Advancing photovoltaic system design: An enhanced social learning swarm optimizer with guaranteed stability

Abstract

Parameter estimation of photovoltaic (PV) models, mathematically, is a typical complicated nonlinear multimodal optimization problem with box constraints. Although various methodologies have been explored in the literature, their performance tends to be unstable owing to inadequate adaptability. In this paper, an enhanced social learning swarm optimizer (ESLPSO) is developed to achieve more reliable parameter estimation in PV models. Firstly, using the non-stagnant distribution assumption, we obtain a sufficient and necessary condition to guarantee the stability of the basic social learning swarm optimizer (SLPSO). Secondly, a nonlinear control coefficient is introduced to balance convergence and diversity. Finally, an interactive learning mechanism is devised to preserve population diversity. The efficacy of ESLPSO is validated using three extensively applied PV models and several scalable optimization problems. Statistical outcomes highlight the robustness and competitiveness of ESLPSO compared to other state-of-the-art methodologies.