Parameter identification of photovoltaic models using a sine cosine differential gradient based optimizer

Abstract

In this paper, an efficient sine cosine differential gradient-based optimization method is proposed for identifying unknown parameters of photovoltaic models. In the simulation, parameter identification is formulated as an objective function to be minimized based on the error between the estimated and experimental data. Based on the original gradient-based optimization method, the proposed method combines the sine cosine method and the mutation crossover of the differential evolution algorithm. Specifically, the crossover operator enables the algorithm to avoid local optima; and meanwhile, the sine cosine strategy encourages the new individual to calculate the worst position. The simulation results demonstrate that the new optimization method can achieve the minimal root mean square error and obtain better optima relative to other algorithms in different photovoltaic cells. Therefore, the proposed optimization method has great potential to be used for estimating photovoltaic model parameters.