Analysis of rheological properties of MWCNT/SiO2 hydraulic oil nanolubricants using regression and artificial neural network

Abstract

In this article, the rheological behavior of MWCNT/SiO2 based nano-hydraulic oil nanolubricant is evaluated using experimental and Artificial Neural Network (ANN) approach. Viscosities of the hybrid nanolubricant samples were measured at temperature and shear rate range of 10–80 °C and 10–200 s−1 respectively. A new regression model is being proposed to predict the dynamic viscosity of nanolubricants. The proposed regression model (R2 0.98338–0.99583) predicts the viscosity of nanolubricants closer to experimental results (least deviation 2.62%). Consistency index (m) and power law index (n) values reveal that nanolubricant samples are non-Newtonian fluid with shear thinning behavior. To improve the accuracy in predicting the viscosity of nanolubricants, the ANN model was designed having input variables among temperature, solid volume fraction and shear rate. In the first phase, temperature and solid volume fraction were taken as input variables, and in the second phase shear rate was introduced as an additional input parameter. The entire data was split into 70:30 proportions for the training and testing phases of the ANN model. The testing results of ANN revealed better accuracy than the proposed correlation in terms of average values of Root Mean Square Error (RMSE) and R2.