Research on PV Power Prediction Model Based on Hybrid Prediction

Abstract

A hybrid prediction model based on wavelet transform (WT) -sample entropy (SE) -improved particle swarm optimization (IPSO) -weighted least squares support vector machine (WLSSVM) -iterative error correction is proposed to solve the problem of low accuracy and poor stability of photovoltaic output prediction under grid-connected conditions. Firstly, WT is used to reduce the noise in the collected power signal, and SE is used to quantify the weather type. Then IPSO is used to optimize the main parameters of WLSSVM. Finally, power prediction model and error prediction model are established respectively, and the final prediction power is obtained by superposition of power prediction value and error at all levels. Finally, the proposed model is compared with other prediction models, and the results show that the method has high prediction accuracy.