Full Text - ASU Digital Repository
Contents
1. 1X PART 1 INDOOR SOILING METHOD ssosoessseeeessesesssssseessssressssressssressssreesssseessssressserees I TAL Ne CID TION oes toc eps etos ete since isti bu Ert eerte LL LED M AL MIL UL EE 2 KIRB GCRBEOUTIGL a essit d DIU MUS 2 1 12 Statementor Be Problem oroscopo tete ve oe x ea e dev ene veces 2 TTE SODDIC CENE O e 3 I2 EEE ERAT URE REVIEW ereen e ai 4 1 2 1 Artificial Soil Formulation and Application Sandia s Approach 4 1 2 2 Reflectance Spectroscopy An Overview ccsscccccceeesseesstteeeeeees 5 LS MEPHODOLOGI Y e E MM MM ME 7 1 3 1 AZ Arizona Road DUSL eitis o ee eere sro e Ryo nope 7 1 3 2 Spray Gun Specifications and Adjustments sessss 7 fs Fo SoN EOfrulatiOD me np e tee eat tdt ue eaus 10 EAT COUPO pe reer eR dead ned tul en 10 1 3 5 Artificial Chamber Set Up ssessseeesssssssssereessssssssreeessssssssreesesssseo 12 1 3 6 Importance of Laser Guided Technique 13 15 7 5o01l Density Measurement acest ao oed aux da au de 14 1 3 8 Poly Si Coupon Good Indicator of Uniformity 15 1 3 9 Spectroradiometer Instrument Overview eeesssss 16 1 3 10 Quantum Efficiency Measurement System ssssss 18 V 1 3 11 Characterization Techniques eese 20 TOL TSESDETS SNPEODESCUSSIONSL distet tu d aca du unten art tu neu ctu t 24 I4 Son Umrtormitty2. 19 1 3 11 Characterization Techniques The test coupon was first cleaned with tap water followed by distilled water and finally with isopropyl alcohol EL imaging was done on a cleaned module to look for any localized defects as QE measurements are to be done on the module Figure 13 EL Image a Poly Si b Mono Si IV characterization was done on a cleaned module to observe whether or not the curve is smooth as it gives an idea about the health of the test module Followed by I V soil is sprayed on the module uniformly and various density soil is obtained by varying the concentration of acetonitrile solution The soiled module undergoes soiled I V characterization to understand the Isc loss for different soiling densities According to EL image two spots on the test module were chosen and soiled reflectance was carried out 20 Figure 14 Reflectance b Mono Si Measurements on a Poly Si After soiled reflectance a soiled QE measurement was performed using QEX12M Solar Module Quantum Efficiency Measurement System PV Measurements Since the polycrystalline module had 18 cells it was important to keep the cell of interest completely in the dark Hence a mask was cut exactly to the size of the cell and a small window was provided so the light source can reach the module as shown in Figure 14 Another existing feature is the shroud but as the usage of this might disturb the soiling layer it was not used Individual v
3. 1 4 4 Relation between Cleaned QE and Reflectance Considering Figure 22 the drop in wavelength below 1100 nm is due to the cell absorption as the absorption region of c Siis below 1100 nm From 1000 nm the reflectance increases as the QE curve starts decreasing The dip at 1700 nm 1s due to EVA Ethylene Vinyl acetate absorption If the property of EVA changes over time then the dip varies accordingly QUANTUM EFFICIENCY Mono Si Cleaned QE Mono Si Cleaned reflectance EVA absorption Absorption regi n of c Si FLECTANCE ON c Si CELL A o c a 2 u E m c g eB Q c o Q a ru eP ecc O 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 Wavelength nm Figure 24 Graph between Cleaned QE and Reflectance for Mono Si 1 4 5 Glass EV A Backsheet Reflectance The white area refers to the area that is sandwiched between Glass EV A White backsheet The backsheet reflectance decreases as the wavelength increases and most of the peaks 29 that are seen are mainly due to the backsheet properties Also any changes in the encapsulant layer backsheet can be determined from Figure 24 White area on cell corner Figure 25 Contact Probe on White Area for Reflectance Measurements see Figure 16 for the full size picture of the 1 cell coupon White spot Cleaned White spot Soiled White Backsheet w Q c o S e ET
4. Sandia s Approach Burton et al from Sandia National Laboratories have reported a means to deposit and characterize artificial soil coatings composed of NIST traceable dust with known chemical and physical properties 3 The process 1s as follows Arizona Road Dust ISO 12103 1 A2 Fine Test Dust nominal 0 80 micron size Powder Technology Inc Burnsville MN USA was mixed with a soot mixture composed of 83 3 w w carbon black Vulcan XC 723 Cabot Corp Boston MA USA 8 3 diesel particulate matter NIST Catalog No 2975 4 2 unused 10W30 motor oil 4 2 a pinene Catalog No AC13127 2500 Acros Organics Geel Belgium in a glass jar and tumbled without milling media in a rubber ball mill drum at 150 r min for 48 to 72 h The composition was varied to include 3 wt 10 wt and 25 wt soot mixture and samples were prepared on a 100 g total solid basis The composition of the varying soot mixture did not represent any specific location but on an average represented the soot content present in a few locations For application onto the samples this grime mixture was combined with Acetonitrile ACN HPLC grade Sigma Aldrich St Louis MO USA in a ratio of 3 3 g to 275 ml and was sprayed on a commercial glass coupon 7 62 cm X 7 62 cm The grime mixture was sprayed using a HVLP High velocity Low Pressure gun held approximately 30 cm from the coupon surface The samples were sprayed from right to left for a duration of 1 3
5. d NA amp NO d ESO NUES e we d gt S S eS R Ss qe Sd S oy SP E 5 e e y g i e e x o e Frequency 9 6 50 00 45 69 33 05 24 71 8 62 23 26 Percent 27 0 24 7 17 8 13 3 4 7 12 6 Cum 9 6 27 0 51 6 69 5 82 8 87 4 100 0 Figure 44 Pareto Chart for Model JVA 64 2 5 CONCLUSION Model J is an 18 year old power plant in a cold dry climatic condition having string level degradation of 0 73 year and module level degradation of 0 55 year The global RPN for the power plant is divided into safety and degradation RPN For model J string level global RPN was determined to be 704 and module level global RPN was calculated to be 470 For cold dry climatic conditions the degradation rate is about 0 6 per year framed to 0 73 per year frameless Encapsulant delamination was the dominant failure degradation mode for frameless modules while interconnect discoloration was the dominant degradation mode for framed modules However both these modes are the result of extent of moisture ingress 65 REFERENCES 1 Travis Sarver AliAl Qaraghuli Lawrence L Kazmerski comprehensive review of the impact of dust on the use of solar energy History investigations results literature and mitigation approaches Renewable and Sustainable Energy Reviews March 15 2013 2 J Zorrilla Casanova M Piliougine J Carretero P Bernaola P Carpena L Mora L pez M Sidrach de Cardona Analysis of dust losses in
6. seconds and to obtain high soiling density multiple coatings were sprayed The glass 4 coupons before and after soiling were weighed with a Mettler Toledo Columbus OH USA XP205 balance with 0 00001 g resolution and characterization tests were also performed Variations in the grime mixture was produced by incorporating major optical components like iron oxide and in house synthesized g thite as primary spectral components After soil formulation and application characterization tests like current voltage and QE measurements were carried out on soiled glass samples and the results are discussed 1 2 2 Reflectance Spectroscopy An Overview Reflectometery or reflectance spectroscopy is used in a variety of metrological applications for determination of chemical composition material identification and measurement of optical properties of materials 4 Reflectance on PV modules can be reduced by the introduction of anti reflective AR coating or texturing of the glass surface In the case of cleaned modules the reflectance between 350 2500 nm on a module surface can measure surface roughness surface cleanliness contamination texturing AR coating properties and metallization parameters For soiled modules as soil is deposited on the surface reflectance curve can be used to determine the physical composition chemical composition and properties of the soil The prediction of various components between 350 2500 nm i
7. CMC CK uto te ee Letti Aol kb et tis 24 1 4 2 Process Sample Size Independent Repeatability 25 1 4 3 Mono Si Better Technology to Characterize QE and Reflectance Sache MN 27 1 4 4 Relation between Cleaned QE and Reflectance 29 1 4 5 Glass EV A Backsheet Reflectance 0 0 0 0 ceeeseeeecccceeeeeeseeeeeees 29 1 4 7 Particle Size Effect on Reflectance cccccsssscccccceessesssseeeeeees 3 1 4 8 Reflectance Measurements A Measure of Soiling Density 33 iy CONCLUSION CUm 35 PEANT T 38 DA NOES TION 13 3 09 Du aus uae asada ude ete uu hareiu uia E E 39 2 T DOCE SPOUTIO UonEoA RA OREL AUAM ORO AAS AAS A 39 2 1 2 Statement of the Problem eene 40 ZU FOD INO edi ed Ede dud qud MN I MN MM M MM 40 ZETTAI WIRE RE VIE eo o RR DE RET aie 42 2 2 1 Field Failure and Degradation Modes sssssss 42 2 2 2 Rehability Fa lure S nnna ERE ERRER 42 2 2 SULA My ATE a S 43 vi 22A NEAT INI A enanada eran RRR ERAEN hinin nihin 43 yrs mile TIO B L618 O I dier NNNM NI NI NN NN UNE 45 2 5 5ystem DeOSCETDLUOD uo e eso E E NEUE MEE 45 2 15 29 16 UU va eee aurea Onde du od t dud end teure 45 2 3 3 Determination Of SeVerlty unio ieetetites ires cista aae eed 0S ER EReoDS 47 2 5 Determipdlom OL D ete OD nenne 48 2 2 9 Det rminaton of OCCULENCE eee entere tante hee dine ee tenus
8. Headers Data Organization Print Header Information Columns Print ONLY Header Information Rows Axis Print Axis Field Seperator 9 Wavelength T O Channel Dutput to a Single File Column Title Print Column Title Print FileName s at Top of Column C Print Collect Time s at Top of Column Print Description Note Figure B10 Exporting of asd files NOTE For any further information about the Spectroradiometer click on the below link to access the user manual http support asdi com Document Viewer aspx id 140 78
9. JVA c Encapsulant Browing and Interconnect Discoloration in Model JVA d Encapsulant Delamination in Model J eeesseeeeeeess 63 43 Pareto C hart for Model J eicit eR era deceencdas era eoe Uno te eo do nuege Ra De vEso CHR Ete SiS 64 44 Pareto Chart for Model JVA c ccccccccecccceccccscsccccecscceccecsccccececsececsececescscensecesscs 64 Xl PART 1 INDOOR SOILING METHOD 1 1 INTRODUCTION 1 1 1 Background Soiling is a major source of energy loss on Photovoltaic PV modules and it becomes difficult when considering the quantification of dust influence Particle size shape composition moisture content deposition pattern and accumulation rate vary from location to location due to the geography climate and urbanization of the region 1 In the context of PV soiling loss refers to losses primarily due to dust deposition Previous studies show that losses due to accumulated dirt on modules can reach as high as 15 for a period without rain 2 Apart from these losses there are increasing number of claims of dust resistant coatings abrasion resistant coatings during cleaning and new dust removal techniques Currently there is no standardized way of verifying the validity of such claims As a first step towards validating such claims a standardized artificial soiling method using a laminated module construction of glass EV A cell EV A backsheet is developed in this report Also the
10. Mohan Shrestha Jaya Krishna Mallineni Karan Rao Yedidi Brett Knisely Sai Tatapudi Joseph Kuitche and GovindaSamy TamizhMani Determination of Dominant Failure Modes Using FMECA on the Field Deployed c Si Modules Under Hot Dry Desert Climate IEEE Journal of Photovoltaics vol 5 no 1 January 2015 14 G TamizhMani and J Kuitche Accelerated Lifetime Testing of Photovoltaic Modules A report of Solar America Board for Codes and Standards solarabcs org 2013 67 APPENDIX A CHEMICAL COMPOSITION AND PARTICLE SIZE OF ARIZONA ROAD DUST 68 E PTI POWDER TECHNOLOGY INC PRODUCT LIST PP2G4 TSO 12103 1 ARIZONA TEST DUST CONTAMINANTS A2 FINE AND A4 COARSE GRADES TYPICAL CHEMICAL ANALYSIS Loss on Ignition 2 5 I50 TEST DUST PARTICLE SIZE DISTRIBUTIONS BY VOLUME o ISO 12103 1 A Fine ISO 12103 1 A4 Coarse sa ex 110 220 320 360 49 80 30 395 88 L 8 oso o mss l 85100 wo m 69 APPENDIX B STANDARD OPERATING PROCEDURE FOR REFLECTANCE 70 Applications This procedure shall be used in all indoor and outdoor Reflectance and Transmittance measurements using HandHeld FieldSpec 4 Wide Res spectroradiometer Procedure Reflectance 1 In the rear portion of the spectroradiometer unit connect the power supply to the input 12 VDC port Also connect the Ethernet cable to the appropriate port with the other end connected to the laptop Ensure that
11. a 1500 Wavelength nm Figure 26 Reflectance Plot for White Area and White Backsheet 30 1 4 6 Reflectance Soiled Cleaned Plots for Various Soil Densities The reflectance plots for various soil densities ranging from 0 18 g m to 1 8 g m is shown in Figure 25 From the graph below it is evident that the reflectance increases as the soiling density increases For all soil densities maximum reflectance was found to be between 400 1100 nm which corresponds to the absorption region of c Si ex oO v c S Q ex UO 2 o 2 qp Q c o O a rum Q 8 6 5 4 3 2 MI alla ab AM o Uil l l Dot i aa lk errs aldad FF GA se S S S Ss Q S S S S S S S S S S Wavelength nm Figure 27 Reflectance Soiled Cleaned Plots for Various Soil Densities 1 4 7 Particle Size Effect on Reflectance When comparing the outdoor and indoor reflectance the outdoor reflectance curve over the absorption region seems to be uniform flat whereas for indoor it was observed to be non uniform Bowers et al showed that reflectance increased with a decrease in particle size and this effect was more prominent for a particle size less than 400 microns which is 31 possibly true in the case of indoor soiling 12 As scattering increases the wavelength decreases The scattering effect due to fine particle size could be the reason for non uniformity patterns observed in indoor refl
12. a direct measure of soil density on the modules Part 2 determines the most dominant failure modes of field aged PV modules using experimental data obtained in the field and statistical analysis FMECA Failure Mode Effect and Criticality Analysis The failure and degradation modes of about 744 poly Si glass polymer frameless modules fielded for 18 years under the cold dry climate of New York was evaluated Defect chart degradation rates both string and module levels and safety map were generated using the field measured data A statistical reliability tool i FMECA that uses Risk Priority Number RPN is used to determine the dominant failure or degradation modes in the strings and modules by means of ranking and prioritizing the modes This study on PV power plants considers all the failure and degradation modes from both safety and performance perspectives The indoor and outdoor soiling studies were jointly performed by two Masters Students Sravanthi Boppana and Vidyashree Rajasekar This thesis presents the indoor soiling study whereas the other thesis presents the outdoor soiling study Similarly the statistical risk analyses of two power plants model J and model JVA were jointly performed by these two Masters students Both power plants are located at the same cold dry climate but one power plant carries framed modules and the other carries frameless modules This thesis presents the results obtained on the frameless module
13. characterization tests that would give all round information about the soiling losses are identified 1 1 2 Statement of the Problem Natural soiling in PV is time consuming and location specific The results obtained in natural soiling cannot be generalized due to varying physical and chemical properties of soils across the globe Hence there comes the necessity to develop to speed up the soil 2 depositing pattern artificially Accelerated and artificial means of soil deposition can help reduce the time taken to estimate the losses due to soiling and help authenticate such claims of dust resistant properties and dust removal techniques Pre characterized soil from different regions can be deposited using this laboratory oriented approach and the losses can be quantified 1 1 3 Objectives One of the main objective is to determine that reflectance and QE measurements can be used as a direct measure to calculate soil density By measuring reflectance on the soiled modules reflectance loss is calculated From reflectance loss soil density g m7 is calculated and the corresponding Isc drop is determined Measure reflectance Measure QE Calculate reflectance loss Calculate QE loss Calculate soil density g m Calculate soil density g m Calculate energy loss I drop Calculate energy loss I drop Figure Goal of the experiment 1 2 LITERATURE REVIEW 1 2 1 Artificial Soil Formulation and Application
14. module surface and the energy loss Isc loss can be calculated using the reflectance loss Recommendations e Instead of single cell coupons modules can be soiled artificially by increasing the distance between the chamber and spray gun e Artificial soil formulation ad application as mentioned above can be extended for various soils across the globe and characteristics can be determined e Similar to outdoor AOI Angle of Incidence indoor AOI measurements can be performed for artificially soiled coupons and results can be compared e Results obtained by measuring reflectance on PID cells can give a brief description about AR coating 34 1 5 CONCLUSION The study presented largely follows the procedure developed by Burton et al and major conclusions resulting from this study are as follows e Gravity assisted and laser guided approach of spraying soil onto coupons helps in improving the soil uniformity pattern and total area of the test coupon for soil application can be further increased by increasing the distance between the module and spray gun e Mini modules can be used to check uniformity by measuring I V curves wherein for characterization tests single cell coupons are more favorable e Properties of encapsulant EVA over time can be determined by carrying out reflectance measurements e Particle size plays an important role in reflectance measurements The smaller the particle size the higher the reflectance A
15. photovoltaic modules World Renewable energy Congress Sweden 2011 3 P D Burton and B H King Application and Characterization of an Artificial Grime for Photovoltaic Soiling Studies IEEE Journal of Photovoltaics vol 4 No 1 pp 299 303 Jan 2011 4 B Sopori Y Zhang R Faison and J Madjdpour Principles and Applications of Reflectometery in PV Manufacturing National Renewable Energy Laboratory October 2001 5 D Summers M Lewis B Ostendorf D Chittleborough Visible near infrared reflectance spectroscopy as a predictive indicator of soil properties May 5 2009 6 http manuals harborfreight com manuals 94000 94999 94572 pdf 7 http www asdi com products fieldspec spectroradiometers fieldspec 4 hi res 5 http support asdi com Document Viewer aspx id 140 9 B Knisley Angle of Incidence and Non Intrusive Cell Quantum Efficiency Measurements of Commercial Photovoltaic Modules MS thesis Arizona State University December 2013 66 10 http www pvmeasurements com Quantum Efficiency Measurements qex 12m solar module quantum efficiency measurement system html 11 Sravanthi Boppana Vidyashree Rajashekar Govindasamy Tamizhmani Working towards the Development of a Standardized Artificial Soiling Method Accepted IEEE PVSC New Orleans 2015 12 S A Bowers R J Hanks Reflection of radiant energy from soils Soil Science vol 100 130 138 1965 13 Sanjay
16. were observed during the characterization tests like reflectance and quantum efficiency on poly Si was noted In order to overcome these above stated uncertainties and to get accurate measurement results the same process was followed on monocrystalline silicon coupons 15 1 3 9 Spectroradiometer Instrument Overview A FieldSpec 4 UV Vis NIR reflectance spectroradiometer from Analytical Spectral Devices ASD Colorado was used for all the reflectance measurements In addition to reflectance the instrument can measure transmittance and irradiance as it is a special kind of spectrometer that can measure radiant energy It is a compact field portable and precision instrument which has a spectral range of 350 2500 nm and a fast data collection time of 0 2 seconds per spectrum The instrument has a fixed optic cable that helps to calibrate to units of radiant energy irradiance and radiance and it is operated using a computer that has RS software installed This instrument is extensively used by the agricultural industry for analyzing soil samples Figure 10 High Resolution Spectroradiometer 7 Figure 11 Contact Probe 7 16 Table 4 Technical Specifications of Spectroradiometer 7 Spectral Range 350 2500 nm Spectral Resolution 3 nm 700 nm 8 nm 1400 2100 nm Sampling Interval 1 4nm 350 1050 nm 2nm 1000 2500 nm Scanning Time 100 milliseconds Stray light specification VNIR 0 02 SWIR 1 amp 2 0 01 The fron
17. 18 Years Cold Dry frameless 300 2 0 Module RPN m Frequency Global RPN 470 1 8 Degradation Rate Yr a 1 6 Safety RPN 280 Degradation RPN 190 Frequency or RPN Degradation Rate Y yes M M sco ot av co cov o N oV ce Ao ov axo N gov Figure 41 Global Degradation and Safety RPN for Module level all defects included The global string level RPN for this site 1s calculated to be 704 and global module level RPN is calculated to be 470 Statistically both string and module level global RPN should 60 nearly match but for this site 1t does not seem to be true Only 46 modules were considered for module level global RPN but to have a CL confidence level of 95 and CI confidence interval of 5 254 modules should have been considered Insufficient data for the calculation of module level global RPN could be one of the reasons for this discrepancy 2 4 5 Comparison Plots between Model J and Model JVA As indicated in the abstract the statistical risk analysis of two power plants was jointly performed by two Masters students Both power plants are located at the same cold dry climate but one power plant carries framed modules and the other carries frameless modules as shown in Figure 41 This thesis presented the results on the framed modules Comparing these two sites would help understand the failure modes and mechanisms for this climatic zone as both the p
18. 49 ZARESULTS AND DISCUSSIONS euam ar EE pe EE e eM EU MEME RE DM 5I 2 4 1 Degradation Analysis of Model J String and Module level 5I 2 4 2 Layout of Degradation and Safety Failures 53 2 4 3 String level RPN Calculation for Model J 56 2 4 4 Module level RPN Calculation for Model J 58 2 4 5 Comparison Plots between Model J and Model JVA 61 P Sx Celio ar tT 65 REFERENCES M 66 APPENDIX CHEMICAL COMPOSITION AND PARTICLE SIZE OF ARIZONA ROAD DUST 68 STANDARD OPERATING PROCEDURE FOR REFLECTANCE 70 vii LIST OF TABLES Table Page 1 Prediction of Various Components in Reflectance Spectra 5 6 2 Sprds Cr nSpecrcations G ient eoe uro ERE ERN ERE ERU oe uni AR ap Sw ERES uiuat 8 3 Specifications for Poly and Mono Si Coupons essen 11 4 Technical Specifications of Spectroradiometer 7 cccccsssesesecceeeeeeeeeeeeeeeeeeeeeeas 17 DeLee Loss for 44cm PO ly Si erren eite oett Decided mete tatc sided 26 6 I Loss for 233 cm MON0 Si cccccessccecsceesccesceesceeseceesecessceetecesceesecesseetseceseetseceaes 26 7 Various Tools to Identify the Failure Modes eese 47 8 Severity S and Detection D Ranking Table ceeccceccccceeeeeeeeseeeeeeeeeeeaeeeeees 48 9 De
19. ENES HEUTE E DOSE E CRANE NN Figure 37 Safety Map for Model J When extrapolating the measured module degradation and including the safety failures 86 3 of the modules are safe and are meeting the manufacturer s warranty with only 5 5 of modules being safety failures and another 8 7 exceeding the manufacturer s typical warranty of 1 year degradation rate as shown in Figure 38 55 Safety Failures gt 1 dr yr 5 5 a 8 7 lt 1 dr yr 86 3 Figure 38 Safety Failures Reliability Failures and Durability Loss for Entire Power Plant Model J 2 4 3 String level RPN Calculation for Model J Of all 68 strings visual inspection and I V s are taken on 62 strings excluding the 6 strings in the burnt combiner box and all the strings present in the plant are evaluated for I V and VI measurements with Confidence Level CL 100 and Confidence Interval CI 0 The sum of each defect that is present in the above stated 62 strings are considered to be the count From the count frequency of occurrence of failure mode is computed using the formula discussed in the previous section A maximum value of 9 is given to four defects as almost all modules along the string had these defects From the degradation data Ra severity is calculated A severity of 10 is given for both broken module and bypass diode failure as a broken module possesses an electrical hazard 56 Failed diodes may lead to backsheet burning so
20. Indoor Soiling Method and Outdoor Statistical Risk Analysis of Photovoltaic Power Plants by Vidyashree Rajasekar A Thesis Presented in Partial Fulfillment of the Requirements for the Degree Master of Science in Technology Approved April 2015 by the Graduate Supervisory Committee Govindasamy Tamizhmani Chair Devarajan Srinivasan Bradley Rogers ARIZONA STATE UNIVERSITY May 2015 ABSTRACT This is a two part thesis Part 1 presents an approach for working towards the development of a standardized artificial soiling method for laminated photovoltaic PV cells or mini modules Construction of an artificial chamber to maintain controlled environmental conditions and components chemicals used in artificial soil formulation is briefly explained Both poly Si mini modules and a single cell mono Si coupons were soiled and characterization tests such as I V reflectance and quantum efficiency QE were carried out on both soiled and cleaned coupons From the results obtained poly Si mini modules proved to be a good measure of soil uniformity as any non uniformity present would not result in a smooth curve during I V measurements The challenges faced while executing reflectance and QE characterization tests on poly Si due to smaller size cells was eliminated on the mono Si coupons with large cells to obtain highly repeatable measurements This study indicates that the reflectance measurements between 600 700 nm wavelengths can be used as
21. age a Poly Si b Mono Si eeeeeeese 20 14 Reflectance Measurements on a Poly Si D MORBOS P arrene 21 15 a Poly Si Cell Covered with Mask An Opening for Light Source b Mono Si COME Wat I SOU ECE Se sene qd Mos eee VER SCUR A Ub ese VRAC ORA eq MERCURIO er a 22 16 Soiled Spots Cleaned for Reflectance and QE sesssssssssserssssssssseeerssssssssserrssssssssees 22 17 Flow Chart Indicating the Characterization Process sese 23 18 Coupon Containing Microscopic Slides cccceccccccessssssseeeceeceeeaeeseeeccceeeeeaaeeeseeeeeees 24 19 oHdedEocatton V s SOT DONS Iy eer atero Deep E o xx aped etude 25 20 Soil Density g m vs Isc Loss 96 26 21 Circle Indicating Contact Probe Diameter gt Cell Area for Poly S1 Coupon 27 22 QE Curve for Poly Si Coupon 1 55 9 M7 o cccccccccccccecescescsscsccsccsessesscessssteessseeesess 28 23 QE Curve for Mono Si Coupon 1 57 9 M ccccccccccecescescssssccscesecsessessessesssesseesees 28 24 Graph between Cleaned QE and Reflectance for Mono Si eeeeesssssss 20 25 Contact Probe on White Area for Reflectance Measurements see Figure 16 for the full SIZE piciure or the I cell COUPON oes deti diri tete uri tdt eem deleted resin ntu edes 30 26 Reflectance Plot for White Area and White Backsheet eeeeeeeeeeessssss 30 27 Reflectance Soiled Cleaned Plots for Various S
22. all four soil densities was found to be 0 02 which is a very good indication that the process is uniform Bottom Figure 18 Coupon Containing Microscopic Slides 24 wa S 90 gt _ Wi c Q Oo Vn bottom left Slide location Figure 19 Slide Location Vs Soil Density 1 4 2 Process Sample Size Independent Repeatability Poly Si modules were coated with different soil densities ranging from 0 6 to 1 55 g m 11 and single cell modules were also coated with densities ranging from 0 18 to 1 8 g m as shown in Tables 5 amp 6 In Table 6 soiling densities 1 57 g m and 1 8 g m almost had the same Isc loss meaning the process is repeatable The uniformity pattern was observed in both Poly Si and mono Si modules of sample size 144 cm and 225 cm respectively This indicates that the soil spraying process holds good for various size coupons 25 Table 5 Isc Loss for 144 cm Table 6 Isc Loss for 233 cm Poly S1 Mono S1 MiniModule Soil Density g m Figure 18 shows a linear dependence in Isc loss with respect to soil density The Isc loss is the difference between cleaned Isc and soiled Isc divided by cleaned Isc This plot indicates that the Isc values can be used to measure the soil density or soil loss for the experimental variables used in this work Isc loss 96 Mono Si Poly Si Soil Density g m2 Figure 20 Soil Density g m vs Isc Loss 96 26 1 4 3 Mo
23. c condition can be manufactured 2 1 3 Objectives The main goal of RPN is to determine the health of the power plant by prioritizing the failure modes and allocating rank accordingly The eventual goal of RPN study would be to classify power plants into classes based on their Safety RPN and Degradation RPN as shown in Figure 31 The class boundaries would be climate specific 40 High Class B Class C Plant Plant SafetyRPN Class A Class B Plant Plant Low Low DegradationRPN High Figure 32 Grading PV Power Plant Conceptual Approach The other objectives of this study are l To determine the String and module level degradation of power year by using the field measured data i1 To generate a safety map for the entire power plant that includes all the modules that had safety issues iil To determine the safety and reliability failures involved iv To determine the RPN for the failure modes present both string and module level and to rank them accordingly 41 2 2 LITERATURE REVIEW 2 2 1 Field Failure and Degradation Modes Field failure degradation mode and mechanisms of PV modules depend on the design packaging construction in which the modules operate 14 These failures can further be classified as reliability and durability losses A few of the field failures are broken interconnect broken cells corrosion delamination bypass diode failure and hot spots The above stated failure modes are ca
24. ccurrence denotes the probability of occurrence of a failure mode for a predetermined or stated time period CNF Cumulative Number of module Failures per thousand per year is to be computed 49 It is given by CNF 1000 Xsystem 96 defects 10 Xsystem operating time where of defects number of particular defect total number of modules 100 Operating time duration for which modules are employed Table 10 Occurrence O Ranking Table Failure Mode Frequency WM OOOO lt 0 01 module per thousand per unlikely year Low Relatively few 0 1 module per thousand per year failures 0 5 module per thousand per year Pm j Moderate Occasional 1 module per thousand per year failure 2 module per thousand per year 5 module per thousand per year High Repeated failures 20 module per thousand per year Very high Failure 15 50 50 module per thousand per year per thousand per year almost inevitable Em m module per thousand per So 50 2 4 RESULTS AND DISCUSSIONS 2 4 1 Degradation Analysis of Model J String and Module level The histogram shown in Figure 32 provides the mean and median string level degradation rate of power for Model J The histogram fits a near normal distribution The average degradation for string level power at this site is found to be 0 73 year Out of all 62 strings 57 strings 92 are meeting the maximum degradation rate of 1 0 year typically provided by the module ma
25. e mode is determined just by looking at the monitoring system If the visual inspection technique is sufficient to determine the failure mode then a ranking of 2 3 is given A ranking above 3 is given depending on the tools ex IR 48 diode line checker IV tracer used and when the failure is impossible to determine in the field rank 10 1s assumed Table 9 Detection D Ranking Table Criteria Likelihood Monitoring System itself will detect the failure mode with Almost warning 100 certain Very high probability most likely of detection through visual Very high inspection Very high probability most likely of detection using non Low conventional handheld tool e g diode line checker 50 50 probability less likely of detection using non Very low conventional handheld tool e g diode line checker Very high probability most likely of detection using Extremely performance measurement equipment e g IV tracer Low 50 50 probability less likely of detection using performance Remote measurement equipment e g IV tracer Detection impossible in the field Absolutely uncertain ig Very high probability most likely of detection using Moderately conventional handheld tool e g IR Megger high 50 50 probability less likely of detection through visual High inspection 5 50 50 probability less likely of detection using conventional Moderate handheld tool e g IR Megger 2 3 5 Determination of Occurrence O
26. ectance measurements Uniform over cell absorption region l NN rf Owl W A 1500 d X 7 S Q X UO 9 a n Q c q S 9 T a Wavelength nm 0 170 g m2 0 263g m2 0 447 g m2 0 7g m2 0 966 g m2 Figure 28 Outdoor Reflectance Plots for Various Soil Density 32 Non uniform over cell absorption region _ eR o Qo em i 9 eR oO S o A sB Q c c Q S9 rum ecc Figure 29 Indoor Reflectance Plots for Various Soil Density 1 4 8 Reflectance Measurements A Measure of Soiling Density Delta soiled cleaned for reflectance measurements is calculated for each soil density The slope of the data is calculated and the R value is determined for all the wavelengths ranging from 400 to 2500 nm The wavelength between 600 700 nm showed an acceptable R value of 0 918 Similarly for outdoor soiling the wavelength between 600 700 nm was determined to be a good fit irrespective of the technology for measuring density The equation is as follows Soil density 25 35 x average reflectance loss 0 36 33 The correlation plot between QE and soil density Figure 29 reflectance and soil density Figure 30 and transmittance Isc loss and soil density Figure 18 is linear This further indicates that both reflectance QE can be confidently used to determine soil density without collecting soil from the
27. epared to have a composition of AZ road dust mixed with acetronitrile ACN in a ratio of 3 3 g to 275 ml While spraying soil it was found that formulated soil from the gun did not reach the test module and resulted in a thin layer of soiling even after many rounds of application Hence the composition was changed to 15 g of AZ road dust for every 1000 ml of acetonitrile By varying the composition of ACN different soil densities were obtained and a density of above 1 8 g m led to clogging in the spray gun 1 3 4 Test Coupons Polycrystalline and monocrystalline silicon coupons with no AR coating were used in this study Polycrystalline silicon mini modules of construction Glass EV A Cell EV A Backsheet having an aperture area of 144 cm was used Each mini module was comprised of 18 polycrystalline silicon cells that are series connected The dimensions of each cell are 5 7 cm x 1 cm and are rated to produce 1 48 W 10 A single cell monocrystalline silicon coupon also had the same construction as Poly Si The area of the cell was found to be 225 cm 15 cmx15 cm As an attempt to characterize the optical properties of soil using spectroradiometer this coupon was laminated such that there was extra space close to the cell Table 3 Specifications for Poly and Mono Si Coupons Variables Poly Si Mono Si Multi cell coupon Single cell coupon Glass EV A Cell EV A Backsheet Glass EV A Cell EV A Backsheet Number of cells 12 cell
28. he degradation RPN contributes to 344 and the safety RPN contributes to 360 Broken glass broken cell interconnect and bypass diode failures belonging to safety should be immediately considered and the modules should be replaced depending on the safety threat they possess String Grading Power Plants for Degradation Safety and Global Risks 18 Years Cold Dry frameless 300 2 0 E String RPN m Frequency Global RPN 702 1 8 B Degradation rate Y 1 6 Safety RPN 360 1 4 1 2 Degradation RPN 342 1 0 0 8 Frequency 96 or RPN Degradation Rate Y noo 0 6 0 4 1 0 2 1 0 0 5 jes T P owe co p s i ae d Figure 39 Global Degradation and Safety RPN Chart for String level Analysis 2 4 4 Module level RPN Calculation for Model J Due to plant layout and time constraint only top modules in the strings were accessible and individual I V s were taken only for 46 modules Modules that had both I V and VI data is considered for the module level RPN analysis The count and detection ranking is 58 the same as string level RPN but severity ranking that is based on the degradation rate 1s going to vary Few defects backsheet scratches corrosion like broken module that are not observed in the calculation of module level RPN are observed in the string level calculation The defects involved in the calculation of module level RPN are as follows Table 12 Mod
29. ing consists of 12 c Si frameless modules in series This results in the rated power of each string being 1 44 kW The operational size of the power plant is 97 92 kWac The inverter is located inside another building within the facility and is operational 2 3 2 Site Survey Visual inspection VI was carried out for all 744 modules at this power plant The VI was performed by using the developed checklist from NREL National Renewable Energy Laboratory Defects that were found on every module were noted in the checklist and a safety map that included all the safety failed modules for the entire site was generated Every diode was checked in all 744 modules to test for failed diodes that are resulting in open circuited strings This was conducted using the diode checker Infrared imaging IR was also done using an IR camera for every module at the site There were no modules that showed any hotspot issues since there was no cell operating at or above 20 C of the 45 average module temperature For the entire power plant there were no shading issues that could be found when the SunEye was used Both string and module level I Vs were measured using the Daystar curve tracer and then corrected to STC for this site The I Vs for all strings excluding the 6 strings in the burnt combiner box for a total of 62 strings was measured The total number of modules in this power plant was 744 operational modules Due to the layout of the power p
30. it 1s considered as a safety issue and a maximum ranking for severity is given Depending on how the defects were detected ranking is given accordingly As most of the failure modes are detected visually a ranking of 2 is given Bypass diode failure alone 1s determined using diode line checker and a conventional handheld tool is used and a ranking of 4 is given Table 11 String level Severity and Occurrence Table for Model J ____Defects_ Frequency Degradation rate Severty s Occurence O a a 2 c7 1 Interconnect Discoloration 336 80 2 5 Cell Discoloration 3844 04 3 9 OverCellEncapsulantDelamination 8468 073 5 9 Near Edge Encapsulant Delamination 7567 059 4 9 Comoson ike 3mmspot 484 013 1 6 BrokenModule 108 03 10 4 Brokencell Interconnect 296 06 8 5 Figure 40 indicates that RPN is calculated in 3 ways l Global RPN gives entire plant RPN includes both degradation and safety issues 2 Degradation RPN considers defects that contribute only to degradation issues safety 1ssues are neglected during this RPN calculation 3 Safety RPN considers defects that possess only safety issues degradation issues are neglected during the safety RPN calculation 57 The global string level RPN for this site is found to be 704 where t
31. lant only top row modules were accessible when taking I V curves so a total of 46 individual module I Vs were taken As the visual inspection data is obtained the RPN is calculated to determine the failure modes The tools and the corresponding accelerated test for the modes are as follows 46 Table 7 Various Tools to Identify the Failure Modes Field failures Causes Mechanisms Characterization Test Broken Thermal expansion and Interconnects contraction repeated Visual inspection mechanical stress Broken cells Mechanical stresses Electroluminescence EL Corrosion Moisture induced corrosion Visual inspection of cell metallization Delamination Adhesive bond sensitive to Visual inspection UV or contamination the material Encapsulant Heat and UV Visual inspection E NEM solder bond Stresses induced by thermal Visual inspection ues aegre Hotspots Operating current l Infra Red scan IR Bypass diode OC diode inspections with failures mm device Backsheet Visual inspection 2 3 3 Determination of Severity Severity depends on two factors Degradation rate Ra and how the failure modes affect the appearance cosmetic and performance of the modules and also safety concerns involved with the failure modes A rank of 1 7 depends on performance factor and a rank of 8 10 depends on safety factor A rank of 8 10 forms the highest severity part as they are involved with safety problems created by the failu
32. lants had modules from the same manufacturer Table 13 Comparison between Model J and Model JVA Place State of New York close to State of New York higher water body temperature influence compared to Model J 6l Dominant Failure Over cell Encapsulant Interconnect Discoloration Mode degradation Delamination and Encapsulant Browning Degradation Rate String level 0 73 year 0 6 year Model J 18 year old plant had a string level degradation of 0 73 year whereas model JVA 19 year old plant had a string level degradation of 0 6 year From Figures 45 and 46 it is inferred that the dominant failure mode for model J is encapsulant delamination browning and interconnect discoloration for model JVA These failure modes are mainly due to moisture ingress For model J intrusion of moisture resulted in backsheet bubbles that further lead to encapsulant delamination Encapsulant delamination resulted in optical decoupling that not only had an effect on Isc drop but also on Voc drop resulting in triggering of bypass diodes Due to high Voc drop the number of diodes failed in model J is higher compared to model JVA Table 15 This comparative study between framed and frameless modules of the same model in cold dry climate clearly indicates that the frameless modules are highly susceptible to backsheet delamination leading to severe encapsulant delamination The severe encapsulant delamination leads to high Isc drop high Voc dro
33. lso scattering effect 1s dominant for smaller particles e Reflectance QE loss can be used as a direct measure of soil density The correlation plot between soil density g m reflectance loss and QE loss 96 varying linearly 1s shown in Figures 29 amp 30 35 ue un 9 Qo Q o o i Qo cc 1 Soil density g m7 Figure 30 Correlation plot between reflectance and soil density g m7 C m N U A A A A N O Soil density g m Figure 31 Correlation plot between QE and soil density g m 36 The indoor and outdoor soiling studies were jointly performed by two Masters Students Sravanthi Boppana and Vidyashree Rajasekar This thesis has presented the results obtained from the indoor soiling study whereas the other thesis presents the outdoor soiling study Key findings of the outdoor soiling study are presented below e If there is an identical soil density on PV modules then the relative optical response at different AOI 1 e f AOD will be nearly identical irrespective of the PV technology type e The power or current loss between clean and soiled modules would be much higher at a higher AOI than at a lower AOI leading to excessive energy production loss of soiled modules on cloudy days early morning hours and late afternoon hours e Based on the results obtained in this study it can be stated that the critical angle shifts from 57 for the clean air glass interface to 40 for the naturall
34. nce optimization is done then click the WR White Reference which is right next to the OPT After collecting the WR you get an image as below A straight line appears at reflectance 1 indicating that the spectroradiometer unit has reflected all the light that it has encountered DATA COLLECTION Then the WR cap is removed and the contact probe is placed perpendicular to the sample for which the reflectance measurements are to be made For saving the measurements go to Control gt Spectrum save gt Dialog box appears gt Give the file name and check for the dates gt Hit Begin Save The measurements start saving and for each and every spot on the sample you will have 10 readings 76 Figure B9 Reflectance measurement on module CONVERSION OF ASD TO TXT FILES Once all the measurements are done the reflectance values are saved as ASD files and the next step is to convert them to TXT files Go to ViewSpec Pro gt File gt Open open the files you want to convert gt Process gt ASCII Export gt In the dialog box just change the Data for asd files only to Reflectance don t change any gt OK The Output path where the processed data gets stored is indicated at the bottom of this software TI Data Format for asd files only DN Reflectance CO Radiance lrradiance Log 1 R Absolute Parabolic Correct Log 1 T O Transmittance Derivative Q Noe Set Derivative Gap
35. no Si Better Technology to Characterize QE and Reflectance Losses For reflectance measurements the contact probe diameter of spectroradiometer was found to be larger than the cell area of the polycrystalline mini module Instead of covering one cell at a time the contact probe covers two cells that are shown using the circle in Figure 19 Hence this resulted 1n noise in the reflectance measurements Figure 21 Circle Indicating Contact Probe Diameter gt Cell Area for Poly Si Coupon When comparing the Quantum Efficiency QE measurements for both poly and mono Si the influence of grain boundaries on poly Si infused noise in the results On the other hand mono Si being a single cell module eliminated these noises Figures 20 amp 21 show the variation in QE curve for both poly and mono Si modules The variation in QE curve for mono Si both cleaned and soiled is minimal and it clearly proves that in order to eliminate noises and to obtain accurate results in QE measurements mono Si samples should be preferred 27 Spot 1 Cleaned Spot 1 Soiled Spot 2 Cleaned Spot 2 Soiled Quantum Efficiency 96 600 800 Wavelength nm Figure 22 QE Curve for Poly Si Coupon 1 55 g m Spot 1 Cleaned Spot 1 Soiled Spot 2 Cleaned Spot 2 Soiled Quantum Efficiency 96 400 600 800 Wavelength nm Figure 23 QE Curve for Mono Si Coupon 1 57 g m 28
36. nufactures The median and average degradations are very close to each other 0 73 year vs 0 69 year indicating a tight quality management system during production However there is an outlier for one string due to the presence of a delaminated backsheet module in the string leading to severe encapsulant delamination causing Isc loss along with voltage loss due to bypass triggering 5I Do not meet rate of 1 0 yr 1 6 1 29 0 89 Degradation of Power 99 yr Figure 34 Model J String level Degradation of Power From 46 module I V curves the average degradation rate of power is found to be 0 55 year Out of 46 modules 40 modules 86 3 are meeting the less than 1 0 degradation rate limit typically provided by module manufactures as shown in Figure 33 The string degradation rate seems to be slightly higher than the module degradation rate due to module mismatch that occurs during string level analysis 52 Do not meet rate of 1 055 yr eet rate of 1 096 yr 7 L a m E m LL 3 0 25 20 1 5 1 0 0 5 0 076 0 5 Degradation of Power 76 yr Figure 35 Model J Module level Degradation of Power 2 4 2 Layout of Degradation and Safety Failures The defects that are present in 744 modules are provided in the graph below Of all backsheet bubbles overcell encapsulant delamination and near edge encapsulant delamination are prominent in almost all the modules Bypass diode failure broken mod
37. oil Densities 31 28 Outdoor Reflectance Plots for Various Soil Density ic eecccccceeeceeeeeseeeeeeeees 32 29 Indoor Reflectance Plots for Various Soil Density essere 33 30 Correlation plot between reflectance and soil density g m eee 36 31 Correlation plot between QE and soil density g m sse 36 32 Grading PV Power Plant Conceptual Approach eeeeeeeesseeeeeeee 4 33 Reliability and Durability Issues of PV Module 14 seeeeeeeeesssss 43 34 Model J String level Degradation of Power esssessssoeeessssssssseersssssssssceersssssssseeeees 22 35 Model J Module level Degradation of Power ssesssssennssssssssseerssssssssseeersssssssseerees 53 36 Detects Identitied am Models Vosa ettet AME OH ERE EDU Eten Au e eae eR Mes 54 27 5ateby MIDO Mode IT iei tutte epe neh in todu idein a elas 55 38 Safety Failures Reliability Failures and Durability Loss for Entire Power Plant 56 39 Global Degradation and Safety RPN Chart for String level Analysis 58 40 Global Degradation and Safety RPN for Module level ssss 60 41 Global Degradation and Safety RPN for Module level all defects included 60 42 clock wise a Backrail Mounting using Adhesive of Frameless Module Model J b Framed Module at Model
38. ollows Table 2 Spray Gun Specifications 6 Air Pressure Range 30 40 PSI Maximum Air Pressure 40 PSI Three factors that need to be considered during spray gun adjustment stage are Fan direction 2 Pattern adjustment 3 Fluid adjustment After soiling the variation in density was found to be high along a particular direction If the fan of the gun was placed along a horizontal direction it was observed that the variation in density was high in the slides that were placed along a vertical direction and vice versa when the fan was adjusted along a vertical direction For this study the fan of the gun was constantly fixed along a horizontal direction and the microscopic slides to calculate soil density were placed along a vertical direction so the variation was controlled Lock Ring Horizontal Fan Bn Nozzle Figure 2 Horizontal Fan Direction 6 Based on trial measurements it is identified that when the pattern knob was adjusted to round closed position the soil spray was more uniform The pattern knob was used to adjust the spray pattern d Round dme aa Figure 3 Pattern Adjustment 6 The fluid knob was adjusted to a fine position and the air pressure was set to 30 PSI to get a fine layer of soil on the test modules Too Coarse Correct Too Fine Increase Decrease air flow air flow Figure 4 Fluid Adjustment 6 1 3 3 Soil Formulation Initially the suspensions were pr
39. oltage bias was applied and by keeping the voltage bias constant multiple QE was taken due to grain boundary effect for each spot and the average was considered to be the QE of that spot The light bias was always maintained at 100 intensity Monocrystalline silicon being a single cell module had no necessity in using voltage bias The light bias was maintained at 100 intensity similar to poly Si Identical QE curves were obtained along any place in the spot and proved to be more accurate 21 Figure 15 a Poly Si Cell Covered with b Mono Si cell with Light Source Mask An Opening for Light Source Figure 16 Soiled Spots Cleaned for Reflectance and QE Also it was made sure that QE was performed on the same spot where reflectance was done The soil layer that was present in the spots were cleaned using cotton gauze dampened with isopropyl alcohol Finally cleaned QE and reflectance measurements were performed 22 Cleaned module EL imaging Cleaned module I V Cleaned Cleaned module module QE reflectance Figure 17 Flow Chart Indicating the Characterization Process 25 1 4 RESULTS AND DISCUSSIONS 1 4 1 Soil Uniformity Check To check if the soil layer obtained using this approach is uniform four microscopic slides were placed on four sides of the coupon A laser guided technique was used during soil application and the slides were weighed before and after cleaning The standard deviation for
40. p due to bypass triggering and high bypass diode failures as they conduct electricity during daytime every day for several years before they permanently fail under open circuit conditions 62 Interconnect discoloration is scarcely found in model J because a frameless module Figure 43 does not have a leakage current path as the mounting rail is on the backskin of PV modules In the case of a framed module current leakage occurs between cell and frame through electrolytic corrosion resulting in interconnect discoloration These modules should have higher series resistance problems leading to local PR heating which in turn leads to encapsulant browning Figure 44 eo f Le Aa d Figure 42 clock wise a Backrail Mounting using Adhesive of Frameless Module Model J b Framed Module at Model JV A c Encapsulant Browing and Interconnect Discoloration in Model JV A d Encapsulant Delamination in Model J 63 Model J Pareto chart of defects 744 modules in total frameless 18 yrs age functional il Beil eh ee MN 1 o c o J g 2 i Frequency 96 84 68 75 67 72 3 38 44 15 19 Percent 29 6 26 4 25 3 13 4 5 3 Cum 96 29 6 56 0 81 3 94 7 100 0 Figure 43 Pareto Chart for Model J Model JVA Pareto chart of defects 348 modules in total framed 19 yrs old not functional estimated to be 4 yrs 100 80 60 o 40 D 20 i 0 s A a8 eU SF xe amp ov RX Re d ex A S Q GS
41. que which is a multiplication of Severity Occurrence and Detection for ranking the failure modes The higher the RPN the worse is the failure mode Arizona State University Photovoltaic Reliability Laboratory ASU PRL has recently evaluated crystalline silicon PV power plants in the State of New York a cold dry climatic condition This site has 18 year old frameless glass polymer modules on a rooftop 41 tilt Sanjay et al 13 was successful in determining the global RPN and in identifying the dominant failure and degradation modes for the hot dry desert climate of Phoenix Arizona In this study apart from the global RPN degradation RPN and safety RPN for both string and module levels have been determined Safety RPN gives information on the order of 39 priority for the defects to be addressed which would cause property damage or personnel electric shock Degradation RPN provides information on the order of priority for the defects to be addressed which would affect the energy production 2 1 2 Statement of the Problem There are various field failure modes affecting the safety reliability and performance of the PV modules These failures are not unique and as they vary according to the climatic conditions there needs to be a Statistical tool to determine the dominant failure modes FMECA is one such tool and by determining the failure modes the module designs that are resistant to those failure modes specific to that climati
42. re modes 47 The degradation rate is calculated using the formula below Degradation rate Rg Pmax drop 100 Rated Pmax age of operation The below table gives the Severity ranking corresponding to the degradation rate Ra Table 8 Severity S and Detection D Ranking Table Detection Criteria Severity Criteria Monitoring System itself will No effect Rd 0 3 detect the failure mode with warning Very high chance of detection through visual inspection High chance of detection through visual inspection Very high chance of detection using a sophisticated tool Very high chance of detection using multiple sophisticated tool High chance of detection using multiple sophisticated tool Very low chance of detection 1 2 3 4 5 7 Insignificant Rd approx to 0 3 Minor Cosmetic defect Rd 0 5 Cosmetic defect with Rd lt 0 6 Reduced performance Rd 0 8 Performance loss approx to typical warranty limit Rd approx to 1 Significant degradation Rd approx to 1 5 Remote safety concerns Rd gt using multiple sophisticated 1 tool Remote chance of detection Remote safety concerns Rd gt 2 in the Safety hazard Catastrophic Good chance of detection using a sophisticated tool Detection impossible field 2 3 4 Determination of Detection The detection ranking is based on how simple complex it is to determine the field failures A rank of 1 is given if the failur
43. rease the accuracy a laser pointer was attached to the tip of the gun as shown in Figure 6 b The laser spot helped in determining the exact center of the test coupon while spraying and maximum deviation in density was reduced to 0 2 g m 13 Figure 8 Spray Gun a Without Laser Pointer b With Laser Pointer 1 3 7 Soil Density Measurements The soil density measurements g m were carried out using commercially available microscope slides 2 5x7 6 cm placed on two sides of the test coupon The density calculations were carried out using Mettler Toledo AG285 resolution 0 001 mg The soil density was calculated by measuring the difference in the weight of the slides before and after soil deposition divided by the area of the microscopic slide The average of these measurements was taken to determine soil density on the mini module 14 METTLER TOLEDO Vid E Figure 9 Soil Density Calculation Soiled Microscopic Slide Circled 1 3 8 Poly Si Coupon Good Indicator of Uniformity To verify the uniformity in soiling pattern using this approach I V curves were taken before and after soiling on the polycrystalline S1 coupon as multiple cells are connected in series If there exists any significant non uniformity then no smooth curve would be expected between Isc and Imp values The soiled poly Si coupons went through a few characterization tests as indicated in the flow diagram Figure 15 Uncertainties that
44. rial issues and the reliability issues are primarily attributed to the design and or manufacturing issues 2 2 4 FMEA FMECA FMECA is a quantitative approach to determine the dominant failure mode that is observed in the field FMECA extends FMEA with an addition of a detailed quantitative analysis of criticality of failure modes FMECA is a method used in a product system device to identify the failure modes as they are manufactured using complicated technologies 43 FMECA is usually conducted in the product design or process development phase or after a quality function deployment to a product but conducting it on fielded systems also yields benefits FMEA FMECA analysis allows us a good understanding of the behavior of a component of a system as it determines the effect of each failure mode and its causes and ranks each failure mode according to criticality occurrence and detectability 13 44 2 3 METHODOLOGY 2 3 1 System Description The system was installed in November of 1996 in a cold dry climate and is currently operational The 18 year old system is located on the rooftop of a facility and has a total of 13 arrays all at latitude tilt 41 Of the 13 arrays 12 are actually in use and 1 array has been removed reason unknown Within the 12 arrays that are present 1 combiner box is burnt up resulting in the entire array producing no power There are 6 strings in 11 arrays and 2 strings within array and each str
45. rum is saved to collect reflectance e The reflectance values are saved as asd files and they are converted to txt files using ViewSpecPro software The detailed procedure for collecting and saving spectrum is provided in APPENDIX B 1 3 10 Quantum Efficiency Measurement System Quantum efficiency QE is defined as the ratio of the number of electron carriers generated to the number of photons of a given wavelength that are incident on the solar cell QEX 12M quantum efficiency measurement system as shown in Figure 10 is a device that measures QE of a cell within a module using a non intrusive approach 9 18 Figure 12 QEX12M Module QE System 10 System calibration and operation To turn the machine on first turn the main power switch on and then proceed to turn on the auxiliary power switches After starting up give the xenon arc lamp about 10 15 minutes to warm up The module bias light should be turned on The intensity of the light can be adjusted by turning the knob Always start by calibrating the system before taking QE measurements Position the monochromatic light on the calibration photodiode Select Calibration PD on the home screen Measure the calibration curve by clicking start After the curve is finished save the curve and select apply as calibration The monochromatic beam is then positioned on the test coupon and by applying voltage bias and light bias the QE measurements are carried out
46. s ii DEDICATION This thesis work at ASU PRL 1s dedicated to my parents Rajasekar Arunachalam and Padma Rajasekar and my brother Karthikeyan Rajasekar for their constant motivation love and support during my Master s program 111 ACKNOWLEDGMENTS First of all I would like to thank my advisor chair Dr Govindasamy Tamizhmani for his constant guidance effort and supervision throughout my Master s thesis work It was a great pleasure and honor to be mentored by such a hardworking and dedicated figure to the solar industry Also I would like thank my committee members Dr Rogers and Dr Srinivasan for their constant support during my work I would like to thank Dr Joseph Kuitche for giving me the opportunity to work at Arizona State University Photovoltaic Reliability Laboratory I value the lessons learned under his guidance My special thanks goes to Sravanthi Boppana who was a source of encouragement Also I was really grateful to work with hard working individuals in the lab and would like to thank Sai Tatapudi Sanjay Shrestha Mohammad Naeem Mathan Kumar Moorthy Neelesh Umachandran Christopher Raupp and Matthew Chicca for their support I would also like to acknowledge the support provided by Patrick D Burton and Bruce H King from Sandia National Laboratories and thank them for their technical expertise 1V TABLE OF CONTENTS Page LSEOF TABLES od nn ee Vill DES OT GO ceca
47. s 1n series 1 cell Cell dimension 5 7 cm by lcm 15 4 cm by 15 4 cm 11 Figure 5 a Polycrystalline Silicon b Single cell Monocrystalline Silicon Mini module 1 3 5 Artificial Chamber Set Up To maintain a controlled environment during soil deposition an artificial chamber was constructed The chamber consisted of a cuboidal mechanical structure to support an air bag from Sigma Aldrich Corporation In order to avoid human errors the spray gun was placed on a mechanical structure and the soil was sprayed The distance between the test coupon and the tip of the gun was ensured to be about 2 5 feet The artificial chamber was initially erected horizontally but on spraying soil it was observed that the soiling pattern was more uniform when the chamber was flipped placed vertically and the coupon was placed on the ground as gravity helps in maintaining uniformity Also it is important to ensure that the spray gun was held perpendicular to the center of the module and a pulse spray approach was implemented to obtain further uniformity 12 Figure 6 Artificial Chamber 60 46 76 in cm Figure 7 Erected Vertically With Erected Horizontally Without Glove Bag Glove Bag 1 3 6 Importance of Laser Guided Technique Initially the laser pointer was not employed during the application of soil and on weighing the microscopic slides the maximum deviation in density between two slides was found to be 0 6 g m To further inc
48. s given in the table as follows Table 1 Prediction of Various Components in Reflectance Spectra 5 Wavelength nm Predicted components 350 400 Eliminated due to noise Anti Reflection coating 600 900 Soil Organic Content SOC 1900 2200 Clay content 1900 2300 Carbonate content 1400 1900 and 2200 Water absorption peaks 1 3 METHODOLOGY 1 3 1 AZ Arizona Road Dust Soils were formulated artificially by mixing standardized soil or particulate matter commonly referred to as AZ road dust ISO 12103 1 A2 Fine Test Dust with HPLC High Performance Liquid Chromatography grade acetonitrile According to the manufacturer the raw material for Ar road dust 1s the dust that settles out of the air behind or around tractors operating in the Salt River Valley Arizona They are recommended to be caught on a canvas cloth and are dried in an oven The dried dust is made to pass through 200 mesh screen 0 0029 in width of openings and the dust that stays on the mesh is discarded The dust that is obtained is finally made to pass through 270 mesh screen 0 0021 in width opening and is collected The chemical composition and the test dust particle size 1s included in APPENDIX A 1 3 2 Spray Gun Specifications and Adjustments The solution is then uniformly sprayed on the test module using a HVLP High Velocity Low Pressure spray gun with a 1 mm nozzle from Centralpneumatic The detailed specifications of the spray gun are as f
49. t panel consists of an accessory power port and a fiber optic cable that is fixed The cable should be handled with care as it tends to break on bending any breakage in the cable can be identified using the fiber optic checker and can be replaced if the Signal to Noise Ratio SNR drops below an acceptable level The instrument back panel has an on off switch that allows the instrument to switch on off accordingly and an Ethernet port that is to be connected to the laptop that has the software installed in it The power port supplies power that is required by the instrument and is connected to the instrument controller A Nickel Metal Hydride NiMH battery is also provided to assist in outdoor measurements 8 17 Setting up and saving spectrum e For reflectance measurements the light source should be switched on for a minimum of 15 minutes The contact probe acts as a light source and also as a receiver e Before taking any reading the instrument needs to be optimized to the current atmospheric conditions or else the instrument gets saturated e The calibrated white reference WR reflector is fixed to the contact probe and the reflectance is collected A straight line at 1 1s observed indicating that all light is reflected as it is white and with respect to this all other reflectance measurements are carried out e The white reference cap is removed the contact probe 1s perpendicularly placed on the coupon surface and spect
50. tection D Ranking Tables isspiss aa a 49 To Occurrence O Rankine D3ble oup coo p a a eed RRUde peu 50 11 String level Severity and Occurrence Table for Model J sss 57 12 Mod lelevel Detects Tor Model Jine ERE tenia iemusenaauaea 59 13 Comparison between Model J and Model JVA 0 cc eeeccccccccccccsssssesecceeeeeeeaeeeseeeeeees 61 Vill LIST OF FIGURES Figure Page Goabot thee XDELIOEE soc no ineo ES artes PR tenebo eese Lu tema ebat tnu ota eins 3 2 Horizontal Pan Directo Fo acc retiro tee os Et buon CREE NERO RES PER ested edo EDU E IE D22O Bd 9 SoPatterm Adus meN O o toes ril enam E nemus en lebat tie Unde oae 9 TEU TAOS Nne O oce Mo edu am rese dada cider tiara dde tran 10 5 a Polycrystalline Silicon b Single cell Monocrystalline Silicon 12 6 Artificial Chamber 60 46 76 in cm Erected Horizontally Without Glove Bag 13 7 Erected Vertically With Glove Bag ot reete tut saved suada Du seges beer te Dub Page eoe 13 8 Spray Gun a Without Laser Pointer b With Laser Pointer eeuusssse 14 9 Soil Density Calculation Soiled Microscopic Slide Circled sssse 15 10 heh Resolution Spectrorddtometer 7 osi ee et EDI ERE RERO Ute pa Fett aec ea ebd 16 MEC OMA t PRODO d S ecpsetctenuertt ette tameteducas iut ient UC MM EL IM cM EE i M E 16 12 OEXIZM Module QE System O ecese ven beer easu d E Eras aEC Pere 19 13 EL Im
51. the laptop is always switched on after the spectroradiometer Figure B1 Power supply to port connection 2 Connect the accessory power port to the contact probe as shown below 3 To connect the fiber optic first remove the screws in such a manner that the grey color screw is placed in the same place Then take the fiber optic and insert it in the screw that has been removed Gently push the fiber optic in the place were the screws were already present and tighten it Handle the fiber optic with utmost care as it is sensitive and tends to break 71 Figure B2 Handling of fiber optic 7 s d i 2 D lt Figure B3 Step by step process of inserting optic fiber to the contact probe Hit the ON button which is on the rear side of the spectroradiometer unit and then click the ON button that is present on the contact probe so that the instrument starts warming up For reflectance measurements the light source should be 12 switched on for a minimum of 15 minutes whereas for radiometric measurements the time 1s extended to an hour Figure B4 Switching and warming up of the instrument 5 Even for outdoor measurements initially use the power supply as the source and then once the instrument 1s warmed up the battery can be used The battery is charged separately by connecting one end of the power cord to the battery and the other end to the supply As in step 1 instead of connecting
52. the power supply to 12VDC connect the battery in its place 79 Figure B5 Charging of battery 6 Once the instrument is warmed up take a small square shaped black colored cardboard sheet and make a circle in the center the same as the size of the lens Insert it to avoid the entry of the stray light and then clean the lens using lens wipes Isopropyl alcohol and a soft cloth Switch on the laptop Figure B6 Cleaning of lens OPTIMIZATION AND WHITE REFERENCE Before taking any reading first you need to optimize the instrument to the current atmospheric conditions If you are doing an outdoor experiment take the instrument outdoors and optimize it as the indoor and outdoor atmospheric conditions differ 74 Optimize the instrument whenever the atmospheric conditions differ or whenever a beep sound comes from the instrument indicating that the instrument is saturating 1 Cover the lens of the contact probe using white reference WR Never touch the central white portion of the WR as it is already calibrated Then hit the RS software in the desktop There are two RS software in the desktop click high contrast for outdoor measurements Figure B7 White reference measurement 2 A dialog box appears Using the drop down menu change the settings to bare fiber and raw DN mode Then hit the OPT Optimize to go ahead with the optimization 75 Figure B8 Dialog box to change settings 3 O
53. ule and broken cell interconnect photographs of a few defects shown in Figure 41 are determined to be the safety issues wherein the rest of the defects are considered to contribute to the increase in degradation rate 53 Defects of Model J 744 modules total un S E o o rn o o c Figure 36 Defects Identified in Model J The safety map below shows the safety failures that are found in the power plant 45 in total The majority of the issues are found to be due to broken cell interconnects 22 modules and 18 modules had failed diodes The 8 broken modules are found to pose a safety hazard due to electrical hazard and mechanical hazard Broken cell interconnects are a safety hazard due to the possibility for arcing to occur Failed diodes have the possibility to cause backsheet burning 54 Broken Modules 8 Failed Diode 15 Broken Cell Interconnect 19 Failed Diode with Broken Cell Interconnect 3 Ww Combior tT C aaner oos Sl Cha Per rruna Tul tt end replace Ihe mde Die Litt LL EEH ER CLE bi I Cimone tcs burst Chect Ver DUNDI Won ant omaes the some bes TI GIH PSS EE LLL A kEES L5E 4 tees ToT EE ye an a luem Al ae a a a Se NE Aa Se n a a a a a a Uer el ommum ae ES eee imm Ge meni eS Ee Ke Se eZ mus a a a el ee ume ee Fam 2 LEE ac cael INSGEO Ec ST Ll Lone i d IER eS CENE DUNS ee ONERE Ponens ee id Sa DUNMET ONENDU ON
54. ule level Defects for Model J Degradation Rate Yr BadshetBibles 72331 02 1 9 2 1 Cell Discoloration 384 02 1 9 2 16 OvercellEncapsuantDelammation 8468 06 s5 9 2 9 Near Edge Encapsulant Delamination 7567 03 3 9 2 s Bypass diode Failures OpenCK 242 000 w 5 4 xo Broken cell Interconnect 2 03 8 5 2 8 Module level RPN is also calculated in a similar manner as string level RPN Figure 44 includes only the defects that are present in 46 modules Figure 42 includes all defects that are present in the strings but severity of O is given if that particular defect is not identified during the module level calculation 59 Module Grading Power Plants for Degradation Safety and Global Risks 18 Years Cold Dry frameless 300 2 0 E Module RPN E Frequency Global RPN 470 B Degradation Rate 9 6 Yr Safety RPN 280 Degradation RPN 190 a alla Backsheet Interconnect Cell Over cell Near Edge Bypass diode Broken cell Bubbles Discoloration Discoloration Encapsulant Encapsulant Failures Open Interconnect Delamination Delamination Ckt 2 a ec ra o ES 2 o c v 23 c w i Li Degradation Rate Y Figure 40 Global Degradation and Safety RPN for Module level Module Grading Power Plants for Degradation Safety and Global Risks
55. used due to the failure mechanisms Failure mechanism cause Failure mode effect The failure mode defects can be identified by performing a visual inspection on modules or by using tools to determine the failure 2 2 2 Reliability Failures A reliable PV module may also be defined as a PV module that has a high probability of performing its intended function adequately for 30 years under the operating conditions encountered 14 All modules degrading at higher than 1 per year excluding the safety failures are reliability failed modules All the reliability failed modules qualify for the warranty claims proportional to the rate of degradation Reliability failures are also called hard failures 42 Degradative Issue 77 Durability Issues of Modules RA Wear Out Issue HA GG 70 a a 60 3 Catastrophic Failure Warranty Limits 5 50 c Wear Out t 8 008 N a 30 9 Reliability Failures of Modules 20 6 8 10 12 14 16 18 20 22 24 26 28 30 Year Figure 33 Reliability and Durability Issues of PV Module 14 2 2 3 Durability Failures If the performance of a PV module degrades but still meets the warranty requirements then those losses may be classified as soft losses All the modules that degrade less than 146 per year excluding the safety failures are called durability failed modules The durability issues are attributed only to the mate
56. y developed air soil glass interface in Mesa Arizona for fall season 0 648 g m soil density e Using an average reflectance measurement between 600 700 nm bandwidth the soil density of the module can be determined By using the empirical formula presented in this work f2 AOI values for any AOI as well as transmission losses can be estimated if the soil density is known measured e If the soil density of a particular region is known the angle of incidence related losses for the whole year can be modelled using PVSyst oT PART 2 OUTDOOR STATISTICAL RISK ANALYSIS OF PHOTOVOLTAIC POWER PLANTS 38 2 1 INTRODUCTION 2 1 1 Background Photovoltaic PV modules employed in the field can experience different types of failure mechanisms and varying degradation rates due to the changes in environmental conditions design installation type electrical configuration and many other factors The industry is in need of coming up with appropriate accelerated tests for new modules and in order to achieve this statistical analysis of performance parameters is necessary to determine degradation modes responsible for the degradation of those parameters Also in order to find the dominant failure modes in the modules employed there is a need to carry out quantitative analyses like FMECA Failure Mode Effect Criticality Analysis FMECA is carried out by ranking and prioritizing the failure modes It uses the RPN Risk Priority Number techni
Download Pdf Manuals
Related Search