Research on ultra-short-term heating load forecasting based on HNN and SDA

Accurate forecasting of heating load can help central heating systems solve the problems of energy waste and source-load mismatch in their long-term operation. On this basis, this paper proposes an ultra-short-term heating load forecasting method based on hybrid neural network (HNN) and similar day approach (SDA), which improves the ability of neural network models to extract data features and enhances the quality of input training sets. Using heating data from a combined heat and power (CHP) unit in Shandong Province, the experimental results show that compared with single convolutional neural networks (CNN), Transformers, long short-term memory (LSTM) neural networks, CNN-LSTM and other models, the HNN model proposed in this paper has higher accuracy in heating load forecasting. Meanwhile, the introduction of SDA improves the forecasting accuracy of the neural network model and shortens the training time of the neural network model.