How kWh/kWp modelling and measurement comparisons depend on uncertainty and variability

Most PV simulation programs (PVSP) use similar algorithm sequences [1] to predict system kWh/kWp and cent/kWh from weather, PV performance and BOS databases with users' best guesses for dirt, shading etc. PVSPs usually estimate PV performance vs. irradiance and temperature from a 1-diode model of either a manufacturer's datasheet or from an indoor or outdoor characterisation of one (or a small number) of samples and often won't consider the variability or uncertainty of module measurements. When modelling kWh/kWp there are a very large number of uncertainties, variabilities and unknowns which dominate the predicted or measured energy yield. Levelized cost of electricity (LCOE) [2] calculations have been performed on the modelled performance data. Important factors are analysed and quantified for locations and suggestions made as to how to improve modelling and performance claims.