Using artificial neural network for investigating of concurrent effects of multi-walled carbon nanotubes and alumina nanoparticles on the viscosity of 10W-40 engine oil

Abstract

The present study used artificial neural networks (ANNs) and experimental data to model the viscosity of the MWCNT (50%)–Al2O3 (50%)/10W40 hybrid nanofluid at different temperatures from 5 to 55 °C and for nanoparticle volume fractions of 0.05% to 1%. An ANN and a new correlation as function of solid volume fraction, temperature and shear rate. The multidimensional MLP-ANN was employed and investigated as the learning algorithm. The R-squared values for the proposed correlation and ANN were obtained to be 0.9973 and 0.9944, respectively. In order to analyze the importance of each term of correlation, p-value of parameters is determined. Also one factor and two factor analysis of viscosity are presented in this study. According to one and two factor analysis results, temperature changes has the highest effect on viscosity. Shear rate and solid volume fraction effects on viscosity are in next level of importance.