Nanofluids as coolant in a shell and tube heat exchanger: ANN modeling and multi-objective optimization

Abstract

In the present study, an artificial neural network (ANN) was developed to predict the thermal and hydrodynamic behavior of two types of Newtonian nanofluids used as coolants in a shell and tube heat exchanger (STHE). Inputs of the ANN model are nanoparticle volume concentration, Reynolds number, nanoparticle thermal conductivity, and Prandtl number. Results indicate that the ANN model predicts the experimental data with very high accuracy. Values of Nusselt number resulted from experiments and those obtained from the ANN have at most 9% difference, this value is 9.6% for the pressure drop. Multi-objective optimization was implemented with the aim of minimizing the total pressure drop and maximizing the nanofluids Nusselt number in the STHE according to NSGA-II algorithm. In optimization procedure nanofluids pressure drop and the Nusselt number (tube-side) was evaluated by the ANN model. To find the shell-side pressure drop method of Delaware was employed. Nanofluids concentration and Reynolds number were selected as decision parameters. The Pareto front was obtained. The best solution adopted from points on the Pareto front by two well-known decision-making methods LINMAP and TOPSIS. The Nusselt number of optimal solutions are about 30% greater than the base fluid and pressure drop of optimal solutions are about 10% lower than the base fluid.