Advances and Challenges in Seismic Data Processing and Imaging for Geologically Complex Areas
Rubén D. Martínez
Petroleum GeoServices (PGS), 15150 Memorial Dr., Houston, Texas 77079
The success of seismic imaging in depth greatly depends on the right choice of the data acquisition and processing technologies based on the good understanding of the geologic problem at hand. These choices are not trivial in geologically complex areas.
Over the last years, we have witnessed a significant evolution of 3D data acquisition designs, for example, the evolution from narrow azimuth (NAZ) in the 1990s to wide azimuth (WAZ) in the 2000s (Long, 2006). Presently, it is widely accepted that the azimuth component provided by WAZ geometries increase seismic resolution and illuminate better the subsurface in the presence of high geological complexity. However, these wide azimuth geometries pose new challenges on how the seismic data should be processed. In data processing, processes well accepted for years, e.g., binning, source designature, debubble, deghosting, demultiple (for surface, ‘peg legs,’ and interbed multiples) (Aaron et al., 2011), are now a new challenge. Very sophisticated algorithms are required to address these effects in the presence of azimuth variations. If the azimuth and offset dependency is ignored at the pre-processing steps, depth velocity model building and depth migration algorithms could not possibly deliver useful seismic depth images for structural interpretation and prospect generation.
Seismic imaging in depth has also evolved. Velocity model building using true azimuth 3D tomography is used to provide not only accurate velocity models for migration but also be able to account for velocity anisotropy, in this case, TTI (tilted transverse isotropy). The estimation of anisotropic parameters is still a challenge and it is the subject of significant research efforts. To handle large volumes of data during the model building stage, a very reliable and fast imaging algorithm is required to iterate with tomography very efficiently and accurately. Beam PSDM (prestack depth migration) is one of the most efficient and accurate algorithms for model building (Sherwood et al., 2008). When it is implemented properly in a combined computer environment, anisotropic depth models can be delivered in a short period of time. This is advantageous as it allows the interpreter more time to make decisions about the geologic model. After several iterations, the final seismic imaging stage can be achieved using hi-end imaging algorithms such as anisotropic prestack one-way wave equation migration or reverse time migration (Crawley, 2010).
As high performance computing evolves, methods such as full waveform inversion (FWI) become more viable to produce high resolution velocity models. 3D FWI is available and current efforts are in the expansion of these methods to account for velocity anisotropy.
AAPG Search and Discovery Article #90158©2012 GCAGS and GC-SEPM 6nd Annual Convention, Austin, Texas, 21-24 October 2012