Experimental Optimization of Mechanical Performance in T6 Heat-Treated Aluminum Alloy A380 with Machine Learning-Based Prediction of Specific Wear Behavior

This study investigates the influence of quenching temperature, aging temperature, and time on the mechanical properties of gravity die-cast A380 aluminum alloys subjected to T6 heat treatment. Samples were aged at 175 degrees C and 275 degrees C for different time periods, with quenching performed at 20 degrees C and 80 degrees C. The effects of heat treatment on microstructure, tensile strength, wear behavior, and hardness were evaluated. The results demonstrated that ultimate tensile strength increased substantially from 165 MPa to 275 MPa at an aging temperature of 175 degrees C, whereas a higher aging temperature of 275 degrees C did not provide any additional improvement. In contrast, hardness showed a strong dependence on aging duration at 175 degrees C, rising from 108 HV to 152 HV. However, at 275 degrees C, hardness declined significantly, reaching as low as 80 HV. The highest wear resistance was achieved in the samples quenched at 20 degrees C and aged at 275 degrees C for 6 hours, exhibiting an 81.2% improvement compared to the untreated sample. These findings provide valuable insights for optimizing the heat treatment process of A380 aluminum alloys in industrial applications, particularly in improving mechanical properties for automotive and aerospace components. The specific wear rate was estimated using various machine learning algorithms within the framework of data-driven approaches using the obtained experimental data. Model performances were compared in training, test, and validation sets; Random Forest and Deep Neural Network models stood out with their high accuracy and generalization abilities. While the Random Forest model showed low error rate and high stability, the Deep Neural Network model attracted attention with the lowest MSE (approximate to 1.7) and highest R2 score (approximate to 0.85), especially in the test set.