PART I: The Swedish Nuclear Fuel and Waste Management Co. (SKB) has recently completed a safety assessment project named SR-Can, related to the KBS-3 disposal concept. In this concept, the waste packages are surrounded by a buffer made of either MX-80 or Deponit CA-N bentonite. Interactions between the buffer and groundwater may modify the buffer composition and thus its containment properties. The Swedish Radiation Protection Authorities (SSI) requested the French Institute for Radiological Protection and Nuclear Safety (IRSN) to perform the present study in support of SSI review of the SR-Can report. The purpose is to assess the geochemical evolution of both potential buffer materials due to the intrusion of different types of groundwater, with a similar modelling layout to that reported in SR-Can and detailed in Arcos et al. (2006). Three main categories of water inflows via a fracture intersecting a deposition hole are considered: the Forsmark reference groundwater, a high-salinity groundwater to account for up-rise of deep-seated brines and a diluted water representing ice-melting derived-groundwater. In addition to this, the redox buffering capacity of Deponit CA-N bentonite and the thermal effect on MX-80 bentonite geochemistry have been assessed. This modelling work has been performed using the reactive transport modelling code HYTEC.
PART II: Elements of the SR-Can project relative to piping and erosion phenomena of bentonite components of a KBS-3 repository are analysed with regard to the experience feedback available at IRSN and consisting in experimental results obtained on samples at the UJF-Grenoble between 2000 and 2004. A synthesis of these tests is presented, with a closer attention to the Argillite/Bentonite tests during which phenomena of erosion occurred. The reference evolution of a KBS-3 repository, the resaturation and swelling kinetics of backfills and buffers and the possibility for a buffer to swell upwards the backfill have been considered. According to the reviewed documents, IRSN notes that the SR-Can project tackles the piping and erosion phenomena with local modellings and “rough estimates”, the latter being based on 3 “key” parameters: the water inflow in an underground opening, the concentration of bentonite in pipe water and the duration of the phenomena. IRSN considers that the reviewed elements do not evidence enough the conservatism of the parameters value, especially for the duration of the phenomena. Additional experimental results, at small and large scale, may be necessary.