Parameter's extraction of solar photovoltaic models using an improved differential evolution algorithm

Improving the performance of solar photovoltaic systems is reliant on accurately modeling the cells parameters. However, the identification of the lumped element model parameters remains a challenge given the non-linear nature of the diode circuit models. In this paper, we propose an improvement of the state-of-the-art of differential evolution (DE) technique. This relies on the Lambert W function and a preliminary meta-heuristic step to select the optimal mutation factor and crossover parameters for each specific Current-Voltage characteristic. We have applied it to three widely known cells and modules in literature. The Root Mean Squared Error of the proposed method is compared with literature results on the same systems. Furthermore, due to the stochastic nature of DE, an analysis of the stability was performed in order to evaluate its consistency. By using the meta-heuristic strategy, we were able to further improve the convergence characteristics of DE. The selection of DE parameters in the preliminary step have been shown to significantly reduce the computation time by half.