Optimized prediction of solar irradiation based on MPC and ELM neural network

Abstract

Based on the characteristics of randomness and instability of solar radiation, this paper presents an optimal prediction algorithm of solar radiation based on Model Predictive Control (MPC) and Extreme Learning Machine (ELM). Firstly, the meteorological factors directly related to solar radiation are selected to determine the input attributes; secondly, the weather forecast and historical irradiance information at the current time are used as inputs, and the irradiance at the current time is used as the output to predict the solar irradiance; finally, the MPC rolling optimization idea is used to realize the optimal processing of solar radiation prediction data. The simulation results show that MPC-ELM algorithm has smaller simulation error than other algorithms, which shows that the algorithm has more advantages.