Accurate module performance characterisation using novel outdoor matrix methods

Indoor PV module characterization is performed with a matrix of 6-23 irradiance and module temperature points (G,T). Separate angle of incidence and spectral measurements are needed. Reducing the number of matrix points would lower measurement costs but may reduce modelling accuracy.Outdoor PV module characterization uses frequent measurements (IV or Pmp) under actual weather conditions which include angle of incidence, diffuse fraction, and spectral effects. Measuring every minute can give ~260k records / year. These need good data sanitizing and filtering for accurate understanding.This paper shows novel methods to analyse matrix data much better than 6-23 points to generate the following :-1)Outdoor matrices (~100 points) for PRdc, Voc, Isc etc.2)Better modelling accuracy and performance checking.3)Degradation rates and causes (e.g. %Voc fall/year).4)Quantified technology differences e.g. cSi vs. Thin Film.5)Pmp values at different conditions STC, LIC, PTC etc.6)Temperature coefficients e.g. gamma(Pmp), dRoc/dTmod.7)Problem finding such as "lower Rshunt than expected".8)Comparison of outdoor with indoor reference IV curves.A mechanistic fitting approach is used for the modelling, 6 parameters are sufficient to model c-Si and HIT well. Residual errors of the matrix points can show if extra mechanistic coefficients need to be added for even closer modelling fits. An enhanced model with meaningful, independent coefficients has been used for any non-linear modules e.g. some thin films where gamma (dPRdc/dTmod) can vary a little with Tmod and/or Irradiance.Much of this improved matrix analysis can be performed with only (Imp and Vmp) or Pmp values.The data analysed is from many different PV technologies (>15 modules both cSi and Thin Film) which have been measured from ~2011-2021+ at Gantner Instruments’ OTF in Tempe AZ