LE TURDU, CAROLINE, Elf Petroleum Norge, Stavanger, Norway, NAAMEN KESKES, FREDERIC JEANJEAN, Elf Exploration et Production, Pau, France, IVAR KRISTVIK, JEAN-CLAUDE RINGENBACH, PHILIP STRAW, Elf Petroleum Norge, Stavanger, Norway

Abstract: New Tools for 3D Fault Detection by Sismage TM: Implications to Interpretation of Complex Faulted Zones of the North Sea

Abstract

Fault interpretation is not fully constrained in complex structural areas. A good approach of fault patterns is a must to clearly understand problems of compartments, fault segmentation, and sealing at prospect scale. New tools (such as Optical flow, Curvature and Geometrical anisotropy) developed by Elf Research in Pau (SISMAGE TM ), using image processing techniques, were successfully applied for fault detection.

These new tools have been tested on complex faulted zones in the Northern Viking Graben and can be applied to seismic volumes without any horizon picking. Major faults are revealed, as well as minor and transverse faults that often play a prominent role in both tectonic history and production behaviour of an oil field.

These methods allow detailed fault patterns to be directly spotted, leading to a better knowledge of the 3D relationships between faults. Another important fact to point out is the great improvement of the imaging leading to an unbiased view of a 3D seismic dataset less and less dependant on the interpreter.

I - Curvature Method

This method consists in computing positive and negative curvatures from the 3D seismic volume. In most of application cases, it allows to detect hanging- and foot-walls of a fault (Fig. IA).As the curvature is applied though the 3D volume, it can easily follow each fault in 3 dimensions using timeslices, sections, or by extracting curvature along a specified path. 3D volume calculation also allows to extract an average curvature slice following the morphological variations of each reflector, leading to a "realistic" image of the structure (Fig. 1 B).2, 7 4

II - Applications

Detailed structural schemes are more and more requested to better understand problems of compartments, faults segmentations and networking, as well as fluid flow at the reservoir scale. This method has been successfully applied to several 3D seismic blocks of the North Sea.

III - Validation Of The Structural Map And Geological Implications

As major faults are generally detected by classical seismic interpretations, our interest was focused on the detection of minor or transverse faults. We compare in Figure 3A, a classical interpretation with the overlain curvature extraction along the interpreted horizon. This comparison shows that minor faults are generally not detected due to their small throw or related to their transverse orientation. Hence, some faults can be elongated (Fig. 3, Case 1), their segmentation is also well imaged (Fig. 3, Case 2) and some transverse faults are also detected (Fig. 3, Case 3).

The next part of the interpretation is to confirm the existence of faults by using seismic sections and time-slices. Following this validation, the fault surface is built and can be exported in a 3D modelling software. In this case, the method helps to focus on vertical and horizontal fault segmentations at each stratigraphical layer (Fig. 3 B and C).

IV - Conclusions

The curvature block allows to have an unbiased view of a 3D seismic block and to observe faults in 3D without any horizon picking. However the most interesting maps are extracted from curvature average slices following the reflector morphology. Thus, all the elongated curvature anomalies are detected such as flexures, small and transverse faults with lower throw. Even of less importance, these latter may play a prominent role at the prospect scale.

Tracking all types of faults within every layer and building them in 3D with increased confidence leads to more detailed and realistic structural schemes.The result is eventually to better understand problems of compartmentalisation at both regional and reservoir scales.

Curvature average maps can be combined together with seismic average slices of the same thickness to compare faults and maximum amplitudes distribution.This leads to the detection of faultcontrolled depocenters and is crucial to link tectonic and sedimentary processes.