Photovoltaic Power Prediction Model Based on Parallel Neural Network and Genetic Algorithms

Abstract

With the wide application of large-scale photovoltaic systems, photovoltaic power prediction can reduce the negative effects caused by the intermittency and randomness of output power for photovoltaic system. This paper proposes a novel photovoltaic power prediction model based on parallel back propagation neural network (BPNN) and genetic algorithms to predict output power, whose input parameters are historical power output data, historical meteorology data, and meteorology data of the objective day. A parallel BPNN algorithm based on MapReduce is proposed to establish a mapping relationship between input and output through studying large amounts of training sample data. Furthermore, a parallel genetic algorithm based on MapReduce is proposed to optimize BPNN initial weights and thresholds. Experiment results show that the proposed model with parallel BPNN and genetic algorithms can significantly improve prediction accuracy and speed, compared with traditional photovoltaic power prediction model.