| dc.description.abstract | The theoretical framework, including a short introduction to the Valhall oilfield, is provided in Chapter 2. To allow for readers with backgrounds other than geophysics, basic terms like anisotropy and the types of body waves are explained in 2.2 and 2.3. Also included in this section is the polarization and splitting of shear-waves and its link to transverse isotropy. Large portions of the thesis rely on understanding the effects of the Thomsen parameters, so special care is taken to properly define them and their relationship with Hooke’s law and the phase velocities of P- and S-waves in 2.4. Definitions of terms involved in AVO and AVOA analysis, such as AVO gradient as well as examples of use and underlying assumptions of the concepts are provided in 2.5 and 2.6.
The Zoeppritz equations describe exactly how seismic waves are reflected at a media boundary as a function of incidence angle, but their complexity obscure insight into the problem at hand. For that reason, Andreas Rüger’s (1997) approximation for the P-wave reflectivity in horizontally isotropic media is introduced in 2.6.1. Two approaches for inverting this equation for the symmetry axis orientation are outlined in the same section. The AVOA signature can be obscured by several error factors which vary with azimuth and/or wavefront normal angle. These include ghost effects, attenuation, source array directivity and heterogeneous/anisotropic overburdens. Brief explanations of these phenomena are given in 2.6.2.
In 2.7, the theory behind some of the processing applied to the seismic data in this thesis is explained – Kirchhoff migration and noise removal through f-x deconvolution and f-k filtering. Finishing off chapter two is a brief description of the finite element model ISAMGEO and the modeling software mech2seis. This software package, with the ISAMGEO-model as input, was used to model AVOA-effects to be compared with observed data.
Chapter 3 contains the complete description of the practical work in this thesis. The first part summarizes the seismic data preprocessing and the AVOA attribute calculation and extraction along the horizons THC and Top Balder, figures of with common offset/midpoint gathers from every processing step and AVO/AVA plots. The second part of Chapter 3 describes the comparing of the extracted AVOA attributes along the horizons THC and Top Balder with:
1. Direction of maximum modeled stress. Theoretically, the direction of maximum stress is approximately equal to the isotropy plane orientation.
2. Unidirectionality of modeled stress. Large maximum stress compared to minimum stress could create small fractures and microcracks aligned in the direction of maximum stress. Also, the alignment of open fracturing can be influenced by the stress field if the maximum stress is significantly larger than the minimum stress – Cracks oriented orthogonal to the maximum stress direction may close, while cracks oriented parallel to the maximum stress direction may open. This will cause azimuthal anisotropy in the form of birefringence.
3. Modeled maximum stress magnitude difference across the interfaces. A certain difference in stress magnitude across the interface may be needed to induce measurable azimuthal anisotropy.
4. Modeled AVOA-effects. Synthetic AVO gradients were computed from modeled stress, density and velocity in an effort to match the observed AVO gradients.
These comparisons confirmed a relationship between modeled stress and AVOA-effects, particularly at reservoir level, and the modeling of the AVO gradient indicated that the azimuthal anisotropy on Valhall is mainly caused by differences in the shear wave splitting parameter difference, Δγ, which points to that the azimuthal anisotropy could be caused by near-vertical fractures.
Chapter 4 features a general discussion of the methodology and conclusions from the three comparisons between modeled and observed data, as well as a discussion of the relevance for using AVOA as an indicator of reservoir depletion based on data from a single survey. Input parameters for all figures, acquisition parameters, and PGS preprocessing are summarized in three tables in Appendix A, while Appendix B contains all MATLAB functions written for computing the AVO gradients from mech2seis output. | nb_NO |
