Laplacian Nelder-Mead spherical evolution for parameter estimation of photovoltaic models

Abstract

A critical aspect of the design, simulation, control, and optimization of Photovoltaic (PV) systems is evaluating the PV model's optimal parameter values based on the actual measured voltage and current values. To address that concern, an enhanced spherical evolution algorithm with no metaphor is proposed for identifying unknown parameters in PV models. The algorithm combines Laplace's cross search mechanism (LCS) and the Nelder-Mead simplex method (NMs), called LCNMSE. In a sense, the goal of LCS is to enrich the diversity of solution sets and make the variety of solutions coarser. The NMs enhances the algorithm exploitation by further scanning more promising ranges in the local region. This idea is developed to improve the local optimal solution's accuracy. In conjunction with both, a balance between exploration and exploitation is maintained. To verify the effectiveness of LCNMSE on high and multi-peaks cases, it is compared with eight state-of-the-art and basic algorithms based on 28 benchmark functions selected from 23 benchmark functions and 30 IEEE CEC2014 benchmark problems. Then, the method is utilized to evaluate the solar cells' parameters and PV modules. Experiments show that the algorithm performs well in evaluating different PV models' unknown parameters than other existing algorithms. Therefore, LCNMSE is an accurate and efficient technique for solar cell and PV models' parameter extraction problems. For further info or any question on metaphor-free LCNMSE, please visit https://aliasgharheidari.com.