Supporting Virtual Aircraft Certification via Provenance

Abstract

Exhaustive physical testing ensures the airworthiness of an airplane and is time-consuming and costly. With the latest advancements in computational models and computing power, testing parts virtually becomes feasible. Ensuring the reliability and trustworthiness of such virtual verification processes is a critical point. We collect and assess the obligations and requirements stated by the certification authorities Federal Aviation Administration (FAA) and European Union Aviation Safety Agency (EASA) and from the contributors towards virtual certification. Then, we discuss the possible applications and benefits of using provenance data, which is the documentation of the origins and history of data. By recording detailed metadata, it can support traceability and thus reliability. We propose two provenance models, the first one captures only the workflow between the contributors and their tools. The second one shows the steps of each tool in detail so that the data flow is traceable. These models can be connected by the inputs and outputs of each tool. In the end, the certification authority can use the workflow provenance graph to request the simulation and tool provenance graphs of each contributor. The certifiers have detailed insights into the computational analyses, while the contributors do not reveal their business secrets to each other. Integration of provenance can support regulatory compliance by fulfilling many of the stated obligations and is therefore a promising approach.