Résumé
To reduce exploration risk and optimize production in structurally complex areas, the geological interpretation must be physically and mechanically reliable. Despite the advances in 3D seismic techniques and in the availability of computationally-robust interpretation software, the challenge associated with interpreting complex structures in seismic reflection data is that highly deformed areas surrounding faults, folds and salt surfaces are often poorly imaged. We present a methodology to help geophysicists quickly check the mechanical strength and weakness of their interpretation and to automatically correct the faulted horizon geometry. The workflow consists on restoring interpreted seismic horizons and relating the concentrations of computed attributes such as stress or strain to areas of interpretation uncertainty. The technique used, based on the finite element method, allows unfolding and unfaulting of 3D horizons using the physical behavior of rock mass. A fast algorithm has been developed to automatically correct the interpreted structures in zones that exhibit anomalous stress or strain concentrations after restoration. The modified horizons are then imported back to the seismic data interpretation software for comparison with the seismic volume. This approach is able to mechanically check and correct seismic data interpretation. Its application to both synthetic and reservoir data demonstrates a high degree of reliability in the characterization of structurally complex reservoirs.