Thermophysical optimization of ND/PG-water nanofluids by NSGA-II coupled with RSM and ANN

Abstract

In this study, an experimental based optimization is performed to reach simultaneous optimized thermal conductivity (TC) and viscosity of ND/PG-Water nanofluid (NF). RSM and ANN modeling were used to simulate the rheological and thermal behavior of mentioned NF. The RSM proposed separate mathematics correlations for TC and viscosity estimation, but the ANN method proposed a comprehensive topology for TC and viscosity prediction. The optimized topology for ANN selected from 400 ANN topologies structures. Based on amounts of coefficient of determination (R 2 ), all behavioral simulations of thermophysical properties by ANN and RSM have high accuracy. The regression coefficients of the RSM model for TC and viscosity are R 2 \u2009=\u20090.9992 and R 2 \u2009=\u20090.9961, respectively. Also, in MLP model, R\u2009=\u20090.9957 and MSE\u2009=\u20090.0015 were obtained. The optimal structure for it has two neurons in the first layer and 5 neurons in the second layer. After prediction of desired parameters, to have a multi-objective optimization, the NSGA-II method was used with the help of RSM and ANN results. According to results, selected simultaneous optimized points to have a suitable viscosity and TC occur at high temperatures but in different volume fractions (VF) that one of them could be selected based on industrials or academic demands.