The thick-walled cylinder test has become increasingly widely used recently as a means of determining minimum mud weights while drilling for hydrocarbons and maximum drawdowns during production, and as a way of validating elastoplastic models of rock behaviour. A set of such tests has been carried out on four sandstones, together with triaxial testing to obtain parameters for elastoplastic modeling of the cylinder geometry. Tests were conducted in axial plane strain and with a range of low pressures applied to the inner surface of the hole. This internal pressure has been found to exert a very strong influence on the closure of the hole. Examination of the shape of the hole, and application of elastoplastic models, indicate that these changes are not consequences of homogeneous plastic deformation, but of symmetry-breaking and the onset of localized deformation. The relevance of the effect to the stability of real wellbores is discussed.
The thick-walled cylinder (TWC) test (where a cylinder with a central axial hole is subjected to increasing external pressure) is a relatively straightforward rock mechanics test that allows investigation of deformation in a geometry other than that of the triaxial test. The geometry is one that is specially relevant to the problems of wellbore stability and sand production during hydrocarbon exploitation, and this has provided the motivation for several recent experimental investigations (e.g., Ewy, 1993;, Addis and Wu, 1993; Tronvoll et al, 1993; van den Hoek et al, 1994) and for the work presented here. Wellbore instability and sand production are strongly influenced by the stress state and deformation of the rock around the hole. It is often found that, in order to predict such problems, elastic-brittle models (where the deformation is assumed elastic up to failure) are inadequate (McLean and Addis, 1990), and elastoplastic models must be used instead. Such models generally require calibration and validation. Calibration parameters are derived from tests where the deformation state is nominally homogeneous (such as triaxial tests); the model can then be validated by checking its predictions against measured deformation in another geometry. The non- homogeneous deformation during the TWC test makes this a stringent test of models, and the cylindrical geometry also allows close comparison with field phenomena. The pressure required to cause failure of thick-walled cylinder samples has been used in the past for model validation (Papamichos and van den Hoek, 1995), but we will not do this; only stable deformation is considered. This paper presents TWC data and calibration for a number of sandstones. These are used to test a semi-analytic elastoplastic model, and also to demonstrate the high sensitivity of the TWC test to the stresses on the inner surface of the cavity.
The model (Bradford and Cook, 1994) was developed to provide a prediction of deformation and stress state around a wellbore or perforation in an isotropic rock subject to a cylindrically- symmetric stress state. It is semi-analytic because it must run quickly, to provide predictions for many depths in a single well.
