Reusable Launch Vehicle Utilization of Predictive Maintenance to Address Risk Reduction Assessment

In the current age of increased private launches and utilization of reusable launch vehicles, risk due to reuse of flight hardware increases with every flight. Rigorous procedures have been developed based on heritage flight operations to ensure mission and hardware safety needs. These procedures take significant time, money and manpower to recertify flight hardware to be ready for launch. To this end, engineers are investigating ways to supplement these heritage risk mitigation processes using new data-driven techniques. The utilization of Predictive Maintenance (PMx) capabilities has increased across aerospace in the last few years. The core capabilities stem from two key areas: enhancing all aspects of the data pipeline to find new information from existing flight data and leveraging increased access to high end computational assets to capture potential points of failure using a unique combination of models (such as Computational Fluid Dynamic simulations with AI/ML neural nets). By partnering aerospace engineers with computer and data scientists, improvements can be made across the pipeline from data conditioning and handling to modeling to how to supplement current operational risk-reduction processes with this new information. With this in mind, this paper focuses on how applying PMx capabilities to reusable launch vehicle data can turn unscheduled maintenance into scheduled maintenance (to help reduce cost and manpower requirements), capture new information using operational assessment of system-level processes, and provide new insights into aspects of the risk reduction process through correlation of physics-based and data-driven models. It will also present a CFD simulation case study using PMx capabilities built upon the GTRI-developed SPEAR data pipeline applied to propulsion system impingement upon ground support structures.