A data-based certification approach for additively manufactured metal aircraft components

Abstract The aviation industry is changing towards more sustainable production, service and maintenance of aircraft. Additionally, more and diverse aircraft are developed, especially considering UAVs and VTOLs. This results in more diverse components. Additive manufacturing (AM) comes in handy to tackle those challenges. Due to the various parameters which influence the component’s properties, the qualification process is very complex. No specific qualification is available among the existing AM part integrators. This paper proposes a data-driven component evaluation process for the certification of aerostructures, considering all available data i.e., design, manufacturing and post-treatment data. Machine learning (ML) algorithms are used to predict the physical properties of components based on the data generated by monitoring their production. Necessary training data comprises non-destructive (NDT) and destructive tests (DT), whereas the quality assurance (QA) in a production environment works just with NDT data. Here a platform is presented that implements the processes necessary to guide a user through the certification process of an AM component made of AlSi10Mg, produced using laser powder bed fusion (LPBF).