Research on Short-Term Electricity Load Forecasting Model Based on SWT-ResNet-LSTM

Short term electricity load forecasting can timely assess the overall power load changes in a certain area, which is crucial for the safe and stable operation of modern power systems. This paper proposes a short-term load forecasting model that combines Stationary Wavelet Transform (SWT), Residual Network (ResNet), and Long Short-Term Memory Network (LSTM). Firstly, the historical load and weather factors are selected as the original data, where the load data is decomposed into multi-level wavelet components by SWT, which are added to the input feature sequence. Then, ResNet-LSTM model channels are established respectively to predict the wavelet components. Finally, the load sequence is obtained by wavelet reconstruction. Compared with LSTM, SWT-LSTM, ResNet-LSTM and other networks, the model proposed has certain advantages in the field of short-term load forecasting research, and the prediction accuracy is the highest among all models.