A Power Prediction Method for PV system Based on Wavelet Decomposition and Neural Networks

Abstract

High penetration of Photovoltaic (PV) systems is variable resource as challenges to the stability and power quality of electrical grids. Accurate prediction of PV power has been recognized as a way to solve this problem. Due to PV power periodicity and non-stationary characteristics, traditional power prediction methods based on linear or time series models are no longer applicable. Methods based on neural networks are widely used for power prediction for PV system. Considering the training time and network accuracy, this paper discusses the factors in the prediction model. The number of nodes in the hidden layer that the network accuracy is high and the training time is guaranteed is determined, rather than just using the formula. After determining the number of nodes in the hidden layer, this paper presents a method combining artificial neural network (ANN) and wavelet decomposition (WD) for power prediction for PV system. Solar irradiance and other six parameters are chosen as the input of the hybrid model based on WD and ANN. The output of the neural network is reconstructed to obtain the final predicted power. The proposed model are validated by experimental data to predict the output power of PV system effectively and accurately, which is useful to enhance the safety and stability of the electrical grid.