Machine Learning for Propulsion System Health Management
AIR7137
This Aerospace Information Report (AIR) presents considerations specific to machine learning (ML) applied to propulsion system health management (aka EHM), illustrated via examples. These examples are used to highlight concerns and approaches that are unique to EHM, including the typical design space for propulsion systems, terminology, data collection and processing methods, requirements, and characteristics of machine learning models that have been developed and are being implemented.
Rationale: Machine Learning (ML), which is a branch of Artificial Intelligence (AI) involving learning from data to make inferences is being introduced in many fields. The SAE G-34 / EUROCAE WG-114 committee was formed to address concerns and provide guidance related to the implementation and certification of AI and ML technologies in aerospace applications. The concerns, which are associated with the broad consensus that existing development assurance standards may not be appropriate for all AI/ML solutions, have been published in AIR6988 - Artificial Intelligence in Aeronautical Systems: Statement of Concerns. Additionally, G-34/WG-114 has collected aviation specific use cases to be published in the information report, AIR6994 - Artificial Intelligence in Aeronautical Systems: Use Cases, and a new standard is being developed, AS6983 - Process Standard for Development and Certification and Approval of Aeronautical Safety-Related Products Implementing AI. This AIR is intended to supplement the guidance from G-34/WG-114, especially AS6983, for use cases related to propulsion system health management (aka EHM).
