An Epistemological Analysis of Metaheuristic MPPT Performance for Photovoltaic Systems under Partial Shading Conditions

Abstract

Metaheuristic-based maximum power point tracking algorithms are widely used in photovoltaic systems to address nonlinear and multi-peak characteristics under partial shading conditions. However, many reported performance claims rely mainly on numerical simulation and therefore require cautious interpretation. This study presents a simulation-based comparative and epistemological analysis of Particle Swarm Optimization and Differential Evolution for photovoltaic maximum power point tracking. Both algorithms are implemented in an identical buck converter-based photovoltaic framework to ensure fair comparison. Performance is evaluated under uniform irradiance and partial shading conditions using convergence time and tracked power as evaluation metrics. The results show that under uniform irradiance, both algorithms reliably converge to the maximum power point with similar steady-state accuracy, while Particle Swarm Optimization converges faster. Under partial shading conditions, Particle Swarm Optimization consistently tracks the global maximum power point, whereas Differential Evolution shows occasional convergence failure or suboptimal tracking. From an epistemological standpoint, these findings constitute coherent and pragmatically useful model-based knowledge, while remaining provisional due to the absence of experimental validation.