Quality prediction of semi-solid die casting of aluminum alloy in terms of machine learning

Semi-solid die casting of aluminum alloy has been successfully employed to manufacture high-performance components with precise net shapes. However, the quality of these components is highly susceptible to variations in both environmental conditions and process parameters, leading to a narrow process window that restricts its widespread application in engineering. In this study, a machine learning (ML) model has been developed to identify defective products through the detection of injection pressure, thereby providing a foundation for monitoring and further optimizing the manufacturing process. Among various ML algorithms, the Multilayer Perceptron (MLP) is the most effective for overall quality prediction. Additionally, the mechanism for identifying defect types based on pressure curves has been revealed: the filling pressure at the gate entrance has been found to exhibit a strong correlation with the internal quality of the casting, while the V-P transition point has been identified as a reliable indicator of the external quality.