Optimization of MWCNTs (10%) - Al2O3 (90%)/5W50 nanofluid viscosity using experimental data and artificial neural network

The present study aimed at predicting the viscosity of MWCNT (10%)-Al2O3 (90%) hybrid nanofluid in the lubricant 5W50 using artificial neural networks (ANNs). Artificial neural network modeling was performed in the temperature range of 5-55 °C at different volumetric fractions of 0.05 to 1% using 174 experimental data. Three inputs and one output (viscosity) were determined for ANN. The MLP-artificial neural network (MLP-ANN) was used in this study. The results showed the accuracy and reliability of the proposed model according to criteria of R-squared and mean square error. According to the results of one factor and two factor analysis, temperature variation had the greatest effect on viscosity in comparison to shear rate and solid volume fractions. Also a correlation was proposed to predict the viscosity of the hybrid nanofluid with 5W50 fluid as its base fluid in terms of temperature, solid volume fraction and shear rate. The R-squared values for the new correlation and ANN were 0.99703 and 0.9998, respectively.