Improved honey badger algorithms for parameter extraction in photovoltaic models

Abstract

Precise estimation of the parameter values of solar models is very essential for optimization of solar systems. Many studies that use metaheuristic algorithms have recently been proposed for parameter estimation and optimization in photovoltaic models. In this study, it is aimed to enhance convergence performance in photovoltaic systems by scanning the search space using two improved versions of the honey badger algorithm. First, the Gauss/Mouse map -based chaotic honey badger algorithm has been considered motivating by the fact that chaotic maps are successful in controlling critical random values in exploration and exploitation phases. The other algorithm is based on hybridization of opposition based learning with honey badger algorithm. Opposition based learning has efficient convergence capability as it scans the search space using opposites of candidate solutions. The performances of these improved methods and recent metaheuristic optimization algorithms are firstly evaluated for CEC2017 and CEC2019 datasets. After obtaining the successful results for these datasets, proposed algorithms are compared for single-diode, double diode and photovoltaic module models which are given as poly-crystalline Photowatt-PWP201, mono-crystalline STM6-40/36, and poly-crystalline STP6-120/36. For each model, optimum parameter values are found minimizing the root-mean-square-error. In addition, three commercial PV panels; Mono-crystaline SM55, Thin-film ST40, and Multi-crystalline KC200GT are also considered for performance evaluation. According to simulation results, proposed algorithm exhibits high performance in terms of minimization of root mean square error. The compliance of estimated and actual values of the parameters are visualized with current-voltage (I-V) and power-voltage (P-V) character-istics. The results show that the proposed methods are effective alternatives for solution of photovoltaic parameter estimation problem and contribute to the parameter optimization of photovoltaic models.