Enhanced fault imaging from seismic and geological model

Lacaze, Sebastien (Eliis) | Luquet, Benoit (Eliis) | Valding, Thomas (Eliis) | Pauget, Fabien (Eliis)


Imaging faults is a complex process, which requires a combination of various approaches. Methods based on the gradient vector field, obtained from the seismic 3D cross correlation, is sensitive to any local variation. Deriving the vector field to local dip, curvature or oriented filters such as variance, is used extensively to enhance structural discontinuities. By analyzing the maximum of variance, a new attribute depicts the probability of fault occurrence. Although it shows a skeleton of the fault network, it remains difficult to use it for automatic extraction.

Another method consists in using derivatives of a relative geological time model, obtained during a comprehensive interpretation process. In such case, the fault image is directly related to the vertical throw and provides a high level of detection even where the seismic variance is limited due to a low signal to noise ratio. To increase the precision of the detection, surface attributes for each relative age are computed in the flattened space and then converted to the seismic domain.

With such technique, the calculation of the extrema values of the deepest descent gradient shows the fault break points at a sub seismic accuracy and is related to the vertical throw. It becomes a complementary attribute to the variance and the fault probability. Applied to the Exmouth data set, located on the North West Australian margin, these various types of attribute were used to interpret complex faulted deposits in the reservoir level.

Presentation Date: Monday, October 17, 2016

Start Time: 4:35:00 PM

Location: 143/149

Presentation Type: ORAL