Hybrid Modeling Method for Predicting the Cooling Water Outlet Temperature in Plate Heat Exchanger

The circulating cooling water system which ensures the production process running safely plays a very important role in the iron and steel enterprise and its energy-saving operation is based on the accurate prediction of the cooling water outlet temperature in the plate heat exchanger. A hybrid modeling method is proposed in this paper to predict the cooling water outlet temperature in the plate heat exchanger. By analyzing the operational mechanism of the heat transfer process in the plate heat exchanger, the mechanistic model is established according to the thermal transfer equation and the thermal balance equation and then the radial basis function neural network (RBFNN) is applied to develop the hybrid model with the mechanistic model, in which the RBFNN corrects the error made by the mechanistic model due to the theoretical assumption and the inaccurate value of parameter. Finally the simulation is made to verify the modeling method and the results show that the hybrid model has the better prediction performance than the mechanistic model, which lays the foundation for the operation optimization of the circulating cooling water system.