System Health Monitoring for Deep Space Electrical Power Systems

NASA is interested in advanced techniques to provide fault diagnosis for increasingly complex space electrical power systems requiring higher levels of autonomy. Okean Solutions proposes to provide system health monitoring including fault detection and isolation/identification (FDI) to enable reliable and resilient space power. The software solution will support autonomous recovery and reconfiguration of space power systems. Our model-based FDI system called MONSID® determines the health state of power system components so appropriate action can be taken whether the system is healthy or not. It can be applied to monitor any electrical power system, including but not limited to Gateway, Deep Space Transit Vehicle, and Lunar power microgrids. The system can be implemented standalone or integrated with an onboard executive/controller. The primary use case will be in support of load shedding and reconfiguration but health state knowledge and history can also be used for power forecasting. Knowledge of failures can be used autonomously or by human operators to minimize downtime by swapping to alternate components or replacement. MONSID utilizes physics models of hardware functionality similar to digital twins. The modeling technique captures nominal behavior so it can uncover unanticipated off-nominal behavior in addition to known faults. Explicit fault models are not required to isolate the source of faults, which reduces model complexity and simplifies validation. The fault diagnosis engine is compact, deterministic, and application-agnostic making it completely re-usable. During the program, through Phase II, Okean Solutions will demonstrate the effectiveness of the modeling approach and FDI technique for NASA Glenn’s autonomous modular power architecture. We will develop MONSID models of power system components and functions, run test cases for nominal and fault scenarios, integrate MONSID with the autonomous power control (APC) architecture, and provide training.