Deep repository for spent nuclear fuel. SR 97 - Post-closure safety. Main Report. Summary

In preparation for coming site investigations for siting of a deep repository for spent nuclear fuel, the Swedish Government and nuclear regulatory authorities have requested an assessment of the repository's long-term safety. The purpose is to demonstrate whether the risk of harmful effects in individuals in the vicinity of the repository complies with the acceptance criterion formulated by the Swedish regulatory authorities, i.e. that the risk may not exceed 10-6 per year. Geological data are taken from three sites in Sweden to shed light on different conditions in Swedish granitic bedrock. The future evolution of the repository system is analyzed in the form of five scenarios. The first is a base scenario where the repository is postulated to be built entirely according to specifications and where present-day conditions in the surroundings are postulated to persist. The four other scenarios show how the evolution of the repository differs from that in the base scenario if the repository contains a few initially defective canisters, in the event of climate change, earthquakes, and future inadvertent human intrusion. The time horizon for the analyses is at most one million years, in accordance with preliminary regulations. By means of model studies and calculations, the base scenario analyzes how the radioactivity of the fuel declines with time, the repository's thermal evolution as a result of the decay heat in the fuel, the hydraulic evolution in buffer and backfill when they become saturated with water, and the long-term groundwater flow in the geosphere on the three sites. The overall conclusion of the analyses in the base scenario is that the copper canisters isolating capacity is not threatened by either the mechanical or chemical stresses to which it is subjected. The safety margins are great even in a million-year perspective. The internal evolution in initially defective canisters and the possible resultant migration of radionuclides in buffer, geosphere and biosphere are analyzed in the canister defect scenario. Releases from the geosphere are converted to doses in different ecosystems. The variation in flow-related data in the geosphere has the greatest impact on the result, followed by data uncertainties for the biosphere. Other conclusions are that our understanding of fuel dissolution needs to be improved, and that the probability and size of initial canister defects that escape quality-control inspection is difficult to estimate. Risk analyses in the form of simplified probabilistic calculations are also performed. The risk analyses show that all sites lie well below the acceptance criterion. The maximum risk for release to a well is never more than 0.5 percent of the acceptance criterion, even when the calculations are extended a million years into the future. The same applies to releases to peat land for times up to 100,000 years, while the maximum risk here grows to about one-tenth of the acceptance criterion at the least favourable site at times after 100,000 years. The consequences of future climate change are explored in the climate scenario. A conceivable sequence of events, including severe glaciation, on each of the three sites is sketched for the coming 150,000 years. In the climate scenario as well, the overall conclusion is that the isolating capacity of the copper canister is not threatened by either mechanical or chemical stresses. As far as the retarding capacity of the repository is concerned, for example in the event of initial canister damage, the most important changes take place in the biosphere. The repository sites are expected to be covered by ice sheets or sea during long periods, and the aggregate effect of climate change will therefore be a reduction of the dose consequences compared with a situation where the present-day climate persists. In the earthquake scenario, the consequences of earthquakes are analyzed by means of model studies where site-specific data are used for the structure of the geosphere and for earthquake statistics. The analysis method is new and includes several highly pessimistic simplifications. The analyses show that the probability of canister damage is comparable with the probability assumed for initial damage in the canister defect scenario. In the evaluation of the analysis method, it is shown how less pessimistic assumptions should lead to no canister damage at all in the model studies. The method will be refined. The scenario that deals with future inadvertent human actions that could conceivably affect the repository is surrounded by great uncertainties, chiefly because the evolution of human society is unpredictable. SR 97 discusses how conceivable societal evolutions and future human actions that affect the repository can nevertheless be categorized to some extent. In an illustrative example, a situation is analyzed where a canister in the repository is inadvertently penetrated by rock drillers. Dose and risk are calculated for the drilling personnel and for a family that settles on the site at a later time. The principal conclusion of the SR 97 safety assessment is that the prospects of building a safety deep repository for spent nuclear fuel in Swedish granitic bedrock are very good. The results of the assessment also serve as a basis for formulating requirements and preferences regarding the bedrock in site investigations, for designing a programme for site investigations, for formulating functional requirements on the repository's barriers, and for prioritization of research. The next stage in the siting of a deep repository entails investigation of the bedrock at a number of candidate sites in Sweden. It is SKBs judgement that the scope of the safety assessment and confidence in its results satisfy the requirements that should be made in preparation for such a stage