Three-Dimensional Fracture Pattern Predictions in Thrust-Related Anticlines by Hybrid Cellular Automata (HCA) Numerical Models

Tavani, Stefano¹, Francesco Salvini¹, Fabrizio Storti¹, Roberto Gambini² (1) Universita’ Roma Tre, Roma, Italy (2) OMV Aktiengesellschaft, Vien, Austria

Substantial improvements of our predictive capability of fracture distributions in hydro­carbon exploration and development require to implement parameters such as rock mechanics in more sophisticated predictive tools. Our approach for predicting the geomet­rical and deformational architectures of fault-related structures includes the use of Hybrid Cellular Automata (HCA) numerical models. HCA algorithms, merging the properties of cel­lular automata and finite elements techniques, are implemented in a real time forward mod­elling technique that allows the numerical simulation of the behaviour of natural rock multi­layers undergoing deformation at shallow crustal levels. Physical parameters describing the mechanical properties of each simulated rock type are derived from seismic attributes. This information is integrated with stratimetric data to create the undeformed numerical multilay­ers, which deform without any externally-imposed velocity field. Numerical outputs from HCA algorithms (FORC 2) include the distribution of the stress-time integral values across the modelled sections. This parameter is self-determined during the model runs and derives from kinematical and rheological constraints. The distribution of the stress-time integral val­ues 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 converted into quantitative fracture predictions by using field analogues. In the simulation of prospect reservoirs, the “tuning process” provides the link for statistically achieving the proper parameters of deformational features and for using them to generate synthetic frac­ture datasets along synthetic wells.