Font Size: a A A

Elastic Wave Pre-stack Reverse Time Depth Migration

Posted on:2014-11-10Degree:MasterType:Thesis
Country:ChinaCandidate:P GuoFull Text:PDF
GTID:2250330401483793Subject:Earth Exploration and Information Technology
Abstract/Summary:PDF Full Text Request
Elastic wave pre-stack depth reverse time migration can utilize compressional and shearwave simultaneously, which plays an important role in fractured reservoir, coalbed methane andshale gas exploration. There are two goals for elastic wave imaging, namely, to producecompressional and shear wave migration results of underearth structure and to provideamplitude-reserved gathers for pre-stack inversion.The multi-component wave-field extrapolated by elastic wave equation includescompressional and shear wave information. To produce compressional and shear wave migrationresults, first an wave-mode separation needs to be done. The traditional wave-separation methodsbased on divergence and curl perform well in isotropic media, while it is not the case inanisotropic media, In this paper, through solving the Christoffel equation, we can get thepolarization vector changing with wave number in wave-number domain, then an inverse fouriertransform is exerted to get a wave-field separation operator in space domain. This wave-fieldseparation operator varied with the elastic parameters in space domain and can be used toimplement the wave-separation in anisotropic media. To test the efficiency of the algorithm, atest is implemented by a homogeneous and layered VTI anisotropic model. The result show thatthe wave-separation methods proposed here can achieve the complete wave separation in VTImedia effectively.Based on the above research, qP wave and qS wave decomposition in transversely isotropicmedia with a vertical symmetry axis is implemented in space domain in this paper. Thedecomposition operators in wave-number domain are constructed using the polarization vectorsevaluated by solving the Christoffel equation, then IFFT is applied and decomposition operatorsin space domain are derived. Wave-field decomposition with multi-component characteristicscan be obtained by convolution operation between spatial operator varied with spatial locationsand multi-component wave-field. Since the algorithm developed in this paper utilizescorresponding decomposition operators in each grid point, it can adjust to the condition thatanisotropic parameter varies spatially. Elastic wave-field of homogeneous VTI media, stratifiedVTI media and anisotropic version of sigsbee2A model have proven the solidness of thisalgorithm.In this paper space domain algorithm of elastic wave-field qP and qS wave separation inVTI media is utilized to implement the wave-mode separation in the reconstruction wave-fieldproduced by reverse-time wave-field extrapolation. Then source-normalized cross-correlationimaging condition is utilized to get the qP and qS imaging results in VTI media. Compared withwave-mode separation based on divergence and curl, migration noise significantly reduced. Interbed reflection plays a key role in the imaging result of reverse-time migration. Thispaper devotes to suppressing the interbed reflection by improvement on the imaging condition.Poynting vector in the electromagnetism is introduced into acoustic area and hence the poyntingvector expression in acoustic equation. On the basis of classic cross-correlation imagingcondition, cross-correlation imaging condition based on poynting vector is concluded, interbedreflection is thus recognized according to the propagation direction of seismic wave energy andattenuated in the process of wave field cross-correlation. To verify the effectiveness of thealgorithm, calculation based on different models applying classic cross-correlation imagingcondition and the algorithm proposed has been done. The comparison of the two kinds ofmigration results shows that cross-correlation imaging condition based on poynting vector cansuppress the interbed reflection effectively and thus a more accurate migration imaging result canbe acquired.Reverse time migration takes off recently, however, there are still some bottlenecks whichhinders the further development of reverse time migration, namely, the massive computation, themassive RAM and hard-disk demand, low-frequency noise introduced by interbed reflection andamplitude affected by transmission effects. The methods for reverse time migration accelerationand elimination of massive hard-disk demand were investigated in this paper. GPU programmingwas used to improve the efficiency of computation and random boundary conditions were usedto eliminate the hard-disk demand.
Keywords/Search Tags:Elastic wave, VTI media, reverse-time migration, wave-mode separation, wave-field decomposition, GPU programming
PDF Full Text Request
Related items