Use of PSDM for improving the accuracy of horizontal drilling program
The January issue of The Leading Edge includes a special section on “Advancements in 3D seismic processing”. Devon Energy and SeismicCity contributed a case study article to this section discussing the use of prestack depth migration in unconventional reservoirs with examples from the Woodford shale play in the Anadarko basin in Oklahoma.
The article: “Use of prestack depth migration for improving the accuracy of horizontal drilling in unconventional reservoirs” is authored by Marianne Rauch-Davies and Scott Sutherland from Devon Energy and Michael Bradshaw, Jeff Codd and David Kessler, from SeismicCity Inc.
The full text can be accessed here: TLE article.
The following is a summary of the important points from the TLE article.
Historically, prestack depth migration (PSDM) was used to improve the image quality in areas of complex geology. In simple geological settings typical of those encountered in most unconventional shale reservoirs, prestack time migration (PSTM) was mostly utilized. However, for unconventional exploration and production programs, PSTM fails to produce the data quality needed. In these cases we are required to perform lateral drilling along thin intervals associated with rapid lateral and vertical velocity and anisotropy variations.
In general, PSDM aims to achieve two goals: (a) to produce a reliable and high resolution image, and (b) to correctly convert the time domain seismic data to depth and specially accurately locate the geological units . Both goals can be accomplished if the right models are used as input to PSDM and if the right parameters are used in the application of PSDM. As detailed in the TLE publication, planned horizontal well paths using PSDM rather than time to depth stretched PSTM have a better chance of following the true geology and staying in the producing zone, thus achieving the production objectives.
To produce a PSDM volume that can be successfully used for optimization of horizontal drilling program the following needs to be done:
Earth Model
An accurate layer based anisotropic model needs to be constructed. This earth model consists of layers that are associated with the main geological units of the subsurface. The earth model consists of both velocity fields, as well as anisotropic parameters fields.
Velocity and Anisotropic fields
The velocity and anisotropic parameters in each layer need to be optimized. This is done by using azimuthally dependent tomographic inversion where gather moveout at each CDP location is measured along several azimuths.
PSDM parameters
Final PSDM is applied using optimized parameters. Application of final PSDM using 10 ft depth sampling will result in a PSDM output that has the same fidelity of seismic data normally provided by PSTM. In the TLE article we demonstrate that the PSDM image actually contained a higher frequency range than the equivalent PSTM image. This enabled detailed interpretation of the seismic data.
Seismic to well calibration
Our goal is to produce a final PSDM volume that ties main geological well formation tops. This includes a final update of the velocity field to ensure that the final PSDM image ties the provided well data. The final update of the velocity (and anisotropic fields) is carried out using a tomographic inversion process. The input to the inversion is the discrepancy between well data and seismic data. The objective of the inversion is to minimize the discrepancy between the seismic data and all well formation tops associated with each geological unit, as well as between well locations. This process ensures that velocity anomalies are not created as part of the final seismic to well calibration process.
The result of the above procedures is a PSDM volume that (a) produces an optimal PSDM stack, (b) produces an anisotropic model that represents the sub-surface geology, (c) yields a PSDM image that is equivalent or superior to any PSTM image, and (d) best ties well data so the final PSDM can be used directly for interpretation and mapping.
Use of a PSDM volumes produced using the above procedures are now successfully used to optimize horizontal drilling programs.
Thank you Bruce !
Great stuff, similar conclusions are seen in the Delaware Basin !
The lateral velocity in the vertical section needs a bit more look. How about a depth section? Could it be due to the local effect introduced by the well data? We used horizontal well tops to control the velocity model at BHP for Eagle Ford shale. I can expect similar effect. How to seamlessly make the velocity transition geologic may require high frequency FWI.
Nice article and supports our premise that even for Unconventionals PSDM is superior to PSTM. I did notice a rather large lateral velocity contrast in what appears to be that same geologic layer. And the reverse pattern in a layer beneath. Without a reference colorbar it's hard to tell the actual differential. How well does the model tie the sonic logs in those parts of the model?