Preparing a site to last longer than the pyramids

The Egyptian pyramids have survived more than 4000 years in a desert environment. The question now is whether the USA knows, or can learn, enough to determine whether nuclear waste can remain for 10,000 years in tunnels in another desert environment – Yucca Mountain, Nevada – without releasing radionuclides in excess of regulatory limits.

The rock mechanics issues confronting the US Department of Energy (DoE), which will design and build

the repository, and the US Nuclear Regulatory Commission (NRC), which is responsible for licensing it, are “whether we really understand enough to be able to predict the behaviour of the rock in and around the drifts [tunnels] where nuclear waste will be emplaced” over a 10,000 year period, Mysore Nataraja from NRC told NEI.

Scientists are agreed that there will always be surprises. DoE currently plans to leave the repository accessible for monitoring and observation for 100 years after it is loaded with 77,000tU of spent nuclear fuel and high-level defence waste, said Mark Board, lead subsurface project engineer with DoE’s chief Yucca Mountain contractor, Bechtel SAIC. Many scientists inside and outside DoE are excited about the opportunities this observation period will offer to gain information that can be obtained no other way. The period also will provide critical insights into whether the drifts are likely to degrade significantly over the long term.

TWO ZONES, TWO SETS OF ISSUES

DoE’s current plans call for repository drifts to be built in two separate rock zones, known as the middle nonlithophysal rock unit and the lower lithophysal rock unit. Although both zones consist of welded tuff, a highly compressed rock formed from volcanic ash, they differ in significant respects, including the number and length of fractures, and the number of voids remaining in the rock from gases trapped during cooling. Geologists and rock mechanics experts are looking at the different ways drifts in each zone could degrade over a 10,000-year period.

In October and November 2003, NRC’s Advisory Committee on Nuclear Waste (ACNW), a five-member independent panel representing diverse scientific disciplines, met rock mechanics experts from DoE and the NRC staff to discuss drift stability issues. The panel outlined those issues in a March letter to the NRC commissioners.

Only a small portion of the drifts (which DoE now estimates at 12-15%) will be in the nonlithophysal zone, a highly fractured layer of welded tuff with few void spaces, but longer and more pronounced fractures. The remaining 80-85% of the drifts will be in the lower lithophysal zone, with fewer and shorter fractures, but a larger amount of void space. Voids represent 10-30% of the rock volume in the lower lithophysal zone, with void sizes ranging from less than a millimetre to more than a metre.

In the nonlithophysal zone, the main issue is the possibility that the rock mass surrounding the excavated emplacement drift will eventually break into blocks that can drop and damage the drip shield.

The impact of large falling rock blocks is not a concern in the lower lithophysal rock unit, ACNW stated in its letter. “However, static loads are another matter,” ACNW said. “The NRC staff analyses indicate that, on average, 75% of the drip shields will buckle under static rock fall loads within 500 years after closure.” These NRC analyses are based on empirical data and observations of coal mining operations, ACNW said, later questioning the validity of the coal mining data as analogues for the repository.

In both cases, ACNW identified the key question as whether interactions between the drip shield and waste packages could accelerate the failure of the waste packages during the 10,000-year compliance period. ACNW also noted disagreement between the NRC staff analyses and DoE analyses relating to the time at which rock fall and rock fall rubble accumulation will become a threat to the integrity of the near-field engineered barriers, primarily the drip shield and the waste package.

JUST 10 YEARS OF EXPERIENCE

The tunnels, which have instrumentation to measure displacement and rock loading, are stable without ground support, Board said. DoE also is conducting extensive laboratory and in situ tests of rock properties aimed at understanding the deformability and ultimate strength of the rock. Data from these tests is used to calibrate numerical models. The Yucca Mountain project is likely to result in the world’s most extensive geologic mapping effort, Board said. The approximately 100 kilometres of tunnels will generate the largest geotechnical database in history, he said. Plans call for an extensive regime of testing, remote photography and instrumentation. Changes in the mechanical and thermal properties of the rock will be evaluated.

Another function of the testing is to determine how the rock will respond to thermal stress from the hot waste packages, changes in humidity over the life of the repository, seismic stress and decreases in rock strength over time.

Board agreed that the 100-year observation period after the repository is filled – but before it closed – would give scientists even more understanding of how the tunnels will respond. “A lot gets learned as you excavate tunnels,” he said. The tunnelling itself offers the opportunity to make observations, conduct geologic mapping and test how the tunnels and the ground supports respond to various stresses.

Board said he was not aware of any plans to continue monitoring or remotely observing the drifts once the 100-year period is over, but, as one civil engineering professor pointed out to NEI, “the decision ultimately will not be ours, but our grandchildren’s and great-grandchildren’s.”

The work of rock engineers at Yucca Mountain is complicated by the lack of past concern with long-term tunnel stability – particularly on the order of thousands of years, Board and several academic engineers pointed out. They said this was still an uncertain area. Mining excavations are not expected to last more than 100 years. Two examples of old tunnels are the oldest section of the London Underground, which is more than 100 years old, and some of the large access tunnels at the Hoover Dam, about 70 years old.

“There has not been a lot of study on how rock strength changes over time,” Board said. The USA is not alone in looking at these issues. Canada, for example, is studying the long-term time-dependence of granite strength at its Underground Research Laboratory, part of the Canadian radioactive waste repository effort.

Rock normally fails by propagation of fractures, known as stress corrosion. Creep tests, subjecting a rock block to extreme pressure until it fails, are the standard way of testing for rock failure in the laboratory. Failure rates depend on factors including the rock grain, mineral constituents and stresses.

Rock stability is inversely proportional to porosity, meaning that the weakest rock at Yucca Mountain will be found in the lower lithophysal zone. DoE estimates that this weakest rock represents only about 10% of the drift area and expects that, at some point, this rock will begin to ‘break out’ along the sidewalls of the drifts, producing rubble that can affect the drip shields, Board said. Current predictions are that the breakouts will widen tunnels to about one and a half times their original size. “We do expect some tunnel damage, but we think it will be mostly confined to areas of the weakest rock,” Board said.

Board pointed out that tuff has a very low rate of strength loss over time because of its high silica content and fine-grained structure. The presence of the voids indicates that the rock has been stable for long periods. Geologists estimate the rock mass at Yucca Mountain to be at least 12 million years old, Board said. They know the temperature at the formation of the voids; they know the temperature at which minerals form; and they understand the temperature and cooling history of the rock.

In some ways, the rock voids act like small caves. Small, delicate mineral precipitates, such as calcites, that formed millions of years ago remain intact inside the voids, Board said. “This indicates that there was not sufficient shaking or disruption [of the rock mass] for these minerals to have been disrupted.” The presence of these minerals provides an analogue for checking the models.

CLOSING THE REPOSITORY

Under current plans, DoE, or whatever successor agency inherits the responsibility, will not be faced with closing the repository for almost 150 years. Closure plans may look very different in 2140 than they do today, but US regulations require that post-closure repository safety be evaluated based on current plans.

The plans call for installation of titanium drip shields over the waste packages in the drifts. The curved drip shields would consist of support beams covered by sheets of titanium several millimetres thick. The drip shields would be put in place using remote gantry cranes and would rest on legs sitting on the drift floor. In each drift, the drip shields would interlock and overlap. Once the drip shields are in place, the access tunnels and the repository shaft would be backfilled with crushed tuff from the tunnelling operations.

The waste emplacement drifts will not be backfilled. One concern was that backfilling would keep the temperature of the waste too high, said Steve Frishman, a geologist with the state of Nevada’s nuclear waste office. The backfill would act as an insulator, possibly allowing the temperature of the fuel rods in the waste packages to exceed 350˚C. At these temperatures, the fuel cladding begins to expand and break. Rubble on the waste package, or the drip shield, would also raise the fuel temperature, but DoE expects such rock fall to occur after the time of major heat generation (about 300 years), Frishman said.

Seismic activity is a possible trigger for rock fall and drift collapse. Board acknowledged the possibility of earthquake activity in the area, and said DoE has analysed the potential damage from such activity after the repository is closed. He said they had considered the worst case, if some tunnels collapsed, including what would happen to temperature and humidity on the waste package.

In both seismic activity and rock fall, the question is whether the drip shields and the waste packages can take the beating and if they cannot, will they fail in such a way that would allow release of radionuclides from the repository above regulatory limits? This last, according to Nataraja, is the key question for NRC. “The regulatory burden is not to get into an argument about when the drifts will collapse, because that may not be answerable, but the objective is to check whether DoE’s engineered barrier designs take into account drift issues and whether this analysis is incorporated into the performance assessment.”

DoE is designing its drip shields to withstand rock fall and static loading, Board said. The department is also looking at the effects of temperature and humidity if a drip shield collapses after the repository is closed and sealed.

The hottest waste packages in the repository will reach their maximum temperature of about 170˚C 20 to 30 years after repository closure. Package temperatures will stay above the boiling point of water for about 1000 years.

If a drift collapses several hundred years after closure, DoE models show that the waste package temperature will go no higher than it would if the collapse had not occurred, Board said. However, the quicker a collapse comes after closure, the more significant its effects. A collapse 50 years after closure would raise the waste package temperature about 50˚C.

State officials also worry that damage to the Alloy 22 waste package could result in corrosion.

Alloy 22 is resistant to corrosion because it relies on a passive film, Frishman told NEI. Any rock fall or collapsed drip shield that mars the surface of the waste package could allow corrosion to start, he said. He also questioned whether there had been any studies on the possibility that a twisted, collapsed drip shield could funnel water onto a waste package. Frishman expressed frustration that little has been discussed about the possible corrosion of the titanium drip shields themselves, which could make them more subject to collapse. Some studies indicate that small amounts of fluoride present in water at Yucca Mountain could cause the drip shields to begin decaying within decades, he said.