Joachim Mispel

Improved applicability of ray tracing in seismic acquisition, imaging, and interpretation

Ray-based seismic modeling methods can be applied at various stages of the exploration and production process. The standard ray method has several advantages, e.g., computational efficiency and the possibility of simulating propagation of elementary waves. As a high-frequency approximation, the method also has a number of limitations, particularly with respect to a lack of amplitude reliability in the presence of rapid changes of the model functions representing elastic parameters and interfaces. Given the objective of improving the applicability of the standard ray method, we present a strategy that does not require specific extension to finite frequencies. Instead, we define each ray-based process as an element of a system that, as a composite process, is able to obtain better results than the ray-based process applied alone. Other elements of the composite process can be finitedifference modeling or numerical solutions for surface and volume integrals, which are basic constituents of Kirchhoff modeling and imaging. We also include among the process elements recently developed techniques for simulating the migration amplitude on a target reflector and in a local volume, e.g., a reservoir zone. The model is decomposed according to its complexity into volume elements, surface elements, or a combination. The composite process consists of a specified interaction between process elements and model elements, which fits well with the philosophy of modern software design. Model elements that will be exposed to ray-tracing algorithms may need appropriate preparation, e.g., smoothing and resampling. We demonstrate specifically, in a tutorial example, that simulating the seismic response from a reflector by ray-based composite processes can yield better results than applying standard ray tracing alone.

3D full-waveform inversion at Mariner - a shallow North Sea reservoir

We applied isotropic acoustic 3D full-waveform inversion (FWI) to OBC data from the Mariner field, a shallow heavy-oil field in the North Sea. This resulted in a multiphase workflow that can be adapted to imaging challenges in similar geological settings. FWI improves the resolution of the velocity field compared to the benchmark velocity field from reflection tomography. The background trend of and details introduced in the velocity model correlate well with the geology from seismic data and with well logs down to reservoir level. Resulting depth images show significantly better well tie in the overburden and improved definition of sand bodies at reservoir level.

Towards better amplitude maps by simulated migration

Illumination maps are a useful tool for survey planning and for QC of amplitudes picked on selected target horizons after depth migration. The illumination amplitude maps, which are used most often, are based on a simple summation of ray-theoretical amplitudes within bin cells of the target horizon. Although better than just counting the hits, this approach is still to simple to give amplitudes directly comparable to PSDM, because it neither takes the seismic pulse nor the Fresnel zone into account. Here we improve an approach based on Kirchhoff migration of ray traced data. We use a local paraxial approximation of the true traveltime field on the target horizon around the reflection point to generate amplitude maps where both the source pulse and the Fresnel zone are taken into account. The ray tracing is fully elastic and includes 3D geometrical spreading and reflection/transmission effects. For the synthetic examples, the results are almost equal to results from prestack depth migration (PSDM).

3D Multiple Attenuation and Depth Imaging of Ocean Bottom Seismic Data

In 3D Ocean Bottom seismic surveys (3D-OBS) both pressure and vertical particle velocity is recorded. This presents the opportunity to decompose the recorded wavefields in up- and down-going components and apply to 3D common receivers gathers a designature/demultiple algorithm to attenuate free surface multiples. This single processing step replaces several steps in conventional processing usually encompassing τ-p predictive deconvolution and Radon demultiple. Using a 3D-OBS data set from the Gullfaks Sor field in the North Sea, the new 3D designature/demultiple approach, together with 3D common receiver depth migration, is shown to result in seismic images of better quality than by the traditional processing sequence.

Full Azimuth Seismic Modeling in the Norwegian Sea

SUMMARY One of the fields in the Norwegian Sea has been imaged several times over the past decades, both with conventiona l narrow azimuth seismic surveys as well as with ocean bottom seismic. The extensively faulted structure of the field and the possible presence of a salt diapir cause imaging problems in some areas. Therefore, a simulation study has been initiated to judge whether or not a marine full azimuth acquisition geometry improves the image of the subsurface. For this simulation study, simplified velocity and densit y models of the field were created, containing the main features characterizing it, as well as the problem areas. The simulation was done with 3D finite difference (FD) modeling. Data sets with and without free surface multiples were generated, and imaging from a full azimuth acquisition geometry was compared with imaging from a conventional narrow azimuth geometry. FD modeling shows that the full azimuth design generally leads to a better suppression of noise in the data, mainly due to increased fold. Depth sl ices show that fault edges are imaged sharper in a full azimuth geometry. Also, the image of and below the salt/limestone structure is improved. However, attenuation of multiple energy due to increased cross line fold is less than expected, except for the first seabed multiple. The low maximum frequency used in FDmodeling may have limited the increase in image quality with the full azimuth modeling.

3D Multiple Attenuation And Depth Imaging of Ocean Bottom Seismic Data

In 3D Ocean Bottom seismic surveys (3D-OBS) both pressure and vertical particle velocity is recorded. This presents the opportunity to decompose the recorded wavefields in upand downgoing components and apply to 3D common receivers gathers a designature/demultiple algorithm to attenuate free surface multiples. This single processing step replaces several steps in conventional processing usually encompassing tau-p predictive deconvolution and Radon demultiple. Using a 3D-OBS data set from the Gullfaks Sør field in the North Sea, the new 3D designature/demultiple approach, together with 3D common receiver depth migration, is shown to result in seismic images of better quality than by the traditional processing sequence.