Testing the abstractions used in total system performance assessments
- INTERA, Inc./PMO, Las Vegas, NV (United States)
Multiple levels of complexity and detail are involved in the performance assessment models used in the Yucca Mountain project (YMP). Included are the detailed process-level models, and various higher level abstractions or idealizations of those process models, that may be used in the simpler total system simulators for performance assessments. Abstractions are applied to reduce the complex process models to simpler overall simulators for more direct total system analyses. Although seldom done in the past, the abstractions require appropriate testing to demonstrate that each is an appropriate simpler system representation. To be adequate, such testing must either: (1) Demonstrate that essential processes, conceptual representations, and the parameter variations in the original process-level models are neither lost nor the performance results unduly altered by applying the abstractions, or (2) Show that any significance lost is appropriately bounded by the abstraction assumptions and that subsequent assessments using such bounding assumptions still provide appropriate margins of safety in the overall repository performance. Failure to satisfy one of the above conditions requires changing the abstraction being test until it, in fact, verifies the representation is adequate. Such testing provides the foundation necessary for technically defensible performance assessments using the abstracted total system models. The objective of this paper is to outline an approach for testing any proposed or specific future abstractions of the process-level models used to obtain simpler system simulators for application in total system performance assessments (TSPAs).
- OSTI ID:
- 127154
- Report Number(s):
- CONF-9504179--
- Country of Publication:
- United States
- Language:
- English
Similar Records
FY2021 Status Report on the Computing Systems for the Yucca Mountain Project TSPA-LA Models and Testing of Selected Process Models
Tailored model abstraction in performance assessments