This shows you the differences between two versions of the page.

sep:research:theses:sep157 [2014/11/26 05:21]
mandyman created
sep:research:theses:sep157 [2015/05/27 02:06] (current)
Line 17: Line 17:
 +This dissertation presents a novel technique for using surface-related multiples to improve imaging in geologically complex areas. It overcomes a challenge of the migration-based approach where crosstalk artifacts appear in the image. In the case of reverse-time migration, these imaging artifacts are the result of cross-correlation between wrong pairs of incident and scattered wavefields. Joint least-squares reverse-time migration, also known as linearized inversion, can coherently focus the reflection energy of primary and surface-related multiples into one image. By posing the imaging problem as an inversion problem, spurious reflectors or noises in the image can be attenuated. In geologically complex areas that contain salt structures, the proposed method not only improves the imaging but also added additional angular coverage in poorly illuminated areas. With respect to improving subsurface illumination, it is particularly advantageous to apply this method to ocean-bottom node acquisition.  
-Wave-equation migration velocity analysis (WEMVA) is a powerful technique for robust velocity model building when the subsurface is complex and the starting model is far from true. However, this traditional isotropic WEMVA technique cannot explain anisotropic wave phenomenon, which has significant effects when the reflectors are steeply dipping and/or the waves are traveling at large angles. Furthermore, errors in +A modified modeling operator in the inversion process was introduced to model the surface-related multiples in the data. This modified modeling operator uses the data as an areal source, which removes the need to estimate the source wavelet. By using the data as a source, only the last down-going and up-going legs of the wavepath have dependency on the migration velocity model. As a result, this operator has the same sensitivity to the migration velocity model as the conventional primary modeling operator. From an imaging prospective, the robustness of the technique is improved when there are deviations between the migration velocity and the true velocity. The surface-related multiple operator can be combined with the conventional primary operator to form a joint operator that properly accounts for the physics of both primary and surface-related multiple reflections. 
-the anisotropic parameters cause similar defocusing as errors in isotropic velocity. Any defocusing caused by anisotropy may be translated to false updates in the isotropic velocity, which lead to further mispositioning and misinterpretations. However, simple extension of the isotropic WEMVA to the anisotropic medium cannot provide an unique and reliable solution due to the nonlinear and underdetermined nature of the anisotropic model building problem. Many anisotropic models with vastly different geological interpretations may explain the surface seismic data equally well. +
-This thesis addresses these issues by including anisotropic effects in WEMVA and by integrating other useful information from geology and rock physics to better regularize the inversion. First, I extend the isotropic WEMVA method to the anisotropic +Several methods are introduced to improve the convergence of least-squares reverse-time migration, particularly for areas with a complex salt structure. To emphasize the shadow zones in the image, a target-oriented data-reweighting scheme is incorporated in the inversion process. To extract the most information from the least-squares reverse-time migration algorithm, salt dimming data-weighting was introduced to down-weight the reflection energy coming from strong velocity contrasts in the migration velocity model. Such energy often dominates the inversion. When least-squares migration is extended to the angle domain, prestack extended-angle domain filtering can be incorporated into the modeling operator to remove unwanted noise in the image.
-medium to evaluate and update the anisotropic Earth models. Instead of the industry +
-standard ray-based tomographic methods, the anisotropic WEMVA technique +
-uses wavefields as information carrier to handle both the complex subsurface and the +
-frequency-dependent behavior of the wave propagation. I include the geological information +
-using steering filters to regularize the gradients. Both synthetic and field +
-2-D examples show that the anisotropic WEMVA technique can resolve the errors in +
-both velocity and anisotropic parameters. Consequently, the migration images are +
-improved with higher resolution and correct depths for the reflectors.+
 +Applications to synthetic and field ocean-bottom node datasets show that, compared to migration or single-mode inversion, joint least-squares reverse-time migration provides the best overall subsurface image and angular coverage. In particular,  areas near and underneath a complex salt structure are better illuminated when surface-related multiples are used as signal.
-Second, I mitigate the ambiguities among the anisotropic parameters using the 
-rock physics prior information. I model shale anisotropy in the seismic scale from well 
-log measurements and the interpretation results from a previous isotropic processing 
-workflow . By varying the key parameters to the rock physics models, I include different 
-geological scenarios and parameter uncertainties in the rock physics modeling. 
-As a result, multiple realizations of the anisotropic parameters are obtained and used 
-as the rock physics constraints for the seismic data inversion. On a synthetic example, 
-I show the correct rock physics regularization accelerates the convergence of the 
-well constrained parameter and brings extra information for the poorly constrained 
-parameters. Anisotropic WEMVA inversion of a 3-D field data set yields models that 
-are consistent with seismic data, geological knowledge, as well as rock physics information. 
-The migration image based on the inverted models shows better-resolved 
-faulting discontinuities and better imaged salt flanks. 
-**Reproducibility and source codes**\\ 
-This thesis has been tested for [[sep:research:reproducible|reproducibility]]. The source codes are made available for [[http://sepwww.stanford.edu/data/media/private/docs/sep154/src.tgz|download]]. The scripts are available for [[http://sepwww.stanford.edu/data/media/private/docs/sep154/chap3.tgz|chapter 3]] and  
-[[http://sepwww.stanford.edu/data/media/private/docs/sep154/chap4.tgz|chapter 4]].\\ 
-[[http://sepwww.stanford.edu/data/media/public/docs/sep154/elita-defense.pptx|Defense presentation]]\\ 
/web/html/data/attic/sep/research/theses/sep157.1416979266.txt.gz · Last modified: 2015/05/26 22:40 (external edit)
www.chimeric.de Creative Commons License Valid CSS Driven by DokuWiki do yourself a favour and use a real browser - get firefox!! Recent changes RSS feed Valid XHTML 1.0