Conditioning of Geomechanical Simulations to Well Data in a Fractured Carbonate Reservoir

Conditioning of spatial explicative variables, or "drivers”, to well data is an important step in validating fractured reservoir models and increasing confidence in their utilisation for reservoir management. A case study is presented of a carbonate reservoir in which fracture orientations and frequencies output from geomechanical simulations are conditioned to distributions interpreted from well data. The distribution of fault-related elastic stress is simulated by imposing a palaeostress field on the interpreted structural framework. Orientation of shear and tensile fractures is output along with the distance to failure according to a combined tensile and shear failure criterion. These are compared with fractures interpreted from borehole image logs, which allows rejection or modification of some possible hypotheses for fault geometries and palaeostress. Once a satisfactory match is obtained, the simulated failure criterion is combined with other drivers in a multivariate analysis to produce a three dimensional fracture probability model which explains over three quarters of the observed variation. The resulting property can be used as input to karst modelling and subsequent reservoir simulation.