| dc.description.abstract | Elastic response of rocks often depends on the rock’s orientation, i.e., most rocks are anisotropic. In particular, shales, which are the overburden rock in most of conventional oil and gas fields, often show strong elastic anisotropy due to alignment and platy nature of its constituent mineral. Various anisotropic rock physics models have been proposed to predict the elastic anisotropy of shales, however, practical applications are limited and most of rock physics models in the past do not follow the observed internal structure of shales and bound water properties suggested by existing studies.
Elastic anisotropy affects seismic data and geomechanical response in various ways. Amplitude-variation-with-offset (AVO) response, stress concentration around the borehole, stress change around a compacting reservoir and associated induced pore-pressure, for example, are influenced by the elastic anisotropy. The elastic anisotropy is, however, often ignored in seismic reservoir characterization and geomechanical studies because it is difficult to measure enough parameters in fields to characterize anisotropy parameters or to estimate them from limited data. Correspondingly, there are only limited discussions in literature about practical application of anisotropic methods in these tasks. (1) Being able to correctly estimate/model anisotropy parameters in shales through appropriate anisotropic rock physics model and (2) having practical methods/workflows to account for the elastic anisotropy in seismic reservoir characterization and geomechanical studies, can therefore significantly improve subsurface evaluation. These are the focus of this dissertation.
A predictive rock physics model is developed based on the recently published Sayers and den Boer approach. This model is consistent with the observed internal structure of shales and allows the finite shear stiffness of bound water to be taken into account. A comparison with existing measurements indicates that the model can be used to make realistic estimates of anisotropy parameters based on limited data.
Practical methods/workflows were also proposed for (1) wellbore stability analysis for anisotropic shales, (2) AVO inversion of an anisotropic data set, (3) AVO projection of an anisotropic data set, and (4) the interpretation of 4D seismic data with time-lapse changes in anisotropy. Each method/workflow can be summarized as follows:
(1) Proposed wellbore stability analysis workflow for anisotropic shales is based on the Amadei solution and anisotropic Skempton’s B parameters. The impact of elastic anisotropy on stress concentration around the borehole and associated induced pore pressure can be accounted for by this workflow.
(2) AVO inversion of an anisotropic data set was performed using the existing concept of pseudoisotropic elastic properties. The pseudoisotropic elastic properties are pseudo properties which give anisotropic reflectivities when used in the isotropic modelling. The industry standard isotropic AVO inversion can therefore be followed by using the pseudoisotropic elastic properties as an input. Proposed workflow includes optimization of anisotropy parameters for pseudoisotropic elastic properties generation through well-to-seismic ties and wavelet extractions.
(3) Proposed workflow for AVO projection of an anisotropic data set is based on the recently introduced three-term (3T) AVO projection and associated 3T extended elastic impedance (EEI). The impact of anisotropy on 3T EEI was taken into account. To obtain inverted 3T EEI whose time profile reasonably represents lithology fraction, the projection angle was optimized through a correlation scan using lithology fraction as a target.
(4) The impact of time-lapse anisotropy on 4D seismic response was assessed through a dynamic rock physics template based on the pseudoisotropic elastic properties. Applications to field data demonstrate that the elastic anisotropy must be taken into account. Main results, in terms of the anisotropy impact, from each method/workflow can be summarized as follows.
(1) Failure regions, modes, and the safe mud weight window given by the workflow were found to be completely different from those given by the industry standard approach where the classical isotropic Kirsch solution is used with the assumption of no induced pore pressure.
(2) AVO inversion using pseudoisotropic elastic properties gave results consistent with input data (i.e. pseudoisotropic elastic properties logs), while the conventional AVO inversion did not give satisfactory results; in particular, density given by the conventional AVO inversion was totally different from the input data (i.e. acquired density log). This is because the relative difference between various lithologies in acquired density is totally different from that in pseudoisotropic density for a given case.
(3) AVO Curvature © was found to be an important component of AVO projection, even if it appears unusable. This is because the noise in AVO attributes is correlated and can be suppressed by choosing certain projection angles. The impact of anisotropy must be accounted for since the relative difference between various lithologies in anisotropic curvature impedance (CI) can be totally different from that in the isotropic CI.
(4) 4D seismic response is significantly affected by the time-lapse anisotropy. Pore pressure depletion causes time-lapse AVO response as if there is a reduction in VP/VS although the vertical VP/VS increases. This is caused by a decrease in anisotropy parameters, which is an expected change under uniaxial compaction
Since anisotropic shales are the overburden rock in most of conventional oil and gas fields and time-lapse changes in anisotropy is expected in most of depletion scenarios, those results are relevant to other fields. Proposed methods/workflows can be immediately applied in many practical situations. | en_US |
