Symbiotic Multi-mode Structural Health Monitoring utilizing Physics Informed Neural Networks

Atomic oxygen (AO) erosion and/or oxidation, micrometeorite impact, “space junk” collisions (e.g. the bolt lost during the STS-115 mission), as well as radiation damage incurred from long duration space exposure, are all serious threats to orbiting Missile Defense Agency (MDA) satellite, interceptor, and kill vehicle hardware. Recently, MDA has expressed a high level of interest in developing satellite interceptor and kill vehicle structures with embedded Structural Health Monitoring (SHM) and Damage Prognosis (DP) functionality. In the past, embedded functionality has typically implied “strapping” sub-systems onto an existing superstructure structure. This processes is now inadequate for the stringent mass and space requirements of several high performance MDA vehicles. Therefore a new innovative and highly integrated, symbiotic, SHM/DP paradigm must be developed to preserve vehicle functionality without compromising operational cost and performance. In this Phase I effort, our overall objective will be to demonstrate an integrated, online ‘state awareness’ approach that performs event/impact detection, local mass/stiffness change detection and source differentiation in order to separate stiffness changes (due to softening) from mass changes (possibly due to a parasite), and specialized local anomaly (e.g., crack or delamination) assessment. This approach will be implemented via a multi-tier hierarchy with an eye on utilizing SHM/DP output diagnostics for online flight and mission control algorithm optimization.