Long Short-Term Memory for Short Term Load Forecasting with Singular Spectrum Analysis and Whale Optimization Algorithm

Abstract

Short-term load forecasting is necessary for the safety and high efficiency of power system. In recent years, a large number of artificial intelligence-based methods have been used for short-term load forecasting by researchers, among which deep learning methods such as long short-term memory (LSTM) has been proved to be better at load forecasting. However, in the process of constructing LSTM models, data processing and hyperparameter selection tend to depend on experience, resulting in relatively slow modeling speed. This study proposes a short-term load forecasting method with singular spectrum analysis (SSA), LSTM network and whale optimization algorithm (WOA). SSA is used to decompose and reorganize the load sequence, and the WOA is used to optimize the hyperparameters of the LSTM network. The results show that the proposed method performs better than the benchmark in the three counties of the United States used for verification and comparison, that is, SSA-WOA-LSTM can be used to improve the accuracy of the short-term load forecasting.