Feasibility of least-square support vector machine in predicting the effects of shear rate on the rheological properties and pumping power of MWCNT–MgO/oil hybrid nanofluid based on experimental data

Abstract

The main objective of the present paper was to investigate the feasibility of the least-square support vector machine (LSSVM) in predicting the effects of shear rate on the dynamic viscosity of a hybrid oil-based nanolubricant containing MWCNT and MgO nanoparticles in different solid concentrations and temperatures. Firstly, measuring the dynamic viscosity of the nanofluid revealed that the nanofluid is a non-Newtonian fluid at the temperatures of 10 °C and 20 °C in all the studied shear rates and solid concentrations while it showed Newtonian behavior at the rest of the studied temperatures. Then the effects of solid concentration and temperature on the dynamic viscosity have been experimentally studied, and it is found that the dynamic viscosity increased as the solid concentration increased; the maximum increase has been observed at the solid concentration of 1.5% and temperature of 60 °C by 52 vol.%, while the minimum increase has been observed at the solid concentration of 0.125 vol.% and temperature of 10 °C by 11%. Based on the experimental data, a new correlation to predict the dynamic viscosity of the nanofluid in terms of shear rate, solid concentration, and the temperature has been proposed. Then, the LSSVM has been employed to predict the dynamic viscosity behavior of the nanofluid considering the shear rate, temperature, and solid concentration as the input variables and the dynamic viscosity as the output variable and the results showed the excellent capability of the LSSVM in predicting the dynamic viscosity. Finally, the effects of adding the hybrid nanoparticles on the pumping power have been studied.