PV parameters estimation using optimized deep neural networks

Abstract

Estimating the parameters of a Photovoltaic (PV) cell is crucial, given the significant integration of the PV systems into electrical power systems. One of the primary challenges in the estimation of PV cell parameters is identifying a generalized method applicable to any PV system, irrespective of environmental variations and power ratings. This paper introduces a novel application of an optimized deep neural network designed to estimate the parameters of the PV systems across a range of temperatures, irradiance values, and PV module ratings. The network undergoes a training process by utilizing data obtained from the PV module block located within the Simulink library. In order to evaluate the effectiveness of the proposed methodology, the network is subjected to a series of assessments. These assessments encompass the utilization of PV cell data from the Simulink library, comparisons with recently developed methods, and practical evaluations using experimental PV cell data to estimate the PV cell parameters. The findings underscore the simplicity and precision of the proposed method across diverse PV cells.