PV Lifetime Modeling an Example: PID ensuring 25 years of service life

PV Lifetime Modeling an Example: PID ensuring 25 years of service life PV Lifetime Modeling Workshop, Apr 7 th, 2015, Orlando, Florida Max B. Koentopp, Christian Taubitz, Matthias Schütze, Marcel Kröber 1

CONTENTS I Introduction II PID-s lifetime modeling: How do we ensure a 25 year service life with respect to PID-s? need to have reproducible tests need to understand progression of PID-s in the field PID-s model need to derive parameters for tests need to establish process control in production III Challenges in Module Reliability and Lifetime Modeling 2

INTRODUCTION PID-s CHARACTERISTICS solar-cell characteristics Current I [ma] 6000 3000 ohmic characteristics degradation 0 electroluminescence image 0 200 400 600 Voltage Spannung V [mv] Schütze et al., 37 th IEEE PVSC, 2011 PID-s Potential induced degradation by shunting of solar cell 3

PID-s MODELING FOR 25 YEARS SERVICE LIFE How do we ensure a 25 year service life with respect to PID-s? need to have reproducible tests need to understand progression of PID-s in the field PID-s model need to derive parameters for tests need to establish process control in production 4

LABORATORY PID TEST METHODS Al foil test climate chamber test (DH) Al foil HV A temperature: 25 C voltage: -1 kv contacting: Al-foil temperature: 60 C voltage: -1 kv contacting: indirect via humidity reproducible, but continuous PID stress 5

PID-s MODELING FOR 25 YEARS SERVICE LIFE How do we ensure a 25 year service life with respect to PID-s? need to have reproducible tests need to understand progression of PID-s in the field PID-s model need to derive parameters for tests need to establish process control in production 6

PROGRESSION OF PID-s LABORATORY ENVIRONMENT VS FIELD CONDITIONS Laboratory PID-stress Field PID-stress PID-stress Module-performance time common PID-tests represent only one part of PID-s kinetics in the field 7

MODELLING PID-s Environmental data 8

MODELLING PID-s Environmental data Measured Computed irradiation ambient temperature relative humidity module temperature humidity at module surface rain data 9

MODELLING PID-s Environmental data R sh -kinetics from lab measurements 10

MODELLING PID-s Measurement of shunt resistance (R sh ) 11

MODELLING PID-s R sh -kinetics Taubitz et al., 27 th EU PVSEC, 2012 shunting (S)-, transition (T)- and regeneration (R)-phase 12

MODELLING PID-s Environmental data R sh -kinetics from lab measurements Classification 13

MODELLING PID-s Classification bias voltage on? (IR mod > 0?) no yes RH mod > 85 %? OR rain? yes no conditions for T-phase fulfilled? no yes S-phase T-phase R-phase 14

MODELLING PID-s Environmental data R sh -kinetics from lab measurements Classification R sh simulation 15

COMPARISON TO OUTDOOR MEASUREMENTS Measurement of shunt resistance (R sh ) 16

COMPARISON TO OUTDOOR MEASUREMENTS ; resistant samples ; prone samples (type A) ; prone samples (type A) prone sample (type B) --- --- calculation h calculation min STC power decrease 17

COMPARISON TO OUTDOOR MEASUREMENTS ; resistant samples ; prone samples (type A) ; prone samples (type A) prone sample (type B) --- --- calculation h calculation min STC power decrease 18

COMPARISON TO OUTDOOR MEASUREMENTS ; resistant samples ; prone samples (type A) ; prone samples (type A) prone sample (type B) --- --- calculation h calculation min STC power decrease 19

COMPARISON TO OUTDOOR MEASUREMENTS ; resistant samples ; prone samples (type A) ; prone samples (type A) prone sample (type B) --- --- calculation h calculation min STC power decrease 20

COMPARISON TO OUTDOOR MEASUREMENTS ; resistant samples ; prone samples (type A) ; prone samples (type A) prone sample (type B) --- --- calculation h calculation min STC power decrease 21

COMPARISON TO OUTDOOR MEASUREMENTS ; resistant samples ; prone samples (type A) ; prone samples (type A) prone sample (type B) --- --- calculation h calculation min STC power decrease 22

COMPARISON TO OUTDOOR MEASUREMENTS ; resistant samples ; prone samples (type A) ; prone samples (type A) prone sample (type B) --- --- calculation h calculation min STC power decrease good qualitative agreement 23

COMPARISON TO OUTDOOR MEASUREMENTS Results for full size modules: Valletta test site (Thalheim, Germany) test setup prone modules positive grounded 24

COMPARISON TO OUTDOOR MEASUREMENTS R sh [kωcm 2 ] 1000 100 10 CALCULATION calculation h calculation h, meas. temp. calculation min calculation min, meas. temp. Ys [kwh/kwp] 1 60 40 20 0 0 500 1000 1500 2000 2500 3000 3500 4000 time [h] STC power decrease PID prone module string reference module string rel. dev. [%] -20-40 -60-80 -100 FIELD MEASUREMENT (Valletta) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 time [week] 25

PID-s MODELING FOR 25 YEARS SERVICE LIFE How do we ensure a 25 year service life with respect to PID-s? Need to have reproducible tests Need to understand progression of PID-s in the field PID-s model Need to derive parameters for tests Need to establish process control 26

LONG TERM PREDICTION POSSIBLE 1000,00 R sh [kωcm 2 ] 100,00 10,00 resistant cell-type prone cell-type 1,00 STC power decrease 0 100 200 300 time [days] parameters necessary for 25 years service life can be derived Thalheim, Germany weather data of 2012 used 27

PID-s MODELING FOR 25 YEARS SERVICE LIFE How do we ensure a 25 year service life with respect to PID-s? need to have reproducible tests need to understand progression of PID-s in the field PID-s model need to derive parameters for tests need to establish process control in production 28

PROCESS CONTROL: PRODUCTION MONITORING module level monitoring (Al-foil test, 168h) weekly sampling: random cells from production are tested on module level monthly sampling: random modules from every conversion site cell level testing (PID cell tester, 24h) weekly sampling: PID-test on cell level (feedback < 48h) fast reaction possible 29

PID-s MODELING FOR 25 YEARS SERVICE LIFE How do we ensure a 25 year service life with respect to PID-s? Need to have reproducible tests Need to understand progression of PID-s in the field PID-s model Need to derive parameters for tests Need to establish process control 30

CONTENTS I Introduction II PID-s modeling: How do we ensure a 25 year service life with respect to PID-s? need to have reproducible tests need to understand progression of PID-s in the field PID-s model need to derive parameters for tests need to establish process control in production III Challenges in Module Reliability and Lifetime Modeling 31

CHALLENGES IN LIFETIME MODELING Identify important failure modes Identify which failure modes determine wear-out: ribbon fatigue? Solder joints? Encapsulants? Do we want to determine wear-out (end of life) or predict degradation rates (e.g. from cracks, yellowing, )? Develop failure mode specific accelerated tests develop appropriate accelerated test methods & standards. Up to now no statistical approach. Make sure failure modes can be distinguished Ensure failure modes in accelerated tests are same as failure observed in the field Often, long test durations are needed, e.g. UV. Component tests and standards? Develop kinetic models for failure modes understand location specific stress levels understand acceleration factor of laboratory tests develop kinetic models for specific failure modes to predict time of failure develop / adapt test methods to feed experimental results into models Test models with outdoor data Long times until defects occur in the field Fast product cycle in PV industry Wear out failures observed in field stem from products w/ very different BOM from today s products 32

CONCLUSION reproducible PID tests are available (Al-foil, climate chamber) laboratory PID-tests represent only part of PID-s kinetics in field studying regeneration properties is essential PID-s model based on laboratory tests and meteorological data can describe long term progression allows for prediction of time of PID-s onset in specific location good qualitative agreement with outdoor measurements continuous production monitoring is essential 33

Thank you Orlando, 07.04.2015 Max B. Koentopp, Christian Taubitz, Matthias Schütze, Marcel Kröber 34