Skip to Main Content
Skip Nav Destination
ASME Press Select Proceedings
Proceedings of the Eighth International Conference on Probabilistic Safety Assessment & Management (PSAM)
Michael G. Stamatelatos
Michael G. Stamatelatos
Search for other works by this author on:
Harold S. Blackman
Harold S. Blackman
Search for other works by this author on:
No. of Pages:
ASME Press
Publication date:

Assessing risk from potential geologic repositories for high-level radioactive waste (HLW) poses a unique challenge because of the large temporal and spatial scales involved. The repository systems typically involve multiple engineered barriers and emplacement in a remote geologic location. It is not feasible to conduct experiments on the scale of a repository system. As a result, scientists rely on inference from shorter-term, controlled laboratory and field experiments, as well as historical data from natural analogs to model the system. This leads to uncertainties in modeling the risk from the repository.

The importance of these uncertainties depends on the context for conducting and using the risk assessments. The U.S. National Research Council has pointed out that the proper role of risk assessments is to inform societal decisions and facilitate deliberation among stakeholders. In the case of HLW repositories, the most important sources of risk information for these decisions are the results of integrated repository system modeling compiled in performance assessments (PA). These PAs inform critical decisions for national HLW repository programs. The important uncertainties in the PA are those to which decisions may be sensitive.

To build confidence given potentially important uncertainties, a typical safety case for a proposed HLW repository is comprised of PA results coupled with various defense-in-depth elements, such as the multiple-barrier requirement for repository design, and insights from supplementary analyses. This paper proposes an additional supplementary analysis, the Strategic Partitioning of Assumption-Ranges and Consequences (SPARC), whose goal is to construct a specific explanation of the PA results of interest, to aid risk-informed decision-making and risk-informed stakeholder deliberation. The method seeks to provide explanations for undesired system behavior that are similar to those provided by a reactor probabilistic safety assessment (PSA); i.e., if we want to know the reasons for possible reactor core damage (a result of interest), we can trace the responsible sequence of events through the event trees, and similarly trace the ways each event can occur through the underlying fault trees in a PSA. Since the system nature of repositories versus that of reactors is very different (continuous slow degradation in passive systems versus largely binary random failures in active systems), the scenarios of reasons for undesired behavior in a repository system would be defined on different terms.

The SPARC method extracts risk information from existing PAs and supporting databases to explain how the repository system may produce undesired behavior, specifically by uncovering what sets of model parameter values taken together could exceed a particular criterion set for the repository (e.g., an instance where a performance measure such as projected dose exceeds a particular goal.) The results can be displayed in SPARC trees. The method could be used: (1) in a safety case to help build confidence in a repository system; (2) to risk-inform decisions on how to allocate resources for future research; and (3) to risk-inform stakeholder deliberation. As a demonstrative example, the SPARC method is applied to a potential HLW repository at Yucca Mountain.

This content is only available via PDF.
You do not currently have access to this chapter.
Close Modal

or Create an Account

Close Modal
Close Modal