Experimental study and computational intelligence on dynamic viscosity and thermal conductivity of HNTs based nanolubricant

Abstract

Purpose This paper aims to evaluate the dynamic viscosity and thermal conductivity of halloysite nanotubes (HNTs) suspended in SAE 5W40 using machine learning methods (MLMs). Design/methodology/approach A two-step method with surfactant was selected to prepare nanolubricants in concentrations of 0.025, 0.05, 0.1 and 0.5 wt%. Thermal conductivity and dynamic viscosity of nanofluids were ascertained over the temperature range of 25–70 °C, with an increment step of 5 °C, using a KD2-Pro analyser device and a digital viscometer MRC VIS-8. Additionally, four different MLMs, including Gaussian process regression (GPR), artificial neural network (ANN), support vector machine (SVM) and decision tree (DT), were used for predicting dynamic viscosity and thermal conductivity by using nanoparticle concentration and temperature as input parameters. Findings According to the achieved results, the dynamic viscosity and thermal conductivity of nanolubricants mostly increased with the rise of nanoparticle concentration in the base oil. All the proposed models, especially GPR with root mean square error mean values of 0.0047 for dynamic viscosity and 0.0016 for thermal conductivity, basically showed superior ability and stability to estimate the viscosity and thermal conductivity of nanolubricants. Practical implications The results of this paper could contribute to optimising the cost and time required for modelling the thermophysical properties of lubricants. Originality/value To the best of the author’s knowledge, in this available literature, there is no paper dealing with experimental study and prediction of dynamic viscosity and thermal conductivity of HNTs-based nanolubricant using GPR, ANN, SVM and DT.