Short-term Load Forecasting Method Based on LSTM-Attention Mechanism

Accurate load forecasting is important to the safe operation of the power system. As the the accumulated load data in power system rapid increasing, the forecasting methods need to deal with massive data information. Thus, a short-term load forecasting method based on LSTM-Attention mechanism is proposed in this paper. Firstly, the correlations among the historical load data, the temperature data and the current load analyzed. And the most correlated data is selected as the input data of the prediction model. Then, a novel short-term load forecasting model based on LSTM-attention mechanism is established. The LSTM layer learns features from the input data, while the Attention layer highlights key features. Finally, the actual data of a certain region in China is taken as an example to verify the feasibility of the model proposed in this paper. The result shows that, compared with other models, the short-term load forecasting model based on LSTM-attention mechanism has the best prediction performance.