BP network predictive model structure for heat supply network

Artificial Neural Network (ANN) is used in heat supply network real-time prediction. The problem of structure design for BP network predictive model is discussed. Back Propagation Network (BPN) is adopted to establish heat supply network real-time predictive model. This model is trained and tested with part of the data from a boiler plant in Xihailin district, Mudanjiang, from Nov. 2000 to Apr. 2001. The result shows that the model has preferable ability of dynamic tracking and predicting.