Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
5 SUGGESTIONS FOR ADDITIONAL RESEARCH At several places in this review, topics are noted on which additional research is recommended. The suggested research, in the panel's opinion, will be required to develop a specific system design, but is not critical for establishing system feasibility. The technical basis for the KBS proposal seems already adequate to support the conclusion that a repository for disposal of spent nuclear fuel can indeed be constructed in Sweden with confidence that human health and safety will not be jeopardized for a million years and longer. Such a prediction about very long future times, however, must always be attended with at least minor uncertainties about some aspects, and additional research could be fruitful in narrowing the uncertainties. Suggested research topics are assembled below in these areas: sensitivity analysis, groundwater movement, properties of bentonite, thermochemical data, geologic investigations, and probabilistic risk assessment. The topics are arranged in roughly what the panel considers their order of importance. Several of the listed topics are part of the continuing research program in which the KBS scientists and engineers are already engaged. Sensitivity Analysis. The KBS has chosen for the most part to use pessimistic assumptions in their safety assessment. However, final repository design will require better evaluation of the effects of various assumptions on doseâespecially when these assumptions must be bal- anced against cost. A formal sensitivity analysis has been performed only for the BIOPATH model. The KBS should develop their calculation plan (see Figure 4-1) into a systematized, formal sequence, using "best-estimate" assumptions as a base calculation much as was done for 61
62 BIOPATH. This entice safety analysis could then be sub- jected to a sensitivity analysis to gauge the relative importance of various sources of uncertainty on the calculated results. Predictions of Groundwater Flow in Fractured Crystal- line Rock. In KBS-3, hydraulic conductivities are inferred from packer tests, on the assumption that the concept of a homogeneous isotropic medium is a valid basis for interpreting the test results. The porous- medium concept is then also used in the finite-element computations of regional groundwater flow. The adequacy of this concept for describing flow in fractured rock is questionable, and the development of more accurate ways of treating flow in such rock is an important subject for research. Examples of questions to be addressed are as follows: 1. To what extent is the porous-medium concept valid? Essential in attacking this question is the use of inter- ference tests for measuring hydraulic conductivity. A possible approach is the experimental design for inter- ference tests described in TR 80-05, using packers rather than the usual pumping tests. Huyakorn et al. (1983) suggest that interference tests may be used to discrimi- nate among four conceptual models for flow in fractured rock. For assessing the applicability of the porous- medium concept, Witherspoon et al. (1983) also suggest the use of fracture geometry statistics. 2. How important is anisotropy? That knowledge of vertical heterogeneity and anisotropy is important for estimating travel times in the fractured conditions examined in KBS-3 has been demonstrated using simulation models (TR 82-01). Interference tests may be useful here also in determining whether anisotropy must be considered in predicting flows through a repository (TR 80-05; Kohut et al. 1983). 3. Are packer tests reliable for determining hydraulic conductivity? How much is their reliability influenced by anisotropy and local heterogeneity? If they are known to be reliable locally, do they give values representative of the region? Calculations in TR 82-06 suggest that heterogeneity and anisotropy should have only a small effect, but that a "borehole skin" effect might be impor- tant. Leif Carlsson (Swedish Geological Survey, personal communication, 1983) indicated that preliminary analysis of the transient pressure response observed in the packer
63 tests showed no evidence of a skin affect, but the poten- tial error is large enough to warrant further investiga- tion. To assess the reliability of packer tests for estimating regional values of hydraulic conductivity, relationships should be explored among values derived from packer tests, from interference tests, and from observations of fracture frequency. Questions about the adequacy of the porous-flow model are important also when the model is extended to con- taminant transport. In the transport calculations for KBS-3, a dual-porosity model was used rather than the standard porous-medium model (TR 83-48), and Neretnieks (TR 82-03) has suggested as an alternative a "stratified- flow" model. These and two other models were tested against available data in TR 83-38, but none of the models could be rejected. This result reinforces Neretnieks' opinion that there is "no experimental evidence at present to assess channeling or dispersion mechanisms over large distances in fissured bedrock" (TR 82-03). Research on the spatial scale over which chan- neling predominates is needed to resolve the question as to what conceptual model is appropriate for describing contaminant transport. Properties of Bentonite. Despite the very extensive investigations by Pusch and his colleagues, some ques- tions remain about the suitability of bentonite for its many intended functions in the KBS plan: 1. Will the properties of bentonite be adversely affected by long-continued heating? Laboratory results to date indicate that heating to temperatures that do not exceed 80°C, which is the maximum planned in a KBS repository, will have no appreciable effect. The experi- ments are necessarily of short duration (months or a few years), however, and slow reactions over very long times still seem possible. Hoffman and Hower (1979), for example, cite geologic evidence for a slow change of smectite to illite over periods up to 50 million years. The change occurred at pressures somewhat greater than those expected in a repository, and whether reduced pressure would accelerate or decelerate the reaction is uncertain. Long-term tests under repository conditions seem desirable to test this reaction and also possible slow reactions of wet bentonite with copper or with minerals of the bedrock. Such tests are in progress at ,
64 Stripa, with electric heaters substituted for fuel rods (Roland Pusch, University of Lulea, personal communica- tion, 1983), and their results a few years hence will be of much interest. 2. Are bentonite barriers in tunnels and shafts adequate for sealing against groundwater flow? Tunnels and shafts may possibly become channels for groundwater flow after a repository is closed, either because the backfill is less impermeable than expected or because rock adjacent to the openings has been weakened and cracked by blasting. To block such movement, seals of compacted bentonite at intervals in tunnels and shafts are planned, each seal to be inserted into slots sawed into the weakened rock on all sides. Doubts have been expressed that such seals would function as expected, on the grounds that sawing the slots, no matter how gently it is done, might weaken the adjacent rock by leaving edges and corners under stress, so that permeable fractured rock would still exist around the seals and the bentonite might not expand to fill all the cracks and open spaces. Some in-situ tests of the proposed seals have been carried out at Stripa, and more are planned (Roland Pusch, University of Lulea, personal communica- tion, 1983) , and the panel strongly endorses this procedure. 3. Will bentonite protect a canister from rupture by sudden rock displacement? The compacted and water- saturated bentonite around a canister will probably serve as an effective cushion for slow deformation, but sudden displacive rock movement in an earthquake might cause both copper and bentonite to act like brittle solids and fracture. How large a sudden displacement could a canister endure without rupture, and how much protection would the bentonite offer? Tests to answer these questions will be difficult to set up, but the effort should be made. 4. What is the mechanism of transport of water and dissolved materials through bentonite? Much experimental work has already been done on the movement of fluids through bentonite, but details of the diffusion process are still obscure. Rates of migration of different ions should be studied under a variety of conditions, and the possible influence of the Soret effect and surface migra- tion should be considered. The research could be extended to the similar process of diffusion of water and radio- nuclides from fractures into the matrix of relatively unfractured granite. Such research will probably not
65 greatly change predictions based on the empirical results already in hand, but a sounder theoretical basis would help to reduce uncertainties. Solubilities, Retardation Factors, Complexes, and Colloids. Laboratory data already exist in great abundance, but some important numbers are still poorly known. Further refinement is needed in simulating repository conditions as closely as possible, and additional field checks of laboratory data are desirable. The effects of complexes, colloids, and slow reactions in modifying predictions from thermodynamic data are particularly in need of further study. Geologic Investigations. Further search for repository sites is obviously desirable, and is part of the current KBS plan. Since at least four fairly satisfactory sites are already known, and since a site for actual construc- tion will not be needed for at least another decade, this activity is less urgent than some other kinds of research. Additional study of evidence for postglacial rock dis- placements in various parts of the country would be of much scientific interest and might aid in the ultimate choice of a repository site, but there is little chance that it would change the general conclusion that Swedish bedrock is sufficiently stable to ensure long-term repository integrity. Probabilistic Risk Assessment. A probabilistic risk assessment of the repository system should be made.
