The dynamic properties modeling of a heat exchanger based on improved Elman neural network

Improved Elman neural network was used to establish a dynamic properties model on analyzing nonlinear characteristic of a heat exchanger. The heat-medium and cooling-medium were input of the neural network and the difference between the real output and the networkpsilas output was treated as self-feedback. In order to train the network, matrix formed back-propagation learning algorithm was used to adjust weights. Compared with the real data, the simulation results indicated that dynamic characteristic of system can be achieved accurately and avoid the simple linear mapping. At the same time, the method can provide a new dynamic modeling method for unknown nonlinear system.