Fabrication of hybrid aluminum matrix composites and prediction of physical characteristics using artificial network

Abstract

This study looks at how to make LM16 aluminum alloy-based Hybrid Aluminum matrix composites and how to forecast their physical properties (HAMC). The goal of this research is to forecast the Brinell hardness number (BHN) and density of Hybrid Aluminum matrix composites (HAMC). The Levenberg-Marquardt algorithm of a multiple-input multiple-output artificial neural network is used to make predictions (ANN). The HAMC with Nickel coated Graphite (Ni-Gr) and Silicon Carbide reinforcement into the LM16 aluminum alloy is prepared using the stir casting technique considering the brake lining applications. The L9 orthogonal array was chosen for testing to see how process parameters such Nickel coated Graphite (Ni-Gr) reinforcement, Silicon-Carbide (SiC) reinforcement, stirring speed, and stirring time affect the results. The Brinell hardness number (BHN) and density of the material are the output parameters investigated. The determination coefficient (R2) and mean square error were used to assess the ANN's prediction performance (MSE). The average R2 and MSE of 0.99 and 0.0041, respectively, suggest that the ANN fit the experimental experimental response variable values effectively. The correlation coefficient (R) of 0.99721 also supports the findings.