ABSTRACTThis paper discusses the technical approach and rationale of the performance assessments (PAs) conducted for the Waste Isolation Pilot Plant that contributed to the success of the certification in 1998. The PAs were iterated: there were four preliminary PAs between 1989 and 1992 and the certification PA in 1996. Although many changes occurred between the 1992 PA and 1996 PA, the preliminary iterations provided a strong foundation on quality assurance, parameter and model databases, documentation, and peer review. The seven general steps of a PA are used to provide a brief overview of their history. The paper then delves into the rationale used for the most computationally intense step of a PA, the consequence evaluation. For this evaluation, Sandia National Laboratories mostly used detailed models when evaluating the probabilistic performance of the repository under the stylized conditions specified by the U.S. Environmental Protection Agency (EPA). A computational tool, CAMCON, was specifically tailored for this approach. Many advantages were gained by using detailed models directly in the PA, rather than using simplified results of these models. One important advantage was that models and parameters remained fairly unbiased (i.e., the analysis had only a few major conservative assumptions rather than many, unquantified conservatisms). This approach allowed Sandia to faithfully provide a “reasonable expectation” of repository performance, as specified by the EPA.
Regulatory issues for Waste Isolation Pilot Plant long-term compliance with U.S. Environmental Protection Agency 40 CFR 191B and 268