Although industrial traceability is more and more digitalized, and concepts such as digital product passports are gaining relevance, current metal-manufacturing practice still rarely connects them into an integrated qualification framework. Traceability remains documentary, while digital twins are typically used for isolated simulation or monitoring tasks rather than for auditable quality decisions. As a result, the link between incoming material variability, process signals, final-part quality, and release decisions remains weak, particularly in foundry environments.
This project implements a process-oriented digital passport in which material identity, process conditions, sensor evidence, and quality-relevant outcomes are connected in a coherent and auditable decision chain. AI methods (ML or DL) are employed to realize a process digital twin as a virtual replica of physical components and processes, enabling high-fidelity modelling, real-time data integration from distributed sensors, and enhanced control over manufacturing precision and efficiency. Digital twins also underpin digital product passports (DPPs), which document the provenance, quality, and lifecycle data of automotive parts, supporting circular economy models by ensuring traceability and sustainability across the product lifecycle . In high-value industrial processes, lack of explainability, weak uncertainty handling, and poor integration with existing workflows still limit deployment. This is particularly relevant when AI outputs influence quality-relevant decisions, process adaptation, or final-part acceptance, where industrial trust depends also on transparency, auditability, and human interpretability.