Improvement of LM12 alloy hybrid metal matrix composites for automotive applications: An in-depth experimental analysis using statistical and predictive ANN approaches

In this study, the mechanical properties and wear behavior of LM12 alloy reinforced with B 4 C and Al 2 O 3 particles were examined. The hardness, yield strength, and tensile strength of hybrid metal matrix composites (HMMCs) were determined experimentally. The sliding wear is studied as per ASTM G99 standard to produce 27 data points based Taguchi design of experiments. Sliding wear experiments are conducted in accordance to Taguchi’s orthogonal array. The experimentally captured wear results are analyzed statistically through ANOVA technique. ANN model is developed and validated with the statistical results. The obtained results revealed that, hardness, yield strength and tensile strength of HMMCs have increased with increase in the weight percentage of Al 2 O 3 and B 4 C reinforcements in HMMCs. ANOVA results indicates that, applied load and % of variation of Al 2 O 3 have significant effect on the sliding wear of HMMCs, whereas distance and velocity have shown least effect. Main effect plots indicate that, the increase of wear parameters have resulted in the increase of wear rate. While, increment of Al 2 O 3% has helped to enhance the wear resistance, thus decreased with wear rate. The presence of hard reinforcements in HMMCs have to form mechanically mixed layer (MML), which is crucial in improving wear resistance of HMMCs. Larger wear debris fragments were observed in HMMC with a lower Al2O3 content (3%), while the fragment size decreased as the Al2O3 content increased. This reduction in debris size may be attributed to plastic deformation occurring before fracture. A Multi-Layer Perceptron (MLP) model was used for result predictions and has recognized a strong correlation within the experimental data, achieving an R² score of 0.99 during training and has consistently predicted wear rate and coefficient of friction values during testing with R 2 score of 0.98. The statistical analysis and ANN model predictions fall within the acceptable limits.