Integrated Forecasting Models Based on LSTM and TCN for Short-Term Electricity Load Forecasting

Abstract

Electricity load forecasting is an important prerequisite for ensuring the stability and reliability of regional power systems. Researchers have proposed many combined forecasting models, but most of them cannot capture the global characteristics of the data properly. To further improve the accuracy of short-term power load forecasting, this paper proposes a combined forecasting model based on long short-term memory (LSTM) and temporal convolutional network (TCN). For the electricity load data, the LSTM forecasting model and TCN forecasting model are first established, and then the output results of LSTM and TCN are weighted together according to the inverse squared error ratio to obtain the combined LSTM-TCN forecasting model. The LSTM-TCN model has more advanced model performance and its error is significantly lower than that of the single forecasting model and other classical network models. The results show that the LSTM-TCN model has higher accuracy in short-term load forecasting.