IV degradation
I-V Curves from 300 Modules Deployed over 5-years in Tucson 1
2
3,4
3
Emily S. Kopp , Patricia L. Hidalgo-Gonzalez , Adria E. Brooks , Vincent P. Lonij , and Alex D. Cronin 1 University of Arizona College of Optical Sciences, Tucson, Arizona, USA 2 Pontificia Universidad Católica de Chile, Santiago, Chile 3 University of Arizona, Department of Physics, Tucson, Arizona, USA 4 Arizona Research Institute for Solar Energy, Tucson, Arizona, USA
ABSTRACT As part of our ongoing study of the degradation of PV systems, we have collected I-V curves from over 300 modules that have been deployed in grid-tied systems for up to 10 years in Arizona. We present families of I-V curves that exhibit different types of imperfection. One system (with Sharp modules) displays a large range of fill factors between the modules. In another system (a prototype CIGS system), the modules produce varying Isc and Voc values. The modules of another (Sanyo HIP) have a high degree of uniformity and a high fill factor. We present results correlating imperfections in I-V curve data to visual inspection findings and to changes in historical energy yields.
1,3
measurements for all of the modules in each system using a commercial I-V curve tracer made by Tahara model TSC-PD01 [2] [3]. Data from several clear-sky days is taken within one hour of noon when the global irradiance 2 is 990 W/m . I-V CURVE DATA FROM 300 MODULES Families of curves from several modules of the same type are presented in Figures 1 – 4. While the I-V curves in Figure 1 appear uniform and high in quality, the others show significant imperfections that we are correlating with degradation.
INTRODUCTION The TEP solar test yard contains over 600 grid-tied PV modules grouped into 22 PV systems, each with their own inverter. Most of the systems are 1 to 2 kW P in size and use a fixed-tilt of 32 degrees for the modules. Each system contains several modules of the same make and model, but often these modules degrade at different rates, as we show here. Originally designed as a test-bed for household-sized systems, the TEP solar test yard has been in operation since 2003 in Tucson, Arizona. We show here data from several modules that have been operating continuously for 5 to 9 years. The variety of I-V curve data shows several different types of imperfections that may result from degradation.
Figure 1 Eight I-V curves from Sanyo HIP-G751BA2 modules. These appear uniform and of high quality. These modules have a measured Pmax of 148 W under the field conditions and a nameplate value of 167 W at STC.
We previously studied degradation rates of PV systems at the TEP solar test yard by analyzing the historical ac daily power output from each system [1]. We found that comparing individual PV systems to the yard average is a sensitive way to measure degradation. In [1] we reported that several PV systems have degraded at 3%/year with a typical uncertainty of 0.5%/year. However, the daily final yield did not identify why some systems degrade, and from ac-power data alone, one cannot unambiguously distinguish between inverter degradation, changes in the surrounding albedo, or module degradation. The present study is aimed at determining the causes of module degradation. We temporarily disconnected each module from its gridtied systems and obtained individual I-V curve
Figure 2 Nine I-V curves from Sharp NEQ5E2U modules. The maximum power for these modules
ranges from 96 to 147 W due to a wide range of fill factors and have a nameplate value of 165 W at STC.
Figure 3 Sixteen I-V curves from prototype CIGS modules. The maximum power for these modules ranges from 15 to 32 W due to a wide range of Isc and Voc values. These modules have a nameplate value of 45 W at STC.
Figure 4 Ten I-V curves from amorphous silicon BP MST-50 modules. The maximum power for these modules ranges from 29 to 39 W and show a These significant range of Isc and Voc factors. modules have a nameplate value of 50 W at STC.
visible module browning, junction box failure, and cell damage with the I-V curves. The results of our correlation analysis will be catalogued along with their I-V curve and ac-power characteristics. Examples of weathering are shown in figures 5 – 6.
Figure 5 Failure of a surface electrode on the CIGS modules whose I-V curves are shown in Figure 3.
Figure 6 Electro-chemical weathering (note the asymmetry associated with polarity) of thin film amorphous silicon modules whose I-V curves are shown in figure 4.
In order to better compare the current module condition with the manufacturer’s rated sticker values we follow the procedure outlined in IEC 60904-10 to translate data taken 2 in the field to standard test conditions (STC: 1000 W/m , 25°C, AM1.5), and we are studying the uncertainty caused by this translation procedure. A previous study found this procedure to produce an error of 3% [4]. We find that some modules operated under noon-time 2 sunlight of 990 W/m on a clear day in January in Tucson perform at 88% of their nameplate value while the average of the 43 various modules presented here perform only at 67% of their nameplate value. VISUAL INSPECTION A visual inspection of each module allows us to correlate observations of electro-chemical weathering, wire failures,
Figure 7 Visible damage to the module surface in the region over a junction box. This effect is seen in several modules deployed in Tucson.
CONCLUSION We observe module degradation in three different ways: ac-power changes, I-V curve variances, and visible damage. The correlation analysis proposed here will be an important step in discovering the underlying causes and diagnosis of degradation. We expect this information will be of value for PV module manufacturers so that they may find ways to prevent various types degradation. This information is also significant for PV power plant designers and agencies that issue warranties for PV modules or system components. REFERENCES [1] A. Cronin, S. Pulver, D. Cormode, D. Jordan, S. Kurtz, R. Smith, “Measuring Degradation Rates Without Irradiance Data”, Thirty-fifth IEEE PVSC, March 2010. [2] http://www.tahara.co.jp/solar.html, last access date 06/02/2012 [3] E. Kopp, A. Brooks, V. Lonij, A. Cronin, “I-V Curves from Photovoltaic Modules Deployed in Tucson”, American Physical Society Four Corners Meeting, Nov. 2011. http://www.physics.arizona.edu/4cs2011/webresources/4C S2011Program.pdf, last access date 06/02/2012 [4] K. Paghasian, G. TamizhMani, J. Kuitche, M. Gupte Vemula, G. Sivasubramanian, “Photovoltaic Module Power Rating Per IEC 61853-1: A Study Under Natural Sunlight”, Arizona State University Photovoltaic Reliability Laboratory, March 2011.
2
3,4
3
Emily S. Kopp , Patricia L. Hidalgo-Gonzalez , Adria E. Brooks , Vincent P. Lonij , and Alex D. Cronin 1 University of Arizona College of Optical Sciences, Tucson, Arizona, USA 2 Pontificia Universidad Católica de Chile, Santiago, Chile 3 University of Arizona, Department of Physics, Tucson, Arizona, USA 4 Arizona Research Institute for Solar Energy, Tucson, Arizona, USA
ABSTRACT As part of our ongoing study of the degradation of PV systems, we have collected I-V curves from over 300 modules that have been deployed in grid-tied systems for up to 10 years in Arizona. We present families of I-V curves that exhibit different types of imperfection. One system (with Sharp modules) displays a large range of fill factors between the modules. In another system (a prototype CIGS system), the modules produce varying Isc and Voc values. The modules of another (Sanyo HIP) have a high degree of uniformity and a high fill factor. We present results correlating imperfections in I-V curve data to visual inspection findings and to changes in historical energy yields.
1,3
measurements for all of the modules in each system using a commercial I-V curve tracer made by Tahara model TSC-PD01 [2] [3]. Data from several clear-sky days is taken within one hour of noon when the global irradiance 2 is 990 W/m . I-V CURVE DATA FROM 300 MODULES Families of curves from several modules of the same type are presented in Figures 1 – 4. While the I-V curves in Figure 1 appear uniform and high in quality, the others show significant imperfections that we are correlating with degradation.
INTRODUCTION The TEP solar test yard contains over 600 grid-tied PV modules grouped into 22 PV systems, each with their own inverter. Most of the systems are 1 to 2 kW P in size and use a fixed-tilt of 32 degrees for the modules. Each system contains several modules of the same make and model, but often these modules degrade at different rates, as we show here. Originally designed as a test-bed for household-sized systems, the TEP solar test yard has been in operation since 2003 in Tucson, Arizona. We show here data from several modules that have been operating continuously for 5 to 9 years. The variety of I-V curve data shows several different types of imperfections that may result from degradation.
Figure 1 Eight I-V curves from Sanyo HIP-G751BA2 modules. These appear uniform and of high quality. These modules have a measured Pmax of 148 W under the field conditions and a nameplate value of 167 W at STC.
We previously studied degradation rates of PV systems at the TEP solar test yard by analyzing the historical ac daily power output from each system [1]. We found that comparing individual PV systems to the yard average is a sensitive way to measure degradation. In [1] we reported that several PV systems have degraded at 3%/year with a typical uncertainty of 0.5%/year. However, the daily final yield did not identify why some systems degrade, and from ac-power data alone, one cannot unambiguously distinguish between inverter degradation, changes in the surrounding albedo, or module degradation. The present study is aimed at determining the causes of module degradation. We temporarily disconnected each module from its gridtied systems and obtained individual I-V curve
Figure 2 Nine I-V curves from Sharp NEQ5E2U modules. The maximum power for these modules
ranges from 96 to 147 W due to a wide range of fill factors and have a nameplate value of 165 W at STC.
Figure 3 Sixteen I-V curves from prototype CIGS modules. The maximum power for these modules ranges from 15 to 32 W due to a wide range of Isc and Voc values. These modules have a nameplate value of 45 W at STC.
Figure 4 Ten I-V curves from amorphous silicon BP MST-50 modules. The maximum power for these modules ranges from 29 to 39 W and show a These significant range of Isc and Voc factors. modules have a nameplate value of 50 W at STC.
visible module browning, junction box failure, and cell damage with the I-V curves. The results of our correlation analysis will be catalogued along with their I-V curve and ac-power characteristics. Examples of weathering are shown in figures 5 – 6.
Figure 5 Failure of a surface electrode on the CIGS modules whose I-V curves are shown in Figure 3.
Figure 6 Electro-chemical weathering (note the asymmetry associated with polarity) of thin film amorphous silicon modules whose I-V curves are shown in figure 4.
In order to better compare the current module condition with the manufacturer’s rated sticker values we follow the procedure outlined in IEC 60904-10 to translate data taken 2 in the field to standard test conditions (STC: 1000 W/m , 25°C, AM1.5), and we are studying the uncertainty caused by this translation procedure. A previous study found this procedure to produce an error of 3% [4]. We find that some modules operated under noon-time 2 sunlight of 990 W/m on a clear day in January in Tucson perform at 88% of their nameplate value while the average of the 43 various modules presented here perform only at 67% of their nameplate value. VISUAL INSPECTION A visual inspection of each module allows us to correlate observations of electro-chemical weathering, wire failures,
Figure 7 Visible damage to the module surface in the region over a junction box. This effect is seen in several modules deployed in Tucson.
CONCLUSION We observe module degradation in three different ways: ac-power changes, I-V curve variances, and visible damage. The correlation analysis proposed here will be an important step in discovering the underlying causes and diagnosis of degradation. We expect this information will be of value for PV module manufacturers so that they may find ways to prevent various types degradation. This information is also significant for PV power plant designers and agencies that issue warranties for PV modules or system components. REFERENCES [1] A. Cronin, S. Pulver, D. Cormode, D. Jordan, S. Kurtz, R. Smith, “Measuring Degradation Rates Without Irradiance Data”, Thirty-fifth IEEE PVSC, March 2010. [2] http://www.tahara.co.jp/solar.html, last access date 06/02/2012 [3] E. Kopp, A. Brooks, V. Lonij, A. Cronin, “I-V Curves from Photovoltaic Modules Deployed in Tucson”, American Physical Society Four Corners Meeting, Nov. 2011. http://www.physics.arizona.edu/4cs2011/webresources/4C S2011Program.pdf, last access date 06/02/2012 [4] K. Paghasian, G. TamizhMani, J. Kuitche, M. Gupte Vemula, G. Sivasubramanian, “Photovoltaic Module Power Rating Per IEC 61853-1: A Study Under Natural Sunlight”, Arizona State University Photovoltaic Reliability Laboratory, March 2011.
