Data-driven multivariable monitoring and control of critical parameters in Vacuum High Pressure Die Casting towards zero-defect structural components

Vacuum High Pressure Die Casting (V-HPDC) is widely used for the production of aluminium structural components requiring high mechanical performance and energy absorption under crash conditions. Compared to conventional HPDC, the V-HPDC process reduces defects such as gas entrapment and improves casting integrity. However, current monitoring approaches are typically limited to conventional parameters, including vacuum pressure, injection profiles, and thermal conditions of the die, providing only partial insight into the actual process performance. This work proposes an advanced data-driven digitalization approach for V-HPDC process monitoring, focusing on the vacuum evacuation phase. The methodology integrates synchronized measurements of vacuum pressure in both the system and cavity, plunger position, extracted air mass flow using a thermal mass flow meter, and thermocouples in contact with the die cavity. This multivariable strategy enables a quantitative characterization of air evacuation throughout the injection cycle, as well as the evaluation of cavity vacuum conditions and filling dynamics through metal front detection.

The direct measurement of extracted air flow, combined with cavity pressure data, introduces a new dimension to process diagnostics. By comparing the evacuated air volume with the theoretical cavity volume, leakage phenomena in the die or vacuum system can be identified and quantified, while deviations may indicate insufficient vacuum generation. Additionally, time-resolved airflow signals provide insight into transient filling events, including correlations between cavity vacuum levels and die leakage, as well as characteristic signatures such as airflow peaks associated with final filling stages or die opening phenomena. The proposed methodology offers a comprehensive digital representation of the vacuum evacuation process, enhancing process diagnostics, die integrity assessment, and monitoring capabilities. Future work will focus on analyzing the stability of these variables using large industrial datasets and applying artificial intelligence techniques to correlate process parameters with operational efficiency and defect formation, enabling the development of predictive and intelligent control strategies.