LSTM-TCN ensemble learning based photovoltaic power prediction

With the increasing importance of renewable energy sources in energy production, solar photovoltaic power has gained widespread utilization as a clean and sustainable energy form. However, the inherent uncertainty and variability associated with photovoltaic power pose a challenging issue in terms of predictability. To improve the prediction performance, a LSTM-TCN ensemble learning is proposed. This method leverages the advantages of LSTM and TCN models in time series modeling and feature extraction through collaborative training, thereby providing more precise photovoltaic power prediction under varying meteorological conditions. Experimental results demonstrate that the proposed approach reduces the RMSE relative to LSTM and TCN by 0.083 and 0.047, showcasing the effectiveness and superiority. This research holds significant theoretical and practical implications for enhancing the reliability of renewable energy sources and the management of power systems.