Experimental Investigation and Machine Learning Techniques on Tribological Characteristics of Blend of Coconut and Mustard Oil Based Nano-lubricant.

Vegetable oil-based lubricants can substitute mineral oil-based lubricants for environmental cleanliness. In this paper, nano-lubricant containing blend of coconut and mustard oil added with CeO₂ nano-particles was developed. Tribological characteristics, viz., coefficient of friction and specific wear rate of nano-lubricants by varying concentration of nano-particles, blend ratio, load and speed, were measured using a pin-on-disc tribometer. Furthermore, three different MLMs, such as random forest (RF), support vector machine (SVM), and feed forward neural network (FFNN) were used for predicting tribological characteristics. The results of model performance statistics showed that RF, SVM and FFNN have greater ability and stability to regress tribological properties of nano-lubricants effectively. The most influential parameter in predicting wear has been found to be concentration of nano-particles, which occupies 60.78% of coefficient of friction and 88.78% of specific wear rate. The study could help to optimize the cost and time required for modelling tribological properties of lubricants. [ABSTRACT FROM AUTHOR]