Load Prediction of Photovoltaic System Based on Transfer Learning and LSTM Network
Due to the influence of solar irradiance,temperature and some random factors,the photovoltaic(PV)power genera-tion has strong intermittency and volatility,so it is difficult to make accurate PV power prediction.In order to improve the ac-curacy of the trained PV load prediction model,a PV system load prediction method based on transfer learning and LSTM net-work is proposed.The transfer learning method based on parameters is selected,and considering that the information extracted by neural network becomes more proprietary as they go to higher levels,a transfer method with fixed first-layer parameters is proposed for PV-DT based on LSTM.PV system load forecasting results show that the proposed method can accurately predict the normal operation conditions of PV power generation,and detect fault existing in the PV system,so that the maintenance staff can take corresponding measures in a short time to minimize the power loss.It may improve the performance of the PV system.
万方数据
维普
