一种非常规改进的猫群算法用于太阳能发电系统中的最大功率点跟踪

Abstract PV systems require Maximum Power Point (MPPT) for efficiency, but traditional methods, and also Meta-heuristic algorithms, struggle with convergence and efficiency under shading and irradiance changes. This article proposes a modified Cat Swarm Optimization (CSO) algorithm with 11 improving steps to enhance convergence precision toward the global Maximum Power Point. The key innovation of this work lies in the fact that these steps represent either the first employment of their kind or, more significantly, their combined application. These steps include the implementation of probabilistic distribution mechanisms to dynamically adjust cat parameters, and introducing new control parameters for optimized position selection and controlled movement. The algorithm was rigorously tested under multiple partial shading conditions with varying iterations, cat populations, and PV configurations. Across 1000 simulations and for four cases of small-scale and large-scale systems, average error rates ranged from 0.55% to 7.52%,. Compared to the original CSO, the modified CSO demonstrated an average of 14.25% improvement in convergence accuracy which resulted in a 2.74% boost in output power of the proposed system. Further simulation results indicate that accuracy improves when increasing cats up to 3 to 4 times the panel count, but plateaus sharply beyond this point while computational load grows; likewise, increasing iterations beyond 400 provides rapidly diminishing accuracy gains. Therefore, the proposed method adapts seamlessly to new PV systems using only the panel count as input, eliminating complex tuning and ensuring broad applicability.