Interpreting strain measurements when drilling anisotropic rocks: return of experience from using CSIRO cells in Tournemire argillite

In Geomechanics, determining the state of stress and the rheology of rock massif can be obtained by measuring the strain response of the ground under the effect of a known stress. A method among others is to use a cell integrated (with strain gauges oriented in different directions) installed in a borehole and secured to the mass through an epoxy glue. This measurement is used, notably, to determine the stress state in situ by the 'overcoring' method and the elastic parameters of the rock from the 'biaxial' test. Between November 2005 and January 2006, a geomechanical testing campaign was conducted in the argillaceous formation of the Tournemire experimental site (Aveyron, France) using CSIRO Hi cells. The strain measurements obtained during overcoring and biaxial tests have shown unusual phenomena, which have made difficult the determination of anisotropic elastic parameters of the rock and the access to the site stress. Therefore, through researches for explanations of the origin of these phenomena, this thesis aimed to improve and contribute to the understanding of the Tournemire argillite behaviour and to upgrade the measurement protocol as well as the interpretation of cells CSIRO's strain. The approach was, firstly, to issue a number of hypotheses to explain certain phenomena observed in literature. In a second step, these hypotheses were tested through analytical and numerical modelling of the biaxial and overcoring tests then through the realization of new experiments in situ within laboratory on argillite, and also on materials tests (cement, sample of glue). It is concluded that the unusual phenomena observed are essentially the result of the conditions for implementing in situ CSIRO's cell. The study particularly focused on the artefacts induced by the visco-plastic behaviour of the epoxy glue when it is incompletely polymerised. The role of damage on the rock generated by drilling operations is also discussed. Some practical recommendations for implementing new tests under similar conditions are drawn