A Reliability-Driven Design and Test (ReDDT) methodology for space nuclear power and propulsion systems

The development methodology employed in the United States’ Rover NERVA program is reviewed as the basis for the Reliability-Driven Design and Test (ReDDT) methodology described herein. Postulated as perhaps the first full programmatic implementation of probabilistic design, the NERVA foundational methodology appears to have never been adopted elsewhere following program cancelation. Nevertheless, much of the framework developed during NERVA remains applicable, and the ReDDT approach described in this work is based upon their approach. The ReDDT approach builds on the NERVA method by introducing a qualitative failure mode analysis to identify which tests are most useful and maximize program efficiency. The qualitative analysis is supported by uncertainty analysis, which is applied to identify the drivers of the system's technical uncertainty and thereby drivers of safety and reliability. The ReDDT approach is distinct from existing reliability-based methodologies including Reliability Based Design Optimization (RBDO) and Design for Reliability (DfR) in that ReDDT focuses on concurrent and synergistic system design, integration, and test planning. ReDDT has been developed for NASA's current Space Nuclear Propulsion effort, but it has broader applicability to nuclear power systems and complex aerospace system developments and upgrades in general. The ReDDT methodology has the goal of transforming the relationship of the design effort to the integration and test effort within a system development from one of test-fail-fix to model-test-evaluate.