Quantitative fracture predictions in reservoirs: A methodology based on the integration of field analogues and HCA numerical modelling (Forc 2).

It has long been recognised that the kinematics of thrust-related folds directly influences the 4-D distribution of their associated deformational features. Geometrical modelling has been used to produce templates of deformation patterns associated to contractional fault-related folds. Geometrically-derived deformation templates, however, are only valid for reproducing the first order distribution of deformational feature in fault-related anticlines for they neglect the role of boundary stress conditions and rock rheology. The mechanical stratigraphy, for example, exerts a first order control on deformation patterns in fault-related folds and can significantly alter the distribution of deformational features predicted by geometrical modelling. The impact of fracture distributions in hydrocarbon exploration and development imposes further efforts for substantially improving our predictive capability by implementing parameters such as rock mechanics in more sophisticated predictive tools. Our approach to the prediction of the geometrical and deformational architectures of fault-related structures includes the use of a specifically developed numerical tool, the Hybrid Cellular Automata (HCA). This real time forward modelling technique allows the numerical simulation of the behaviour of natural rock multilayers undergoing deformation at shallow crustal levels, by merging specific properties of cellular automata and finite elements techniques, and by providing the possibility of simulating bed thickness smaller than one centimetre. Input parameters used for describing the mechanical rock properties in the undeformed numerical multilayers are directly obtained from seismic datasets. Numerical outputs from HCA algorithms (FORC 2) include the predicted distribution of the stress-time integral values across the modelled section, which is self-determined during the model run and merely derives from kinematical and rheological constraints, since the velocity fields for hangingwall deformation are not imposed. This model output allows the recognition of hangingwall sectors expected to be more deformed than the adjacent ones but does not provide quantitative information on fracture type, orientation, and frequency. To achieve this result, stress-time integral values predicted by FORC 2 are tuned to field analogues where similar rock types are exposed. In the simulation of prospect reservoirs, the tuning process of FORC 2 outputs provides the link for statistically achieving the proper parameters of deformational features and for using them to generate synthetic fracture datasets along synthetic wells. Tadpole diagrams of fracture populations and their statistical analysis diagrams are then produced, thus providing support for exploration wells location and drilling, including their trajectories. In reservoir development, more direct information is available on fault-fold geometry, kinematics, and on fracture patterns. This allows a more precise tuning of numerical outputs, that can be used for planning locations and trajectories of new exploitation wells.