Artificial Intelligence Optimization of Turning Parameters of Nanoparticle-Reinforced P/M Alloy Tool

Abstract

In this research, the powder metallurgy- (P/M-) based metal matrix composites were prepared to compose the machinability characteristics. Therefore, the Al2024 and boron carbide (B4C) were the base and strengthened reinforcements. During the powder metallurgy process, weight fractions of boron carbide are 4, 8, and 12%; the compaction pressure is 300 to 400 MPa, and the sintering temperature is 420 to 540°C, respectively. These parameters were planned with Taguchi L9 array for achieving the proper design. After the processing, composite specimens were utilized to conduct the turning process. For all the nine specimens, depth of cut, speed, and feed rates were maintained constant with optimal parameters. The surface roughness and material removal rate responses are successfully achieved in the optimal turning parameters. Then, the artificial neural network (ANN) model was implemented to analyze the predicted values with back propagation algorithm. In this ANN, three input layers, 6 and 4 hidden layers, and two outputs were created as per the design. Finally, the minimized surface roughness and maximized material removal rate were achieved at the process parameters like 8 wt. % of boron carbide, 300 MPa of compaction pressure, and 480°C of sintering temperature. All the predicted values are slightly maximum than the experimental values.