Improved particle swarm optimization algorithm for photovoltaic system under local shading

Abstract

ABSTRACT To solve the prediction accuracy problem of the commonly used photovoltaic power generation model in industry, the theoretical model of photovoltaic power generation is improved based on the experimental data in this paper. Furthermore, the influences of the photovoltaic module output characteristics and local shading on the power generation efficiency of series photovoltaic modules were analyzed using MATLAB simulations. Finally, to solve the maximum power point tracking problem of the traditional maximum power point tracker algorithm under local shading, an improved particle swarm optimization (PSO) algorithm based on a linearly decreasing inertia weight w and a learning scale factor δ is proposed in this paper. The simulation results showed that the proposed algorithm tracked the power 1.45 times faster than the PSO algorithm, had 5.8 and 5.09 times more power stability at the working points than the PSO and perturb and observe (P&O) algorithms, respectively, and had 1.12 and 1.25 times more power at working points than the PSO and P&O algorithms, respectively.