Comparison of Static and Dynamic Neural Network Models in Predicting Outlet Temperature of Shell and Tube Heat Exchanger

Abstract

This paper presents the comparison of static and dynamic neural network (NN), model to predict the exit temperature of the heat exchangers. Feed forward NN was used as a static network while Time delay NN was used for a dynamic network. Experimental data was collected from a shell and tube heat exchanger to provide sufficient data processing, namely training, test and validation data to develop the models. The static and dynamic network models of the heat exchanger have been developed using Matlab. The performances of the models were evaluated by their statistical validity using the correlation co-efficient and the mean squared error. For time series predictions, the dynamic NN has shown better results than the static NN.