Advanced Soft Computing Tools/Optimization Techniques Used For the Analysis of Metal Matrix Composite Materials

Abstract

The processing methods used to create composites have been examined thoroughly in this article, and artificial neural network models have the potential to replace costly experimental research in a variety of manufacturing processes, including casting procedures. To evaluate the sensor data and optimize the design parameters, one must have a thorough grasp of the interrelationships between the input variables. Regression, artificial neural network (ANN), chi-square automatic interaction detection (CHAID), extreme gradient boosting (XGBoost), and random forest were among the supervised ML models that were trained and tested using 409 experimental data points to map the predictor variables to the response variables. To obtain a uniform distribution of carbon nanotubes in the matrix, the processing methods employed for the composites' production have been evaluated rigorously. Important aspects of the stir casting process include the furnace design, the composites' characteristics, the difficulties in producing composites, and possible research prospects. Investigated were two distinct copper powders, dendritic and spherical, as well as their composites with conventional MoS2, nanometric MoS2, and graphene nanoplatelets.