Future human spaceflight missions beyond low Earth orbit will pose novel operational challenges as the distance from Earth makes real-time ground support impractical for some situations and impossible for others. Thus, future missions will require the crew and habitat to function with increasing autonomy, especially during time-critical events such as a system anomaly with a short time-to-criticality response constraint. While current fault management systems and architectures are well-suited to perform threshold-driven anomaly detection and automated safing actions, tasks like diagnosis, planning, prognosis, and the implementation of corrective actions are primarily ground operator (and subject-matter-expert) driven. To enable effective anomaly response in deep space and reduce the risk of catastrophic events, a novel system architecture, which can detect, diagnose, and respond to failures autonomously must be developed. In this paper, we propose an anomaly response system architecture that will allow future crews to resolve anomalies in deep space and apply a model-based systems engineering (MBSE) simulation methodology to evaluate the expected performance of the architecture. This approach is being demonstrated on a case study scenario that involves the transition between two phases-of-flight (habitat uncrewed with crew in transit and habitat entry). Four habitat state transitions (uncrewed nominal, uncrewed degraded, crewed degraded, crewed nominal) and an unknown interdependent ECLSS fault are addressed. For this paper, we present an MBSE model of the system architecture, an approach to integrating novel autonomous technologies, and the simulation results for the case study. Finally, design trade-offs derived from the simulation results are discussed.
Pischulti et al. (Sun,) studied this question.