Predicting 4D fracture patterns and related permeability evolution in fault zones by numerical-analytical modelling (FRAPtre).

Fracture patterns and related permeability properties along fault zones result from the complex interplay between the state of stress and rock deformation during faulting. Rock fracturing is controlled by a number of parameters including the fault geometry and kinematics, the fault slip mechanism, the rock type and rheology, the boundary stress conditions, the pore fluid pressure, and the structural inheritance. The behaviour of fault zones has been simulated by a numerical-analytical tool (FRAPtre) that links in both ways the boundary stress conditions and the deformation fabric for a given fault geometry and kinematics. Results include also the evaluation of the state of stress along the fault in static stress conditions (i.e. before fault activation), as well as permeability predictions. The FRAPtre modelling parameters include the clay content, the amount of displacement, the stress/strength condition along the fault surfaces, and the path of the displacement on it. Fault surfaces are gridded and stress/strength conditions at each cell are analytically computed by a series of stress tensor additions. The strength values are then computed (at each cell) by projecting the resulting stress tensor on the fault cell surfaces, as well as the normal pressure and the shear. At this stage it is possible to compute the deformation function DF, that is the difference between shear and strength on the fault cell surface. The DF parameter measure the attitude of rock to break into a cataclastic rock (or to grind existing grains in clastic rocks) and to start the fault gouge formation process. The standard variation of rock rheology around the given mean values is included in the modelling procedure to obtain a more realistic statistical approach. The Fault Gouge Index FGI is computed by cumulating the deformation function of all cells along the displacement path, conveniently weighted by the corresponding portion of the displacement, and multiplied by the clay content fraction at that cell. The fault gouge thickness is computed taking into account the amount of displacement and the deformation function. Thicker fault gouge layers are predicted to occur where their formation is easier. Fault permeability is computed by a generalised equation that matches permeability values available in literature on fault cores. The resulting spatial distribution of predicted permeability is then output to interface with commercial software for fluid flow modelling in reservoirs.