FAQ Faults & Fractures
Q: I'm trying to find a workflow to explore the Faults and Fractures plug-in. Do you have a guide that will lead me through my first try at this plug-in?
A: The Faults & Fractures plugin is divided into four sections: Attributes, Filters, Planes and Tools. The interface is organized in such way that first fault Attributes are calculated. Among the most popular are: Curvature, Semblance, Similarity, and Thinned Fault Likelihood (TFL) - known to give superior results. Proceed further with Filters, which filter the calculated fault volume (Thinning) or the seismic data, steering it with the dip and fault volumes. A user can also enhance faults in the seismic data (Dip-Steered Diffusion Filter) and then go back to the Attributes to calculate the fault volume again, providing the fault enhanced seismic as an input.
The Planes section allows extracting fault bodies using either Extract Planes tool or Voxel Connectivity Filter. Tools offer you to calculate Fault dips, which can further be incorporated with dips from the seismic volume to get better results for HorizonCube and Unconformity Tracker.
Q: Cannot find much information on parameter choices. Can you give some guidance on how the range, step and Velocity parameters affect the output?
- Stepout calculation should always be smoother in the vertical direction to catch the fault trends. 1-1-16 is our best practice, but you could also try 1-1-8 or 1-1-32
- Fault Strike Scan range - the 0 deg reference line is along the Cross-line of the survey (not true north coordinate) from smaller In-line values to larger In-line values
- Fault Dip Scan Range - the 0 deg reference line is along the In-line of the survey from smaller Cross-line values to larger Cross-line values. The algorithm scans for the fault dips and strikes within a given range of the dips to maximize the fault likelihood
- Nr steps is how many scans should be done. The wider the dip and strike ranges are, the more information will be extracted from the data
- Velocity needs to be used in time surveys to convert angles. Use an average value for your data set, it will not affect the results much.
Q: What parameters can be altered to increase processing speed?
A: Decrease the number of steps can speed up the processing, but it will affect the results as well. It is also recommended to use machines with more RAM, and higher number of CPUs. E.g. TFL processing of 63 GB input volume on a machine with 128 GB RAM should end within 2-3 hours.