A Hybrid RIME Optimization Framework for Enhanced Feature Selection in Photovoltaic MPPT

Abstract

Maximum power point tracking (MPPT) is an essential box of the photovoltaic (PV) systems. The most common maximum power MPPT methods do not work well in a highly varied and dynamic situation, and phenomenologies of multi-peaks is observed in the curve of the output characteristic of the solar system due to the dissimilarities of temperature and light concentration. The paper suggests a combination of RIME optimization method that improves exploratory nature of the method in the initialisation stage, which involves combining tent mapping. It is aimed at enhancing the feature selection activities and MPPT of PV systems under the partial shading condition. It employs the piecewise mapping to optimize parameters in the algorithm and assault a reasonable stability between global exploration and local exploitation. The search strategy is dynamically modified and an adaptive inertia weight is added which further enhances the speed of convergence, effectiveness of the search as well as flexibility of the algorithm. To help decrease the cost of computations and maximize classification probability, the hybrid approach uses natural-inspired metaheuristics when performing feature selection which produce optimal subsets. In terms of monitoring speed, accuracy and stability in the PV MPPT environment, the technique is superior to PSO-BOA, traditional RIME, IRIME and EHRIME technique.