A strain-softening numerical model of core discing and damage

There has been a debate in the rock mechanics community regarding the mechanisms causing what is known as core discing. This phenomenon occurs when diamond drill cores are retrieved from rock masses in which high in situ stresses relative to rock strength are present. The interest in that phenomenon is to use it to estimate the in situ stresses from the shape and frequency of the failures along the core axis. In the present paper we argue that discing is only an indicator of high stresses, and that estimating in situ stresses from fracture observations is much too inaccurate.As most of the literature found on the subject has tackled the problem using elastic numerical models, it is shown that the stress distribution in the core being formed obtained from such models does not exist once failure has been reached. Numerical analyses using Flac2D with an elasto-plastic cohesion softening friction hardening model show that for a given stress state, discing or core damage may involve tensile failure, a combination of shear and tensile failure, or only shear failure, depending on the stress state and ratio of tensile to shear strength of the rock. The numerical model used is validated by replicating core discing observed under controlled laboratory conditions. Parametric analyses involving changes in mesh density, deformability parameters, dilatancy, drill bit pressure, drilling fluid pressure and applied stress states are also performed. Finally, it is shown that drilling-induced core discing or damage store important residual stresses in the core which may explain why recovered core samples tend to show a deterioration of their mechanical properties with time.