLICATION

A. Kaul, B. Kumar, S. Khatri, and N. G. Dhere, Outdoor
Monitoring High and High Voltage bias Testing of Thin Film PV Modules,
Poster Presentation, FLAVS, March 2006.