Considerations de l'IRSN relatives à la reconnaissance de discontinuités géologiques à proximité d'une installation de stockage profond de déchets radioactifs
Titre du congrès :3rd AMIGO Workshop on "Approaches and Challenges for the Use of Geological Information in the Safety Case"
Ville du congrès :Nancy
Date du congrès :15/04/2008
IRSN has carried out the safety review of ANDRA’s 2005 study on the feasibility of deep geological disposal of high level, long-lived radioactive waste in the Callovo-Oxfordian clay formation (COX) located in the Meuse/Haute-Marne area (MHM) in the eastern Paris Basin, and investigated by ANDRA in the Bure URL. At the scale of a potential repository, the question raised by IRSN was the survey methods to identify potential fractures in the COX, which could be transmissive or be hydraulically activated by the excavation and operation of disposal facilities, and thus could play a role on radionuclide transport.
To determine the geometry of the geological layers and to detect faults at the URL scale, ANDRA has carried out a 3D seismic survey (4 km2), which shows seismic subvertical structures crossing the Triassic to Middle Dogger formations, but nothing above, notably in the COX. These structures that IRSN has interpreted as possible secondary deep faults were not detected in the 2D seismic profiles. At the IRSN Tournemire experimental station (Aveyron, France), in an ancient railway tunnel crossing a 150-m thick Toarcian clay formation at 250 m depth, a secondary fault zone was intercepted. This hectometre-sized subvertical fault set, displaying a 2-m vertical offset and larger horizontal offset, varies in space, from a dry meter-sized fault gouge showing the same aspect as the intact rock, to a hydraulically active breccia associated with 10’s m wide fractured zone. From the surface, IRSN has tested a 3D high-resolution seismic survey which showed faults in the clays’ underlying limestone layer and at these two units’ interface but not within the clay layer itself. Also, some locally transmissive faults may be not identified by 2D-3D seismic surveys.
In order to investigate the Bure seismic indices, ANDRA has drilled subhorizontal boreholes above these indices. Only reef formations and joints in the Late Dogger as well as a few joints in the COX were observed. In the investigated zone, these structures neither propagated to, nor through, the COX. Furthermore, for the future repository site survey, a 3D seismic survey combined with a set of oblique boreholes in the COX, drilled above the most pronounced seismic indices in the underlying formations, should provide sufficient data to assess structural characteristics of this site.
In the boreholes drilled by ANDRA as well as in the URL’s main and auxiliary shafts, IRSN noted that fracturing was rare in the COX, which contrasts with the fracture density observed in surrounding Dogger and Oxfordian limestones. IRSN considers that a “differential fracturing” phenomenon could be a plausible explanation for how fracturing is organised in the MHM formations. This phenomenon reflects fracturing expressed differently in limestones, which have a breaking mechanical behaviour, and in the more plastic clays. This well known phenomenon at inframetric scale was revealed at a plurihectometric scale in Tournemire, where boreholes in the limestones underlying the Toarcian clays have intercepted a normal fault re-activated as strike-slip associated with a thin fractured zone; this structure continues into the clays as a wide diffuse zone of thin, strike-slip faults (described above); this fractured zone has a far clearer impact on the overlying limestones and widens as it fades toward the surface. The COX probably has great potential to dampen fracturing in MHM, which suggests that sound blocks of size consistent with possible repository designs exist in this layer in the 200 km2 zone of interest considered by ANDRA.
However, South-West of the MHM study area, several of ANDRA’s boreholes have intercepted open fractures without any vertical offset, both in the Dogger and the Oxfordian limestones, which hydraulically influence the entire sedimentary pile. They are located in a western extension of regional faults and correspond to supposed structures on a 2D seismic profile. IRSN believes that these structures, showing the same direction and size as the deep ones observed in the Bure 3D seismic block, might relate to strike-slip secondary faults that have propagated across the COX. Also, caution is necessary when extrapolating data acquired in the URL area to a wider scale, given the co-existence of faults of similar origin to, but with varying influence on, the hydraulic properties of the COX.