Research on prediction of photovoltaic power generation based on SSA-BP

Abstract

Accurately and timely predict photovoltaic power generation is a major challenge for photovoltaic power generation. This study proposes a deep fusion algorithm of sparrow algorithm (SSA) and basic neural network (BP) to improve the prediction ability of photovoltaic power generation. The core idea is to use the powerful global search ability of SSA to optimize the key initial parameters (mainly weights and thresholds) of BP neural network, so as to overcome the shortcomings of traditional BP neural network, such as easy to fall into local minimum, sensitive to initial values and slow convergence speed. The experimental results show that SSA-BP photovoltaic power generation has achieved good results in both normal weather and abrupt extreme weather, and its effect is better than that of traditional BP, PSO-BP, PSOEM-BP and other prediction methods, especially in abrupt weather, its adaptability is more significant, which provides a strong support for the safe and efficient production of photovoltaic power generation.