Using Formal Methods to Help Explain Real-Time Operator Support in Metal Additive Manufacturing

Metal Additive Manufacturing (MAM) offers significant advantages in design flexibility and material efficiency, yet industrial adoption is limited by process instability and the complexity of integrating machine learning. To bridge this gap, the I-Form Centre for Advanced Manufacturing has developed a novel recommender system that represents MAM processes as a real-time knowledge graph. This system leverages formal methods, specifically, logics, to express operator knowledge about a manufacturing process. These rules are used to reason about the state of an in-process knowledge graph and generate actionable operator recommendations. In this way, the system translates complex data-driven insights into the operator’s domain vocabulary, ensuring that the logic is both deterministic and trivially explainable. This approach builds operator confidence and removes the need for deep data science expertise.