Evaluation and prediction of frictional behavior of AA 2024 based hybrid composites using ANN model

Abstract

The paper investigates the frictional behavior of AA 2024 hybrid aluminum matrix composites reinforced with Zirconium dioxide (ZrO2) and Graphite (Gr). The hybrid composite is fabricated by stir casting technique by using fixed 6 wt. % of ZrO2 and varying 1.5, 3.0 and 4.5 wt. % of Gr reinforcement. The test specimens were fabricated in the form of pins to carryout frictional testing by using Pin-on-disc tribometer under dry and lubricated conditions. Tests were conducted at 15 N load corresponding to 2000 m sliding distance and varying sliding speed of 1 m/s, 2 m/s, 3 m/s and 4 m/s. Coefficient of Friction (COF) increased with the increase in sliding speed for both dry and lubricated conditions but hybrid composite with 3.0 and 4.5 wt. % showed a decrease in COF at 4 m/s sliding speed under lubricated conditions. At higher sliding speed, greater amount of frictional heat is developed which leads to the softening of Gr particles and a solid lubricant layer of Gr along with Polyalphaolefin(PAO) oil reduces the metal-to-metal contact and thus reduces the COF. Maximum COF was observed for 1.5 wt. % Gr reinforcement under dry condition whereas under lubricated conditions pure AA 2024 resulted in maximum COF. Artificial Neural Network (ANN) technique was used for predicting the friction behavior and the confirmatory tests were also performed corresponding to sliding speed of 5 m/s. From the investigations it was revealed that the error in predicted and experimental results is in the acceptable range.