Optimal prediction of process parameters by GWO-KNN in stirring-squeeze casting of AA2219 reinforced metal matrix composites

Abstract

Abstract Rapid globalization and demand for the advanced material, hybrid materials are possessing an up marketed strategy in the field of material research. Due to the demand for lightweight and strong nature of materials, aluminium is considered as largely utilized material for manufacturing components in industrial and other manufacturing sectors. However, aluminium materials are being researched through the decades for improvising the mechanical properties. Therefore, hybrid matrix composites are introduced to improve the mechanical properties by the addition of reinforcements. One of the recent advancements in hybrid metal composites is the introduction of nanoparticles in the composition mixture. In this research work, an aluminium hybrid metal matrix composite is prepared by the nano reinforcement particles. Aluminium is considered to be an efficient element for manufacturing and numerous level of hybridization is carried out in the metal processing. In this current research work, aluminium alloy 2219 is hybridized with nanoparticles of Titanium Carbide, Aluminium Oxide and Silicon Nitride with the proportion of 94% of the primary material with 6% of reinforcement material. The reinforcements are fabricated by the conventional processing of advanced method of casting process namely stirring-squeeze casting process for manufacturing the proposed reinforced material. The advantage of this squeeze casting process, it can limit the defects than conventional casting process. The fabrication process develops eight sets of cylindrical specimens of dimension 50 × 150 mm for testing under varied melt and die temperatures which are maintained at 800 °C and 400 °C and squeeze pressure of 100 Mpa during each level of fabrication. Mechanical testing is carried out after the fabrication process and tested the properties of prepared specimens under loading conditions. Tensile testing and hardness of the material are tested on universal testing machine and Brinell hardness with intender ball of 5 mm diameter under 250 kg loading condition. The statistical method of analysis is one of the easiest way analysis the error rate of the result obtained in the experimentation through machine learning process. Therefore, this research develops a statistical analysis method for optimizing the process parameters for better improvement in mechanical properties. The K-nearest neighbour (KNN) and grey wolf optimization (GWO) are proposed for optimizing the process parameters of stirring-squeeze casting process. Conclusively, the result concludes with a comparative study of experimental and predicted outcomes. From the result of comparison, the proposed composition of material has highest level of mechanical properties for manufacturing components in a real-time scenario respectively.