Application of neural network model for predicting fouling behaviour of a shell and tube heat exchanger

Abstract

The artificial neural network (ANN) is used as a significant tool in various industrial applications dealing with thermal problems during past few years. The purpose of the present work is application of ANN for heat transfer analysis of a shell and tube heat exchanger subjected to fouling. With the available experimental data, the network configuration was trained and tested. The outlet temperature differences both on shell and tube side and the exchanger efficiency were predicted by using neural network approach. This paper uses cleanliness factor as a tool for investigation of the performance of a heat exchanger due to fouling. Information criteria have been used for selection of relevant input variables and determination of optimal NN model structures. The root mean square error (RMSE) of the predictions in tube and shell side temperature difference and heat exchanger efficiency of 3.1°C, 2.2°C and 7.26% respectively illustrate that the model is adequately accurate. The closeness of experimental results with predicted results justify the correctness of neural network in the heat transfer analysis of a heat exchanger. Hence, it can be recommended that the ANN methodology can be successfully used to predict the performances of a shell and tube heat exchanger in industrial applications.