Sandia LabNews

Labs team ponders future safety of Yucca Mountain

How safe will a potential Yucca Mountain nuclear waste repository be in 10,000 years? That’s the question scientists have been trying to answer for almost two decades.

Sandia geologist Peter Swift (6851) is part of a multi-organization team that is homing in on the answer by using a combination of data, expert judgment, common sense, and refined computational models, together known as "performance assessment." Peter recently reported on the Yucca Mountain Project’s progress in scenario analysis at a meeting of the Organization for Economic Cooperation and Development (OECD) Nuclear Energy Agency (NEA) in Madrid, Spain.

"It’s difficult to assess future risks," he says. "But if you use an approach of identifying all potential normal and disruptive processes and events at the site, screening to determine which matter and which don’t based on your performance criteria, and analyzing what’s important in computational models, you can get something meaningful to base decisions on."

Yucca Mountain, a ridge of volcanic rock located about 100 miles northwest of Las Vegas in a sparsely populated desert locale, is under consideration as a site for a DOE-operated high-level nuclear waste repository. Tunnels excavated in the heart of the mountain — about 1,000 feet below the surface and 1,000 feet above the water table — would hold spent nuclear fuel from commercial power plant reactors and various forms of high-level radioactive waste generated by defense programs.

The radioactive waste would be stored in sealed, corrosion-resistant metallic containers deep inside the mountain. Both the containers and other elements of the engineered system would be designed to take advantage of favorable aspects of the natural environment, such as the relative dryness of the area.

Working on the Yucca Mountain Project is a collaborative team of 26 subcontractors, including private companies such as TRW and Duke Engineering, four DOE laboratories, and the US Geological Survey. Operations are directed by DOE’s Civilian Radioactive Waste Management office.

Regulations set by the US Environmental Protection Agency and the US Nuclear Regulatory Commission require that project analysts show the system will protect humans from the vast majority of the radiation for at least 10,000 years.

"We have no direct experience concerning how man-made materials will react with the environment over such a long time period," says Holly Dockery, Manager of Yucca Mountain Project Performance Assessment Dept. 6851. "However, the analyses used to model the future performance of a waste repository system aren’t based on ‘guesswork.’ Instead, their development involves a rather complex approach that uses judgment, data, and computations."

She adds, "We are required to determine all the possible things that could happen to the system and be able to show the regulator and the public that we thought of ‘everything.’ "

The performance assessment method that the Yucca Mountain Project and most other geologic repository programs use to evaluate future safety of the nuclear waste repository begins with identification of all processes and events (scenarios) that could possibly influence the system. From this list those scenarios that can be shown to be truly important to describing system behavior are systematically selected for inclusion in computer analyses.

Screening techniques fundamental to this approach can be traced back to work done in the late 1970s and early 1980s in support of the WIPP project. Felton Bingham (6800) and George Barr (6851) developed the earliest approach. A few years later, Jim Campbell (6411), Bob Cranwell (6411), and others devised a similar, but slightly different method.

The synthesis of these two techniques pioneered by Sandians has become the international standard for assessing geologic repository systems in the US and throughout the world. Today, the team members in both the Yucca Mountain Project and Waste Isolation Pilot Project (WIPP) are further refining the scenario-analysis method and making strides in forecasting the future.

The first step is to identify potential scenarios that might alter the system, allowing radioactive materials to escape. These scenarios include any number of expected processes such as water seepage, earthquakes, or container corrosion. They also include unlikely, though possible calamities like a volcanic eruption or a meteorite impact.

The next step is to screen the list using well-defined criteria to distinguish between those "events or processes that matter and those that don’t." The criteria are established by the regulatory agencies.

After the screening process, a probability is estimated for each remaining scenario to indicate how likely it is that the scenario will occur. Then, the consequence of the scenario on the system is calculated.

"If an event has less than one chance in 10,000 of happening, regulators do not set limits on its consequences. If it is greater than that, then we have to consider it," Peter says.

Probabilities are determined using various methods and sources. For example, a panel of experts might estimate the probability of a volcanic eruption or of an earthquake based on evidence from past events in the geologic record. The possibility of water dripping onto a waste container could be based on studies of the current site. The categories of scenarios studied are somewhat general to all repository systems and have been established in an international database. However, the particular scenarios selected as important and the probabilities assigned to those scenarios are unique to a given repository system.

Several scenarios have been analyzed specifically for the Yucca Mountain system.

One scenario greatly studied over the years is a water table rise that could perhaps result in flooding the repository, allowing a massive release of radiation from breached containers. However, research has shown that the likelihood of this event is minute. While it is considered in the scenario development, extreme water table rise will not be included in the final analysis because of the very low probability.

Another scenario being considered is volcanic eruption. The rock forming the mountain — called tuff– is composed of volcanic ash from eruptions that occurred mostly between 11 million to 14 million years ago. Very small eruptions associated with basalt cinder cones occurred much later and continued until as recently as 100,000 years ago. Because of the way the geologic setting has changed in the past million years, the probability of more volcanism occurring and affecting the repository is very small. But the probability is just above the boundary of the regulatory limits, so some consequence analyses will be run to look at the consequences of volcanic activity.

A scenario of greater concern involves groundwater derived from rain and snow at the surface infiltrating down to the level of the repository, corroding those containers that become wet, and releasing radiation when the containers are breached. The contaminated groundwater could eventually reach the water table and flow southward about 20 kilometers until it could mix with the drinking supply for future inhabitants.

This scenario is sufficiently likely that it must be included in the detailed assessment.

"It poses the most credible way that radiation might reach humans," Peter says. "Even so, future human exposures are estimated to be small, and acceptably below the limits set by the NRC.

Relatively little water is expected to reach the waste, and the waste packages will be designed to resist corrosion even in damp environments for many thousands of years."

Peter says the group is far along in identifying scenarios in the assessment process and creating computer models. Members still need to do more work to screen the scenarios and determine which are important to consider and model.

The selected scenarios will then be used to construct the performance assessment analyses. The results of these analyses will form part of the basis for determining if the site performs well enough to be recommended for a repository, and then to demonstrate the system behavior for regulatory licensing.

"We need to be confident that the information we provide regulators, the public, and critics is comprehensive and sound," Peter says. "We want to provide future humans living in the area the same kind of protection from radiation that we have and want."