Prediction short-term photovoltaic power using improved chicken swarm optimizer - Extreme learning machine model

Abstract

Photovoltaic power generation is greatly affected by weather conditions while the photovoltaic power has a certain negative impact on the power grid. The power sector takes certain measures to abandon photovoltaic power generation, thus limiting the development of clean energy power generation. This study is to propose an accurate short-term photovoltaic power prediction method. A new short-term photovoltaic power output prediction model is proposed Based on extreme learning machine and intelligent optimizer. Firstly, the input of the model is determined by correlation coefficient method. Then the chicken swarm optimizer is improved to strengthen the convergence. Secondly, the improved chicken swarm optimizer is used to optimize the weights and the extreme learning machine thresholds to improve the prediction effect. Finally, the improved chicken swarm optimizer extreme learning machine model is used to predict the photovoltaic power under different weather conditions. The testing results show that the average mean absolute percentage error and root mean square error of improved chicken swarm optimizer - extreme learning machine model are 5.54% and 3.08%. The proposed method is of great significance for the economic dispatch of power systems and the development of clean energy.