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.
SUMMARY In the judgment of the panel, the KBS-3 plan provides an adequate technical basis for the conclusion it reachesâ that there is reasonable assurance that a waste repository built according to the KBS-3 plan will not permit an unacceptable rate of radionuclide escape. The Swedish plan for disposing of spent nuclear fuel was first reviewed in 1979 by a subcommittee of the Board on Radioactive Waste Management (BRWM; then called the Committee on Radioactive Waste Management) of the National Research Council (NRC). The present review by a newly formed panel of the BRWM is a follow-up to the earlier one, made necessary by the extensive research accom- plished during the past five years by the Swedish organization responsible for the disposal plans, the Karnbra'nslesakerhet (KBS) . In 1978 the Swedish proposal was described in a document entitled Handling and Final Storage of Unreprocessed Spent Nuclear Fuel and commonly referred to as KBS-2; a largely rewritten and consid- erably expanded version of the KBS-2 document, entitled Final Storage of Spent Nuclear Fuel, or KBS-3, is the subject of the present review. Both KBS-2 and KBS-3 are supported by a large number of detailed technical reports, most of which were critically examined in the course of the two NRC reviews, which were undertaken in response to requests from the Swedish Ministry of Industry for external evaluation of the work of their own scientists and engineers. The current NRC review, as was true in the case of the earlier review, relies almost exclusively upon information developed by KBS. The panel members did not undertake to perform research or to gather evidence. Essentials of the Swedish plan have not changed between KBS-2 and KBS-3. Both are concerned with the
disposal of spent fuel rods from power reactors (in con- trast to KBS-l, which considered the disposal of radio- active waste recovered from the reprocessing of spent fuel and subsequently incorporated into glass cylinders). Like many proposals for managing high-level waste, the Swedish approach involves multiple barriers, some natural and some engineered, to keep the long-term escape of radionuclides within acceptable bounds. The Swedish plan differs from most others in its heavy reliance on engi- neered barriers, specifically thick-walled copper canis- ters to enclose the spent fuel rods, surrounded by buffers of compacted bentonite. These are to be placed in holes drilled into the floor of a cavity excavated in bedrock 500 m below the land surface. The bedrock, which is expected primarily to provide stable conditions in which the corrosion of the copper would be very slow, could also serve as a secondary barrier to radionuclide escape should some of the canisters fail. The canisters, how- ever, are expected to last for a million years or more, and the function of bedrock as barrier is envisioned only as insurance against remotely possible accidents early in the life of the canisters. These major aspects of the plan have remained the same between 1978 and 1983, but details have been altered, and much research has been carried out to test predictions about the long-term behavior of canisters and buffer and about the movement of radionuclides through bedrock. Evaluation of the recent research results is the focus of the present panel's review, and discussion from the earlier KBS-2 review is repeated only when necessary for background and continuity. The panel has devoted its effort primarily to judging how well the new work of the KBS scientists and engineers has supported their con- clusions about the long-term safety of the planned dis- posal system and how convincingly they have answered questions about weaknesses in earlier versions of the plan. Emphasis is on the adequacy of the enhanced technical data base with respect to two elements of the disposal plan: (1) the long-term stability of copper canisters enclosed in a bentonite buffer and (2) the availability of deep geologic disposal sites with the necessary dimensions, stability, and groundwater properties for maintaining a benign canister environment and providing an effective secondary barrier to radio- nuclide migration. As with the earlier KBS-2 review, shortage of time for this review precluded consideration of some aspects of the Swedish planânotably costs,
facilities for predisposal handling of the fuel rods, comparison with other disposal methods, and the vulner- ability of the waste repository to later exploitation for copper or fissile materials. The current review is essen- tially a subjective evaluation by the panel members of the quality and completeness of the Swedish research and the logic relating research results to conclusions drawn. Like its predecessor KBS-2 subcommittee, the KBS-3 panel finds that the conclusions reached in the plan were warranted by the evidence submitted and is convinced that an adequate technical basis has been developed to provide reasonable assurance that radionuclides will not escape into the biosphere in unacceptable amounts from a reposi- tory located and constructed as specified by the KBS-3 plan. A few gaps and uncertainties in the evidence remain, despite the newer work, but the panel finds little reason to question the overall soundness of the KBS con- clusion. The principal reasons for the panel's judgment, together with notes about the remaining uncertainties, are summarized below. Canisters. Two kinds of canisters are described in KBS-3, both of copper with walls 0.1-m thick, reduced from the 0.2-m thickness proposed in KBS-2. One kind is fabricated as in KBS-2, with molten lead filling around the fuel rods and a cap sealed by electron-beam welding; the new second variety is to be filled with copper powder that is compacted and sealed by hot isostatic pressing. A choice between the two has not yet been made. Fabrica- tion of both kinds has been demonstrated, the former at full scale and the latter at about one-third scale. Recent research has shown that pit corrosion is less rapid than was conservatively assumed in KBS-2; hence the wall thickness can be safely reduced. The research has further amply corroborated the KBS-2 conclusion that stress corrosion cracking is not a significant hazard for the kind of copper that will be used. Experts at the Swedish Corrosion Institute had cautiously estimated a guaranteed canister life of 105 years in KBS-2, but they now extend this to more than 106 years. The panel concludes that the larger estimate is fully justified. The Buffer (referred to as "overpack" in KBS-2). The compacted bentonite to be placed around each canister in its deposition hole is expected to (1) provide a cushion for the canister in case of mechanical disturbance; (2) keep groundwater stationary or almost stationary near the
canisters, so that access of corrosive agents to the canister walls will be only by slow diffusion through the bentonite; and (3) assist in maintaining slightly alkaline conditions near the canisters. The second and third functions have been amply demonstrated by experiment, but the ability of the bentonite to cushion against sudden shear, as might happen in an earthquake, is not yet adequately proved. This is a gap in available evidence, but not a serious one, because the probability of fault movement of the right magnitude and orientation to cause disruption is exceedingly slight. Questions raised in the earlier KBS-2 review about possible difficulties in emplacing the bentonite without leaving open channels for groundwater have now been answered by full-scale demon- strations in the underground experimental facility at Stripa. Existence of Repository Sites. Identification of suitable repository sites is made difficult by the need to extrapolate fracture patterns in bedrock from surface exposures to depths of 500 m, aided only by measurements in a few boreholes and by geophysical observations. Such extrapolation is to some degree necessarily speculative, but in KBS-3 the speculative aspect has been diminished by the use of additional deep boreholes and by improve- ments in geophysical techniques. Three new sites have been found that seem at least as promising as Karlshamn, the site that survived detailed study in KBS-2. Bach site has an area of rock between major fracture zones large enough, though only barely so, to accommodate a repository, and each has secondary fractures that would have to be avoided or sealed by grouting. None of the four studied sites is ideal, but any one of them could be made usable by proper choice of locations between frac- ture zones and by grouting of minor fractures. There is little reason to doubt that better sites will be found by the ongoing Swedish program of geologic exploration in the decade or so that remains before a site will actually be needed. Stability of Sites. Recent geologic and geophysical studies of faults in the northernmost part of Sweden have shown Quaternary displacements of as much as 30 m. Else- where in the country, evidence for even small recent fault displacements is scanty, and the general stability of the crystalline bedrock is shown by the infrequency and small magnitude of Swedish earthquakes in historic
time. Much new information has been gained in the past five years about the relative importance of postglacial rebound and tectonic forces in the slow, long-term, country-wide up-and-down motion of bedrock, and doubts once expressed by some Swedish geologists about general tectonic stability have been largely dispelled. Quantity, Movement, and Chemical Composition of Groundwater. Study of groundwater from boreholes at the newly explored repository sites has shown relations much like those found earlier at the sites described in KBS-2: wide variation in hydraulic conductivity from high values in zones of fracturing to very low values in relatively unfractured rock, and a general decrease in conductivity downward from the land surface. Average values of hydraulic conductivity at different depths, combined with estimates of hydraulic gradient, make possible estimates of rates of groundwater movement. Such estimates are necessarily uncertain, especially since much groundwater flow may be concentrated in narrow channels; but even the higher estimated rates are not large enough to affect a repository adversely, and the water adjacent to a canister is kept practically motionless by the buffer. Many new chemical analyses, aided by improved techniques, confirm earlier results showing that groundwater at depth is uniformly slightly alkaline and slightly reducing, with only minor amounts of possible corrosive agents. In con- tact with water of this nature, the corrosion of copper canisters would be exceedingly slow. Temperature in a Repository. Repository temperatures depend on the length of time fuel rods are stored before disposal and on the loading and spacing of the canisters. These variables can be adjusted to achieve any desired temperature, so that no difficulty is expected in holding the temperature of canister surfaces to the intended maximum of 80Â°C. Recent calculations show that convec- tion currents produced by this temperature are too weak to disturb groundwater flow, and heater experiments at Stripa have demonstrated that the effect on fracture apertures is small. This temperature has also been shown to be too low to affect appreciably the properties of bentonite or the rate of corrosion of the canisters. Repository Closing. Mixtures of bentonite and sand are to be used as backfill in tunnels and shafts; and the emplacement of such mixtures has now been demonstrated at
full scale in the underground facility at Stripa, thus answering a question about feasibility that seemed especially serious to the earlier KBS-2 subcommittee. Recent experiments with compacted bentonite-sand mixtures have shown that densities can be readily attained at which permeabilities are comparable to the permeability of most of the bedrock in which the backfill will be placed. For preventing rapid groundwater flow through the rock adjacent to tunnels and shafts that has been fractured by blasting, seals of compacted bentonite are to be placed in slots sawed into the fractured rock. Demonstration experiments have been planned but not yet carried out. There is reason to doubt that such sealing could be entirely successful, in the sense that natural hydraulic conductivities are completely restored; but movement of groundwater could almost certainly be slowed enough to prevent harm to the canisters in their holes beneath the tunnels. A method of sealing boreholes has been demonstrated recently both in the laboratory and at Stripa: bentonite is packed in perforated copper tubes that are inserted into the holes, and absorption of water by the bentonite causes it to expand through the per- forations and so to fill the holes completely. Canister Failure. Should canisters somehow fail before their radionuclide content has decayed to harmless levels, thus exposing fuel rods directly to groundwater, escape of dissolved nuclides would be retarded by the insolubil- ity of the uranium oxide pellets and by sorption and ion exchange on mineral surfaces in rocks through which the groundwater passes on its way to a point where it may be tapped for human consumption. Details of the movement of released nuclides have been studied intensively during the past five years, especially with respect to oxidizing conditions that may be produced locally by radiolysis, enhancement of radionuclide movement by formation of complexes and colloids, and the modeling of groundwater flow through fractured rock. New models have been devised as alternatives to the commonly used hydrodynamic dispersion models for porous rocks, but the new models have actually been used only to a limited extent in estimating rates of radionuclide migration. Modeling of transport through fractured rock is a complex problem that remains one of the least satisfactory aspects of the effort to predict radionuclide releases. The abundance of experimental data, the partial check of such data by field tests, and the consistent use of pessimistic
assumptions all indicate, however, that estimates from the models should not be grossly in error. The further estimates of doses to humans from radionuclide releases are made by generally accepted techniques, and the additional uncertainties inherent in such estimates would not cause excessive risk to health and safety. Suggested R&D. Research during the past five years has gone far toward answering questions raised in the NRC review of KBS-2. The new work has been especially con- vincing in that much of it has involved field tests under repository conditions, a notable deficiency in the earlier studies. Some troublesome questions remain, for example those regarding the effectiveness of proposed shaft and tunnel seals, the validity of models for groundwater flow in fractured rock, and the adequacy of treatment of the many variables that influence retardation of radionuclides in their migration through granitic rock. Uncertainties about such questions will doubtless be narrowed in the course of continuing research, although final answers to some will probably remain elusive. Current knowledge is sufficient, however, to ensure that uncertainties will be manageable and that conservative assumptions regarding them have led to reliable estimates of repository performance.