Influence of different parameters on the rheological behavior MWCNT (30%)-TiO2 (70%) / SAE50 hybrid nano-lubricant using of response surface methodology and artificial neural network methods

Abstract

In this study, the effects of different parameters as the volume fraction of nanoparticles (φ), temperature, and shear rate (γ̇) on viscosity (μnf) of MWCNT (30 %)-TiO2 (70 %)/SAE50 hybrid nano-lubricant were investigated. The temperature is the most effective parameter while γ̇ has less influence on μnf. RSM and ANN methods are used for the prediction of MWCNT (30 %)-TiO2 (70 %)/SAE50 hybrid nano-lubricant. Results reveal that a network with 2 hidden layers with 5 neurons and a logistic sigmoid transfer function for the first layer, and 2 neurons with a tangent sigmoid transfer function for the second layer has a minimum error and maximum efficiency. Also, using statistical regression analysis considering training and test data (R = 0.9999) and comparison of ANN estimated values with empirical data have shown the good capability of μnf prediction by improved ANNs (Mean square error of 1.3367 and mean absolute error of 0.7663). By comparing the two presented models, it was found that the ANN model can predict the viscosity of the investigated nanofluid better than the RSM model.