Data Orchestration for AM Digital Twins to Establish Part Provenance

Various critical industries are looking towards additive manufacturing (AM) to produce their complex components as these advanced manufacturing technologies provide benefits in terms of design freedom, material waste reduction, limited tooling requirements, assembly reduction, and a wide range of material processing capabilities. These benefits do not come without their drawbacks. These are primarily related to the complexity of the process in that there are many variables that can be altered or be poorly controlled. The use of sensors and big data to feed into models such as digital twins (DT) are being investigated to manage these complex processes. Component DTs aim at capturing all data that are critical to the efficacy and quality of the component throughout its useful life. The architecture and development of digital systems to realize such DT models and manage their associated data is challenging. Often such datasets and models are in different formats, hosted in different locations, and lack interoperability.

To integrate these data and models, we have developed a SysML V2 information model that allows users to concatenate, register, and query the information related the component in spatial-temporal space. The information model is built on two of our recent standardization efforts, namely, ISO/ASTM DIS 52951 for AM data packages and ISO/CD 10303-238 for model-based manufacturing (STEP-NC for PBF). The NIST AMBench datasets are used to demonstrate the model for a laser powder bed fusion test artifact. Part provenance and traceability is established that spans data from design, process planning, manufacturing, testing, and inspection phases and allows engineers to verify part quality. Product technical data is configuration controlled, packaged for compliance purposes, and can be exported for performing simulations, all controlled and orchestrated by the model.