Application of ANN for modeling of heat exchanger with concentration as variable

Abstract

Artificial Neural Networks (ANN) are effective in modeling of non-linear multi variable relationships and also referred to as black box models. Generally, for modeling of heat exchangers the various parameters to be taken into account are inlet and outlet temperatures of shell, tube side fluids, and their flow rates. In the present paper, the concentration of flowing fluids is also considered as one of the variable parameters for heat exchanger modeling. For the study three different fluids are used, (i) water, (ii) 20% glycerin and (iii) 40% glycerin. Heat exchanger model is developed using optimized ANN architecture¹. ANN model is trained using a water-water² and water-40% glycerin³ system. The trained networks are then used for prediction of shell and tube side exit temperatures for water-20% glycerin³ system. It is observed that predicted values of water–20% glycerin system are in close agreement (98-99%) with the actual values.