Single Solution Optimization Mechanism of Teaching-Learning-Based Optimization with Weighted Probability Exploration for Parameter Estimation of Photovoltaic Models

Abstract

This article presents a novel optimization approach called RSWTLBO for accurately identifying unknown parameters in photovoltaic (PV) models. The objective is to address challenges related to the detection and maintenance of PV systems and the improvement of conversion efficiency. RSWTLBO combines adaptive parameter w, Single Solution Optimization Mechanism (SSOM), and Weight Probability Exploration Strategy (WPES) to enhance the optimization ability of TLBO. The algorithm achieves a balance between exploitation and exploration throughout the iteration process. The SSOM allows for local exploration around a single solution, improving solution quality and eliminating inferior solutions. The WPES enables comprehensive exploration of the solution space, avoiding the problem of getting trapped in local optima. The algorithm is evaluated by comparing it with 10 other competitive algorithms on various PV models. The results demonstrate that RSWTLBO consistently achieves the lowest Root Mean Square Errors on single diode models, double diode models, and PV module models. It also exhibits robust performance under varying irradiation and temperature conditions. The study concludes that RSWTLBO is a practical and effective algorithm for identifying unknown parameters in PV models.