Kinematic Viscosity Prediction of Nanolubricants Employed in Heavy Earth Moving Machinery Using Machine Learning Techniques

Abstract

Recent researchers widely used nanoparticle additives for improving thermal and rheological properties of machine lubricant. In present study the effect of Al2O3 and CeO2 nanoparticles on transmission oil (SAE30), hydraulic oil (HYDREX100) and gear oil (EP90) of heavy earth moving machinery is investigated. Nano-lubricant samples are prepared in 0.01–4% nanoparticle volume fraction range. Four machine learning techniques namely decision tree (DT), random forest (RF), generalized linear models and neural network (NN) have been used to predict the kinematic viscosity for Al2O3 and CeO2 nanolubricants. Further, multi-criteria decision-making technique named technique for order of preference by similarity to ideal solution have been used to find the best predictive method in each category of the nanolubricants. DT, RF and NN methods are found to be most accurate in kinematic viscosity prediction of transmission oil (R2 = 0.861), hydraulic oil (R2 = 0.971) and gear oil (R2 = 0.973), respectively.