Photovoltaic system performance statistical analysis

PV system performance evaluation is not clear today. Performance Ratio (PR) or other performance indicators are site-dependant and in many cases systems reveal to have poorer performance than expected. Thus, the evaluation of factors of influence is very vague and a generalised method is necessary. We present a new methodology addressing performance for the overall PV system, for which all inputs are treated as stochastic variables. The methodology is explained and examples are given for different kinds of PV systems (grid connection, pumping SHS or SAPV systems with storage even PV/thermal systems). Additional applications are also presented (irradiation data generation, islanding probability, PV module matrix): Two main applications illustrate the methodology: the calculation of the battery SOC probability density function to assess the design of the PV system, and a way to evaluate the representativeness of the battery tests sequence in comparison to the field operating conditions the battery will face.