Design of Experiments for Process Optimization and Scrap Reduction in High Pressure Die Casting

This thesis focuses on the analysis of process factors that influence the formation of defects in aluminium alloy components produced with the high pressure die casting process, with specific application in the automotive sector. The main objective is to understand which variables have the greatest impact on production scraps rate, in order to optimise the process, improve casting quality and reduce the costs associated with the production of defective parts. The industrial context of this work is characterised by an increasing need for efficiency, repeatability and quality in the production of complex structural components. Die casting, although a well-established process, presents numerous challenges related to defect control, which may manifest as porosity, cracks or surface defects. To achieve its objectives, the thesis began with an initial analysis of historical production data, during which a data-driven scrap monitoring system and a parameter tracking system were developed to closely monitor the process, capturing daily machine parameters and defect types for each component of 2A S.p.A. Following the implementation of these monitoring tools, the attention was focused on components exhibiting higher scrap rates. Based on the insights gained from this initial phase, a controlled experiment was designed and conducted using the Design of Experiments (DoE) methodology for a specific automotive component. The experimental Taguchi DoE allowed for the systematic variation of key process parameters, enabling the isolation of main effects and interactions between different variables. This approach provided a rigorous assessment of each parameter’s impact on defect formation, ensuring statistically significant and repeatable results. The results obtained offer concrete indications for process optimisation: the combination of stable machine parameters, accurate thermal control and balanced lubrication significantly reduces the scraps rate and improves the overall quality of the components. Furthermore, the applied statistical approach makes it possible to develop guidelines for repeatable production tests and to propose targeted intervention strategies in case of deviations from the ideal process. The methodology developed in this work not only enabled an understanding of defect formation in the component studied but also provides a framework that can be applied in the future for further analyses, such as the investigation of mechanical properties, or the study of additional quality characteristics.