A route for qualifying/certifying an affordable structural prognostic health management (SPHM) system

There is a growing interest in developing affordable SPHM systems that use artificial intelligence (AI) techniques and existing flight parameters to track how each individual aircraft is used and to quantify the damaging effects of usage. Over the past four years, Smiths and BAE systems have launched collaboration work to evolve a practical SPRM system. The collaborative work has built on BAE systems vast experience of operational load monitoring (OLM) that has spanned more than 30 years. The collaborative work has also built on the unique experience of Smiths over the past 20 years that has produced automatic data correction algorithms, mathematical networks (MNs), dynamic models and flight and usage management software (FUMS™). Smiths and BAE systems have been also carrying out extensive investigations that lead to establishing and agreeing a route for qualifying/certifying AI based systems, an essential work element to avoid a delay in introducing SPHM into service use. The investigations have covered assessing the adequacy and quality of the truth data required to train/configure an AI method. They have also addressed the regulatory authority (RA) concern that any volume of data gathered to train an AI method would not capture the truth and some operations/configurations may produce novel data outside the training data. Guidelines for management approaches based on individual aircraft tracking (IAT) have been also investigated. The paper presents the results of the Smiths and BAE systems collaborative work and presents preliminary guidelines for qualifying/certifying AI methods.