Performance Assessment for Radioactive Waste Disposal at Yucca Mountain, Nevada
Project Description and Significance
Some radioactive wastes now in existence will remain dangerously radioactive for thousands of years. The United States has decided to isolate these high-level wastes in repositories located deep underground. The regulatory standards that govern such repositories have not been finalized, but are expected to require sufficient isolation of the waste that the health and safety of the public is protected for many thousands of years. Before a repository can accept wastes for isolation, assessments of its long-term performance must show a reasonable expectation that the repository will meet this standard. The assessments are therefore a vital part of the nation's efforts to reduce potential contamination of the environment from radioactive wastes.
Sandia's Contribution
Sandia participates in several projects that involve the study of deep repositories for radioactive waste. Three of these projects are described in other fact sheets: the Waste Isolation Pilot Plant, Greater Confinement Disposal, and studies for the Idaho National Engineering Laboratory. This fact sheet describes Sandia's assessments of the performance of a potential repository site at Yucca Mountain in Nevada.
Sandia's performance assessments for Yucca Mountain follow the same basic steps described in the above mentioned sheets. The accompanying schematic shows the elements that are considered to be important influences on the performance of an underground repository. On the left-hand side of the schematic, several simple drawings show an underground repository at Yucca Mountain under different sets of conditions that might occur during its period of performance. Arrows suggest paths by which radioactive material might escape. For example, one drawing shows paths for release if the conditions at the repository do not change; another shows paths created by increased rainfall; another shows paths formed by drilling into the repository. The final performance standards for a high-level waste repository may be expressed probabilistically, as is illustrated on the right-hand side of the figure by the shaded region. To show compliance with the standard, the assessors must estimate the probabilities of all the conditions that might produce significant releases from the repository. Then they must estimate the consequences (the releases) associated with those conditions. Finally, they must combine the probabilities and the consequences into the single curve shown on the right-hand side of the schematic. If this complementary cumulative distribution function (CCDF) lies to the left of the shaded region (because the normalized releases are low), or below the region (because the probabilities are low), the repository complies with the standard.
The data describing the system and the conditions contain significant uncertainty, partly because they must describe conditions far in the future and partly because the characteristics of the system are too complex to be known with certainty. As a probabilistic measure of system performance, the CCDF incorporates the uncertainties and allows them to be taken into account explicitly. Using it as the measure of compliance is an example of the use of risk-management techniques in regulatory practice.
Workers at Sandia are developing methods for generating CCDFs. They develop mathematical descriptions of the repository system for use in calculating releases under all significant conditions. They determine the effects of uncertainties in the data that describe the conditions and probabilities of occurrence. In this way the workers guide the collection of data that will reduce the most important uncertainties. Because the standard has never before been implemented, these tasks have few precedents. To carry them out requires skills in research and development as well is iterative, carefully documented routine calculations of releases.
Sandia contributes research and development activities of four types: model development, engineering analysis, database construction, and data collection. The mathematical descriptions of the flow of water through the repository system are examples of model development. An example of engineering analysis is the study of stresses induced in the rock surrounding the buried waste. Such stresses, in principle, might affect the system's ability to isolate the waste. Data describing the site come from many sources and in many formats. To use them effectively requires a dedicated effort to build a special database that includes not simply the data, but also the products of data analysis-probability distributions, for example. These three types of activity are generally combined in the production of Total-System Performance Assessment (TSPA) analyses. The outcomes of these analyses include estimates of the potential suitability of the site, guidance for data-collection activities, and insights into how future TSPA analyses can be improved.
Future Work
Because of the importance the DOE Yucca Mountain Project places on TSPA analyses, they are receiving intense scrutiny. Both regulators and the scientific community are insisting that the assessments incorporate all available data, currently understood parameter and model uncertainty, and conceptual models that reflect our interpretations of site data. These elements are embodied in the physical models used in the TSPA analyses. Sandia's expertise in model development, uncertainty analyses, and data collection allows us to make a meaningful contribution to these activities. Furthermore, some of the tools developed for Yucca Mountain TSPA analyses have been used for similar analyses on other projects.
Studies of Yucca Mountain will continue for the next several years. Any decision on whether to build and operate a repository at the site will be strongly influenced by past and future TSPA analyses.
For further information, contact:
Holly A. Dockery
Sandia National Laboratories, MS 1326
Albuquerque, NM 87185-1326
Phone: (505) 848-0730
email: hadocke@sandia.gov
Submitted October 1996 Layout design by Wanda Mar.
Submitted October 1996