ML20212H458
| ML20212H458 | |
| Person / Time | |
|---|---|
| Issue date: | 08/09/1999 |
| From: | Garrick B NRC ADVISORY COMMITTEE ON NUCLEAR WASTE (ACNW) |
| To: | Dicus G, The Chairman NRC COMMISSION (OCM) |
| References | |
| NACNUCLE-R-0145, NUDOCS 9910010138 | |
| Download: ML20212H458 (47) | |
Text
C \\ J~
a ct?
/
' UNITED STATES ACNWR-0145 8
NUCLEAR REGULATORY COMMISSION I'
PDR a
ADVISORY COMMITTEE ON NUCLEAR WASTE kk,,g 8[
WASHINGTON, D.c. 20555-0001 August 9, 1999 The Honorable Greta Joy Dieus Chairman U.S. Nuclear Regulatory Commission Washington, D.C. 20555-0001
Dear Chairman Dieus:
SUBJECT:
COMMENTS ON DOE'S LICENSE APPLICATION DESIGN SELECTION PROCESS (LADS) AND RECOMMENDED REPOSITORY DESIGN This letter conveys our observations and recommendations regarding the Department of Energy's (DOE's) License Application Design Selection (LADS) process and the Management and Operations Contractor (M&Os) recommended repository design for the site recommendation (SR) and license application (LA). The letter also transmits the attached
" white paper" by Charles Fairhurst titled, " Engineered Barriers at Yucca Mountain -- Some Impressions and Suggestions." In his white paper, Dr. Fairhurst examines some geotechnical aspects of the repository design in the setting of Yucca Mountain with particular attention to two issues - (i) reduction of water inflow to the waste emplacement drifts and (ii) pre-and post-closure stability of the drifts. A concept of an innovative repository design not presently being considered by the DOE is described, together with some impressions of the currently favored repository design. We hope that the paper will help the NRC as it prepares to conduct a thorough and critical safety review of the final repository design and the projected overall performance of the Yucca Mountain high level waste (HLW) disposal facility.
The observations and recommendations we make here are based on briefings we heard on
)
July 20,1999 on DOE's license application design selection, during the 111* ACNW meeting in Rockville, Maryland. The basis for the attached white paper is derived from a variety of sources, including the DOE's viability assessment, and interactions with the NRC and DOE staffs, the Center for Nuclear Waste Regulatory Analyses (CNWRA), the M&O, the ACNW, and others.
White Paper on Engineered Barriers at Yucca Mountain in the attached paper, Dr. Fairhurst examines a repository shield concept that appears to have the potential to greatly reduce water infiltration into repository drifts. The shield acts like an umbrella above the repository to divert water around drifts by taking advantage of the vertical fractures and predominantly vertical flow system in the vicinity of the repository horizon. The shield system may also help reduce near-field flow uncertainties in designs such as the Enhanced Design Alternative-ll (EDA-II) currently recommended by the M&O to the DOE. The shield concept is shown to be most effective when used in conjunction with a multi-layered repository to minimize the surface area contacted by infiltration. Dr. Fairhurst suggests that if the shield can be demonstrated to be effective with high confidence, it may be possible to avoid the need for the yery costly ($4.6 billion) titanium drip shield used in the EDA-ll.
)L iI 9910010138 990809 c7 i
PDR ADVCM NACNUCLE hp- [ y R-0145 PDR
)
e et,
'. The purpose of the paper is not to promote or endorse a specific design. Rather, the paper is intended to demonstrate that there may be innovative ways to engineer the natural setting such
- that the overall performance of the repository is improved. Current DOE designs appear to concentrate exclusively on engineering options within the drift itself We believe that exploration of such ideas supports the NRC in its mission and in its vision of " enabling the safe and efficient use of nuclear materials." Consideration of the repository shield and a multiple level repository and other design concepts can provide insights into approaches for reducing critical uncertainties and for modifying the degree of reliance placed on natural versus engineered barriers. Exploration of altemative design concepts may also provide insights to help the NRC avoid placing constraints on DOE's repository design that might inadvertently limit possible future beneficial design changes and innovations, that would lead to greater confidence in the safe disposal of HLW at Yucca Mountain.
In its July 9,1999, letter to Lake Barrett (DOE)', the Nuclear Waste Technical Review Board (NWFRB) expresses concem about the uncertainties associated with the above-boiling-temperature EDA-Il design recommended by the M&O, and the lack of transparency in the process and rationale used to select this design. The EDA-Il design is a "high temperature" design having a peak drift-wall temperature (160'C) above the local boiling point of water (96*C), with the space between drifts below boiling. To reduce uncertainties, the NWTRB urges DOE to consider modifying the EDA-Il design to achieve below-boiling temperatures everywhere in the rock by increasing the rate or duration, or both, of ventilation before repository closure.
The ACNW believes that further analyses must be done before a determination can be made on a choice between a " totally below boiling" temperature repository and one in which some boiling takes place. Dr. Fairhurst points out that the recommended EDA-Il design has some merits but also some disadvantages. Although a cooler repository design may simplify modeling of water redistribution, the potential for a higher temperature repository design to reduce the quantity of l
water reaching the drifts should not be abandoned without further assessment. It is possible j
that the existing EDA-Il design, possibly modified to include multi-layered emplacement drifts, in conjunction with the infiltration shield concept, can be shown to reduce the uncertainties of water refluxing associated with a hot repository while maintaining the advantage of the hot repository to drive moisture away from the canisters.
We hope that you find Dr. Fairhurst's white paper to be of interest.
i
' July 9,1999 letter from Jared L. Cohen, Chairman, Nuclear Waste Technical Review Board, to Lake H. Barrett, Acting Director, Office of Civilian Radioactive Waste Management, U.S. Department of Energy.
4 u
C
~ - p.-
} t Observations and Recommendations Regarding the DOE's Design Selection Process and the Recommended Repository Design Observation 1 l
l Over the past;10 months, the M&O contractor has been conducting a study of alternative l
repository designs for the proposed Yucca Mountain repository. As noted earlier, the M&O recently recommended that DOE select the EDA-II. The DOE has not yet made a decision l
about adopting the M&O's recommendation. The recommended EDA-Il design differs l
significantly from the repository design presented in the DOE's viability assessment. As noted l
above, the NWTRB has expressed its dissatisfaction with the design selection process as well as with the recommended EDA-Il design. Such recent and rapid changes suggest that the fundamental design and the many design-related details are likely to continue to change until such time as DOE submits its LA to the NRC. DOE's repository design must be regarded as a I
work in progress.
Recommendation 1 A:
The NRC should plan for continued change in the repository design up until the time the LA is i
submitted, it follows that the NRC staff should adopt realistic expectations about the turnaround time that may be required to conduct a thorough review of the SR or LA design. The NRC should also develop a license review strategy that allows the DOE maximum flexibility to implement beneficial design changes and other innovations before its submittal of the LA as l
well as times throughout the preciosure period of the repository.
Recommendation 1B As noted in the attached white paper, the preclosure period of the repository could last as long as 300 years, and, because of this, the NRC staff must be careful to avoid placing constraints on the design that might preclude future beneficial design changes or innovation.' The NRC staff must ensure that it is prepared to recognize such innovation during its review of the LA. Further, as part of a strategy to develop review capability and insights into repository systems, the NRC and the CNWRA staffs should conduct independent evaluations of attemative, cost-effective l
_ designs. In evaluating such innovative designs as part of its preparation to review the LA, the NRC staff would gain insights into the relative importance of various design features, alternative strategies to reduce critical uncertainties, and attemative strategies for demonstrating defense in depth. The insights gained through the evaluation of attemative design concepts will enhance the NRC staff's capability to assess repository safety.
Observation 2 NRC's proposed rule goveming HLW disposal (10 CFR Part 63) requires monitoring of repository performance. The 50- to 300-year repository preclosure period presents a major opportunity to establish the validity of design assumptions. Monitoring will require " performance l'
l I
0 c-j
- confirmation drifts". Such drifts, appropriately located, could also serve as part of the flow diversion system proposed in the white paper.
Recommendation 2 The ACNW endorses the sentiment expressed recently by the U.S. Geological Survey (USGS),
"that a careful description of the proposed monitoring strategy, as well as a detailed and complete list of what is to be monitored-and why, where, how, and for how long-should be developed expeditiously." 8 We encourage the NRC staff to consider long-term monitoring needs and strategies for how DOE may factor performance confirmation monitoring into its final design.
Observation 3 As noted above, in its July 9,1999, letter to L. Barrett, the NWTRB expresses concem over the lack of transparency in the assumptions and value judgments made in the design selection process as well as the recommended design. Implicit in the NWTRB's letter is that the Board is uncomfortable with the M&O's selection of the EDA-Il repository design because of the current uncertainties associated with high repository temperatures. It is not clear to the ACNW how the uncertainty associated with the various design concepts and features has been quantified and factored into the M&O's process for selecting a preferred design. The M&O's identified evaluation criteria do not include uncertainty as a criterion for making a selection. The conceptual model and assumptions for the various design concepts and features will drive the results of the evaluation and comparison of attematives.
Recommendation 3 The ACNW believes that the M&O's approach used to evaluate and compare quantitatively the various EDAs has not been made transparent. We encourage the NRC to ensure that the rationale, approach, and assumptions used in the evaluations and in coraparisons of attematives are appropriate. In addition, as noted in recommendation 18, the NRC and CNWRA staffs should conduct their own independent evaluations of attemative, cost-effective designs, similar to the evaluation of the innovative design described in the attached white paper.
Sincerely, o Q lNh g
i B. John Garrick Chairman 8 Viability Assessment of a Repository at Yucca Mountain, Preliminary Design Concept for the Repository and Waste Package, USDOE, Volume 2,1998, p. 4-111 8L USGS Circe!zr 1184,1999," Yucca Mountain as a Radioactive Waste Repository."
j
1
^
g/ * ~
?
m Summary
' Yucca Mountain was initially recommended as a potentially suitable site for a high-level waste repository because it was anticipated that it would be dry. The repository would be situated in the unsaturated zone at a depth of 300 m below the surface and approximately 300 m above the current water table.- It was also proposed as a " hot repository," in which rock temperatures would rise above 200 C and would remain above the boiling point of water for several thousands of
' years. The intent was to prevent any liquid water from reaching the waste packages during that
- period.
Recent studies suggest that infiltration rates in the unsaturated zone may be higher than originally anticipated, and may increase substantially 20,000 years or so into the future. This information has prompted a redesign of the repository placing greater emphasis on engineered barriers within the waste emplacement drifts, e.g., a drip diversion (Richards) barrier; corrosion-resistant waste package; titanium drip shield (cost $4.6 billion); active ventilation during the 100- to 300-year i preclosure period; and lower repository temperatures.
The viability assessment (VA) published by the U.S. Department of Energy (DOE) in December 1998 indicates that these engineering measures should suffice to meet the 10 CFR Part 63 -
requirements of the U.S. Nuclear Regulatory Commission (NRC) over the 10,000-year regulatory period, although doses are predicted to rise considerably beyond 10,000 years.
These notes, prepared after review of the VA, focus on geotechnical aspects of the repository design. The author has profited from discussions with colleagues of the Advisory Committee for Nuclear Waste (ACNW) and NRC, as well as from participation in numerous meetings and discussions with staff of DOE and its Management and Operating (M&O) contactors. The notes emphasize (1) a repository shield concept and (2) prediction of drift stability during both the (100
. yr ~ 300 yr) preclosure and postclosure periods.
This paper does not promote or endorse any specific repository design. Rather, its purpose is to -
stimulate the NRC's thinking as it prepares to conduct a thorough and critical review of the repository design used in DOE's license application. The paper attempts to demonstrate that
" consideration of such innovative ideas as the repository shield concept and triple-layer repository can redefine the problem by reducing or eliminating critical uncertainties, or altering the degree of reliance placed on natural versus engineered barriers.
Given that decisions regarding f' mal closure will not be made until the end of the operational period of the repository, the NRC must be careful to avoid placing constraints on the project now
' that would inadvertently limit possible future advantageous design changes and innovation. It is 1
incumbent on the NRC to have the capability and be prepared to recognize the possibilities for such innovation during its evaluation of the license application.
i The repository shield acts as an umbrella above the repository, taking advantage of the (dominantly vertical) fracture and flow system of the site to divert water away from the
E:gineered Birriers ct Yucc2 Mount:In. Some impressions ands:ggestions 2
\\
l repository drifts. The shield uses natural material (rock) only, augments an existing design, can be developed at any time during the preclosure period, and can serve to house a remote-
{
monitoring network for the repository.
For a repository shield to be most cost effective, the repository should be a multi-level (three-tier or two-tier) design. (Figure 2 shows a three-tier design.) The shield appears to have the potential of greatly reducing water infiltration to the repository drifts -with attendant reduction of doses and simplification of performance assessment calculations. Construction of a flow diversion barrier in the (radiation-free) slot excavations above the drifts would be simpler than remote placement around the unshielded waste packages in the repository drifts -as currently proposed by DOE. If water infiltration is reduced to the extent predicted by analysis to date (see Appendix I), the expensive titanium drip shield (see Figure 4) may not be required. The presence of the drainage slots directly above the emplacement drifts may also simplify near-field fluid-flow and reflux processes during the thermal cycle. The concept deserves serious examination by DOE and its contractors.
With respect to drift stability, the repository environment is unique in that substantial thermo-mechanical stresses may be generated in both the reinforcement support and the rock. From information available on the mechanical properties of the Topopah Springs formations, it appears that stable excavations can be designed in both the lithophysal and the non-lithophysal units. It is believed that rock reinforcement using fully grouted bolts, mesh, and (if possible) shoterete is preferable to the use of concrete or steel set supports for the repository drifts. Attention will need to be given to pH control of the cement used, but this problem does not appear to be an insuperable problem.
For the postclosure period, it must be assumed that any rock reinforcement or support system will no longer be effective. Recent developments in the numerical modeling oflong-term progressive degradation of the mechanical properties of rock masses can provide more realistic assessment and prediction of the behavior of rock around excavations that are not back-filled than were possible in the past. Progressive disintegration and collapse of the rock may, in fact, result in a
" natural back-filling" process that could be as effective, eventually, as standard back-fill. Of course, this does not preclude the use of a " chemically tailored" back-fill in the drift section below the waste packages, which could provide significant radionuclide " capture" benefits.
l
g A.
~ E::gineered B:rriers at Yzcc2 Mountain, Some Inpressions end S:ggestions 3
l Introduction The goal of geological isolation of highly radioactlye waste is fundamentally simple - to place the waste at depth in the subsurface such that the radioactive elements or radionuclides in the
- waste will never return to the biosphere in concentrations sufficient to pose a significant health risk to humans.
Given the very long half-life of some radionuclides, the times for which isolation is required may be on the order of several hundreds of thousands of years.'
The primary vehicle for transport of the radionuclides from the initial underground location or repository is moving water that comes into contact with the waste. Radionuclides become entrained in the water (by dissolution or by colloidal suspension) and move to the biosphere, either directly or in water that is pumped from the aquifer and used for drinking and/or irrigation.
Thus, one of the main criteria in repository siting is to minimize the probability of radionuclide uptake by water and transport t.o the biosphere. Some radionuclides have very low solubility in the groundwater, others may be very soluble. The physical and chemical characteristics of the rock may also greatly retard the overall rate of movement of particular radionuclides in relation to the rate of groundwater movement. The concentration may also be reduced by dilution (e.g.,
. in water or air) so that release to the biosphere via large bodies of water (i.e., seas or oceans) can also provide an added measure of safety.
The first formal report on the feasibility of geological disposal was published by the U.S.
National Academy of Sciences / National Research Councilin 1957 (NAS/NRC,1957). The L
report noted that:
Wastes may be disposed ofsafely at many sites in the United States, but,
?
conversely there are many large areas in which it is unlikely that disposal sites can befound,for example, the Atlantic Seaboard. The research to ascertain
. feasibility ofdisposal hasfor the mostpart notyet been done....
The report concludes with the following two General Recommendations on Corollary Problems:
1.
The movement ofgrosc quantities offluids through porous media is reasonably well understood by hydrologists andgeologists, but whether this is accomplished byforward movement ofthe wholefluid mass at low velocity or whether the transfer is accomplished by rapidflow in " ribbons" is not known. In deep disposal of waste in porous media it willin many cases be l
1 The " half-life" of plutonium 239, for example, is 24,000 years, i.e., the specific radioactivity will decline to (%)(i.e.,0.001 or 0.1%) ofits initial activity in 24,000 x 10 = 240,000 years, and to (0.001)(0.00l) or 0.0001%
in 480,000 years. Other very long-lived radionuclides that contribute to the potential dose at various (long) times at
. Yucca Mountain are technetium 99 (half-life of 212,000 years), uranium 234 (245,000 years), neptunium 237 (2.14 million years), and iodine 129 (17 million years).
e,
.g Engineered B:rriers ct Yucc2 Mountiin. Some impressions andSuggestions 4
essential to know which ofthese conditions exists. This willbe a difficult problem to solve.
2.
The education ofa considerable number ofgeologists and hydrologists in the characteristics ofradioactive wastes and its disposalproblems is going to be necessary.
Today, more than 40 years later, there are many hydrologists and colleagues in related disciplines worldwide who have studied groundwater flow in considerable detail. Significant advances have been made, but characterization of water flow still involves large uncertainties, especially in fractured rock masses. It remains "a difficult problem to solve."
Geological repository siting and evaluation programs are currently underway in approximately 30 countries. Of these, all but the Yucca Mountain project in the USA are sites below the groundwater table. For these, the host rock is usually oflow intrinsic permeability with a low regional hydraulic gradient (i.e., the overall rate of water movement from the repository is expected to be very low). A number of countries are considering repositories in crystalline rock.
Characterizing groundwater flow in fractures is frequently a serious issue for these sites.
In addition to understanding the natural system at Yucca Mountain, i.e., groundwater flow and radionuclide transport, NRC's proposed high level waste (HLW) disposal regulation, 10 CFR Part 63, indicates that an engineered barrier system (EBS) consisting ofone or more distinct barriers is required in addition to natural barriers. The proposed rule states that the Commission continues to beleive that multiple barriers, as requiredin the Nuclear Waste Policy I
Act of1982 (NWPA), must each make a definite contribution to isolation ofwaste at Yucca Mountain. Thus, DOE must design and demonstrate quantitatively that the total repository system relies upon and balances the contributions of both natural and engineered barriers to isolate waste.
The preclosure period of the proposed Yucca Mountain repository is expected to range from 50 to 300 years. Given that final repository closure will not occur until the end of the preclosure period, the NRC must be careful to avoid placing constraints on the project now that would inadvertently limit possible future beneficial design changes and innovations. It is incumbent on l
the NRC to have the capability (and be prepared) to recognize the possibilities for such innovation during its evaluation of the license application. One way to develop such capability is for the NRC to conduct an independent evaluation of viable, cost-effective designs. To conduct such evaluations, the NRC needs to have competent scientific and engineering expertise available over the broad spectrum of disciplines involved in repository design and long-term performance assessment. With the much larger complement of technical staff available to DOE and the recent and rapid changes in repository designs proposed by the DOE, the NRC faces a formidable challenge.
l This report focuses on geotechnical aspects of the proposed Yucca Mountain repository. A design concept consisting of a repository shield used in conjunction with a multi-tiered repository is outlined. Particular attention is. iven to two issues: (1) diversion of groundwater before it reaches the waste-filled drifts and (2) drift stability. The paper then considers prediction of drift stability during the preclosure and postclosure repository periods. The paper compares the respository shield concept to the DOE's current, preferred repository design, which has l
L
. Engineered B:rriers ct Yteca M:untzin, Some Impressions cndSuggestions 5
t l
changed significantly from the design presented in the DOE VA. 'Ihe purpose of the paper is to stimulate the NRC's thinking as it prepares to conduct a thorough and critical review of the repository design used in DOE's license application. The ACNW may also use the ideas in the.
paper in preparing'its specific comments on the DOE site recommendation and license
~
application. The paper attempts to demonstrate that consideration of alternative, innovative design concepts, such as the repository shield / multiple-layer repo~sitory, may take better
' advantage of the geological characteristics of the proposed repository site at Yucca Mountain.
- Critical, persistent uncertainties may possibly be reduced substantially and the degree of reliance
~
placed on ' atural and engineered barriers can be varied. The proposed " shield drifts" can also n
- serve the role ofperformance-confirmation monitoring drifts (see VA, Vol. 2, p. 4-111)._
' Groundwat'er Flow at Yucca Mountain -
I a
At Yucca Mountain, the proposed repository horizon is in the unsaturated zone, approximately 250-300 m below the surface of the Amargosa Desert and 300 m above the water table.
Tectonically, the region is currently undergoing extension (i.e., the rock mass is tending to extend horizontally).- This implies that, at least near the surface (i.e., within the region of concern with respect to the repository), the lateral stresses in the rock are less (~3 MPa) than the vertical (gravitational or overburden) stresses (~7 MPa at a depth of 300 m). This situation has given rise to high-angle (i.e., almost vertical) fracturing (see VA, Vol. 2, Figure 2-9, p. 2-17). As a result of this situation, the fractures tend to be highly transmissive, so that rainfall and surface waters i
drain rapidly through the fractured mass into the groundwater. However, these fractures are generally not single, continuous planar features. Individual fractures are oflimited extent, so that
- connected pathways, allowing flow through the fracture network, will be considerably less frequent than the individual fractures.
In initial planning for the repository (Roseboom,1983), it was felt that the annual percolafon flux (i.e., precipitation less the amount of surface evapo-transpiration) was very small (on the order of 1 mm/yt) and that little or no moisture would drain into the repository (i.e., the repository would be " dry"). In addition, it was decided to adopt a " hot repository" design (i.e.,
l such a disposal layout that the rock temperature in the vicinity of the repository would remain l
well above 96 *C, the local boiling point of water, for hundreds or thousands of years, so that no 2
liquid water could reach the waste canisters ),
More recent studies indicate that the total infiltration may be higher, and that a considerable portion of this may flow through the interconnected fracture pathways. As noted in the VA:
I i
Estimates ofaverage percolationfluxfrom these various studies rangefrom about 0.1-18 mm (0.004-0. 7 in)peryear. Because ofPaintbrush attenuation most of thefluxprobably requires hundreds to thousands ofyears to reach the repository horizon. However, isotopic (chlorine-36) data suggest that at least afraction of theflux reaches the repository level in tenyears or less. Thus, while some ofthe 2 The high-temperature design is feasible in an unsaturated high permeability zone, such as exists at Yucca Mountain, where the pressurized water vapor in the rock in the vicinity of the excavations can " leakoff" readily l
Ll toward the surface.
i
.e, Engineered B:rriers ct Yucc2 Mount:In, Some impressions endsuggestions 6
water moves downward quickly, much ofit travels more slowly. (VA, Vol.1, p.
2-38 ).
Studies oflong-term climate change in the Yucca Mountain region over the past 500,000 years (see Figure 1) indicate that the climate in the region will very likely become colder within the nextfew hundreds or thousands ofyears (VA, Vol.1, p. 2-30). Annual precipitation and infiltration are then likely to increase considerably. DOE performance assessment calculations consider a mix of dry and wetter climates extending up to several hundreds of thousands of years into the future (VA, Vol. 3, Sect 3.1.2.1, p. 3-15). These periods include dry climate conditions, as now, with an assumed base infiltration rate of 8 mm/yr, a long-term average period with a base infiltration rate of 42 mm/yr; and superpluvial periods with a base infiltration rate of 110 mm/yr (VA, Vol. 3, Table 3-5, p. 3-15). Increased infiltration rates will increase the proportion of total flow through fractures.
5:-3xygen isotope Records From Nevada sad North Atlantic V
IV lli Il ~ 7w9 Devils Hole
- +2 Coteif l
g i
W, t3 r
5
-a
~
- +1 SPECMAP' i
mm
.V,
IV.
III 13 -
~l i,,
.425 AC 475 -350 325 300 275 250 225 200 175 150 -125 -100 -75 40 25 0 anus', me Figure 1 Stable-Oxygen Isotope Recordsfrom Nevada and North Atlantic as Indicators of Past Climate Variation in the Vicinity of Yucca Mountain I
L
p e
i q Engineered Birtlers at Yucca Mountain. S:me Ingressions and S:ggestlo:s The overall conclusion with respect to repository design at Yucca Mountain is that a significant fraction of the total infiltration through the unsaturated zone will be by flow through interconnected fracture pathways. The precise location of these pathways cannot be predicted, and the amount of flow may vary considerably from place to place in the repository. The rates of flow in these fracture pathways can be high, on the order of tens of meters per year.
A fraction of the flux arriving at the drift horizon is assumed to drip onto the waste packages, causing corrosion of the package and, eventually, contact with and dissolution of some of the waste. Details of the calculation procedure are outlined in the viability assessment (VA, Vol. 3, Sec. 4.1.3, p. 4-4 et seq.).
Repository Design and Yucca Mountain Waste isolation poses unique problems for both geoengineering and geoscience. These problems center around the time frames involved, with at least semi-quantitative answers needed over 4
6 2
times on the order of 10 or 10 years - far longer than the 10' or 10 years for which engineers are accustomed to provide quantitative solutions. The geoscience issues have received more attention to date, so there is a good awareness of the uncertainties associated with predictions presented with respect to waste isolation over such times. With engineering design now
{
receiving more attendon, it is important not to overlook the time element. Repository design considerations place severe constraints on the use of " engineering experience" and require an unprecec.ented reliance on predictive (often numerical) analysis.
Development of a convincing prediction of the performance of a waste package alloy thousands of years into the future, when that material may have been known for less than 100 years or so, is an example of the challenges involved.
Time Frames of Concern in Repository Design
' The following three periods ofinterest can be distinguished in the design and assessment oflong-term performance of a repository at Yucca Mountain:
Preclosure'- Between 100 and 300 years (i.e, M period from the start of repository excavation until the decision is ma/e yo "close" the filled repository).
. Although it would not be impossible to retrievt. waste from the closed repository, retrievability at Yucca Mountain is currently enwsaged to be accomplished only
- during the preclosure period. The drift support system should be designed for the preclosure period.
10,000 years beyond closure - This is the regulatory period specified in 10 CFR Part
- 63. If the total system performance assessment (7?A) computations presented in the license application submitted by DOE are deemed by NRC to provide reasonable assurance that individual doses to a reference critical group located 20 km from the 3 The 300-year upper limit was apparently chosen because it corresponds to ten half-lives of radioactive decay for cesium 137 and strontium 90.
e}
E:gineered B:rriers ct )?cc2 Mount:In, Some Impressions tnd Suggestions 8
repository do not exceed allowable limits at the end of 10,000 years after closure, the repository can be licensed.
Beyond 10,000 years - Although this period is strictly not part of 10 CFR Part 63, DOE acknowledges in the VA that doses will continue to increase significantly beyond 10,000 years, approaching the order of natural background radiation (Fig. 4.12 in VA, Vol. 3 shows a peak dose of 0.2 rem, at 200,000-300,000 years), almost an order of magnitude greater than the 25-mrem maxmimum dose allowed during the 10,000-year NRC regulatory period.
The U.S. National Academy of Sciences / National Research Council 1995 report, Technical Basesfor a Yucca Mountain Standard (TYMS,1995), recommended that the regulatory period be sufficient to cover the period of peak dose. As noted above, this period extends well beyond 100,000 years.
Some estimates indicate much higher doses than those given in the VA, as is illustrated in the following extract from a recent article by Carter and Pigford (1998)*:
Calculations by the project show that in 10,000 years the annual dosefrom drinking contaminated waterfrom the repository will be about 0.02 rem peryear.
Il7 ten the dosefrom eatingfood contaminated by irrigation waterfrom these same wells !s added, the total dose will be about 0.13 rem. This is 13 times the annual dose limit established by the U.S. Nuclear Regulatory Commission (NRC) two decades agoforpersons living near a nuclearpowerplant. It isfive times the
' Pigford, T. H., and E. D. Zwahlen, " Maximum Individual Dose and Vicinity-Average Dose for a Geologic Repository,
- Scientific Basis for Nuclear Waste Management XX, W. J. Gray and J. R. Triay, Eds, Materials Research Society, Pittsburgh, PA,1996, Vol. 465, pp.1099-1108.
Professor Pigford also recently provided the writer with the following details concerning the doses mentioned in the quotation:
For the dose calculations, we reliedfirst on the dose calculations in TSPA-95 (Akins, J.
E., J. H. Lee, S Lingineni, S Afishra, J. A. AlcNeish. D. C. Sassani, S D. Secoughian, "TotalSystem Performance Assessment-1995: An Evaluation ofthe Potential Yucca Afountain Repository," TRW, November 1995.) These doses uere calculated onlyfor drinking contaminated well water. Additional dosesfromfood chains were not included in TSPA-95. We utili:ed the graphs showing the cumulative complementary distribution functionsfor 1,000,000 years andfor 10,000ycars. We selectedthe drinking-water doses at a CCDF of 0.05, corresponding to a 95% confidence level. The 95% confidence level is commonly used in ergineering practice, it has been recommended by Britain 's NRPB, it was recommended in my dissent appearing in the National Research Council's TYAfS (1995) report, and it was incorporated in draft legislation proposed by Congress for Yucca Afountain.
From other graphs in 75PA-95 we idennfied which radionuclides were the principal contributors to these doses. From EPRI data (Smith, G. Af., B. M. Watkins, R. H. Little, H. Af. Jones, A. Af. Afortimerk, "Blosphere Afodeling andDose Assessmentfor Yucca Afountain, " EPRI Report TR-107190,1996) we derived the ratio of total individual dose to drinking-water dosefor each ofthe principal radionuclide contributors. Afuluplying the drinking-water doses derivedfrom TSPA-95 by the appropriate ratios yielded the doses reportedin our article in the Bulletin ofAtomic Scientists, I
l j
e Engineered Barriers at 12cca M:nt:In, Some ingpressions cnd Suggestions 9
two decades agoforpersons living near a nuclearpowerplant. It isfive times the annual dose the NRC allowsforpersons making unrestricted use ofa nuclear facility whose license has terminated. (The dose calculations allow a 5 percent probability ofdoses higher than those cited here.)
After 10,000 years, the calculated annual dose at a well three miles distant rises rapidly. Indeed, after 30,000 years, the annual dosefrom iodine 129 and technetium 99 will have increased about 80-fold, to 10 rems. Then the longer-term annualdosefrom neptunium 237 appears and rises to about 50 rem by about 100,000 years, amounting in less than a decade to an exceedingly high, hfe-shortening cumulative dose.
The energy department recognizes that these doses exceed reasonable standards forpublic healthprotection - hence the pressing needfor deeper analysis and a searchfor a more promising strategy.
It is likely that a license application showing a dose that is in compliance over a 10,000-year regulatory period, but that indicates significantly increasing doses beyond that time, will be subject to legal challenge even if considered acceptable by NRC. A repository design that could avoid this difficulty, if sue?. design is feasible, should be given serious consideration.
Engineering design considerations will differ depending on the period of concern. The pre-closure period, although considerable, is comparable to the usual time for which engineered structures (e.g., bridges, tunnels) are designed to perform. Primary concern will likely be occupational exposure of workers involved in construction and maintenance of the open repository and its contents.
As noted earlier, the much longer postclosure period (to 10,000 years and beyond) requires a less traditional engineering design approach. ' However, it is worth recalling that the decision to use underground (geological) settings for waste repositories was made, at least in part, because rock is a natural material that is known to have existed in stable form for many millions of years.
Prediction ofperformance for a small fraction of this time into the future involves much less uncertainty than is the case for fabricated materials that have been available on the order of 100 l
years only. (The Swedish [SKB] decision to select copper as their waste-package material was based in large part on the fact that native copper deposits are known to have survived for millions of years in groundwater environments similar to those proposed for their waste repository.)
Primary Attributes of a Yucca Mountain Repository Design DOE's viability assessment (VA) lists the following four main attributes of a repository at Yucca Mountain that can influence the release of radionuclides to the biosphere:
a water contacting the waste package; a
waste-package lifetime; a
mobilization rate of radionuclides; and a
concentration of radionuclides in water.
r E:gineered Barriers at 12cc3 Mou:t:in. Some Impressio:s cnd Suggestions 10 These attributes serve as primary guides for DOE in establishing its repository safety strategy (RSS). Each attribute has beenfurther subdivided intoprincipalfactors ofthe so-called reference design. Altemative design features have also been defined as possible contributors to an enhanced design (i.e., to improve th: overall safety of the repository). The inter-relationships among these :lements are all contributors to the RSS (see VA, Vol. 2, Table 8-3, p. 8-5).
Clearly, if water percolation into the waste-filled drifts could be avoided (i.e., if no water contacted any waste package), then the remaining three attributes become oflittle or no significance. All are dependent, in large measure, on contact of the groundwater with the waste package.
As noted by Shoesmith and Kolar (1998) in summarizing their study of the corrosion resistance of metallic alloys and the possibility oflong-lived waste packages:
ifthe contact ofseepage drips with the waste package is avoided, then extremely long lifetimes, in excess of10' years, are predicted. This wouldsuggest that the adoption ofany engineering option to avoid contact between drips and waste packages would be a goodidea.
Given the potential benefits of elimination of water contact with the waste package, it is surprising that little consideration has been given in the VA to:
(1) diversion ofinflowing water before it reaches the repository horizon, and j
(2) use of a multi-level design (i.e., to reduce the repository plan area, orfootprint, in order to minimize the potential for dripping into the drifts.
If, as appears to be the case at Yucca Mountain, flow through the unsaturated zone is predominantly vertical, at least in the southem portion of the proposed repository location, then elimination, or at least major reduction, ofinfiltration to the drifts seems technically feasible.
If net infiltration could be eliminated, major TSPA uncertainties would be removed, and doses would be reduced dramatically, especially beyond 10,000 years.
J Elimination of WaterInfiltration I
The following two engineering options are within current technology and offer the possibility of eliminating water inflow to the repository:
(1)
Surface modification (i.e., engineered fill), and l
(2)
Underground repository infiltration shield.
Surface modification is mentioned briefly in DOE's viability assessment (VA, Vol. 2, Sec. 8.2.2,
- p. 8-7 ). The repository shield concept is not considered.
I i
l i
l
E:gineered B:rriers at Wcea Mount:In. Some Inpressions end Suggestions
]1 l
Surface Modification Net infiltration into the mountain could be sigmficantly decreased ifthe surface of the mountain were modified.. Likewise,facilitiesfor drainage ofwater to enhance runo.[fcould be designed. Because these effects couldpotentially eliminate net infiltration at the site, thepotentialimportance toperformance could be high (VA, Vol. 2, p. 8-7, emphasis added).
Standard procedures of surface mining and site rehabilitation could be used to cover the repository site with an impermeable cap and drainage. As noted in the viability assessment:
Surface modifications and near-field rock treatment can be independently evaluated (i.e., without affecting other features of the design] so this alternative concept was not retainedforfurther consideration as an alternative design concept. Howe,er, the merits ofthesefeatures will be evaluated in a separate study (VA, Vol.2, Sec 8.2.4.2, p. 8-12).
' Surface modification treatments (e.g., several meters of thickness of an impermeable barrier, such as clay, overlain by a drainage layer oflarge river gravel covered by, say,10-15 m of alluvium) are well within cunent surface mining technology. However, the surface topography above the proposed repository is variable, so that this surface treatment could be costly and
. environmentally objectionable.
One of the potential shortcomings of surface modifications alluded to in the viability assessment (VA, Vol. 2, Table 8.5, p. 8-30), is the questionable longevity of such a barrier, due to erosion.
However, erosion rates at Yucca Mountain are estimated (VA, Vol.1, p. 2-26) to be less than 1.1 cm per 1000 years, or 11 m in 1 million years. DOE has given preliminary consideration to a more limited treatment of the surface, including a cover of alluvium over the existing surface
~
(E. L. Hardin, personal communication,1999), but this has not been pursued to date. Lack of permanence of the cover was one of the concerns cited.
Underground Repository Infiltration Shield, with Multi-Level Repository An underground infiltration shield is particularly well suited to a repository in the s.nsaturated zone in fractured rock, where groundwater flow is predominantly vertical and the rock mass is anisotropic, both hydrologically and mechanically. At Yucca Mountain, fracturing (subvertical) is such that the vertical hydraulic conductivity is significantly larger than the horizontal conductivity. Similarly, the modulus of deformation of the rock mass is larger in the vertical direction than in the horizontal direction.
The infiltration shield concept is illustrated in Figure 2. In the example shown, the repository is laid out as a three-level system.5 This alone, by reducing the plan area (footprint) of the 8Note that this would also reduce the probability of penetration of a vertical igneous dike intrusion by a 4
4 similar factor, e.g., from a probability of I x 10 /yr as currently estimated by NRC scientists to 3.3 x 10 /yr.
I Engineered B:rriers ct Y ccs Meurtain, SomeInqpressioxs c:dSuggestiozs 12
- repository to one-third of a single-level design, reduces the exposure of the drifts to vertical'
- infiltration by a factor of three. Although the shield principle can be applied to a single-level
[
repository des'ign, it is obviously more cost effective to use a multi-level design.
A numerical analysis of the effect of placing a fourth row of drifts (left open, for example, as ventilated observation and performance confirmation drifts a VA, Vol. 2, p. 4-45) above the three repository levels was carried out by Professor Pierre Puochet, University of Neuchatel, Switzerland, using the numerical (hydrological) code FEFLOW. The analysis, with assumptions and results, is outlined in Appendix I to this paper.
A single typical column of drifts was analyzed. This corresponded to the central column shown in the upper diagram in Figure 2, but with the upper slot replaced by a circular drift (see diagram in Appendix I). The flow conditions and rock mass properties were considered to be representative of those in the unsaturated zone at Yucca Mountain. It was assumed that the rock mass could be considered to behave as an anisotropic continuum (i.e., discrete fractures were not -
considered). A uniform vertical infiltration of 50 mm/yr (1096 x 10 m'/d over the 80 m 4
2 potential capture area (per meter of drift) was assumed to occur 30 m above the top row of drifts.
A wide range of hydraulic anisotropy was examined. For all anisotropies considered, at least 94% of the top infiltration bypasses the lower three (rows of) drifts. The fluid pressure head above the lower drifts is reduced because of the proximity of the overlying drift, thus enhancmg the potential for diversion 'of water around the lower drifts.
.This calculation can be criticized in that it assumes the drifts to be circular and smooth, thus enhancing flow deviation around the drifts - as indicated in Figure 3 (after Philip et al.,1989; Philip,1990). The presence of discrete fractures in the roof would increase the potential for 1 water to drip into the drifts compared to the case analyzed - viz. that of a smooth opening in a continuum.
This criticism can be circumvented if the upper drift is replaced by a slot, say,2 m high and
~10 m-20 m wide. Each such slot could be inclined slightly, as shown in Figure 2, and backfilled so as to establish aflow diversion barrier, to ensure that any infiltration from above the slot would drain into the rock mass outside the perimeter of the repository. Excavation of the 15'm-20 m slot would serve a dual purpose. A zone of enhanced fracturing would tend to develop above the slot (this would be further enhanced during the thermal cycle after the i
repository is filled with waste.) Any water infiltrating into the zone would drain into the slot; any
' remaining flow would be directed into the rock mass away from the drifts. Thus, both mechanisms (capillary diversion around and fracture flow into the slot) act to prevent flow into the drifts.
i
EngineeredBarriers at l'ucca Mountain Someimpressions andSuggestions
]3 overhead slots 10 rn - 20m wide surface
/
N impermeable cap
,i wm,.
pg n
/
_f
._. _,fa~
~
22.
l; C
l nmarwmewr~
~
l V
/ 1.0 m g..p;W:, crushed tuff p *!
tem Slot i
- gravel's
)'
1.0 m [A w-r, F
.w, r m Note: verbcal slots, if needed, would be placed on
(
o%~9 coarse ^*
all four sides of repository to depth of bottom tier of dnfts (outline of two vertcal slots shown as wtute dots)
\\
M f
Richards barnerin slots s
a)
Perspective View s
/
N
/
ver0 cal penmeter slot, broken rock slots containing 2 m wide, if needed, Richards barner to ensure no lateral flow zone
- W*****'
Surface
\\
l4 A
A' 2*
.A.
y I
ka=*"4
==
a==s
==== }
.m:n
(
,==,
Q o ["
l
'%p 2 m tsgh j
o o
o o
o o
l i Jo40m o!
o o [
o o
o o
p
!Z l
Q o V' 3Ddom g
[,,
o o
o j
o o
o l
4[
5 i
e g
i l
- a t
l 1.6 km
- l y
b)
Cross section normal to long axes of drifts (20x 3 ter columns) s A
A' l:
a 1 krt
- l
,Richards barner 3% rade 3% rade 2m
$ 3M m waste emplacement dnfts
$30J.0m 3 ter
(
verbcal penmeter slot (-2 m width, if needed) to ensure no lateral flow of groundwater c)
Section perpendicular to AA' Note: The G0 km of dnfts in the EDA 11 design could be accommodated as 3 layers of 20 dnfts per layer, each onft i krn long Figure 2. Underground Repository Infiltration Shield i
Engin:ered B:rriers ct l'ucc2 Mou:tain, Some Impressio:s tndSuggestions
].$
f Doundary
{, stynpto-lC te ps absa Upr.1rea r 2 = -Sd ret ar::ed Smnda y ID00
.fn i ri
.nyre p
oin,,,.g p
and 0- 4,,
j %'
,/
gy 3
Cf 10!?
^!O X
$!N
$h U5 thin inlllUO; 01:0; R:<t-4!**0 0; IU0 lIj %-A.
drip IUUIZ erJp lobe !Ef5N 161350 :ot>s
'OT.U*+0!
C1$.C,
+ + ~
wf OINI Downst reeni 5
'0!!00 re'erded
\\ owncrearn region G w ndary
/
D of COA'r.stresrn StSQ'16tlCn retsrded region Dry shedow D0irl. V
- O hes last inside
% J<1 t.rc d = V m dry snadow bouridery l
+-. g y ) "
Bound &ry abyt"Clutic j
I to parsbola e g zo sx*
Figure 3.
Seepage around cylindrical cavities (schematic diagram illustrating criticalpoints and regions oftheflowfield)
Richards Barrier This flow-diversion system incorporates two layers of material with contrasting hydraulic conductivities - a fine-grained porous layer overlying a coarser-grained layer, also porous (see j
EPRI (1996), pp.1-2 et seq. for details). The capillary pressure established within the pore space in the upper layer material at the interface with the lower layer acts to prevent flow into the lower layer and promote flow laterally in the upper layer. Currently, the DOE is engaged in considerable study of the Richards barrier. The intention is to cover the waste packages with a
" tailored backfill" possibly designed as a Richards barrier to divert water drips from the roof of the drift away from the packages (see Figure 4)6 Figure 2 shows a similar two-layer arrangement l
of backfill for the slots in the proposed repository shield.
'It may be that the behavior of a Richards barrier over very long times (i.e.,10,000 years and longer) could i
be considered doubtful. It is believed that a simple drain, consisting of graded, more or lessuniformly sized granite boulder (river gravel) would suffice to establish free draining of the slots.
i Engineered Barriers at Yucca Mountain, Some Impressions cnd Suggestions
]5
., v Ib.
f,
3 Ceramic Coating
'o 1
Backfill
,. 3,. Drip Shield f,.
e J
- 1 C9
.<h
~
e s
- /
[
_,tv q
Yy
\\
~
4 s..
ry T,,
/ra L
i e
e, a r>
I
- l.
\\
_,; j w,n.e +. p.p g, -
r Figure 4.
Near-Field Engineering Measures to Prevent Dripping on to Waste Packages (It isplanned to place the backfillin two layers as a Richards Barrier, withfine-grained rock rnaterial overlying a coarser-grained rock material.)
Potential for Lateral Flow at the Repository Horizon The repository shield design described above is designed to be effective against vertical infiltration. It will fail if there is significant lateral flow across the repository. Lateral flow is possible, and is known to occui both above and below the proposed repository horizon. Within the proposed horizon (particularly, the southern region), flow appears to be dominantly vertical.
As noted in the DOE viability assessment (VA, Vol.1, p. 2-38),
... evidence indicates that surface infiltration generally moves downward rapidly infractures through the Tiva Canyon tuffuntil it encounters the non-welded Paintbrush ruff Flow in the non-welded unit appears to be predominantly in the rock matrix althoughfastflow paths alongfaults, fractures and other high permeability :ones are present locally. In general, it appears that the Paintbrush non-welded unit attenuates (slows) and distributesflow downward, perhapsfor periods ofup to thousands ofyears. After migrating through the Paintbrush tuff water moves into the welded Topopah Spring tuff (Note: The proposed tepository horizon is in the Topopah Springs formations] whereflow again appears to be dominantly in thefractures. The distribution offlow is heterogeneous; in some areas characteri:ed by widely dispersed or poorly connectedfracture systems, percolationflutes may be very low. In areas with highly transmissivefeatures i
W
~ ~..
I E:gbeeered &rriers at Y:ces Mourt:in, Some impressions cad Suggestions 16 such asfaults or densefracture networks, sigmficant volumes ofwater may move downwardrapidly.
l l
This discussion suggests that lateral flow across the repository is likely to be minimal, so that (horizontal) slots above the waste filled drifts will eliminate most, if not all, of the potential infiltration into the repository. It is entirely feasible technically, if deemed advisable to further 1
. reduce uncertainty, to construct a vertical perimeter shield around the entire repository, as shown l
. in Figure 2. This would require a single vertical column of four 5-m-diameter drifts, located on the same level as the repository drifts and slots, along each side of the repository periphery. A
)
narrow vertical zone of enhanced permeability could then established by blasting, using the VCR (vertical crater retreat) method (or a similar stopping procedure). Blasting would be conducted in vertical holes drilled downward from each overlying drift. The blasted rock would fill the L
underlying drift such that little,' if any, of the broken rock would need to be removed. The aim is to establish a highly transmissive vertical flow pathway around the periphery of the repository; it is not necessary or desirable to create a vertical excavation. Alternate, less expensive techniques (e.g., creation and propping of hydraulic fractures from vertical holes along the drifts) could also
' be considered.
The horizontal slots and perimeter drifts could be used for monitoring (e.g., by microseismic and other geophysical techniques) repository performance during the preclosure period and beyond, if necessary. Since these openings would be ventilated during this period, any infiltration would be carried out as vapor in the air stream.
' Additional Excavation Required for the Repository Shield i
NorizontalSlots Only-The total excavation to develop 20 m-wide x 2 m-high slots would be i
the equivalent of 40 km of 5 m-diameter drifts. The EDA II repository design envisages a total of 60 km of waste-filled drifts, Thus, addition of the 20 m excavation slots would result in a total excavated volume less than the 110 km of drift excavation contemplated in the VA repository design.
' " Full" Shield - The four drifts along the entire repository perimeter, if needed, would add a further 21 km of excavation (i.e.,4 x 2(1.6 + 1.0) km). It may be possible, in view of the reduced concern over reflux pathways between the (columns of) drifts, to reduce the spacing between
- drifts (currently 81 m).. This would reduce the extent of the repository footprint plus the cost of generating the high-permeability vertical fracture zone between the drifts.
However, it is considered unlikely that construction of these vertical high-permability zones will be needed provided the repository horizon is selected appropriately, i.e., where the two sub-l vertical joint sets are both well developed. They are orthogonal to each other, thus forming an effective barrier to lateral flow across the repository.
The preceding discussion suggests that it is technically feasible to ensure that essentially no infiltration into the repository ever occurs, for a cost that would not significantly exceed that of
, the VA repository design. This does not consider the added cost of a three -level repository compared to the VA single-level repository. DOE has considered a two-tier or split-level repository option, but did not examine the potential for water diversion. An increased cost of construction of 19% compared to'the VA reference design was indicated (CWRMS/M&O Report
. =,
Engineered Birriers K! Yucca Mo:nt:In, Some Impressions cndSuggestions -
17 t-Design Feature Evaluation #25, Repository Horizon Elevation, April 2,1999). It is also worth noting that the repository shield requires no reliance on the long-term performance of manmade materials. It should be relatively easy to establish the very long-time reliability of the repository i.
shield.
The distinct possibility that the repository shield concept could reduce drift infiltration sufficiently to make the titanium drip shield (Figure 4) unnecessary - for a cost saving of $4.6 l
billion - strongly suggests that the repository shield concept deserves detailed study by DOE.
Such a study should examine the implications of the multi-level arrangement (with overlying slots) on the optimum repository design.
Location of a Multi-level Repository at Yucca Mountain l
l A three-tier repository, as shown in Figure 2, would occupy a vertical interval of approximately 60 m-80 m in the Topopah Springs formation. Since the horizon proposed currently for the l
single-level repository is approximately at elevation 1080 m it appears that a three-tier interval from 1,040 m to 1,120 m in the central third of the current repository (see VA, Vol. 2, Fig. 4.21,
- p. 4-40) will remain well within the " groundwater surface plus 100 m" lower limit and within the "200 m cover" upper limit. The slot horizon would be some 30 m or so above the upper row of 1
drifts, but this too will have almost 200 m of rock cover. Since the slot would contain no waste, l
a cover slightly less than 200 m is considered adequate.
Optimum Repository Layout The VA reference design was a " hot repository" in which rock temperatures in excess of 200 C were envisaged. A main intent was to prevent access ofliquid water to the waste packages, at least for much of the regulatory period. Concern over the uncertainties associated with two-phase fluid flow behavior in the near-field of the repository and associated complexity of coupled (thermo-hydrological-mechanical-chemical) effects, especially in the near-field around the drifts, led to calls to revise the design to one in which the rock temperature was lower, preferably below the boiling point of water for much of the duration of the thermal cycle. The EDA II " lower temperature" design responds to these concerns.
I The two designs are compared in Table 1 (from the presentation " Current Status of Repository Design," by Daniel G. McKenzie III, to the Drift Stability Panel, April 13,1999).
The EDA II design has some merits, but also some disadvantages. Although the lower temperature system may be simpler (perhaps/) for purposes of analysis of near-field fluid (liquid l
water and water vapor) movement, the possibility that the high-temperature design may inhibit access ofliquid water to the drifts is a feature that should not be abandoned lightly. Center for j
Nuclear Waste Regulatory Anslysis (CNWRA) staff (R. Green, personal communications,1999) suggests that some counter-current flow may occur, -whereby water vapor may ascend within a fracture while liquid water may descend into the drift via the same fracture. The importance of this possibility in the context of a repository shield design would need to be assessed.) Also, as noted in the EPRI report (EPRI,1996, p.1-2):
l
i
\\
E:gineered Rxrriers it Yuccz Mount:in, Some Impressions end Suggestions 18 1
The proposed DOE schemesfor lower thermal loadings would not eliminate completely any ofthe coupled thermal effects causing concern at Yucca Mountain, although the proposed schemes would reduce the magnitude ofat least some ofthese effects. For example, loweringpeak temperatures below the boiling point does not eliminate the potentialfor evaporation of liquid waterfrom the rockfollowed by buoyant convection and subsequent condensationfarther afield.
In order to reduce dramatically thermal effects in the very nearfield around the containers, the amount ofspentfuel contained in an individual container would have to be dramaically reduced or the decay time ofthe spentfuel would have to be sigmficantly extended (well beyond 100 years). Neither ofthese approaches seems so practical since both would dramatically increase disposal costs.
Table 1. Comparison Between the EDA 11 and VA Repository Design Options J
EDA II Design DOE VA Design 60 MTU/ acre 85 MTU/ acre 1,050 acre-layout 741 acre-layout 60,000 m of emplacement 117,000 m of drifting for statutory waste emplacement drifting capacity for statutory waste capacity 2-5 m' /s/ drift airflow 0.1 m '/s/ drift airflow 81 m drift spacing 28 m drift spcing Line load Point load (3 m between packages)
It is instructive, in this regard, to consider the performance of a multi-level EDA II design, as illustrated in Figure 2. The switch to a "line load" of waste packages (i.e., with the packages placed essentially adjacent to each other along the drift) compared to a " point load"(packages separated by several meters along the drift) and a much increased spacing between drifts (81 m for EDA II; 28 m for the VA Reference Design), together with some (low) velocity ventilation of the EDA II drifts, was intended to simplify the convective flow paths with reflux via the cool region in the center of each pillar.
With the multi-level design, the rock temperatures are likely to be increased, principally along the vertical axis between the drifts. The region between the pillars will be less affected, although raised somewhat. Convection cells of heated water and water vapor would form, driving the fluids upward into the slots, where it would tend to condense on the coarser rock in the lower portion of the Richards barrier, flowing along the inclined drifts to drain outside the repository.
Continued heating would eventually dry out the rock between each column of drifts. This
g
{ :.*, ;
_ ; Ergineered B:rriers ct hecs Mouxtain, Sonee Ingpressio:s c:d Suggestions
]9 h
pathway provided by the slots would tend to eliminate the need for a pathway for the condensed reflux between the pillars, although concentration of the overburden stress through the pillars -
[
would induce a small tension tangential to the central vertical axis of the pillar, thereby tending to open the reflux pathway. In this regard, it should be noted that the intensity of the vertical
- stress concentrations in the pillars will persist to a greater depth than in the case ofisotropic and unjointed rock (i.e., the " aperture opening" effect may be more significant in the jointed rock (see Goodman,1989, Figs. 9.10 and 9.11 pp. 352-361). Shears induced at the corners of the slots o
could also cause fracture dilation, especially during thermal cycles.
It should be possible to reduce the 80-m drift spacing of EDA II somewhat (say, to 50 m). This would increase the temperature along the center-pillar axis, but the stress concentration in the now narrower pillar between the slots would increase, which may increase shear and dilation of I
fractures. The reduced pillar size would reduce the plan area of the repository, thereby either reducing the extent of any vertical perimeter shield or increasing the capacity of the repository.
Chemical dissolution of minerals species (e.g.,' silicates) in the rock by the hotter fluids in the i
near field, with condensation upon reaching the slots would tend to develop a low-permeability 1
" skin" along the slot floor during the thermal period. This would be beneficial to drainage of condensate along the drift.
Obviously, more detailed analysis and optimization studies are needed to establish the merits of
'the multi-level design with the repository shield in order to establish the merits of this concept
]j vis-a-vis the proposed single-level designs.
I Control of Repository Temperature Reference has already been made to the perceived benefits of reduced repository temperatures in order to simplify the near-field fluid flow regime. Low temperatures are also desirable to reduce corrosion of the waste packages. The EDA II waste package involves a 2-cm-thick outer cylinder of C-22 alloy steel, with a 5-cm-thick inner cylinder of stainless steel (316NG).
l Shoesmith and K.olar (1998) argue that pitting and crevice corrosion of C-22 are unlikely to occur at temperatures below 150 'C and 102 C, respectively. The authors present detailed
' discussion of the corrosion processes, but conclude that a conservative design limit is to take 80 'C as the temperature below which crevice corrosion of C-22 can not occur (Shoesmith and Kolar,1998, p. 5-8, para.1). Also, it is noted that water must be present for significant waste package corrosion to occur. A relative humidity less than 70% and a temperature below 80 C are sufficient to reduce the possibilities of corrosion of the C-22 alloy to insignificant values, i.e., yielding estimates of waste-package lifetimes considerably longer than the 10,000 years of the regulatory period.
i
r1 O
se.
E:gineered B:rriers et l'uces Mourt:in. Some Impressions and Suggestions 20 l
Heat Removal by Continuous Ventilation (2 m*31s)
For EDACase 4 icoo
-_~
I I
(-. - :-- - - --:-- -...:... --. - - -- 4 -. - - 4.. - --
- *~--
~
ano g
Heat Removal try Ventilat60n:
N'.
l l
1-l 39.88% for 60 ears Y
800
---d 4.
(--Rate of Wenas Heet 43.13% for 100 years t70#
...-y:..\\oisput > g Drm 50.40% for 300 years 18
/'
l 1
Ea
...% t. - 4........ -. - - 4...... g.. --.m...y.~.
.y.;..;......,.... 4...... ;..... g....
v.;,,
y
!
- s. g:
I V
L ' :- - : ---- - ------ --l --- '----
200 RM e
Rp. e of.
, 'eovd. m e
.M-
. ' _ - 1
' 5-1.-. - L.
1* - l =
.l w, :
p o L.. -
. s.+ ~.
0 30 dc IC 40 100 120 140 100 140 203 230 240 Dec 280
'EC verstamon 7tme. Year Heat Removal by Continuous Ventll-tion (6 m^3/s)
Fw EDACase 4 8000
-i r.
soo 4.----t..
^--s*
- 3 Heat Removal try Venstetiom
\\
a
.g
\\.:
as.ss%:or so years an asas on waste Hogs 69.D% for 160 yea s 74
. W.
j
. d la. bre i ET.M% for 300 years 8-pa
. 3- -...%.. <.......g,.... q..... g.....
,.....;...= p.g.:....,.......,.......... q.
en 1
- \\*
l
{
/
h%'
A r
w' cf est.
- - -- - * *- k-4* E * *.y* * *.; * * * * *k * * * - * * *
- I * - I.
-
- 5' I.
200 ' -**
' - - ! w !
L R.0mova.lin a D.rft e
t t
t t--
1 i
i i
0 20 40 fm SC 100 120 14C 180 180 203 230 3e0 300 300 300 venues woe Yw Figure 5.
Heat removed by continuous ventilation of waste-filled dn'fts during the pre-l closure period (2m '/s airflow in a 5-m diameter dnft corresponds to an air velocity of20ft/ min, or 0.23 mph)
L l
Engineered Barriers at Yucca Mountain. Some Inpressions andSuggestions 21 Temperatures for EDA II (2D, ANSYS, TL-60 MTU/ac, DD-5.5m, DS=81m, O=2m^3/s, No Aging)
-WP Suttace (50yr)
-Dntt Wau (50yr) 250 j
"""""Mddled PWlar (50yr) g
-WP Surface 95yr)
J mO f
O I
-Dntt Wan F5yr) iI (D
-Eddle of-Pillar 95yr)
$ 200
-WP surtaee (tooyr) 200
[
-Detft Wan (100yr) o
-uddieser ocoyo
+
,-f- - -
150 g 150,- - -
\\.
2 i
l E
k100 100 E
b 50 30 t
q 0
0 1
10 100 1,000 10.000 Time (Years)
Figure 6.
Effect ofbackfillon the evolution oftemperature in the repository dnfts (EDA H design)
Figure 5 (kindly provided by the DOE, courtesy of R. Craun,1999) indicates that drift ventilation can remove significant quantities of heat from the waste packages, especially during the first 50-100 years, when heat generation is most intense. Figure 6 (courtesy of D.G. McKenzie, April 5
1999) indicates that, with the EDA 11 design, drift ventilation of 2 m /s, and no aging of the waste:
(1)
The waste package surface and the drift wall both exceed 150 C for several years after installation, and (2)
Active ventilation of the drifts, either natural or forced, can also reduce the humidity.
Stellavato and Montazer (1996, pp. 25-26) have used the atmospheric / hydrologic code ATOUGH to model heat removal from a ventilated repository.
They advocate design of the repository to allow air to flow continuously and indefinitely through the waste-filled drifts driven by natural ventilation. In their report, the authors conclude that:
By considering a naturally ventilated repository (after construction) and taking advantage ofthe thermal drive ofthe waste package, the repository may be kept dry during at least thefirst 10,000 years ifnot longer. The amount ofmoisture removedfrom the rocks during this nme will create a thick low-saturation skin
r 9
4
. Engineered Barriers at Y. recs Mountain, Some impressions c:d S:ggestions 22 around the drifts that will require thousands ofyears to re-saturate. Ventilation can also remove large amounts ofheat generated by the waste canisters.
The authors' analysis indicates that the rock temperature never exceeds 25 *C during the ventilation period. The topography and surface layout of the proposed repository at Yucca Mountain is favorable to natural ventilation (and ventilation produced by waste heat generation),
but it seems likely that the drifts will collapse over time, increasing the resistance to ventilation.
Partial filling of the drifts with " moderately large" boulders to ensure some air access to the packages could be considered, but the resultant overall resistance to flow would be considerable.
Clearly, there is merit in preclosure ventilation of the repository with respect to limiting temperatures. Ventilation also tends to develop a "dryout" zone in die rock. Measurements over the past several years suggest that a region of approximately 100-mm radial thickness is dried out annually. Although the radial extent may not increase linearly with time, it appears that a region not greater than 10 m from the drift excavation will be " dried" over 100 years. With interruption
' of ventilation, this region will resaturate, probably at a comparable rate, so that the drift will be resaturated (i.e., partially) after the order of 200 years from installation of the waste. Thus, for almost all of the 10,000 years of the regulatory period, the waste packages (and backfill?) would
^
be subjes to a humid environment. With the C-22 alloy outer cover of the packages, and a package temperature not significantly above 100 *C, the alloy will corrode very slowly, if at all.
This resaturation rate would be slowed considerably if the repository shield concept was used.
The preceding calculations suggest that, if one would hold the temperature of the C-22 waste package below "O 'C, some combination of waste form " blending"in the drifts, aging of the waste in surface facilities before emplacement in the repository, and active vigorous ventilation of the packages for at least 50-100 years may be necessary in open drifts. An open drift implies that the waste package will not be covered. -i.e., the waste package surfaces should be accessible to the ventilation. Tailored or " getter" backfill in the drift invert below the waste package could still be used.
Design considerations such as those outlined above suggest that it is entirely possible to engineer the natural setting of the unsaturated zone at Yucca Mountain to ensure that a high-level waste repository will be demonstrably safe for an indefinite period into the future. The umbrella principle of the repository shield is simple and can be comprehended easily by the general public.
Drift Stability It is planned to locate the repository in the Topopah Springs tuff formations. For purposes of drift support / reinforcement and stability analyses, the formations can be divided into two general categories:
i 1
(1)
Non-lithophysal tug These formations contain three relatively well-developedjoint sets.
(Two are subvertical: joint set No. I has a dip of 77* and a dip direction of 40 ; joint set No. 2 has a dip of 80* and a dip direction of 130. One is sub-horizontal: joint set No. 3 has a dip of 25 and a dip direction of 300 ); and (2)
Lithophysaltuff. -These formations contain three-dimensional voids-approximating spheres or ellipsoids in most cases - or lithophysae generated as gas pockets during the
.o 4
- Engineered B*rriers et Yxeca Mount:In. Sime ingpressions ands:ggestions 23 period of deposition of the volcanic tuff. Some of the lithophysae can approach 0.5 m in
'. diameter, although most are smaller (predominantly 7-15 cm in diameter). Also, j
fractures in the lithophysal rock are shorter and less persistent than in the other units, and.
often terminate (or originate?) at the lithophysae >
It seems likely that the lithopyhsal zones will be stronger and stiffer (i.e.,
higher rock mass modulus) than the non-lithophysal zones because of the
. lesser influence of through-going joints. The higher modulus would result
> in higher thermally induced stresses for a given temperature, so that the extent ofdamage during the thermal cycle could be comparable for both
- lithophysal and non-lithophysal tuffs.
it seems to the writer that excavations with rock reinforcement should be stable in both
' formations.' The following discussion will examine the likely mechanical response of the two types of formation to loads generated in a repository. The stability of the repository drifts is of i
particular importance for the preclosure period, and can have consequences for the long-term
)
performance of the repository, especially if the drifts are not backfilled.
l Preclosure Stability l
Although there is a wealth of experience in designing and constructing tunnels of the general dimensions of the repository drifts, and there are examples of tunnels that have remained stable
- for much longer than 100-300 years, design of a repository is unique in that a major thermal cycle is involved. For the case of a hot repository, this heating imposes substantial additional 4
stresses on the rock and any rock lining. The likelihood that a concrete lining would be seriously 4
and adversely affected by the I igh temperatures is - in part, at least - the reason why an Expert Panel on Drift Stability has recently recommended the use of rock bolts and wire mesh as being a more suitable support system than a concrete liner.
Postclosure Stability i
DOE lists the following information needed with respect to performance assessment (PA) for ground support / drift stability (R. Howard, Yucca Mountain Drift Stability Panel, April 13,1999):
Ground Support /Dnft Stability Information Needsfor PA (FEPs)'
a masses and spatial distribution of ground support materials e
nature and rates of continuous degradation processes e
nature and probability of disruption by rock fall 4
e nature and probability of disruption by seismic motion Of these, the first can be answered as soon as a support system is selected. The remaining three require an understanding of the long-term, time-dependent behavior of the rock mass only 17the dnyts are not backfdled. If the drifts are backfilled, then these issues are no longer of concern.
I-
' FEPs are features, events, and processes that are considered to influence repository l performance.
e s lEngineeredDirriers at Y:ces Mortnin. Some Inpressions andSuggestions 24 1.,.
3-
. ' No' firm decision has yet been made concerning whether to backfill the drifts after waste
emplacement.
Numerical (discrete element) models currently in use to assess drift stability at Yucca Mountain have a significant limitation in that the rock blocks in these models, although deformable, are assumed to have infinite strength (i.e., they cannot break). This results in significant over-estimation of the consequences of rock falls on to waste packages.' Considerable improvement in prediction of both (1) the consequences of heating on spalling of the drift walls and (2) the behavior of falling blocks can be obtained using a code such as the micro-mechanics numerical code PFC (Potyondy and Cundall,1999) that allows the blocks to break under applied loading.
' Some indication of the difference that may be expected is demonstrated by the simple example of a rock block falling 2 m from the roof of the drift onto a waste package, as shown in Figure 7.
The resultant force-versus-time history during the impact is shown in Figures 7(b) and 7(c) for
. the two cases in which (b) the block has infinite strength, and (c) a similar block has the (finite) strength of Yucca Mountain tuff. Fragmentation of the block (Figure 7(c)) traps a substantial proportion of the kinetic energy and momentum of the block with the result that, in this c se, the peak force on the waste package is reduced to approximately one-third of the value indicated with the infinitely strong block.
- Thermal loading and seismic effects can be considered in the PFC code. The rate of degradation over a long time can also be estimated, but this would require laboratory data on the strength of tuff (and joints in tuff) as a function of applied loading conditions (and possibly thermal conditions).' Such data may not be available. Thepattern of collapse with time can be examined for various assumed strength-degradation models. If this indicates that the pattern is relatively
)
- independent of rate of degradation, knowledge of the degradation pattem may suffice for PA purposes. Another approximate approach is to assume that thejoint cohesion declines progressively in time toward zero. Frictional properties may decline somewhat, but are likely to remain significant.
. It is anticipated that an analysis using PFC would indicate progressive spalling of the drift wall and collapse of relatively small rock blocks on to the packages. This would further reduce the severity of any rockfalls on to the waste packages.
Time-dependent deterioration of rock strength (and possible collapse) can occur whenever rock is loaded in compression beyond 40% to 50% ofits ultimate compression strength. Stresses I
significantly above this level could be generated in the rock during the thermal pulse period of repository operation. (In the case of Yucca Mountain, the stress induced in the rock by
)
temperature increase is approximately 0.5MPa/ C for an assumed modulus of deformation of the rock mass of E = 6 GPa.)
)
Recognition of the limited value of classical geotechnical engineering design approaches in prediction of rock mass behavior for repository design has stimulated studies to obtain a more fundamental understanding of the physical principles that control time-dependent failure in rock.
The report by Potyondy and Cundall (1999), describing studies being conducted for the Canadian nuclear-waste isolation program (and including the influence of heat in degrading rock strength with time) outlines valuable developments on this topic.
c:
.s E gineeredB:rriers ct Yuces Mr tain, Someimpressioxs c dSzggestions 25 f
5 i
l
(.>
_ __ _m_.
T&c=:ss o
T&:: :=: :
g_-%
yunwijjip-g lg r
lgd"@ Kg!.@El:..,
_m y
j
,.unrraum.%
sustra:n.n
[e 7 _,
l [ i4g..U
~
1 4- - - - - - r
- Aviglif, 7"
m g ryg y @l.l$ j,h g:,
p
.,.ne:,,
1
____ z
- s ig-yfy
,ggy wg s
5
,h j&i.
o (b)
(c)
Figure 7.
Effect offinite rock strength on the impulse generated by freefall of a rock block onto a wastepackage (PFC model) (The block in Fig. 6(c) has the same deformability as the (infinite strength) block ofFig. 6(b), but a strength corresponding to that of Yucca Mountain tuff)
Effect of Heating on Drift Stability Figure 8 illustrates the change in stresses produced in the periphery of an unsupported drift as the result of heating, in this case to 145 C, assuming that the rock mass has properties almost equal to those ofintact tuff (i.e., RMQ 5). The initial insitu stresses were assumed to be approximately 10 MPa vertical. (This is equivalent to a depth approaching 400 m and 3 MPa horizontally).
Under these stress conditions, the tangential stresses around the drift preceding heating would reach a maximum compression of approximately 26 MPa acting vertically across the central horizontal axis. Assuming a rock mass modulus of 32 GPa (i.e., RMQS rock properties), the effect of heating to 145 C is to' add compression on the order of 120 MPa more or less uniformly around the tunnel wall if the rock retains the RMQ5 properties and remains elastic.
.g -
E:gineeredCcrriers at Yzcc:s Mou:tain. Some Ingpressions andSuggestions 26 If the rock properties are degraded to those of rock of RMQ1 quality, the stresses shown by the solid lines in Figure 8 are developed. The effect of a total of 50 years of heating, after which high temperaturer (and stresses) have penetrated further into the rock, is shown by the dotted stress distribution. This results in a zone ofinelastic deformation such as indicated in Figures 8 and 9. It is seen that the stress distribution and extent ofinelastic deformation depend heavily on the rock properties. Recent results ofinsitu modulus measurements in the heated drift experiment indicate that the rock mass modulus (of deformation) increases from the order of 6~7 GPa at ambient temperature to higher values at higher temperatures. This is due, very likely, to expansion of the rock and consequent closure of the rock joints with increase in rock temperature. It is unlikely that the rock mass modulus in the jointed rock will reach the laboratory value for intact rock (32 GPa). In the lithophysal tuff, however, the modulus can be expected to be higher than in the non-lithophysal jointed tuff.
e mg u 10 MPa
- - twr 3 Mg l
~ 40 4
~
3o em:
I
- 20 taar.was.tasse pews:
~ faaw erag: p:pm:.1ssm;eq j
- " DVra4 HW; prpaond slre1;s.;
14 m Figure 8. Effect ofHeating on Stresses Around One ofa Series ofExcavations Figure 9 shows the extent ofjoint slip that occurs before (Figure 9(a)) and after (Figure 9(b))
heating when a PMQS-qualityjointed rock mass is subject to heating as described for Figure 8.
Figures 9(c) and 9(d) show the results of numerical modeling in which a 5-m-diameter l
unsupported open drift is subjected to two identical seismic events, one that occurs before heating (Figure 8(c)); the other (Figure 8(d)) that occurs after 50 years of heating of the rock to a maximum temperature of 145 *C at the tunnel wall. The regions ofjoint slip are shown in Figures 9(a) and 9(b), and the rockfall due to the two seismic events in Figures 9(c)and 9(d). It is seen that the rock fall is considerably reduced for the heated rock. This is because the increased l
temperature superimposes a high compression all around the tunnel, tending to " clamp" the rock l
blocks together, and preventing fallout, i
p 4
Engineered Barriers at YCcc3 Mountain, Some impressions cad Suggestions 27 m
14,- ~.: g... :6.
~ T,
~~
'~
- g i -
.---.r l' '.""..
- g. -T.I.;p;'m* '.p-7 cl.
[.==== - p-[.rt. j L
..g d: g T 4. 2.r..h.s, mI e...T -' ;
,o
.. e : n. -r
.J. T- ? -
1 i
, - O;-'., m.
,fy 7 7e t
-.$.g.r., r p
. ;.-,,,,, fp.,.P 4r; q> -
' -.. m:a,, J:;in.L r
d;xj;r 7o..g [;gr pT' dJ..i.,f 4-
~
c
- y W '.g!. M,Q, b.F T,
.. V..hr. c).
l UQ u qJ.
..i..
r-i.'d.d ".
h.f.fr{.k.y"y M+.p~fS5I3, r-r '$. i v.
kh-..'j rh".
]:),{f Y
- ilt.fr(=1 ';
(iI'rd
..J' 2
d.
I
'y ;1,4;t..j;-y1(.f{,l N
J i
t Jl
..r i
-].l. tnt.fiW t,p r
[
., s g
i cy-f j.s.J.
[
i t.
t I
tk.
.k p.I /4ings' 1..
T I
d$1&.~hf V
~ h-fh&@.f.;Lh&
- j@r'lg;
' p st.el,t
- .c5.,y O
a s.r.ftr L ' $.e g lnl @l &.hhjk$~y $
. J
~
7sw l h%
@!pfst.:'. a$pr1 3.jy..J --
U$1fr
.f.i.1 r.d"NJ:!r<-
m FI
- ]
- ;;%h.,. %. J:r'.p:!rei.,rd. sh-M
_1.#.c t.s_
.,_f ~ >
y5 7 iQ..e'& # ti M. 2,. L L d J h. t r. p..i1
' p..
gi 3
4 y
d.by in.~{ Y1 Y
- y
. 3E[).7-m m
o.
m J
i
[f,i[.d tr].tl.,E.byI*[h
- j
.~.-.,r
, q,r;7,$g.fh((.Y5f.h -
1 ~
I
!73 r 3.g Jd I
,t',
7 j,.-
,J FM '-rt*'iiI 3..yT4 i
rF
,..l7rt,3 1
ti'T.al rfg.
T 1
d rfl* '
I!-t
'JT"f'"
J T IJ y
T r
i i ph rlI F'[T ~h ST'M $I*-
I l'
drf[d_I-",b
'b.
YI
-\\ '
T T
- l fJ
.r 4"
I g.L,.
Myr 4 L-[-
'".!!'.# TTIM4 di P
frMyiTgJ.,iyl k' g,,Q.'
at,([p.fTr,()fMW,.,.4-tr..,.T a. ;ggi{ A
-t *gh i yrj}lt.l_,I afTI J pll-p-lf i
~
Myrig,
7 f
.,d T]gph 7f JT 5 T
,-r t
7 Tr* t).r H~ f a yr-
.q -
L t
i 'AI
' I t r*
- j,.I
.t
-iT r
rr IT'.$,M r 11 5 riT,'.alrry(h.d(i$,Myr",,.t]qr'I d
t
' t [,.d-T J.'rrP U
i.ip'g.TTg..prn A_r!T!l1Qr T
7 T
!r uf
- iT..ar.rc_ y %. pdrt;ag-lj.
c a
r i
r%_ e. o. 4r..rn_
m, a.4 Figure 9.
Effect ofHeating (a,c) on Drift Stability and Seismic Event Before (b) and During (d) Heating Thus, the consequences of a seismic event will depend very much on when the event occurs with l
respect to the thermal loading produced by the waste package. Upon cooling, the induced thermal L
stresses will disappear, and some additional collapse could occur. It was found, during the study of seismic effects mentioned above, that the second seismic event in each case caused little additional rockfall. However, the effect of time-dependent weakening of the rock mass was not considered. It seems probable that additional collapse would occur if this factor were added.
Figures 10(a)-10(d) show the results of numerical modeling to simulate various support options and assumed rock conditions.
E:gineeredCarriers at Lcca Mountain, Some impressions c:d Suggestions 28 The rock joints are assumed to have an initial (high) friction angle of 56 and a cohesion of 0.07 MPa. Other properties are those for RMQ5 rock (as defined by the M&O contractor). The reinforcement (grouted bolts (c)],or support [ concrete (d) ] is then installed, or the drift is left unsupported ((a), (b)] depending on the case considered. The rock is then heated to 100 *C. Joint slip and rock failure occur. Then, in order to simulate time-dependent degradation of the rock joints, the joint friction angle is reduced to 35. Except for case (b), the joints are all assumed to be continuous. In case (c), thejoints are non-persistent, consisting of alternate 1-m-long segments ofintact rock and joint, for which the friction angle is degraded to 35.
Results indicate that the extent of the damage zone depends primarily on the frictional properties of the joints. Non-persistent joints (case (b)) behave essentially as intact rock, so that the extent l
of the damage zone is significantly reduced compared to that produced with continuous joints I
(case (a)); see the discussion of the lithophysal rock zone, below. Grouted rock bolts (case (c))
reduce considerably both the slippage onjoints and the extent of the damaged region. Case (d) indicates that the elastic liner installed with a gap between the crown of the drift and the top of the liner to simulate a noncontinuous liner / rock contact does little to reduce the extent of damage compared to the case in which there is no support (case (a)), although the liner does, of course, prevent the rock fallout that would be very likely to occur in case (a).8 Lower Lithophysal Rock Zm. )
A brief analysis of the mechanical properties oflithophysal tuff (see Figure 11 and related discussion) suggests that the overall mechanical response to stresses (including thermal stresses) in these zones may be less influenced by joints and joint slip than is the case in the non-lithophysal zones. Thus, the rock mass strength in the lithophysal tuff may be somewhat higher, j
but the modulus of deformation will also be higher. Because the induced thermal stresses are directly related to this modulus, the ratio of stress: strength will change less. It seems, therefore, that from the mechanical stability perspective, drifts (e.g., for a multi-level repository) may be located in either or both lithophysal and non-lithophysal regions.
Both the Nuclear Waste Technical Review Board (NWTRB) and the NRC have criticized DOE for its failure to determine the insitu mechanical properties of the lower lithophysal rock, in which approximately 70% of the repository will be located. (Most of the rock properties have been deternJned for other, non-lithophysal units.)
l An analysis was conducted to assess the influence of the lithophysae (assumed to be spheres) on the strength of the rock mass. Since, as noted in the discussion of Figure 10 case (b), non-persistent joints tend to exhibit the same strength as the intact rock in which they are found, the analysis assumed that the rock around the lithhophysae had the same properties as those defined by RMQ5. As stresses are increased (in this case, due to heating) on the rock, the lithophysae behave essentially as interior (spherical) excavations, i.e., stress concentrations occur around the I
l
)
8 These analyses were made available, counesy of Dr. R. Hart ofItasca Consulting Group l
Inc. Dr. Hart is a member of the Drift Stability Panel, for which the analyses were conducted.
)
L
i 4
4 1
E:gineered Barriers at Yucca Mou:tiin. Some Ingpressions c:dS:*ggestions 29 i
_.~..*,. g_<"7,.
.y-
.-.%-._,_p-,-,
..,_._,c.=*
j 1
5-L p.-
p.-
p +7 2..-
,p 4
... +:..*~t...
-c., g"..
- s,.
."4.s*,,
-t# 3 7 *,.J <,
1
,e
< eg
- 1
>s u.., u..., ',, ".,
p.c q....s;,..a
- 4... -
j..c,p-e r. c' 1
5.
4 e
f,. 1 f.-k~Tf; h:..n,c.-LTj$.? a.1)1' t v4 -
2.m y.r-N... i. -1.'.< p. v..<Y Yl. d .?
."v' r c?
T, W INA W
?
.....t
.p
.e o
5.
4 r:. :gm.-p.ii A.,2, y..h,<;;Q...-.-p; -
q4 a -r 3 2..M,ll +c:i.%v.
.e c.
i To h.c:,vL;fi, 4 1 h
m.
wG-I
,, n.,p n
,wta1 er:-
.J,.yr,t;,w q.1 n,r.I.-r.z.m.-
p" r.. sw. g(I'}-
a jJ...* 'l.2 ~
r
+3. a e
4 3r v1 e, g.r#
r y
+ -r i G.. " ";.e
'~
hh.L hh~
- Nhkl;...M
- q JiJh!A C.
rN.Nf
~
g t.:tb t
psef;#r7 -
'lm r teg;.?jir% ~~7W!rfi.rlL.
r :
i...r,lJ.c.rilk.;":d.p~!7.J.. e ~ '.,
pr e-1.e...,,......, 1.. r t. 2.- t m...., r
'.. s, r........... -.......e 2
u,.4
-r w.r.. t..e.
..u,.
- , m..w n...ua :
r Ay-6w c
s,. L,...,V' :..;e*... r ".. w; :.>
- u, 1.c.-r
. -1., y..+-f i,;., "....-. - r i
- s -r 4
- 4. u J. s-r i.;.
,....r, 1
2
,,, ;..c
- T - r_.
.. (f,,.r....
.g._,,_ w.,J,a r a
h.T.;s J.J ? L; s 4_. *. ',f.1.
- 4 _T.r.
v 1,n _g _,g,.
t
(*)
(b)
,.r...
"*'f-<,-.
g
- #'L*'O..r. 9 %e 7 ~
h
[~j'., 27 3; 3 p =,,-
~"TU,.1 n,.'c s
eva:.
![T )r h ~,m.";n.sh.w;afli$.-
r,;l4 wI 1.u &.-
]
- ,-r"
- 4,.c~;.*-1'""rlr
~!
Ar ' h, i;~~ T..,.,u 4' - t e..L;.,.a. : ~
ras
.. :.:e' 'R r
. -e r-I'.
.w,e '
-r I
SJr~U 7.
~~~-; i
' * =
r %,L
,t u1 +,.Jr -
H.r
- i.t i~> f.a.
, J.,a,. r tJ.
j &l5,1&
.:r' ~
+;
i lw,,rU.r7n.,h,a.,J ie Q a - r. 7",,',.4 ie t
G**r*i\\ \\ f
,!.! w.9.Gn.r i*
,e...
4 q,-rf c-r 4
T r, y ),;o.;t,T ji G
. {l0...u.',JryY ~
g, 4.! e. +-,. r L, ?7':? < 7,.J:l n
i;
.{
i j.sj;.a l, 1
...Ndr*j,i, '
[yJL; 1Ne.+ r'i' f..y,.,},.l>, ~t..
t
-e J,o.? i s" ? > pI
&r'Ll -
- 2., + -r
/
by4..i r. d7"m:i C
4r'77!.'.ll l g,,r,a arr r*:t,:
1 !L"-Jh~jd,., ~.".lr~N,.m a..
,GI.u.e,+ r,
a<
' s d ",T -
,.,[i'.
T-
,1 l
u t...
jm "*, T " D:.. H !?e"Q,. "=,,....~ v-. a..,
~~~~~[,
r l
- s.t,
- , 6-~.
n :.,. :.L i
u 1,i,
a,...
,..- r.
s
. w -r.7" b.a,...;. a. IL, h.... 44*II],~
+.pr*p "L.a..:.a rr. J. ? r*t.,es,'Ti.. : a.s 1r c - r * ?,>
-e
. m.
t
-r aw Y O!!!]
t..,,
i v,, r< t,c,..,. _ a..r:1..,.: 7'.
- ..
- ,.:y G e r,,... -
c 1
c/t.:7'.s'; []T^I r.~.'<%:aas
- .J*rT.-
=-
u, i..se.y.
r
~.; :,.,. -
r
...i a
,.8 ~.....
..,.-Q.
~ ~ 2~m
~
n*,, ;, s,, -
m r
~-
~
~
~, ~. _
(c)
~'m (d) 1 l
I Figure 10.
Effect ofLong-term Degradation ofRock Joints Properties on Extension of Inelastic Failed Rock Zonefor (a) Unsupported, Regularly Jointed Rock; (b)
Unsupported, Non-persistent Jointing; (c) Reinforced by Jointed Rock Bolts; and (d) Supported by Elastic Concrete Support i
L i
i I
t l
i t
~
4 E7gineered D" triers at Yxccar Mountzin, Some impressions andSuggestions 30 000000 O ]o o o -O o o
c c
o o
/ o g f'
\\
NC I Un
\\c o o
\\
h\\b ok D
O o
y
,o chj n
C
\\.
'q o
-o.
s Q
Excavation d
)
/
D o
o k? o o
V
\\
o\\4 ]-
/ o(o\\%o
\\
C o
o o/oo RMO 5 rock Q
-o-o \\0 ko g$ C P"P***5
\\0
/
o e
jo e
ke o k e -~~ -o. 0 o/
kq o g \\o r
g g
g 8
oo o+
c c o J
o i
(a) idealized representation of excavation (b) Model analyzed in calculations.
in Lithophysae Tuff.
15.0 15.0 12.5 12A 10.0 j
/
', ' X, 10.0 -
Y I
j j
4
[
5 a
5.0 -
4 l
5.0 ]
1 D/B a 0 l
2 D/B = 0.15 j
l '
D/ =
l 5 DiB = 0.60 0.0 ~
0.0 0.0 0.5 14 1.5 2.0 2.5 32 0A 05 1.0 13 24 23 3.0 0
0 3
3 4
Q.
(c)
- Tributary area' yield envelopes..
(d) FLAC* yield envelopes.
Note. Tnbutary area strength in (c) is calculated from the expressions, b"1
~
~
ane where q,, is the unconfined compressive strength and (> is the friction angle [with these intact rock properties, the yield errielope is given by line 1 in (c) and (d)).
The FLAC3D analysis also yielded the following rasults for the influence of the lithophysse on the overall modulus of deformation (E) of the lithophysae rock compared to the modulus of the rock wrthout the lithopysse (E*).
D/B 0.0 i 0.3 l0.45 0.6 E/E' 1.0 l 0.96 j 0.89 0.78 ' ]
Note E' in the FLAC3D analysis was 7.76GPa.
i
\\
Figure 11. Predicted Rock Mass Strength (Mohr-Coulomb) Envelopes and Modulii of Deformationfor Lithophysal Tuff (Intact rock between lithophysae is assumed to have RGM S mechanicalproperties.)
1 l
4 E;gineered Barriers :t Yucca Mo ntiin. Some impressions cndSuggestions 31 t
lithophysae and, eventually, the rock around the spherical periphery will begin to " collapse" into the lithophysal cavity.
The model analyzed is shown in Figure 11(b). B is assumed to be the width of a cubical region containing one cavity, diameter D. Various ratios of B:D were considered. The reduction in strength of the cube of rock containing the cavity, compared to the strength of a cube without a cavity (B/D = 0) is shown in Figures 11(c) and11 (d).
Two approaches are taken. In the so-called tributary area method (frequently used for room and pillar design in mines), it is assumed simply that the strength is reduced in proportion to the reduction in cross-sectional area of the center section of the cube containing the spherical cavity.
In the second approach, a three-dimensional numerical analysis (FLAC3D) was carried out. The strength limit was assumed to be reached when inelastic deformation started at the wall of the sphere. Results are shown in Figures 11(c) and 11(d). Although the FLAC3D results indicate slightly higher strengths for a given cavity size, the difference between the two approaches is small (maximum about 18% for D/B = 0.6), and the tributary area approach is conservative (i.e., it underestimates the strength of the rock). Thus, it seems sufficient to use the tributary area method in calculations involving the rock-mass strength of the lithophysal zone.
The FLAC3D analysis also yielded results for the influence of the lithophysae on the overall modulus of deformation (E) of the lithophysal rock compared to the modulus of the rock without the lithophysae (E*). Results (tabulated in Figure 11) indicate that the reduction in E is also small, and follows a similar trend to that of the strength reduction.
It is recommended that laboratory tests be carried out on intact samples (taken between lithophysae) to establish the envelope corresponding to D/B = 0, and then to estimate an average value of D/B from exposures in drift walls. This information can then be used, with Figure 10(c),
to establish an envelope for the rock mass strength.
Actual lithophysal voids tend to be ellipsoidal rather than perfectly spherical. Although it is feasible to generate ellipsoidal cavities and analyze them numerically, the effect of such cavities will depend on their distribution in size and orientation with respect to each other and to the applied stress field. As a first approach, over-conservative but simple approximation, the voids could be assumed to be " replaced" by spheres of diameters equal to the major axis of the ellipsoid.
(A less conservative option would be to assume spheres of diameter equal to the mean of the major and minor axes of the ellipsoids.) The approximate expressions presented in Figure 11 could then be used.
Use of Concrete for Excavation Support l
l Concern has been expressed that the use of concrete, as is popular, in concrete and "shoterete" linings and in the cement grouting of rock bolts' would result in a high pH of water entering the drift. This could have numerous adverse consequences (for example, on the radionuclide retardation capability of materials that may be placed below the waste packages, or i
f i
i
' Note that resin grouts are not favored, as they are organic compounds.
l
.j
4 Engineered Birriers et Yuccs Mount:in, Some Ingressions and Suggestions 32 that exist below the repository, e.g., zeolites) in order to retard the movement of radionuclides e.g., neptunium.
Discussions with concrete technologists reveal that it is possible to avoid high-pH water (e.g., by carbonating the cement, using carbon dioxide). The carbonation reaction has been studied extensively (it occurs naturally in concrete due to the effects of carbon dioxide in the atmosphere),
and it appears possible to engineer a solution to avoid high-pH water. Also, the strength (and ductility) of concrete can be increased considerably compared to standard concretes traditionally used in construction. Although care should be taken to ensure that adverse effects are avoided, it is recommended that drift support designers not be prevented from taking advantage of the merits of shotcrete and grouted bolts, both of which could play a valuable role in drift support at Yucca Mountain.
Most of the designs showing precast concrete lining or steel sets in the (circular) drifts (admittedly idealized) indicate that the linings / sets are in intimate uniform contact with the drift wall. In reality, of course, there will be inegularities in the wall profile. Normally, these would be filled with cement grout to ensure that the lining is uniformly loaded. Sand backpacking can be substituted, but it is important that analysis of the lining support include consideration of the influence of such irregularities and fill methods on the bending stresses generated in the support during the thermal cycle.
The writer believes that a well-designed system of grouted rock bolts, mesh, and shotcrete will be sufficient to ensure stable openings during the preclosure period. Precast concrete linings or steel set supports, which would be very expensive, will not be needed.
Upper-Bound to Collapse Region A simple estimate of the maximum extent of collapse around an unsupported tunnel can be made as follows.
Consider a circular tunnel, of radius a, surrounded by a circular zone of damaged rock, radius V.
When rock is damaged, slip along joints and dilation occur, rock may collapse into the tunnel, etc.
(i.e., the damaged rock will occupy a greater volume than when it was intact and undisturbed; it is said to undergo " bulking"). Let us assume that the rock is damaged to a radius b (b > a). If we assume that the broken rock has a bulking factor (i.e., unit volume of unbroken rock occupies a volume (1 +k) in the broken state), we may determine the volume of unbroken rock in the annulus (b - a) that, upon breaking, will fill the excavation. Thus, we have n (b2 _,2)(1 + k) = n b 2 f
from which we obtain -
b, ( 14 )u a
k For a value of k = 10% (10% to 25% is considered to cover most mining collapse situations), we find b/a = 3.3. For k = 25%, b = 2.2.
L
4 Engineered B: triers ct Y:ccx Mount:in, Some impressions andSuggestions 33 Thus, the maximum possible extent of the damage zone around a repository tunnel will be of the order of three tunnel radii. Beyond this region, the rock will contain joints and fractures similar to those in the virgin rock mass. Hence, for calculation of post-thermal cycle water influx to the tunnel, such a model should suffice.
Heated drift experiments and niche tests are unlikely to resolve several important post-thermal cycle inflow issues. The effect of the thumal cycle on the mechanical properties of the rock mass, information that would have been very useful in drift stability analysis, appears to be a secondary consideration in these experiments compared to the hydrological issues. There has been no modeling of the effect of discrete jointing on rock mass behavior, for example.
(Appendix Il shows a preliminary study to illustrate what is possible.) Acoustic emission (micro-seismic) studies have only recently been added, and an opportunity to observe the rock-mass re-sponse from the onset ofloading has been missed. Some microseismic equipment has now been installed, and data are being collected. Collection of such data can be very valuable in establishing which joints are slipping, and this information can be used to calibrate numerical models that contain such discrete features. (Figure 12 illustrates the microseismic network set up for the mine-by experiment at the Underground Research Laboratory in Can-Ja, together with the locations of the microfracturing (detected by acoustic emission) induced by excavation. The network was installed before the mine-by excavation was started.)
1 I
i
r 1
I si i
Engineered B:rriers ct Yucca Mountain, Some impressions cnd suggestions 34 7
b Sit'Stripss g* T* w r
w ',
4 ~~.
's,\\y stun es.
SJH_ -
l ne m.co l
~~lL y -
.Q.o* '._.~-)1
.I'9 l
I
- .e O lh'/g' g
Test ~
yg brued
'O l
e c*
Pavenc \\
ai
',Yl n
f'\\
! [,.<,s 9
e
/l kM8 e '.
,.t_.
em t h ~._ _
7~~r*
l Wictmastric 3 %
Agnaten:meder ag3 l
Tigu:e h 1 ment d the JN lewd ehminE ric bcarten of the ),Giety tem umnd. the amcrcueirmi insdualcg system nod Fimm 40*% :.le locatan of the bce-irol tenske m:cty.
l-It,
,hhfA W,.4 z....T--. L,
.q.
-4
'Y lt,.,.'g *y[.
Iscene(40
}'~
M'"
V*
p%-
F.
e 0, a 14 MPa d -66 MPa j
.r.
W Erd View '.*
..o l
1 Vs;me 13: locuou of mk to# ante evtcrz reca6 d er Om e xtavatha d a 1-c:rloeg rmerul bs the test tua:el Figure 12. Mine-by Experiment, Underground Research Laboratory, Pinawa (Read and Martin,1996) l l
E:gi:eered B:rriers et &cc2 M.zntrin. Some impressions and S-ggestions 35 References Carter, Luther J., and Thomas H. Pigford 1998. "Getting YUCCA MOUNTAIN RIGHT," Bull.
Atomic Scientists, 56--61 (March-April).
Electric Power Research Institute (EPRI) 1996. Analysis and Confirmation ofthe Robust Performance ofthe Barrier System within the Yucca Mountain Site. Flow Division, Report TR107189 (3294-17), December.
Goodman, R. E. 1989. Introduction to Rock Mechanics,2"d ed. New York: Wiley.
Philip, J. R.1990. "The Watertightness to Unsaturated Seepage of Underground Cavities and Tunnels,"in GEOVAL-90 (Proceedings ofthe Symposium on Validation of Geosphere Flow and Transport Models, Stockholm, May 1990), pp. 301-308. Stockholm: SKI.
Philip, J. R., J. H. Knight, and R. T. Waechter. 1989. " Unsaturated Seepage and Subterranean Holes: Conspectus, and Exclusion Problem for Circular Cylindrical Cavities," Water Resources Research,25,16-28.
Potyondy, D., and P. Cundall. 1999. "Modeling of Notch Formation in the URL Mine-by I
Tunnel: Phase IV - Enhancements to the PFC Model of Rock," ICG, Report to Underground Research Laboratory. City: Atomic Energy of Canada Limited, March.
Read, R. S., and C. D. Martin. 1996. TechnicalSummary ofAECL's Mine-by Experiment, Phase 1: Excavation Response, AECL-11311, COG 95-171. City: Atomic Energy of Canada Limited, February.
)
i Roseboom, E. H.1983, " Disposal of High-Level Nuclear Waste Above the Water Table in Arid Regions," Circular 903 City: U.S. Geological Survey.
Shoesmith, D. W., and M. Kolar. 1998 " Waste Package Performance," in Alternative Approaches to Assessing the Performance and Suitability of Yucca Mountainfor Spent-Fuel Disposal, Ch. 5, pp. 5 5-67, Report TR108732, City: Electric Power Research Institute (EPRI), November.
Stellavato, N., and P. Montazer. 1996. bye County NVirRP Office AnnualReport 1995/96.
j U.S. National Academy of Sciences / National Research Council (NAS/NRC).1957. "The Disposal of Radioactive Waste on Land," Publication No. 519. Washington, D.C.
U.S. National Academy of Sciences / National Research Council (NAS/NRC).1995. Technical Basesfor a Yucca Mountain Standard, TYMS. Washington, D.C.
U.S. Nuclear Regulatory Commission, Disposal of High-Level Radioactive Wastes in a Proposed Geological Repository at Yucca Mountain, Nevada, Proposed Rule, Federal Register Notice, February 22,1999.
l 4 -.
I Engineered Barriers at Y:cc2 Mou:tain, Some Ingpressions and Sxggestlo:s 36
-- 1 l
U.S. Department of Energy (U.S. DOE). 1998. VA reference design (CWRMS/M&O Report Design Feature Evaluation #25, Repository Horizon Elevation, April 2,1999).
L l
i i
i i
r
- a 1
AppendixI Steady Vertical UnsaturatedInfdtration Through an Array ofHorizontalDrifts i
.n.
Fz'
. : Barriers at Yuces M:unsann, Sonne Impressions and Suggesnons.
I2
, A M --- - i n Validity of Richard's equation with equivalent unsaturated properties. Isothermal medium.
V-( KK,(p)V(y+ z)) = 0 y: relative pressure head (y < 0 unsaturated zone, y> 0 saturated zone) [m]
z : vertical coordinate [m]
K : hydraulic conductivity sensor [m/s]
K,(y): relative hydraulic conductivity (K, < 1 unsaturated zone, K, = 1 saturated zone) [-]
Parametric anodel for unsaturated conductivity
( ! + lawf )"' ( 1 - ( 1 - I + layf )~ )
m=11 van Genuchten : K,(y) ' =
Exponential:. K,(y) = e*
Assuaned snaterialproperties Matrix porosity : 0.1 d8m' Matrix permeability :410 Fracture fmquency : 4.51/m Fracture aperture: 54 pm Matrix hydraulic conductivity (isotropic) K.i : 410-" m/s Fracture hydraulic conductivity (cubic law) K : 5.85 10-7m/s Homogeneous saturated and residual moisture 4, 4 : 0.1,0.01 van Genuchten model parameters a, n : 41/m,2 Exponential model parameter a : 101/m 2D vertical equivalent hydraulic conductivity tensor (assuming vertical fractures)
K=
0 K.
Anisotropy ratio varied from 1 to K
/K.,
Geosmetry Drifts: diameter - 5 m ; spacing - 80 m (1n, i-nial),30 m (vertical)
Potential capture zone (per unit width) for a column of drifts : 80 m' Boundary candklams Unnaturating infiltration rate at the surface (i): 50 mm/y (1.3710" m/d,1.5910 m/s) 4 AtiiQ ic seepage at the drifts Static watertable level at - 200 m
'
- i.e.1096 x 10-8 m'/d over the 80 m potential capture area.
2 i-
F-a Engineered Barriers at Yucca Mountain Some Impressions and Suggestions 1-3 Sketch of the flow towards drifts la columns (with symetry conditions)
Infiltration hU hU U
80 m I
Surface 1i v
RowI o
30 m Row 2 o
s Row 3 o
Row 4 0
h 5
?
z E
1r Watertable level 8
Discharge rates under steady state unsaturated conditions [10-5m/d]
Case 0 : = I (isotropic)
Case 1 : $= 10-'
Case 2 : = 10-2 Top infiltration 1096.0 1096.0 1096.0 Drifts at -30 m 0.0 0.0
- 35.6 Drifts at-60 m 0.0 0.0 0.0 Drifts at - 90 m 0.0 0.0 0.0 Drifts at-120 m 0.0 0.0 0.0 Bottom drainage
- 1096.0
- 1096.0
- 1060.4 Case 3 : = 10-8 Case 4 : = 10"(=KdK.) Ca. s.1:
=0 a
Top infiltration 1096.0 1096.0 1096.0 Drifts at - 30 m
- 58.0
- 60.3
- 68.5 Drifts at -60 m 6,1 7.0 0.0
h
! 4 l:
Engineered Barriers at Yucca Mountain. Some imprwssions amf Suggestions 14 L
3.2' 2.4 0.0
. Drifts at-90 m Drifts at-120 m 2.9 2.0 0.0 Bottom drainage
- 1025.8
- 1024.3
- 1027.5 Remarks Under unsaturating vertical infiltration, buried cavities may behave as obstacles to the flow and so increase water relative pressure head at parts of the cavity surface. Gravity dripping into the cavity occurs only at those points where the pressure head reaches the pressure inside the cavity (e.g., atmospheric pressure). Under uniform infiltration the first point reaching this pressure is the highest point of the cavity roof.
An analytical solution exists for horizontal cylindrical cavities in an isotropic medium (Philip and Knight,1989, Water Resources Research, 25, 16-28). Assuming an infmite vertical medium submitted to constant. uniform unsaturated seepage, the question as to whether or not water drips into a circular section centered at the origin is
. caswered in a straightforward manner with the following simple rule if I<
E
- no dripping in the cavity, dripping otherwise d==(s)
In this (exact) formula / [m/s] is the specified uniform infiltration rate, K,, [m/s] is the saturated isotropic hydraulic conductivity and d_ is a normalized Kirchoff potential. Its value is maximum at the top of the circular section and can be approximated with excellent practical accuracy by 2s + 1, for small values of s d (s) =
s = E S.
2(s + 1), for large values Of s 2
where, s is a dimensionless quantity defined by the decay parameter (a) of the Exponential model for the relative conductivity, and by the cavity diameter (D). Small s indicate capillarity dominated seepage, tending to divert water around the cavity, whereas gravity is dominant for larger values. Moreover, the larger the cavity the more vulnerable.
it is to water entry.
i
< r O
/
/
Roof-drip lobe 6-
/
(stagnation pointt t
\\
u s
ir
/4 iso-6 around the cavity and scepage flow lines In the present situation (s = 105/2) no dripping occurs into the cavity since the above inequality is satisfied (1.5910' j
' < 5.5810* /52). The infi.'tration rate could actually be increased by, roughly, a factor 10 before droplets form at the
)
n Enynnered Barriers at Yucca Mountain, Some impressio s and Suggestions 15 top of the cavity. Alternatively, the isotropic saturated hydraulic conductivity could be reduced, or the cavity diamepincreased by the same factor, to produce dripping into the cavity.
These theoretical consulerations explain why the drifts remain dry in Case 0 and to a certain extend in cases with mild anisotropy (i.e., Case 1). As anisotropy becomes larger, horizontal capillary flow becomes less significant and
- water cannot be diverted around the cavity surface with the same magnitude any more.
As a result, saturation increases and dripping starts in the first drift (Case 2), while the drifts below remain dry because the roof-drip lobes coming from above are too diffuse (capillarity is still active) to generate saturation conditions there.
- At larger anisotropy ratios (Case 3 and Case 4) the lower drifts become g/adually active, but in a manner that is not l
straightforward to understand. There are obviously highly non-linear effects (the decay coefficient a is rather large) combined to the anisotropy ratio. Numerical effects due, for instance, to mesh orientation and refinement around the drifts may also be present. However, several grid size were enforced (the finer with node spacing of the order of 0.2 m around the drifts) yielding the same type of results. More investigations (including analytical ones) are needed to understand the flow processes (e.g. use of finer meshes and various solution schemes, columns with more drifts, etc), particularly at high anisotropy ratios.
With zero horizontal conductivity (Case 5) the first drift theoretically captures the quantity of water given by iD (i.e.,68.5 107 m'/d in the present case) and by-passes the drifts vertically below.
\\
]
)
p
!o l.
I
{
l Appendix H Numerical Simulation of the Effects of Heating on the Permeability of a Jointed Rock Mass i
l B. Damjanac, C. Fairhurst and T. Brandshaug Itasca Consulting Group, Inc.
J 708 South Third Street Suite 310 Minneapolis, Minnesota 55415 th To be presented at the 9 ggg gg j g l
9
r f#Numarical simulation of tha eff cts of haating on the pormeability of a jointed
~
rock mass Simulation numstique det effets d'une augmentation de tempsrature sur la permsabilits d'une masse rocheuse fissurse Numerische Simulation der Hitzeeinwirkung auf gekl0ftetes Gebirge B.DAMJANAC, C.FAIRHURST & T.BRANDSHAUG, Itasca Consulting Group, Inc., Minneapolis, Minnesota, USA ABSTRACT; One of the objectives of the Drift Scale Test (DST), currently underway at Yucca Mountain. US A,is to assess the effect oflarge-scale heating on the permeability of the rock mass. The DST is simulated using continuum and discontinuum models to predict the change in permeability in the rock mass surrounding the heated drift. The simulations show that heating will cause both reduction in permeability (in regions of increasing mean stress) and increase in permeability (in regions of non-linear shear deformation-s Although the elasto-plastic (ubiquitous joint) continuum model and the distinct element model (DEM) indicate similar regions of joint slip in the rock mass, the resulting change in permeability can be calculated much more easily from the DEM.
R8 SUM 8: Un des objectifs de l'essai DST (Drift Scale Test). en cours au site de Yucca Mountain, Etats Unis, est I'dvaluation de l' effet d'une variation thermique sur la permdabilitd de la masse rocheuse, h I'dchelle de la galerie. L'essai DST est simuld numdriquement l' aide de modhles continu et discontinu af n de prddire le changement de permdsbilitd de la masse rocheuse entourant la galerie lo est soumise h une augmentation de temp 6rature. Les simulations numdriques montrent que l'ichauffement cause a la fois une r6 duction (dans les rdgions d' augmentation de la contrainte moyenne) et une augmentation de permdabilitd (dans les rdgions de ddformation non-lindaire en scisaillement-glissement). Bien que les modbles continu dlastoplastique (ubiquitousjoint) d'une part et d'dlements distincts (DEM) d' autre part prddisent des zones similaires de glissement dejoint dans la masse rocheuse,la mdthode DEM se pr6te plus aisemen au calcul des changements de permdabilitd.
ZUS AMMENFASSUNG: Eine der Aufgabenstellungen des " Drift Scale Tests - DS I", dergegenwnrtig im Yucca Mountain Projekt in den US A circhgefuhrt wird, ist es, den Effekt von groBruumiger Erhitzung aufdie Permeabilitut des Gebirges zu untersuchen. Der DSTwurde durch Kontinuums-und Diskontinuumsmodelle simuliert, um die hnderungen der Permeabilitht im Gebirge um den erhitzten Teil zu prognostizieren. Die Simulationen zeigen, dab die Erhitzung sowohl eine Reduzierung der Permeabilitut (in Regionen erh6hter mittlerer Spannungen) als auch eine Erh6 hung der Permeabilitut (in Regionen nicht-linearer Scherdeformationen " slip") bewirkt. Obwohl das clasto-plastische (verschmierte K!Ofte) kontinuumsmechanische Modell und das Distinkt Element Modell (DEM) 5hnliche Bereiche von Scherbewegungen auf K10ften ausweisen, kann die resultierende knderung der Permeabilitlit Ober die DEM wesentlich einfacher bestimmt werden.
1 INTRODUCTION meability can be represented much more directly in models (such as the distinct element method), that simulate explicitly the effect A main objective of the ongoing Drift Scale Test (DST) at Yucca ofjoints on deformation and fluid transport.
Mountain, Nevada, US A, is to assess the effect of large-scale heat-Given the actual geometry of the excavations and joints, rig-ing (intended to simu.' ate the heating produced by stored high level orous interpretation of the effect of heating on joint aperture and nuclear waste) on the permeability of the rock mass. The DSTis permeability changes and flow in the drift experiment requires a conducted in fractured. densely welded, ash-flow tuff at the pro-three dimensional model. 3DEC (Itasca Consulting Group, Inc.
posed repository horizon in Yucca Mountain. The permeability of 1998a) was used to consider this influence. However, since a cou-
)
this rock mass is controlled primarily by natural fractures in the pied thermo-mechanical-hydrological analysis of a fractured rock reck: the matrix permeability is very small-mass is computationally intensive, the main part of the analysis in
{
This paper discusses the results of numerical analyses carried this study has been carried out using the two-dimensional Univer-out to examine the effect of heating around the DST on the change sal Distinct Element Code. UDEC (Itasca Consulting Group,Inc.
of permeability in the surreunding rock. Continuum models of a 1996). The continuum code FIAC (Itasca Consulting Group,Inc.
fractured medium (e.g. the ubiquitousjoint model) provide reason-1998b) was also used to estimate the regions of non-linear defor-able approximation of the rock mass when: (1) thejoint spacing is mation (i.e. the regions where the rock permeability changes)in-small relative to the characteristic dimensions of the problem, and duced in the rock mass by heating. Comparison of results obtained (2)thejoint properties are uniform (i.e. there are nojoints in the set using different models and codes (continuum; discontinuum, two-that have an aperture and transmissivity substantially greater than dimensional; three-dimensional) has proven to be very valuable that of other joints). Determination of the constitutive relations verifying the assumptions used in development of the analyses and needed to allow accurate prediction of the change in permeabil-may guide the use of particular models in further analysis, ity of such a rock mass when deformed is especially difficult with continuum models. The relationship between deformation and pet.
i i
e at J q ll;lll?lll[ lllljl[?l1lllllll ll. i lllllll!. sl! ! ' 'lllllli i slt ' ' ' ' ' W
' lllllllll Illlllll3 Tabic 1. Properties of the rock e
m Ji:'llizls
- n.
- $;2_ $lr lll!l lllll!; jfll Denssty p 2540 k g/m' ll, llllllllll111-lllllll Young's modulus. E 32.4 GPa l l l l l l l l !'l,llL' l i l l l l l l l ll,l l l l l ll
=
Poiuor's ratio. V 0.17 llllllQ< ;l$42ill/llllgl gllll Thermal conductivity. kr 1.67 W/m*K ll lllli ! ! ' llllliitt Specihe heat. cr 928 ilkg* K J' ' l l l l l l l l ; ' l l l l l ll l l l I li ' ll l l lllj iri Coefncient of thermal expansion. a' ID'S s.
wriiriir;'i,,iririr!!!
iiiu Fllllllll[illlllllll! ;;llll[
Zl'l'/lll,
'rllf
" Fast paths", joints or fracture zones with much higher initial O flllll llllllllllll
]l l i !!! ',' ! l ll lll l lll !. ; l, ll ll[u Permeability (i.e. initial hydraulic aperture) than the otherjoints.
> lllll i '!Q are known to occur at Yucca Mountain. It is also expected that 2
m
~~"""
a-fast paths will be more compliant and weaker than the otherjoints.
Figure I. Geometry of the two-dimensional discontinuum model Two cases were considered in the discontinuum models: (1) all joints have the same properties, and (2)" fast-paths" are assumed 2 DESCRIFTION OF THE MODELS to exist at sevemt different locations relative to the heated drift. The properties of " typical" rock joints (Olsson & Brown 1997) used in The heated drift is a 5 m x 5 m excavation of" horse-shoe" cross-the analpis are shown in Table 2.
section (see Fig.1). The observation drift is rectangular. 5 m x 4 m in cross-section. A three dimensional model of the DST was Table 2. Propenies of the rock joints generated using 3DEC. Figure 2 shows the lower half of this model Normal stiffness kn 200 MPa/mm j
(i.e. from the drift horizon downward). Threejoint sets are repre.
Shear stiffness. k, 150 MPalmm i
sented. Joint set I has a dip of 77* and dip direction of 40*; set 2 Cohesion. c 0.23 MPa has a dip of 80* and dip direction of 130*: set 3 has a dip of 25 Fnction angle, d 42' and dip direction of 300* (Wagner 1996a). The joint spacing in I
cach set is 10 m. The vertical cross section. perpendicular to the For this analysis, the mechanical properties of the fast paths axes of the drifts (from the 3DEC model), coincides with the plane (shown in Tab. 3) are simulated by reducing the properties of"typ-of the two-dimensional models used for simulation of the DST.
ical" joints-as can be seen by a comparison of Tables 3 and 2.
Figure ! shows the joint sets I and 3 in the two-dimensional UDEC model. Thejoints in the two-dimensional model are spaced Table 3 Properties of fast paths 2 m apart-i.e. much closer than the 10-m spacing in the three-dimensional model. (The coarser spacing in the 3DEC model is Normal stiffness. kn 50 MPa/mm dictated by the heavy computational demands of three-dimensional Shear stiffness, ks 50 MPa/mm analysis.)
Cohesion, c 0.05 MPa The rock was considered to be linearly clastic and isotropic.
Friction angle. $
25' and to have the properties (Birkholzer & Tsang 1996) shown in Table 1. The response of the joints to deformation normal to the The initial state of stress in the rock mass was assumed to be joint plane is assumed to be linearly clastic for compressive stresses ch = -5 MPa. av = -10 MPa at the drift level. De initial (Joints can not sustain tension.); the response to shear deformation stresses vary as a function of elevation due to gravity, with a con-is assumed to be linearly clastic-perfectly plastic according to the stant ratio maintained between the horizontal and vertical normal Mohr-Coulomb slip condition. Slip of the joints is associated with stresses. The initial temperature in the rock mass was taken to be constant. at 25'C throughout the model.
~ ~~~
Thermal analysis of conductive heat transport was carried out for 4 years. An 800-W/m heat source, provided by heaters located in the square block at the f'.oor of the heMed drift, was simulated as v......
a heat flux uniformly distributed along the boundary of the heated drift. The wing heaters are located symmetrically relative to the 2
axis of the heated drift: a planar source of 125 W/m is distributed between 4 m and 9 m from the drift axis, and a planar source of 2
175 W/m is distributed between 9 m and 14 m distance from the drift axis (Wagner 1996b).
17:,:: i: -
- - 1Otu
. =.~?.I Y.
3 JOINT DILATANCY Joint (normal and shear) stiffness and strength (cohesion and fric-Figure 2. View of three-dimensional model(blocks above the drifts are tion angle) are properties that affect the dependency of the per.
hidden) meability (of thejoints and rock mass) on the imposed mechanical loading. However. thejoint dilation angle p has the most profound
1 d
'f effect on the dependence cf the permeability to shear deformation
~ ~ ~ ~ ~""~ ~ ~ ~
cf a rock joint.
?
Thejoint dilation angle, the measure ofjoint opening as a result ofjoint slip, is a function of:
f@$[h N
1 shear deformation (Dilation is usually large during the initial slip deformation, decreasing with slip accumulation.); and d
9 u..
- 2. stress normal to thejoint plane (con 6nement). (Dilation is a
%4 vigg f'
consequence of joint roughness. The relative movement of
%, hjC
$p
' 7p g (
Jg rock blocks cannot be strictly parallel to the pla,:e of the joint yy gj between them, sincejoint roughness enforces some displace-
~
,f. g n
~--
ment normal to thejoint plane. At very hi h normal stresses, F
the joint asperities can be sheared-off, resulting in a reduced or zero dilation angle.)
Olsson & Brown (1997) reportedjoint dilation angles measured Figure 3. PFC model of a shear box test on samples taken from the TS w2 geological unit at Yucca Mountain rrc2o no
""~ ~ ~
for different confmements. (TSw2 is the repository unit.) The
~'
measured dilation angles show large dispersion, varying between
!. I l' and 33.4*. As a result, the relationship between confinement U
and dilation angle is unclear. Therefore, the first-order analyses E *=
were conducted using an upper value, y; = 30*, and an average i;.,l=,,,,,,,,,,,
value, p = 14*, for the dilation angle. It was further assumed in
~
these analyses that the dilation angle was constant, independent of the shear deformation or normal stress. The dilation angle for the fast paths was assumed to be equal to the dilation angle of" typical" joints.
~
3.1 Numerical Experiment in order to establish a clearer understanding of the dependence g,
i.'
3, s.
4.
4.
of dilation to shear deformation and confinement for the range of values expected to occur in the mooel, numerical experiments Figure 4. Venical (m) versus shear displacements (m) were conducted to simulate shearing of a roughjoint using a shear box--in a manner similar to that described by Cundall (1999).
tests were conducted for normal stresses of 2.5,5,10,15,20, and The results from the numerical experiments (i.e. the relationship 25 MPa. Figure 3 shows the specimen after a significant amount of between peak dilation angle, joint shear displacement and normal shear for a normal stress of 5 MPa. The short black lines indicate r
stress; for TSw2 rock and joint conditions were then used in the. locations of tensile cracks in the specimen, while the dark gray lines UDEC simulation of the DST.
indicate shear cracks. Note that a significant amount of dan age The micro-mechanical model of the shear box experiment us.
can be attributed to tensile cracking (i.e. particle contacts failing ing the Particle Flow Code-PFC2D (Itasca Consulting Group, in tension). The test in Figure 3 predicted a peak shear strength of Inc.1999), is shown in Figure 3. The bonded assembly of par-6 MPa after 0.2 mm of shear displacement. Figure 4 shows normal ticles (Particles are bonded at contact points.) con be envisioned displacement (m) (i.e. dilation) versus shear displacement (m) for as a synthetic rock. By adjusting the contact stiffness (shear and the test in Figure 3. (Figure 4 suggests a peak dilation angle of normal) and strength (shear and tensile), this " rock" was made me-28'.)
chanically similar to the TSw2 rock. The length of the specimen The results from the numerical experiments were simplified as in Figure 3 is 0.10 m. and the height is 0.04 m. The joint trace is a bi-linear relationship between dilation and joint shear displace.
indicated by the continuous black lines transecting the specimen ment. (This relationship is defined by a constant dilation angle and from left to right. The particles at or adjacent to this line are left a shear displacement at which dilation becomes zerp.) The depen-unbonded. The black particles along the boundary of the specimen dence of the dilation angle and the zero dilation shear displacement are designated as the shear box. The shear box particles below the on the confmement, as obtained from the numerical experiments joint trace are fixed, while those above the trace are assigned a con-(shown in Table 4), was implemented in the UDEC model of DST stant horizontal velocity. The joint trace was produced using the to provide a better approximation of the dilation behavior of the following decreasin,g power law power spectrum (Brown 1995):
joints.
G(k) = Ck" (1)
Table 4. Approximate relationship between joint dilation, zero dilation where C is a constant; k = 2x/A;A is the wavelength; cr =
5 + 2D; and D is the fractal dimension of joint surface. Joint Normal stress (MPa) 2.5 5.0 2.5 10.0 15.0 25.0 topography data provided by Olsson & Brown (1997) for specimen Dilanon angle (')
42 28 16 15.5 13.0 12.0 Disp acement(mm) 1.0 1.5 2.5 2.5 2.5 2.5 YM30 taken from the repository unit TSw2 were used. Numerical l
i O
4 MODELING RESULTS shear slip. Both the extent cf the region where joints are opening and the value of the maximum opening increase as a funesion of f
4.1 Temperaturefield3 the duration of heating. De maximum opening is more than twice it was assumed in all simulations (FIAC. UDEC and 3DEC) that as large as the maximum closure. The effect of dilation angle is conduction is the only mode of heat transfer in the rock mass. In significant. The maximum opening in the model with a 30* con-fact, boiling of pore water is likely to occur in the rock around the stant dilation angle (1.5 mm)is two to three tirnes larger than in the heated drift because of the high temperatures. This effect has been model with a 14* dilation angle (0.6 mm). De maximum opening analyzed in models of heat and fluid transport by Buscheck (1998).
in Figure 7. which shows results for variable dilauon angle calcu-The temperature distributions due to heat conduction are al-lated from the PFC model, is 1.2 mm. The regions of slip along most identical for the continuum and discontinuum models. De joint set 1. as calculated in UDEC.ag ee remarkably well with the contours (*C) after 4 years of heating are shown in Figure 5.
regions of plastic deformation indicated by the FLAC ubiquitous j
joint model.
The actual position of possible fast paths relative to the heated 4.2 Deformation in the two-dimensionalcontinuum models drift is unknown. However. the effect of the fast path was assessed The ubiquitous joint model is a continuum, clasto-plastic model by performing a series of sunulations for three different assumed in which an anisotropic strength of the rock mass i taken into locations of the fast paths:
h account-i.e. there are predehned planes of weakness. The Case 1. The fast path passes through the heated drift.
strength in the planes of weakness was assumed to be equal t the joint strength as given in Table 2. The markers shown n Fig-Case 2. The fast path is offset opproximately 15 m to the left of the axis of the heated drift.
ure 6 indicate slipping along the planes of weakness corresponding to sub verticaljoint set from Figure 1.
Case 3. The fast path is offset approximately 15 m to the right of The ubiquitous joint model predicts the deformation and the the axis of the heated drift.
region ofjoint slip in the rock mass. To assess the change in per-The analysis shows that the effect of the fast path in case ! is meability produced by this deformation and slip. it is necessary to insignificant. The effects of the fast paths in cases 2 and 3 are establish a constitutive relation between deformation (volumetric dramatic. De joint opening and closure for case 3. after four years I
and shear) and the change in permeability, in the case of the dis-of heating, is shown in Figure 8. The maximum joint opening tinct element method, the joint deformation is calculated, and it is caused by slip in cases 2 and 3 is about 6 mm. compared to 1.5 mm usually assumed that the change in the joint hydraulic aperture is in the model with uniform joint properties.
equal to joint normal displacement (i.e. closing and opening).
4.4 Deformation in the three dunensionaldiscontinuum model 4J Deformation in the two-dunensionaldiscontinuum models The results of the three-dimensional model show that the two.
The discontinuous model of the rock mass in which thejoint prop.
dimensional model is an acceptable approximation of the defor.
erties are taken to be uniform shows a complex response to the mation in the midd!c of the heated drift. However, deformation perturbation induced by heating (Fig. 7). In general. it is possible of joint set 2. which is neglected in the two-dimensional model, to identify two regions exhibiting significantly different responses.
becomes important in the region close to the drift ends, where the in the immediate vicinity of the drift. the joints tend to close as temperature field is also three-dimensional.
a consequence of an increase in the compressive stress normal to the joint planes. Both the maximum closure and the region over 5 CONCLUSIONS which thejoints close increase with the duration of heating Joints from both sets (sub-vertical and sub-horizontal) tend to close. but Comparison of the results of different computational models used the sub-vertical joints close more. Above and below the region of to predict the thermo-mechanical response of a jointed rock mass joint closure, the sub-verticaljoint set dilates (opens) as a result of
-. m w
- ~. -
KJG W DU 4 sea. W 8maan.
- b'N I
res l
conm nas
"'S
!.M ""
~. l" C '".n".,
~
" ',"J.*J".
e a %_
As
+
- g* ",,,,g,,,. -..
p.~-
~ g p
4 1
g 3';
f khC
$3h5
- i.. m
,5
- r:
Ih. J ij 0 880
}j s-.~, s ~
~
~*
Figure 6. Indicators of slip in the ubiquitousjoint model after 4 years of Fig:re 5. Temperature contours (*C) after 4 years of heating heating
l D
i j
q e
in the vicinity of the DST, indicate the followmg.
.soo mm ost ion,.nr. a.uw, s n v a
UoEC (WRSION s Oe) y
- 1. Continuum clasto-plastic ubiquitous joint models give a uaano good prediction of the regions in the rock mass over which
,n, thejoints slip. However, to calculate permeability change as y ga a result of calculated deformation, a constitutive model that
- -=m '='. = 12m +a =
,=
relates both volumetric and shear (clastic and plastic) strains 6=.arr pu to permeability change is required.
q =.aae h'I C
- 2. Discontinuum models are the most effective way to simulate
""j,. "'.*, 7,"
the effects of heating (or any mechanical deformation) on f
change in permeability of a jointed rock mass. Constitutive
$ ",,.%"7
{
relations are also required, but they are more straightforward than in the case of the equivalent continuum. Joint dilation angle and its dependence on accumulated slip and normal stress are important parameters that defme the change in per-Figure 8. Fast path, case 3 dilation angle 30* - opening and closure (m) meability produced by joint slip.
ofjoints after 4 years of heaung j
- 3. The two-dimensional model is an acceptable approximation of the deformation in the middle of the drift, even for the case dilation angle of 30' is assumed, the opening of the fast path is of
{
j in which orientation of thejoints relative to the drifts' axes is the order of six millimeters.
slightly oblique.
- 4. Three-dimensional effects (particularly the deformation of ACKNOWLEDGEMENT f
the joint set n(glected in the two dimensional rnodel) become important close to the end of the drift.
The advice and interest of Professor P.A.Witherspoon, who sug-gested these studies,is gratefully acknowledged.
The various analyses described above have been used to illus-trate the effects of large-scale heating on the hydrological condi-
. tions in the rock mass around the drifts in the DST. Increase in REFERENCES l
temperature produces different effects on the deformation of the rock joints 0.e. both closure and separation) m. different regions Birkholzer, J. T. & Tsang. Y. W.1996. Forecast of thermal-hydrologica!
of the rock mass. In general, shear stresses cause slip on the sub-conditions and air injection test results of the single heater test. Tech-vertical joints away from the drift, while increase in confmement nical report, Earth Sciences Division. Lawrence Berkeley National Laboratory.
causes closure of the joints (The sub-vertical joints close more.)
Brown S. R.1995. Simple mathematical model of a rough fracture. J.
in the vicinity of the heated drift. Both regions of opening and Geophy Res. /00,5941-5952.
closure, and the maximum values of opening and closure in these Buscheck. T. A.1998. Private communications.
regions are functions of several parameters. including: (1) inten-sity of thermal loading, and (2) properties of the rock mass and Cundall, P. A.1999. Numencal experiments on roughjoints in shear using rock joints (e g. stiffness, strength, dilation angle, orientation and a bonded particie model. In l aw Notesfor Earth Sciences. Springer.
To be published.
sp ^g ofjoints). The effect of the deformation on the permeabil-Itasca Consulung Group.17c.19%. UDEC (Universal Dirriner Element ity of the rock mass is even stronger in the case when a fast path c,g,) Versinn 3.0. Mmicapolis.
crosses the regions oflarge shear stresses induced by heating. The Itasca Consulting Group, IM.1998a. JDEC (J. dimensional Distinct Ele-shear deformation and slip localize along the fast path. If a con-rnent Code) Versi<m 2.0. Minneapolis.
i stant (independent of the magnitude of slip and the confmement)
Itasca Consulting Group, Inc.199Bb. Fl.AC (Fust higrangian Analysis of Continua) Version 34 Minneapolis.
. sos trrun ost we,= >.m. san,.w.
- a.mmp=, m. am.a gg3,.ca Consulting Group, Inc.1999. PFC2D Panicle Flow Code in 2 l' 'C Dimensiens Versinn 2 0. Minneapolis.
Olsson, W. A. & Brown, S. R.1997. Mechanical properties of fractures eve. uno from drillholes UE25-NRG-4, USW NRG-6, USW-NRG-7, USW-m.
um.m -
SD-9 at Yucca Mountain. Nevada. Technical Report SAND 95-1736.
,,,.,,,,,. 3.
Geomechanics Department Sandia National Laboratories.
Wagner, R. A.19%s. Characterization of the ESF thermal test area. Tech-
.,ffp.rmfffv ^ a nical Repon B00000000-01717 5705 00047 REV01, TRW Environ-a.o. e oxem anuc.d
. c.~n. r owem me r,.o o.:oooem mental Safety Systems, Inc.
r*.=.w.
Wagner, R. A.1996b. Test design, plans and layout. Report for the
.. i ma am pion.e a
. o,a
. i.am.ca ESF thermal test. Technical Report BAB000000-Ol717 4600-00025 REV01 prepared for DOE, TRW Environmental Safety Systems, Inc.
Figure 7. Uniform joint propenies, variable dilation angle - opening and closure (m) of joints af ter 4 years of heating
7_ -
- [CMAIL, ACNW 8/99 Letter Report for Full-Text on NUDOCS Page 1l From:
Ethel Bamard (CMAIL) i To:
DOCDESK Data:
Wed, Sep 29,1999 5:50 PM
Subject:
ACNW 8/99 Letter Report for Full-Text on NUDOCS Place:
DOCDESK
- Attached is an ACNW letter report for August,1999...
- Thanks, Ethel i
i 016ezh-
!