Abstract: The heat load forecasting provides data support for urban district heating systems, which is the basis of need-based heating. The change of heat load is greatly influenced by various exterior factors, especially the outdoor temperature. To meet demand of heating system, save energy and balance the comfort of human body, a kind of improved BP neural network method is proposed by temperature and date type. The temperature and date type are quantified and the heat load forecasting model is established by using BP neural network. To guarantee prediction accuracy, the genetic algorithm is used to optimize the weights and thresholds of the neural network, and from which the predicted value of the heat load in the next 24 hours is obtained. The predicted results from the model show that the proposed method can accurately predict the future heat load;and the goals of on-demand heating, energy conservation and environmental protection are achieved.

Key words: heat load forecasting, genetic algorithm, temperature, date type