Study On Seismic Diffraction Separation And High-precision Imaging Method

With the exploration and production of coal and petrolem,the geologic condition of seismic exploration changes gradually from relatively simple structures to complex features.Precise carefully data processing and interpretation is one of the goals of seismic exploration.The theory and methods of traditional seismic data processing are designed for reflection data.However,due to the limitation of Rayleigh rule,the reflection imaging results are inappropriate for exploration and identification of small-scale geologic objects in the subsurface.Small-scale discontinue inhomogeneous structures,such as,faults,fractures and pinch-outs,are widely distributed in the subsurface.These objects have a close relationship with the safety of coal mining and the production of oil and gas.The accurate exploration and location of these objects have a significant effect on the safety of coal mining and the delineation of oil-gas rich area.Diffractions are the response of the small-scale geologic objects with respect to source wavefield.Diffractions carry the key geologic information about the small-scale features,and obey Huygens' Principle,which makes diffractions have superior illumination.Diffraction imaging can overcome the limitation of Rayleigh rule,and thereby diffractions posses the potential of high-resolution imaging.Therefore,diffractions can be used to the high-resolution imaging of small-scale discontinue features in the subsurface.However,the energy of diffractions is much weaker than that of reflections,and diffraction attenuation is rapid with the increase of traveltime.In addition,because of the different propagation rule,traditional reflection methods are inappropriate for diffractions,such as,stack processing and velocity analysis,which makes diffraction implementation difficult.This manuscript aims at the problem of small-scale geologic structures in the subsurface.We use diffractions to do some works composed of diffraction separation,velocity analysis and diffraction imaging.The manuscript proposes the improved plane-wave destruction(PWD)method using the regularization constrain based on the tradional PWD method.The improved method can accurately estimate the local reflection slope,and help to obtain the high quality diffractions.The PWD method is based on the plane-wave assumption,and uses the local reflection slope to predict and suppress reflections,and then separate diffractions.However,the estimation of the local slope is easy to be affected by random noise,which makes the local slope inaccurate.Moreover,when small-scale objects exist,the plane-wave assumption is no longer met.Thus,the estimation of the local slope is unstable.The improved PWD method uses the sparse discrepancy between the local reflection slope and the random noise,and considers the sparse property of the local slope as prior information,which makes the esitimation more accurate.In the meantime,the new objective function adds a regularization term with the local slope to improve the stability and reduce multiple solutions.The new method can effectively estimate the local reflection slope and obtain the high quality diffractions.In fact,most of the current diffraction-separation methods only ultize the kinematic difference between reflections and diffractions,and ignore the dynamic property.These methods are not beneficial for diffraction imaging.This manscript presents a new rank-reduction diffraction separation method based on multichannel singular-spectrum analysis(MSSA).This mehod considers both differences between reflections and diffractions in kinematic and dynamic properties.On one hand,in common-offset or poststack domain,reflections and diffractions have different linear properties;on the other hand,the energy of diffractions is generally much weaker than that of reflections.The differences in linear property and energy are used to predict the reflections with strong energy and separate the diffractions with weak energy.The MSSA algorithm can express the Hankel matrix of seismic data with singular values and singular vectors by singular-value decomposition(SVD)technology.The linear signals with strong energy correspond to the large singular values;on the contrary,the nonlinear signals with weak energy correspond to the small singular values.In accordance with the properties of reflections and diffractions,we can choose the singular values and vectors corresponding to diffractions to construct the original diffractions by rank estimation on the curvature of the smoothed singular-value curve.The rank-reduction diffraction separation method based on the MSSA uses the dynamic property and can preserve the amplitude and phase of diffractions,which helps to diffraction velocity analysis and imaging.Diffraction velocity analysis is directly associated with the high-resolution imaging of small-scale geologic structures.This manuscript proposes diffraction velocity-analysis method based on the combination of the migrated dip-angle gather and adaptive minimum variance beamforming(AMVB).In theory,diffraction events are much sensitivity to the migration velocity.When the migration velocity is a little large or small,the diffraction events behave as “frown” and “smile” shape,respectively.Only if the migration velocity is accurate absolutely,diffraction events behave as horizontal lines.Thus,the sensitivity of diffraction events in the dip-angle domain can be used for diffraction velocity analysis.However,diffraction generally behaves weak energy,and is easy to be affected by random noise.The flattened diffraction events cannot be well focused in velocity spectrum.The AMVB algorithm incorporates the minimum variance,coherence factor,and correlation properties to well stack the diffraction events in the dip-angle domain.The new method can pick up the accurate migration velocity for diffractions.To achieve the imaging of small-scale discontinue inhomogeneous,we combine diffraction separation and velocity analysis,and propose high-precision diffraction imaging scheme with the high-precision traveltime calculation and the approximate expression of the traveltime correction equation for diffractions.The wavefiled examples illustrate the performances of reflections and diffractions in the migrated dip-angle domain and the imaging domain,respectively.The applications verify the science and feasibility of the proposed scheme.Reflections and diffractions have the similar behaviors in both domains,and the only difference is the diffractions show different open directions in both domains,respectively.This property helps to link diffraction events and diffraction imaging.The traditional correction equation depends on several parameters and it's difficult to apply the scheme for real data.Thus,we assum every position exists small-scale diffraction points in the subsurface,and the original equation can be approximated.The approximate equation is reasonable and practical,and helps to the implementation and extension of the new scheme.Several applications of synthetic and real data demonstrate the good performance of the high-resolution diffraction imaging scheme in accurately locating and identifying small-scale geologic objects,and provide more detailed geologic information for seismic interpretation.