Genetic algorithm multiobjective optimization of a thermal system with three heat transfer enhancement characteristics

Abstract

A heat transfer enhancement system including CuO/water nanofluid in a corrugated tube equipped with twisted tape was modeled by two well-known artificial neural network techniques. The multilayer perceptron and group method of data handling neural networks were employed to predict thermal-hydraulic characteristics as functions of main operating conditions. In addition, the genetic algorithm (GA) approach was used to develop applied empirical correlations. The purpose of the models is to estimate Nusselt number (Nu) and friction factor (f ) in the investigated heat exchanger. The main effective parameters investigated in this study are volume fraction of nanoparticle, twist ratios of twisted tape, and Reynolds number. According to the conflicting relationship between heat transfer and pressure drop, the more accurate model was selected as the objective functions for multiobjective optimization by GA. The optimum operating conditions of the investigated heat exchangers that lead to a trade-off between Nu and f were proposed.