16 PRA Applications in Space Shuttle Program Risk Management (PSAM-0467)

The Space Shuttle Program (SSP) has traditionally relied upon qualitative risk assessment techniques such as Failure Modes and Effects Analysis (FMEA) [1] and Hazard Analysis (HA) [2] to identify and characterize Shuttle program technical risks. These risks comprised a safety baseline for Space Shuttle operations. It was thought that Shuttle risks could be effectively managed through careful management of the established risk controls and comprehensive pre-flight control verification actions. Additionally, the SSP relied upon a highly detailed requirements base, with extensive waiver documentation, to define and control risks related to requirements non-compliances. Overall, this approach to risk management was useful, but had significant limitations.

One of the primary difficulties posed by this approach was complexity. The safety baseline for the program at any given time was represented by thousands of hazard causes, critical items, and program waivers — making it difficult to understand the level of risk that flight operations posed. It was also difficult to understand the integrated effect of proposed changes to the shuttle systems, operations, and procedures. Efforts to prioritize safety risks were confounded by this risk management model as well, since it was difficult to compare competing failure modes, hazards, and requirements non-conformances based on the potential risk posed through hazards, failure modes, and waivers.

The SSP began developing PRA models during the 1980s on a piecemeal basis, but did not attempt a system-level PRA until 1995. This landmark study made some important conclusions regarding Shuttle risk that have remained true throughout later assessments. However, the analysis included only some of the Shuttle's critical systems and operations, and did not include significant risks associated with orbit phase operations. Subsequent studies were conducted to gradually increase the completeness and fidelity of the PRA. While these efforts suffered from a lack of consistent standards, limited program sponsorship, and limited funding; additional progress was made in developing a system-level PRA. The latest iteration of this activity, initiated in 2003, has enjoyed a greater level of support and sponsorship from within the program than any previous analysis and has resulted in a higher fidelity, peer-reviewed study, reflecting all three unique mission phases: Ascent, Orbit, and Landing. While additional areas remain for improvement in the SPRA, the current iteration has provided the SSP with a system-level model that can be used to support decision making. Additionally, smaller scope, focused PRAs have been spun off to address specific technical decisions, and for the first time, probabilistic analysis is being used to resolve technical issues, define flight rationale, and inform real time decisions during Shuttle missions.

This paper describes the development of the SPRA and how the SSP has adapted its risk management paradigm to include PRA. It describes how the previous SSP safety risk management paradigm has evolved and how PRA is being used to influence program decisions. Lastly, this paper proposes a model for integrating PRA into program risk management and identifies areas that need further improvement in SSP practice.