Computation and validation of the Expected Value of Power of Two Terminal Series–Parallel PV arrays

Abstract

PV arrays are susceptible to various types of failures such as partial shading that can negatively affect their performance and efficiency. Studying the impact of these circumstances in the performance of the PV array is key for the development of more efficient PV systems. In this study, the definition of the expected value of power (EVP) that a PV array can produce under a random failure scenario is revised and improved. An algorithm to compute EVP that minimizes the number of simulations is developed and tested in three different PV arrays. Results show a 96.6% reduction on average of the number of simulations needed to compute the EVP for PV arrays made up of 9 modules. The model is validated through an experiment that reproduces the random fault scenario and determines the mean power produced by the PV arrays. For all three experiments, the computed EVP fits the experimental data with $R^2>0.995$.