Validation and Optimization of Thermophysical Properties for Thermal Conductivity and Viscosity of Nanofluid Engine Oil using Neural Network

Abstract

In this study, the thermophysical properties of thermal conductivity and viscosity of a motor oil nanofluid were investigated using experimental data and artificial neural network. NSGA II optimization algorithm was used to maximize thermal conductivity and minimum viscosity with changes in temperature and volume fraction of nanofluids. Also, to obtain the viscosity and thermal conductivity values in terms of nanofluid temperature and volume fraction with 174 experimental data, neural network modeling was performed. Input data include temperature and volume fraction, and output is viscosity and thermal conductivity. Various indices such as R squared and Mean Square Error (MSE) have been used to evaluate the accuracy of modeling in the prediction of viscosity and thermal conductivity of nanofluids. The coefficient of determination R squared is 0.9989 indicating acceptable agreement with the experimental data. In order to optimize and finally results as an objective function, the optimization algorithm is presented and the Parto front and its corresponding optimum points are presented where the maximum optimization results of thermal conductivity and viscosity occur at 1% volume fraction.