WIXLIB

Accelerated Life TestingRob SorensenSandia National Labsnrsoren@sandia.gov505-844-5558Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department ofEnergy’s National Nuclear Security Administration under contract DE-AC04-94AL85000This work was performed under funding from the DOE Solar Energy Technologies Program.Inverter Reliability Workshop

Accelerated Aging for Inverters No PV specific industry standard existsHALT testing is spotty; independently appliedSeparate needs identified for residential and commercial scale invertersFailure modes identified but not in a uniform program applicable across theindustrySystem predictive models will require inputs for invertersInverter Reliability Workshop

Laboratory testing provides vital information forPV system reliabilitySystem performance model mustinclude wear out (end of life) informationALT Field Data (O&M, Accelerated Testing / Lab TestsFailures, )Accelerated Aging ofTape Joint – ThermalCyclingLin k toPerformanceAcceleration FactorsInverter Reliability Workshop

What is ALT & why?What? Component life tests High stresses Single or combined Activate “appropriate” failure modes Measureable Failure analysisWhy? Time Full system is expensive and complicatedInverter Reliability Workshop

‘Accelerated Aging’500 C, 5 minTemperature100 C, 12 min40 C, 30 days25 C, 45 daysReaction Coordinate5Inverter Reliability Workshop

High T data are extrapolated to “use” conditions(room temperature)6Inverter Reliability Workshop

How you extrapolate can influence lifetimepredictions. 50% in 20 years 50% in 100 years 50% in 1000 years7Inverter Reliability Workshop

Two approaches to accelerated testing are usedthroughout industry Qualitative Accelerated Tests HALT tests HAST tests HASS testsSmall sample sizeSevere level of stressIncrease reliability(product improvement)Qualify new designsDesign quantitative ALTReliability under normaluse conditions Quantitative Accelerated Life Tests Controlled application of acceleratedstress Produces acceleration factors (AF) Usage rate acceleration(Time compression)Overstress accelerationInverter Reliability WorkshopUsed to determine TTFDetermine reliabilityLong TimeNeed degradation / failuremechanisms

The Goal of an ALT program is to produce accelerationfactors Often empirical correlations Limited root-cause analyses MTTFfield AF MTTF test Cumulative Failure Percentile999590755025MTTF10514050Time-to-Failn E 1 1 RH a bV AF exp a k To T RHo a bVo Inverter Reliability Workshop80100

Empirical relationships may not cut it!Atmospheric Corrosionof MicrelectronicsALT must capture validdegradation / failuremechanisms106410AFFive accepted environmentalmodels.All agree at 85% RH (ALTconditions).Orders of magnitude differenceat 30% RH (use conditions).exp [B(T RH)](RH) nexp (Ea/kT)2exp [Ea/kT B(RH) ]exp [Ea/kT B/(RH)]exp [A/kT B(RH)/kT C(RH)]1021020406080RH (%)Calibration datacollected hereInverter Reliability Workshop100

Issues with ALT Unknown failure mechanisms Unknown / variable use environment Changing mechanisms as functionof environmental stress Difficult to control andcharacterize defects Long duration experiments Evolving / improving technologyInverter Reliability Workshop

What are the likely stresses that lead to Inverter Failure? VoltageTemperatureThermal cyclingThermal ShockVibrationMechanical ShockHumidity?Inverter Reliability WorkshopHeatOtherLightMoisture

How do we apply ALT to predicting end-of-life (wear out)?Determine FailureMechanisms (fieldexposures)Understands stressesthat activate thefailure mechanismsDevise ALT conditionsthat acceleratedegradation / failurePerform ALT (range ofconditions)Qualify ProductDetermineacceleration factorsEstimate or PredictReliabiltyInverter Reliability WorkshopImprove ProductReliabilityValidate NewDesigns

Example: Al bondpad corrosion: corrosion requires moistureand contamination & is accelerated by temperature.Three environmental variables (T, RH, [Cl])1ug/cm2, 30C, 80% RHPEM 4023000d1-w 1Resistance (m-ohm)2500d1-w 3d1-w 42000No Failures!d1-w 111500d1-w 12d1-w 141000d1-w 15d1-w 2250000100020003000Exposure Time (hrs)14Inverter Reliability Workshop

Increasing contaminant level causes failure.Distribution of failure timesNot all failedClear effect of [Cl]20 ug/cm2, 30C, 80% RHPEM 4143000d1-w 1Resistance (m-ohm)250020X chlorided1-w 3d1-w 42000d1-w 111500d1-w 12d1-w 141000d1-w 15d1-w 2250043 ug/cm3, 30C, 80% RHPEM 41600501001503000200d1-w 1Exposure Time (hrs)Resistance (m-ohm)2500d1-w 3d1-w 42000d1-w 111500d1-w 1240X chloride1000d1-w 14d1-w 15d1-w 2250000123Exposure Time (hrs)15Inverter Reliability Workshop

Statistical treatment (life-data analysis) provides ameans of analyzing the bondpad dataReliaSoft Weibull 7 - www.ReliaSoft.comReliaSoft Weibull 7 - www.ReliaSoft.comReliability vs Time Plot1.000Reliability0.800Data 1Weibull-2PRRX SRMMEDFMF 16/S 15Data PointsReliability Line10 ug/cm2, 30C, 30%RHx4Reliability vs Time Plot1.000Reliabilityx2Data 1Weibull-2PRRX SRMMEDFMF 14/S 0Data PointsReliability Line20 ug/cm2, 30C, 80%RH0.800x3Reliability, R(t) 1-F(t)Reliability, R(t) 200.000400.000600.000800.000Rob SorensenSandia National Labs5/28/201010:46:25 AM1000.0000.0000.000Rob SorensenSandia National Labs5/28/201010:50:59 AM18.00036.00054.00072.00090.000Time, (t)Time, (t)β 0.2360, η 943.6290, ρ 0.8798β 0.7073, η 5.2052, ρ 0.9889ReliaSoft Weibull 7 - www.ReliaSoft.comReliability vs Time Plot1.000ReliabilityData 1Weibull-3PRRX SRMMEDFMF 13/S 0Data PointsReliability Line40 ug/cm2, 30C, 80%RHx20.800Reliability, R(t) 1-F(t) MTTFfield AF MTTF test x20.600x30.4000.200x20.0000.000 Provides distributions Includes suspension results Basis for models {Pfail f(T, RH, [Cl])}Rob SorensenSandia National Labs5/28/201010:57:34 AM0.6001.2001.8002.4003.000Time, (t)β 2.0355, η 0.6531, γ 0.3818, ρ 0.9576Inverter Reliability Workshop16

The distributed wirebond property (probability of failure) is input into anelectrical system model & other component outcomes (reliability,performance threshold) can be determinedVoltage ComparatorCircuitPrediction of the effect of corrosion onLM185y reliabilitypas f(storage location)1Arctic0.995Desert0.990.985Gulf Coast0.98Bondpad corrosion is modeledas an additional series resistor0.9750.970Inverter Reliability Workshop204060Time (years)80100

Laboratory testing provides vital information forPV system reliabilitySystem performance model mustinclude wear out (end of life) informationALT Field Data (O&M, Accelerated Testing / Lab TestsFailures, )Accelerated Aging ofTape Joint – ThermalCyclingLin k toPerformanceAcceleration FactorsInverter Reliability Workshop

Example: Conductive metal foil tape (Module)Generate ALT dataDetermine performance effectModuleLoadE I RP I V I 2 RApply acceleration factors to field 2σMeanDevelop “acceleration factors”R 10 ( 0.028(Inverter Reliability Workshopt ) α )-2σ

Use the ALT data to predict long-term performancedegradation (wear-out?)20Inverter Reliability Workshop

Accelerated Life Testing Rob Sorensen Sandia National Labs. nrsoren@sandia.gov. 505-844-5558. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of . No PV specific industry standard exists HALT testing is spotty; independently applied .