CNN-Wavelet-Transform-Based Model for Solar Photovoltaic Power Prediction

Abstract

Solar power is one of the abundant renewable energy sources. But the power generation capacity of photovoltaic power plants fluctuates significantly due to changes in weather conditions. In this paper, a new model is proposed, which consists of convolutional neural network, wavelet transform and support vector machine (CWS). Firstly, the features of the original data are expanded through the convolutional neural network (CNN). And then the wavelet transform is introduced to suppress the noise in the expanded data. Finally, the output power of the photovoltaic power station is predicted by the support vector regression (SVR) method. The experimental results show that the prediction accuracy and training time of the new model show obvious advantages compared with the previous BI-LSTM (Bidirectional Long Short Term Memory), LS-SPP (LSTM-Based Solar Power Prediction) and LSTM under different prediction time ranges.