Analysis of seismic anisotropy in 3D multi-component seismic data

摘要

The importance of seismic anisotropy has been recognized by the oil industry since itsudfirst observation in hydrocarbon reservoirs in 1986, and the application of seismicudanisotropy to solve geophysical problems has been keenly pursued since then. However,uda lot of problems remain, which have limited the applications of the technology.udNowadays, more and more 3D multi-component seismic data with wide-azimuth areudbecoming available. These have provided more opportunities for the study of seismicudanisotropy. My thesis has focused on the study of using seismic anisotropy in 3Dudmulti-component seismic data to characterize subsurface fractures, improve convertedudwave imaging and detect fluid content in fractured reservoirs, all of which are importantudfor fractured reservoir exploration and monitoring.udFor the use of seismic anisotropy to characterize subsurface fracture systems, equivalentudmedium theories have established the link between seismic anisotropy and fractureudproperties. The numerical modelling in the thesis reveals that the amplitudes and intervaludtravel-time of the radial component of PS converted waves can be used to derive fractureudproperties through elliptical fitting similar to P-waves. However, sufficient offsetudcoverage is required for either the P- or PS-wave to reveal the features of ellipticaludvariation with azimuth.udCompared with numerical modelling, seismic physical modelling provides additionaludinsights into the azimuthal variation of P and PS-wave attributes and their links withudfracture properties. Analysis of the seismic physical model data in the thesis shows thatudthe ratio of the offset to the depth of a target layer (offset-depth ratio), is a key parameterudcontrolling the choice of suitable attributes and methods for fracture analysis. Data withuda small offset-depth ratio from 0.7 to 1.0 may be more suitable for amplitude analysis;udwhilst the use of travel time or velocity analysis requires a large offset-depth ratio aboveud1.0, which can help in reducing the effect of the acquisition footprint and structuraludimprint on the results. Multi-component seismic data is often heavily contaminated with noise, which will limitudits application potential in seismic anisotropy analysis. A new method to reduce noise inud3D multi-component seismic data has been developed and has proved to be very helpfuludin improving data quality. The method can automatically recognize and eliminate strongudnoise in 3D converted wave seismic data with little interference to useful reflectionudsignals.udComponent rotation is normally a routine procedure in 3D multi-component seismicudanalysis. However, this study shows that incorrect rotations may occur for certainudacquisition geometry and can lead to errors in shear-wave splitting analysis. A qualityudcontrol method has been developed to ensure this procedure is correctly carried out.udThe presence of seismic anisotropy can affect the quality of seismic imaging, but theudstudy has shown that the magnitude of the effects depends on the data type and targetuddepth. The effects of VTI anisotropy (transverse isotropy with a vertical symmetry axis)udon P-wave images are much weaker than those on PS-wave images. Anisotropic effectsuddecrease with depth for the P- and PS-waves. The real data example shows that theudoverall image quality of PS-waves processed by pre-stack time migration has beenudimproved when VTI anisotropy has been taken into account. The improvements areudmainly in the upper part of the section.udMonitoring fluid distribution is an important task in producing reservoirs. A syntheticudstudy based on a multi-scale rock-physics model shows that it is possible to use seismicudanisotropy to derive viscosity information in a HTI medium (transverse isotropy with audhorizontal symmetry axis). The numerical modelling demonstrates the effects of fluidudviscosity on medium elastic properties and seismic reflectivity, as well as the possibilityudof using them to discriminate between oil and water saturation. Analysis of real dataudreveals that it is hard to use the P-wave to discriminate oil-water saturation. However,udcharacteristic shear-wave splitting behaviour due to pore pressure changes demonstratesudthe potential for discriminating between oil and water saturation in fractured reservoirs.