Modeling and prediction of tribological properties of polyetheretherketone composite reinforced with graphene and titanium powder using artificial neural network

This present research deals with the prediction of tribological properties of polyetheretherketone (PEEK)-based composites filled with different wt% of graphene (C) and titanium powder (Ti) using an artificial neural network (ANN). The PEEK/C/Ti composite was produced by injection molding at 400°C for applications such as shock absorber bearings, gears for oil and gas companies, seals, cams, ball bearing cages, and hybrid bushings in high-pressure diesel fuel injection pumps. A wear test was performed on a polymer composite using a pin on a disk machine against a steel disk of Rc 60 for varying loads and sliding velocities. A lower wear rate was obtained for the composite of PEEK with 10 wt% C and 5 wt% Ti powder. Prediction by ANN indicated that a more optimal composition of PEEK with 10 wt% C and 6 wt% Ti powder is estimated. 3D plots for wear rate and coefficient of friction related to filler fraction and load were established. ANN accuracy is measured by the coefficient of determination B and its value for prediction of specific wear rate is 92.6% and coefficient of friction is 98.3%.