|ADVANCED DEPTH IMAGING WORKFLOWS FOR PRE-SALT TARGETS IN DEEPWATER ANGOLA|
Following a series of significant pre-salt discoveries in Brazil, deepwater Angola is now a focus of intensive exploration. The presence of complex salt bodies means that pre-stack depth migration (PSDM) is essential. This case study from the Kwanza Basin provides an interesting example of depth imaging workflows to successfully image the pre-salt targets. Figure 1 shows the location of this multi-client case study. It covers nearly 4000 sq km on the outer slope of the basin with water depths varying between 1590m and 2370m. Most modern acquisition in deep water salt environments uses long offsets as well as multi-azimuth techniques but here we have been able to obtain good imaging results down to 12km depth with only moderate offsets (4.8km) and conventional narrow azimuth streamer data.
|MODEL BUILDING WORKFLOWS|
The basic workflow for building velocity models with salt bodies consists of multiple passes of sediment tomography, followed by a top salt interpretation. Some scanning is normally performed for an optimal salt velocity, then a salt flood migration is done to allow a base salt interpretation. To complete the model a sub-salt tomography is run if there is adequate resolution or possibly sub-salt velocity scanning for poor data quality areas. Advances in recent years that we will discuss include the use of TTI (tilted transverse isotropy) anisotropy as standard, integrated interpretation tools for top and base salt, scenario testing for difficult salt areas and different migration algorithms, many of which have been used in the Gulf of Mexico for a good number of years but not all of which may be appropriate for Angola.
Boudou et al, 2008 demonstrated the importance of TTI velocity anisotropy for offshore West Africa showing shifts in oil-water-contact positions of 100m. This case study confirms that TTI in the sediments impacts the pre-salt image. Grid tomography was an essential part of the model building workflow. We updated the sediment velocities with full multi-offset picking of pre-stack migrated offset gathers. Test results demonstrate the importance of QC of the residual moveout (RMO) implied by the picks as well as the RMO predicted by different wavelengths of tomographic updates.
Typically interpretation of top and base salt is a vital and time-consuming part of the model building. In this project we incorporated complex overhanging salt geobodies using a Marching Cubes technique (Lorensen & Cline, 1987). This proved to be more efficient than previous methods of salt body representation. Figure 2 shows an example of the salt bodies that were constructed with this method.
We used an interpretation workstation, Petrel, for both the interpretation and the velocity model building incorporating WesternGeco’s i2i software. The integration of these tools shortened the cycle time by avoiding lengthy data reformatting steps
|SALT SCENARIO TESTING|
New migration algorithms such as Rapid Beam Imaging (RBI) and Gaussian Packet Migration (GPM) provide fast scenario testing of salt geometry. This involves either zero offset demigration of one salt geometry followed by remigration to a new zero offset image with revised salt geometry or a decomposition of pre-stack data followed by fast remigrations with different salt models. In some cases we found even a simple depth conversion very useful for a quick QC of a salt model. Figure 3 shows an example of vertically stretching a sediment flood migration with two different salt models. The simpler base salt structure with the interim salt model on the right, suggests this is an improvement.
|DEPTH MIGRATION ALGORITHMS|
We observe higher frequencies in the Angolan pre-salt than in the Gulf of Mexico. For example, we required 40 Hz imaging compared to perhaps 25 Hz in the GOM. Some migration algorithms such as WEM (wavefield extrapolation migration) and RTM (reverse time migration) become exponentially more expensive with higher frequencies and the presence of TTI anisotropy adds further cost. Full Gaussian Beam migration such as ABM (adaptive beam migration) is also frequency dependent but can be tuned to provide pre-stack offset gathers in an efficient manner. This case study shows the use of ABM for both top and base salt interpretation, making use of both its steep dip and multi-pathing capabilities. Figure 4 demonstrates why depth migration is essential. It shows a comparison of a PSTM result with an ABM result using a salt flood model. There is already a significant improvement in the pre-salt even though the model is not optimum yet.
We also compare and contrast data examples from four different migration algorithms -TTI Kirchhoff, ABM, WEM and RTM – to show both their respective advantages and their drawbacks. All still have their place but we believe that the dip limitation of WEM is proving to be a drawback for imaging in certain scenarios.Some initial tests were performed on 2D regional lines to show the potential benefits of each alogorithm. In Figure 5 we show a comparison of a Kirchhoff migration against RTM. Both migrations were run with the same velocity model which was also built with TTI anisotropy. The salt flanks are much more clearly imaged with RTM and there is much less migration swing.
Figure 5 Comparison of Kirchhoff migration (left) with RTM on a regional 2D line
Returning to the 3D data, Figure 6 shows a comparison of Kirchhoff and ABM from the final migration volume. In this case the ABM result consists of only half the fold of the Kirchhoff but it still shows a good uplift in both salt flanks and pre-salt imaging.
This case study highlights the improvements available today in salt and subsalt imaging even with conventional narrow azimuth data. The key to this is building more detailed anisotropic velocity models and using a wider choice of appropriate depth migration algorithms.
We acknowledge Sanangol E.P. and WesternGeco EAF Multiclient for the use of the data and for all examples shown. We thank Gill Brown for technical advice, Mike Hughes for 2D processing, Clara Rodriguez and Clara Abu for interpretation and other WesternGeco staff for assistance with different aspects of the workflow.
Boudou, F., Y. Le Stunff,J. Arnaud,P. Esquier,A. Kenworthy,P. Whitfieldand C. Soufleris, 2008, Benefits of tilted transverse isotropy prestack depth migration for reservoir evaluation offshore West Africa:?First Break, 26, 37-41.
Lorensen, W., H. Cline, 1987, Marching cubes: A high resolution 3D surface construction algorithm: ACM SIGGRAPH Computer Graphics, 21, 163-16