Using CAD feature semantics to enable learning of AM capabilities

Additive Manufacturing (AM) is currently being utilized for manufacturing of more complex parts but its use of metal AM has been limited owing due to high costs. The capabilities of AM vary with process, geometry and material selection thus leading to a complex combination of parameters with successful as well as unsuccessful builds. Each unsuccessful build increases cost and hence reduces the reliability of metal AM which can be avoided with a knowledgebase of known prior failures. To achieve this goal of warning about potential problems in design, we introduce a CAD feature based semantic framework with the Adaptive Query Algorithm (AQuA). This framework enables the capture of information about AM builds as well as infers any potential similarity of design being evaluated with known problems in knowledgebase.