Poststack Seismic Inversion Constrained With Geological Features:Methodological Research And Applications

Seismic inversion is a key that helps to obtain subsurface lithological information and a significant method employed to improve seismic resolution.The ulitization of geologic information in current seismic inversion methods is in the form of the construction of initial models,and seldomly involves the geological information into the inversion system to constrain the whole process of inversion,and the construction of the commonly used regularizion term as the constraint is based on mathematical considerations which lacks diverse geological expressions,and the result may makes offences to the geological background.The real geologic bodies are spatially related in geometric shapes and property parameters.Based on this idea,the geological feature information that can be used to characterize stratal and structural features is incorporated to construct geological structural constraints spatially,and the inversion is jointly driven by geological structural feature and seismic signals.Extraction of accurate geological feature information is a prerequisite for the seismic inversion methods in this thesis,and the geological feature information currently used are mainly the strata dips and coherence attributes.When seismic data are of a low signal-tonoise ratio(S/N),large analysis window should be used in coherence computation which leads to decreased computational efficiency and blurred coherence images.To address this issue,an efficient and robust eigenstructure-based coherence method is proposed.In the method,the seismic traces are assigned different weights according to their relative distances from center trace using the inverse distance weighting method,and the weighted seismic traces are stacked on each direction toward the center trace,and then the coherence value is calculated.In this way,the coherence attribute is mainly determined by the center trace;thus,the faulted points are more accurately positioned;meanwhile,the computational efficiency is increased and the robustness to noise is improved.To address the issue of inaccurate delineation of line-shaped discontinuities with varied directions such as arcuate faults and channels,a geosteering coherence method based on cross-correlations in orthogonal directions is proposed to compute coherence attributes with consideration of structural trends,and the structures are more clearly and accurately delineated.However,this method is inferior to the efficient and robust eigenstructure-based coherence method in robustness to noise,and it’s recommended to choose which method is to be employed in industrial applications according to the S/N of seismic data to extract more stable and accurate geological feature information.For seismic inversion in areas where the strata slopes are variable and steep,a multichannel reflectivity inversion method based on dynamic time warping(DTW)is proposed by exploiting the spatial relationships among adjacent seismic traces using an improved DTW algorithm and constructing a lateral constraint operator with the spatial relationship and difference operator which renders the inverted results high S/N and better lateral continuity.However,this method may smooth some small-scale structure-related discontinuities since structures are not taken into consideration.Finally,the seismic inversion method constrained with geological features is proposed with consideration of both strata and structure,and the method is applicable to reflectivity inversion and impedance inversion,and can be used for structure-oriented smoothing.The reflectivity inversion tests show that the result obtained via the proposed method has higher S/N,better lateral continuity and relatively higher resolution,and the structural edges are well preserved,and the result is more consistent with the geological background;The acoustic inversion tests show that the result obtained via the proposed method is similar to that obtained via the conventional method when the S/N of seismic data is high,and the advantage of the proposed method is obvious when the S/N is low,which indicates that that although the improvement is subtle for acoustic impedance inversion of post-stack data that generally have high S/N,prestack seismic inversion method constrained with geological features may has a promising perspective for the inversion of prestack data with relatively low S/N.