Artificial neural net estimations of gasketed plate heat exchanger performance based on experimental analysis

Abstract

In this study, artificial neural networks (ANNs) are used to estimate the performance of gasketed plate heat exchangers (GPHE) with three different plates. Experimental data is needed to estimate the performance of GPHE and they are obtained from an experimental set-up designed to test plate heat exchangers. [1] Two critical properties, Nusselt number and friction factor, are predicted using ANNs. The results are compared with the correlations obtained from the experimental data for the same plates. The data is divided into two groups; one is for training the network and the other one is for testing the network. Different networks with various numbers of hidden neurons and layers are used to find the best configuration to predict Nusselt number and friction factor. The commonly used feed-forward neural networks based on back propagation algorithm is used in the estimation.