ABSTRACTS

Geophonearrayparameterdesignundermulti-frequencycondition.XUFeng1,LIUFulie1,WANGYu1,MINXinlin1,andXUJisong1.OilGeophysicalProspecting，2020，55(2)：235-241.

Geophone array is used to suppress random noises and regular noises and improve signal to noise ratio,especially in those prospects with low signal to noise ratio.Array parameter design is commonly performed under the assumption that seismic waves are harmonic waves at a specific frequency.Element interval,number of arrays,weighing factor,and time delay have not been taken into overall consideration; this results in significant low-pass filtering and distorted AVO effects.We discuss the relations among above 4 parameters and orientation factor at complex frequencies.To calculate the theoretical signal to noise ratio,we formulate an equation and then take it as the objective function for array parameter design through global optimization.As per model and field data tests,linear noises and random noises could be greatly reduced using the time-delay weighted arrays with above 4 optimized parameters; the behavior of low-pass filtering and AVO distortion could also be remedied.

Keywords：phase control theory，time-delay weighted array，signal to noise ratio，array parameter，directional reception

1.School of Geoscience and Technology，Southwest Petroleum University，Chengdu，Sichuan 610500，China

Secondarypositioningofdeepoceanbottomnodes.WANGZhongcheng1,ZHOUHuawei1,2,TONGSi-you1,3,FANGYunfeng4,andCAOGuobin5.OilGeophysicalProspecting,2020,55(2)：242-247.

Due to the influence of bottom topography,ocean current,et al.,OBNs may deviate from their original positions in the process of data acquisition.To guarantee imaging quality,it is necessary to perform secondary positioning.Current velocity,which is an important parameter in secondary positioning,is usually given as a constant or fitted as the function of offset.But actually,current velocity often changes with depth and almost remains unchanged within a specific range of offset.This means that above velocity definitions are both incorrect.We use model tests to investigate the accuracy of secondary positioning for different velocity definitions,and then we propose to substitute depth-varying velocity in the vertical direction with equivalent velocity,which is derived from the inversion of receiver coordinates.Model tests show that this equivalent velocity is better than constant velocity and fitted velocity in higher accuracy of secondary positioning; such improvement facilitates subsequent NMO correction and imaging.

Keywords: OBN,secondary positioning,current velocity,travel time,equivalent velocity

1.Key Lab of Submarine Geosciences and Prospecting Techniques,MOE,Ocean University of China,Qingdao,Shandong 266100,China

2.Department of Earth and Atmospheric Sciences,University of Houston,Houston,Texas 77204,USA

3.Functional Laboratory for Marine Mineral Resources Assessment and Prospecting,Qingdao National Laboratory for Marine Science and Technology,Qingdao,Shandong 266061,China

4.R & D Center,BGP，CNPC,Zhouzhou,Hebei 072751,China

5.Sinopec Geophysical Corporation,Dongying,Shandong 257100,China

Asimulationofacquisitiondesignanddatapro-cessingforoffshorecompressivesensingseismic.HUANGXiaogang1.OilGeophysicalProspecting，2020，55(2)：248-256.

Compressive sensing seismic acquisition aims at more efficient operation and cost reduction.Due to offshore equipment and technical limitations,compressive sensing random sampling may only be performed in some directions.We develop a Jitter-based system for offshore compressive sensing seismic acquisition geometry design and supporting modules for data reconstruction.We accomplish compressive sensing acquisition design for two offshore surveys and subsequent data reconstruction.The results show that data reconstruction,both for synthetic data from forward modeling and real data simulation,yields good results; whereas,acquisition cost is reduced by 1/3.This means that Jitter-based random sampling is feasible for offshore compressive sensing seismic acquisition geometry design.

Keywords： compressive sensing，Jitter random sampling，acquisition design，seismic data reconstruction

1.CNOOC Research Institute Co.，Ltd.，Beijing 100028，China

Near-surfacevelocitymodelingbasedonmicro-logandazimuth-weightedinterpolation.JINChangkun1,2，WANGYanguang3,SHANGXinmin1,CUIQinghui1,andWangDongkai1.OilGeophysicalProspecting，2020，55(2)：257-265.

To build a good surface velocity model,we present a method using micro-log tomography combined with azimuth-weighted interpolation.The velocity at each observation point is calculated using tomography of micro-log travel time,and the depth-velocity relation is normalized through depth transformation constrained using horizons.The weight coefficients at main azimuths are inverted using a radial basis function and azimuthal basis function,followed by weighted velocity interpolation.After inverse depth transformation,we obtain the final near-surface velocity model.Two case studies,i.e.shooting depth design in Miquan and detailed near-surface velocity modeling in Chuanguanzhuang,show that micro-log tomography and azimuth-weighted interpolation yield good results with higher resolution than those derived from conventional well log interpretation and Kriging interpolation.

Keywords：micro-log，tomography，velocity modeling，near surface，azimuth，interpolation

1.Geophysical Research Institute,Shengli Oilfield Branch Co.,Sinopec，Dongying，Shandong 257022，China

2.Postdoctoral Scientific Research Workstation,Shengli Oilfield Branch Co.,Sinopec,Dongying，Shandong 257002，China

3.Shengli Oilfield Branch Co.,Sinopec,Dongying，Shandong 257000，China

Hydrophonemoving-relatedmoveoutcorrectionformarinestreamerdata.MAGuangkai1,ZHOUZhengzheng1,GENGWeifeng1,QIANZhongping1,YANWei1,andRENXiaoqiao1.OilGeophysicalProspecting,2020,55(2):266-271.

Hydrophones move with the streamers in marine seismic data acquisition.Such movement causes travel time error,which has a negative impact on subsequent data processing and data matching for time-lapse seismic survey.To correct the moveout caused by hydrophone moving,we use offset,vessel speed,and stacking velocity to calculate time-varying travel time error,which is proportional to offset and vessel speed and inverse to the square of stacking velocity.Field data processing shows that after hydrophone moving-related moveout correction,residual moveout in CMP gathers after NMO correction is further reduced,and velocity spectra become more focused.This lays the data foundation for subsequent processing.

Keywords: streamer data,hydrophone moving-related moveout correction,travel time error,time varying,error analysis

1.R & D Center,BGP,CNPC,Zhuozhou,Hebei 072751,China

Seismicdatadenoisingbasedonmulti-layerperceptron.WANGQiqi1,2,3，TANGJingtian1,2,3，ZHANGLiang1,2,3，LIUXiaojia1,2,3，andXUZhimin4.OilGeophysicalProspecting，2020，55(2)：272-281.

Seismic exploration has played an important role in tectonic analysis and prospecting of hydrocarbon and other mineral resources.Due to the influence of environment and instruments,seismic data contain random noises,which have a negative impact on processing and interpretation.We propose a multi-layer perceptron (MLP) method to reduce random noises.Seismic data are sampled using a moving window and then converted into a 1D vector,which is utilized as training samples to establish a multi-layer neural network model.The weighting factor of neurons in each layer is calculated using the back propagation algorithm until the mean square training error reaches a minimum.Synthetic or measured noisy seismic data are imported into this established model,and the output is calculated using the weighting factors after training.We compare the denoising results derived from MLP and curvelet methods and conclude that MLP result exhibits higher signal to noise ratio and better signal preservation,especially for structural details.

Keywords：random noise，multi-layer perceptron，denoising，back propagation，curvelet-based denoising

1.School of Geosciences and Info-Physics，Central South University，Changsha，Hunan 410083，China

2.Key Laboratory of Metallogenic Prediction of Nonferrous Metals and Geological Environment Monitoring (Central South University)，Ministry of Education，Changsha，Hunan 410083，China

3.Hunan Key Laboratory of Nonferrous Resources and Geological Hazard Detection，Changsha，Hunan 410083，China

4.Chengde Petroleum College，Chengde，Hebei 067000，China

RandomnoisesuppressionusingadaptivethresholdinShearletdomain.XUELin1,2,CHENGHao1,2,GONGEnpu1,2,andCHENYijun1,2.OilGeophysicalProspecting,2020,55(2):282-291.

Due to its optimal sparse representation and multi-scale and multi-direction characters,the Shearlet transform has a good performance for seismic noise reduction.Conventional Shearlet-based thresholding method involved the scale of sparsity,but it does not involve the direction of sparsity.It means that noise cannot be efficiently removed.We investigate the signal variation with directions in Shearlet domain and present scale and direction adaptive thresholding based on scale adaptive thresholding.The L2norm is calculated on a same scale in different directions to investigate the distribution of effective signals.The direction adaptive term is added to thresholding to realize scale and direction adaptive thresholding simultaneously.Model and real data tests show that this simultaneous adaptive thresholding exhibits better performance than the conventional method in random noise reduction and the utmost of signal preservation.

Keywords: Shearlet transform,random noise,direction adaptive thresholding

1.Key Laboratory of Ministry of Education on Safe Mining of Deep Metal Mines,Northeastern University,Shenyang,Liaoning 110004,China

2.School of Resources and Civil Engineering,Northeastern University,Shenyang,Liaoning 110004,China

Azimuthanisotropicwide-azimuthseismicdataprocessing：Acasestudyofthree-componentdatafromDaqingPlacanticline.ZHANGLiyan1,LIAng1,LIUJianying2,YANGJianguo1,andCHENZhide2.OilGeophysicalProspecting，2020，55(2)：292-301,310.

For HTI media,P-waves run slower in the direction perpendicular to fractures than in the direction parallel to fractures; meanwhile,reflection strength and frequency decrease with propagation.S-waves split into fast and slow waves.Bosed on the wide-azimuth three-component data in Lama-dian,Daqing Placanticline,we analyzed the perfor-mance characteristics,influence of azimuthal aniso-tropy,and attribute variation with azimuth.Ellipse fitting is then utilized to quantitatively predict the azimuth and intensity of anisotropy in this area.Our results are consistent with those obtained using shear wave splitting.To mitigate the influence of azimuthal anisotropy on the resolution of imaging,we calculate the accurate azimuthal velocity of HTI media through ellipse fitting to accomplish NMO correction.This corrects the time difference between fast and slow waves caused by azimuthal anisotropy and consequently improves seismic imaging.

Keywords：wide azimuth，anisotropy azimuth，anisotropy intensity，time difference between fast and slow waves，azimuthal velocity，ellipse fitting

1.Shenyang Geological Survey Center，China Geological Survey，Shenyang，Liaoning 110000，China

2.Exploration and Development Research Institute，PetroChina Daqing Oilfield Company，Daqing，Heilongjiang 163712，China

Anovelconstantfractional-orderLaplaciansviscoacousticwaveequationanditsnumericalsimulationmethod.CHENHanming1,2,3，WANGYilin4,andZHOUHui1,2,3.OilGeophysicalProspecting，2020，55(2)：302-310.

We develop a viscoacoustic wave equation with fractional-order Laplacians,which is better than the traditional integral-order viscoacoustic equation because the new equation more accurately describes the widely used constant-Qmodel.The operators related to amplitude attenuation and phase change are explicitly independent of each other,which is important for the robust reverse time migration with attenuation compensation.We formulate the first-order velocity-pressure viscoacoustic wave equation with the constant fractional-order Laplacians based on the second-order displacement viscoacoustic equation with the constant fractional-order Laplacians in time domain.To better model amplitude variation,space-varying density is involved in the new equation.To avoid spurious reflections caused by the periodicity of the Fourier transform,a convolutional perfectly matched layer (CPML) is employed as the absorbing boundary for the fractional-order Laplacian viscoacoustic equation.Numerical simulations are fulfilled using staggered-mesh pseudo-spectral method.We compare the numerical solution with the analytic solution for the homogeneous medium,and we find the new equation accurately describes the constant-Qmodel.We also verify its feasibility for complicated media through seismic wave field simulation using the BP salt dome model.

Keywords：viscoacoustic wave equation，numerical simulation，fractional-order Laplacians，staggered mesh，pseudo-spectral method，perfectly matched layer.

1.College of Geophysics,China University of Petroleum(Beijing)，Beijing 102249，China

2.State Key Laboratory of Petroleum Resources and Prospecting，Beijing 102249，China

3.CNPC Key Lab of Geophysical Exploration，Beijing 102249，China

4.School of Ocean and Earth Science，Tongji University，Shanghai 200092，China

NumericalsimulationofdetectingseismicsignalsinDASwells.MAGuoqi1,2,CAODanping1,2,YINJiaojian3,andZHUZhaolin1,2.OilGeophysicalProspecting,2020,55(2):311-320.

Owing to the Rayleigh scattering effect,distributed acoustic sensing (DAS) could detect seismic vibrations in optical fibers,which meanwhile also function as the carrier for signal transmission.Thus,this system is suitable for borehole seismic acquisition with low cost,high resolution,and high performance of anti-electromagnetic interference.Based on the discrete Rayleigh scattering interference model,we use numerical simulation to model borehole seismic signals in the DAS system; we also discuss the impacts of source strength,pulse width,and spatial fiber sampling interval on the waveform and signal to noise ratio of DAS signals.In this process,we do not consider the influence of ambient pressure,temperature,and borehole wall-fiber coupling.The results show that ①DAS signals vary with source strength.A strong source may cause waveform distortion or increased side lobes; this may lead to signal distortion; ②small pulse width is usually associated with strong noises.In contrast,large pulse width may facilitate high-frequency noise suppression and improve signal to noise ratio; but the resolution will be inevitably sacrificed to some extent; ③increased fiber sampling interval by multi-trace stacking may be useful to the improvement of signal to noise ratio.Thus,we may choose a proper sampling interval to improve signal to noise ratio and signal quality.DAS signals usually exhibit slightly higher frequencies than the original seismic signals; there are also inherent high-frequency noises.

Keywords: distributed acoustic sensing,Rayleigh scattering,source strength,pulse width,spatial fiber sampling interval,numerical simulation

1.School of Geosciences,China University of Petroleum(East China),Qingdao,Shandong 266580,China

2.Functional Laboratory for Marine Mineral Resources Assessment and Prospecting,Qingdao National Laboratory for Marine Science and Technology,Qingdao,Shandong 266071,China

3.College of Science,China University of Petroleum(East China),Qingdao,Shandong 266580,China

Elastic-wavenumericalsimulationandreversetimemigrationinpseudo-depthdomain.ZOUQiang1,HUANGJianping1,LIQingyang2，YONGPeng1,andXINTianliang1.OilGeophysicalProspecting，2020，55(2)：321-330.

Wave field extrapolation using finite-difference method is usually accomplished in the Cartesian coordinate system.Oversampling may occur locally when subsurface velocity varies greatly; this will increase the workload of computation.To address this issue,we fulfill finite-difference numerical si-mulation in pseudo-depth domain.In accordance with the chain rule of coordinate conversion,we formulate the mapping relation between the Cartesian coordinate system and pseudo-depth domain as well as consequent elastic wave equation (PDD-EWE) in pseudo-domain.Memory consumption may be significantly reduced for wave field extrapolation in pseudo-depth domain,provided that the accuracy is guaranteed.A variable difference operator is adopted,and its length varies with velocity.The model tests show that for the similar accuracy,elastic-wave forward modeling and imaging with a variable difference operator in pseudo-depth domain is better than conventional finite-difference method in depth domain in view of less memory consumption and higher computational efficiency.

Keywords：elastic wave，numerical simulation，reverse time migration，pseudo-depth domain，variable-length difference operator，memory consumption，computational efficiency

1.Geophysics Department，School of Geosciences，China University of Petroleum(East China)，Qingdao，Shandong 266580，China

2.Geophysical Exploration Research Institute，Si-nopec Zhongyuan Oilfield Company，Puyang，Henan 457001，China

Seismicdatastackingbasedontime-varyingmapping.YUEYubo1,ZHANGJianlei1,ZHANGChao-yang1,SHIYunyan2，XIONGYanrong3,andSUNPengfei4.OilGeophysicalProspecting，2020，55(2)：331-340.

Due to the existence of large source-receiver elevation differences in ocean-bottom-node (OBN) and mountain data,conventional common-midpoint (CMP) stacking based on an assumption of horizontal land surface is no longer feasible.We propose common reflection point (CRP) stacking based on time-varying data mapping as an alternative.We use stacking velocity to calculate time-va-rying CRP sampling points in output sections.At each CRP sampling point,two-way time is calcula-ted using a modified double-square-root formula for NMO correction.We can use this method to improve stacking quality for OBN and mountain data,and we also can obtain accurate stacking velocity for subsequent processing.Model and field data tests show that CRP stacking based on time-va-rying data mapping is better than CMP stacking.

Keywords：common reflection point，time varying，data mapping，stacking

1.R & D Center,BGP,CNPC，Zhuozhou，Hebei 072751，China

2.BGP Seismic Data Processing Center，Zhuozhou，Hebei 072751，China

3.BGP Acquisition Technology Center，Zhuozhou，Hebei 072751，China

4.School of Earth Science and Engineering，Hebei University of Engineering，Handan，Hebei 056038，China

Source-independentfrequency-domainfullwaveforminversionbasedondatasimilarity.XINTianliang1,HUANGJianping1,XIEFei1,2,ZHOUBing3,andLUZizhuo1.OilGeophysicalProspecting,2020,55(2):341-350.

In terms of single-frequency data features in frequency domain,we formulate an objective function (SOD) in complex number field,which is similar to the function describing data similarity in real number domain,to mitigate the non-linearity of inversion.This method is practical because amplitude matching between synthetic data and observed data becomes less important.The source wavelet is naturally avoided in the objective function; besides,the reference trace is not required.But for standard trace normalized (STN) full waveform inversion,the convergence may slow down or even suspend if the reference trace is not properly estimated.Model tests show that ① SOD inversion is independent of source wavelet,and it is superior to additional two source-independent algorithms (average trace normalized (ATN) and STN) in the accuracy of inversion; ② SOD inversion is based on data similarity,and it is tolerant to energy discrepancies between different data sets; thus,the inversion is less dependent on the initial velocity model; ③ random noises may account for similar proportions in the numerator and denominator through multiply and add operations.This means that the objective function of SOD is robust; ④ SOD could be integrated with source encoding to effectively improve computational efficiency.

Keywords: full waveform inversion,source wavelet,data similarity,reference trace,initial model

1.School of Geosciences,China University of Petroleum(East China),Qingdao,Shandong 266580,China

2.East China Branch,Sinopec,Nanjing,Jiangsu 210011,China

3.CNOOC (China) Tianjin Company,Tianjin 300452,China

Finite-differencecontrastsourceinversionbasedonhybridfastconjugategradientmethod.WANGDoudou1,2，WANGShoudong2，ZOUShaofeng1,GAOYanxia1,andLIUHan1.OilGeophysicalProspecting，2020，55(2)：351-359.

Finite-difference contrast source inversion (FDCSI) is a solution to inverse scattering.The background model remains unchanged in the inversion,and forward modeling is performed for only one time; this reduced the workload of computation.FDCSI translates the problem of inverse scattering into a problem of optimization.The cost function is optimized using the conjugate gradient method,which suffers from small convergence rate and low efficiency.After the study of FDCSI using acoustic wave equation in frequency domain,we develop a FDCSI algorithm based on a hybrid fast conjugate gradient method to improve the efficiency of inversion.The hybrid fast conjugate gradient method is modified from the fast iterative shrin-kage thresholding algorithm and is feasible for FDCSI.The cost function could converge quickly without more computation in a single iteration; this guarantees fast and robust convergence of FDCSI.

Keywords：inverse scattering，contrast source inversion，acoustic wave equation in frequency domain，fast iterative shrinkage thresholding algorithm，hybrid fast conjugate gradient method

1.Sinopec Geophysical Research Institute，Nanjing,Jiangsu 211103，China

2.State Key Laboratory of Petroleum Resources and Prospecting，China University of Petroleum(Beijing)，Beijing 102249，China

Sourcelocationthroughmicroseismiccross-correlationimaging.ZENGZhiyi1,2andZHANGJianzhong1,2,3.OilGeophysicalProspecting，2020，55(2)：360-372,388.

Microseismic events may be located through migration and stacking,which improves the signal to noise ratio; it is unnecessary to pick arrival time.But such stacking may be affected by the polarity reversal of first arrivals,and noises cannot be efficiently suppressed.In view of similar waveforms from a same source recorded at different geophones,we propose a method of source location through microseismic cross-correlation imaging.An imaging function is formulated by multiplying cross-correlation coefficients of microseismograms after moveout correction,and then imaging energy at all nodes is calculated.The microseismic source is finally located using the maximum energy.Model and field data tests show that noises and the impact of first-arrival polarity reversal on source imaging could be reduced,which results in improved resolution and location accuracy of source imaging.

Keywords：surface microseism，imaging function，source location，signal to noise ratio，first-arrival polarity，waveform cross correlation

1.Key Lab of Submarine Geosciences and Prospecting Techniques，MOE，Qingdao，Shandong 266100，China

2.Functional Laboratory for Marine Mineral Resources Assessment and Prospecting，Qingdao National Laboratory for Marine Science and Technology，Qingdao，Shandong 266061，China

3.College of Marine Geosciences，Ocean University of China，Qingdao，Shandong 266100，China

Experimentalstudyoncross-frequencywavevelocityanddispersioninrocks.LIMinlong1,LIUHaojie1,YANGHongwei1,WEIGuohua1,andSHIJianxin1.OilGeophysicalProspecting，2020，55(2)：373-378.

We measure the elastic properties of two samples,A and B,saturated with different fluids (water and glycerol) at different frequencies (2～2000 Hz and 1 MHz).The results show that the frequency of dispersion increases with pore fluid mobility.The frequency of dispersion is positively correlated with porosity and permeability and negatively correlated with fluid viscosity.P-velocity increases with saturation at high frequencies; P-velocity first increases and then decreases with saturation at low frequencies.This means there is a critical saturation,the value of which is closely related to fluid mobility.

Keywords：dispersion，characteristic frequency，critical saturation，fluid mobility，viscosity

1.Shengli Geophysical Research Institute of Sinopec，Dongying，Shandong 257022，China

Applicationofmulti-thresholdBIRCHclusteringtofacies-controlledporosityestimation.SUNQifeng1,DUANYouxiang1,LIUFan1,andLIHongqiang2.OilGeophysicalProspecting，2020，55(2)：379-388.

In view of the importance of lithofacies and porosity study in hydrocarbon exploration,we present an approach of multi-threshold BIRCH clustering for lithofacies classification,based on which we estimate porosity using ridge regression.The heuristic initial threshold is set in terms of wave impedance distribution,and the number of thresholds is increased dynamically according to inter-clustering volume inconsistency.Global Agglomerative clustering is then employed for lithofacies classification.For each lithofacies,a modified ridge regression algorithm is used to predict porosity based on well porosity.Model tests show that multi-threshold BIRCH clustering exhibits good robustness and computational efficiency for accurate lithofacies classification.A field data test shows that porosity could be estimated accurately using this method.

Keywords: lithofacies，multi-threshold,BIRCH clustering，ridge regression，porosity

1.College of Computer Science and Technology,China University of Petroleum(East China)，Qingdao，Shandong 266580，China

2.Drilling Technology Research Institute of Shengli Petroleum Engineering Corporation Limited，SINOPEC，Dongying，Shandong 257000，China

High-resolutionBayesiansequentialstochasticinversion.LIQixin1,LUOYaneng2,ZHANGSheng3,ZHANGLu1,YANGYadi4,andHUANGHandong5.OilGeophysicalProspecting,2020，55(2)：389-397.

Model-based inversion and sparse spike inversion,with some constraints,are two common algorithms for deterministic inversion.The output is a single solution; thus,it is hard to evaluate its uncertainty.Geostatistical inversion is usually accomplished using geostatistical simulation combined with Markov Chain Monte Carlo; but seismic inversion,log constraints,and stochastic simulation have not been integrated within a uniform theoretical framework.We integrate seismic inversion,log constraints,and geostatistical information within the Bayesian framework to formulate the simultaneous equations which involve logarithmic impedance and log data.Sequential Gaussian simulation is then employed to sufficiently sample the equations.Numerical studies show that our method is better than conventional least-square inversion because the resolution is high,the inversion is constrained by priori statistical data,and a number of impedance realizations could be used for uncertainty evaluation.Compared with sequential Gaussian simulation entirely based on log data,our method uses seismic data as constraints to reduce the uncertainty of inversion.In accordance with field data inversions,Bayesian sequential stochastic inversion is better than model-based inversion and sparse spike inversion in high vertical resolution and feasibility of uncertainty evaluation.

Keywords: high resolution,impedance inversion,Bayesian theory,sequential simulation,uncertainty evaluation

1.CNOOC Research Institute,Beijing 100028,China

2.R & D Center,BGP,CNPC,Zhuozhou,Hebei 072751,China

3.College of Mining Engineering,Taiyuan University of Technology,Taiyuan,Shanxi 030024,China

4.Research Institute of Petroleum Exploration and Development,PetroChina,Beijing 100083,China

5.College of Geophysics,China University of Petroleum(Beijing),Beijing 102249,China

Multi-dimensionalpre-stackseismicinversionbasedonsparserepresentation.YANGSen1,WUGuochen1,2，ZHANGMingzhen3,DUZeyuan1,SHANJunzhen1,andLIANGZhanyuan1.OilGeophysicalProspecting，2020，55(2)：398-410.

It is difficult to characterize thin reservoirs through conventional pre-stack and post-stack inversion.Interference effects generated by stacking may fade,distort,and suppress effective signals; thus,we suggest avoiding stacking to preserve signals.Pre-stack attributes vary with offset and frequency; this means that FAVO inversion based on high-resolution spectral decomposition may be capable of addressing the issue of fluid detection.We present multi-dimensional pre-stack seismic inversion based on sparse representation,and the point is no stacking.The first step is to extract angle gathers at the zone of interest and then perform high-resolution time-frequency decomposition for each single angle gather extracted in terms of sparse representation.The second step is to establish the mapping relationship between impedance and seismic data in accordance with the Bayesian theory and add some perturbation,calculated using non-linear optimization,to the initial model.The last step is to take the result at the previous frequency as the constraint to perform inversion at the next frequency until the final output at each angle is obtained.The case study in Prospect A shows that constrained inversion at each angle and frequency generates multi-dimensional results with high precision; this facilitates thin reservoir characterization.

Keywords：multi-dimension，pre-stack inversion，sparse representation，interference suppression，F(xiàn)AVO，reservoir prediction.

1.School of Geosciences,China University of Petroleum(East China),Qingdao,Shandong 266580，China

2.Laboratory for Marine Mineral Resources,Qingdao National Laboratory for Marine Science and Technology,Qingdao,Shandong 266580，China

3.Shengli Geophysical Research Institute of SINO-PEC，Dongying，Shandong 257022，China

Researchonstratigraphicstructurebasedonseismicforwardmodeling：AcasestudyofthethirdmemberoftheYanchangFormationinHeshuiarea，OrdosBasin.WANGWenfeng1,2,YUEDali1,2，ZHAOJiyong3,WANGWurong1,2,LIWei1,2,andWANGBo3.OilGeophysicalProspecting，2020，55(2)：411-418.

It is necessary to fully understand stratigraphic structure before isochronal stratigraphic correlation and reservoir characterization.For the third member of the Yanchang Formation in Heshui,Ordos Basin,we use seismic forward modeling to investigate its configuration.In view of clinoform seismic reflections,we design three conceptual models to describe the stratigraphic structure.After forward modeling,we compare synthetic reflections with observed reflections and conclude that two models are feasible in this prospect.In accordance with log responses,one model is finally determined to best describe the stratigraphic structure in this prospect.This method may also apply to other prospects with similar geologic setting.

Keywords：clinoform reflections，stratigraphic structure，seismic forward modeling，Heshui area，Yanchang Formation

1.College of Geosciences，China University of Petroleum(Beijing)，Beijing 102249，China

2.State Key Laboratory of Petroleum Resources and Prospecting，China University of Petroleum(Beijing)，Beijing 102249，China

3.Research Institute of Exploration and Development，PetroChina Changqing Oilfield Company，Xi’an，Shaanxi 710021，China

Quantitativecharacterizationofgashydratebasedonforwardmodeling.YANGJing1,ZHANGLei1,WANGJiushuan1,XIENan1,MANing1,andZHANGHui1.OilGeophysicalProspecting，2020，55(2)：419-425.

Gas hydrate may be characterized qualitatively,semi-quantitatively,and even quantitatively using seismic data; but how to realize quantitative prediction in an undrilled prospect is a hard nut to crack.Forward modeling is a solution to this problem.Our case study focuses on a prospect in the South China Sea.The area with plausible gas hydrate is sited using several seismic attributes.The velocity and density of the sedimentary layer,after calibration using ocean drilling program (ODP) data,are then used for forward modeling to generate synthetic seismogram highly correlated with observed seismic data.After pre-stack elastic inversion,we predict hydrate distribution and its petrophysical properties in terms of the relation between hydrate saturation and P-impedance.Our work provides information for the estimation of gas hydrate reserves.

Keywords：gas hydrate，seismic attribute，forward modeling，pre-stack impedance inversion

1.BGP Seismic Data Processing Center，Zhuozhou，Hebei 072751，China

TectonicevolutioninChangshalingtransferzoneanditsrelationwithhydrocarbonaccumulation.GAOYang1,ZHANGWenping2,SUNXuedong2,ZHOUZaihua3,ZHANGBaoquan2,andBAIJun2.OilGeophysicalProspecting,2020,55(2):426-434.

Fault block-lithologic reservoirs in Changshaling transfer zone exhibit complicated fault-formation intersections due to the occurrence of multi-phase tectonic movements,and it is hard to investigate the fault system and hydrocarbon accumulation using seismic data with poor quality.Based on preceding studies,we focus on tectonic evolution history,structure interpretation,episodes of faulting,and trap features to establish the relations among structures,sands,faults,and hydrocarbon migration and accumulation.We conclude that ①Changshaling transfer zone is the product of two-phase extensional tectonic movements in the Early Cretaceous extension and fault depression period.Basement-convolved extensional tectonic movement occurred at the early stage,when west-inclined normal faults came into being.Overburden-slippage extensional tectonic movement occurred at the late stage,when east-inclined plough-type normal faults appeared and intersected early faults; ②early west-inclined syndepositional faults gave rise to fault-bench belts and thus dominate the lateral distribution of sedimentary systems.Accommodation faults gave birth to an ancient land form with fault troughs alternating with fault ridges.Sands migration was dominated by fault troughs,and sands distribution was dominated by the ancient land form at the depositional stage of the Cretaceous Xiagou Formation; ③early west-inclined faults played a role in hydrocarbon preservation,and late east-inclined faults played a role in hydrocarbon migration.Fault-horst reservoirs are the product of these two-phase faulting activities.In summary,hydrocarbon accumulation is dominated by structures; each fault block has a reservoir; oil and gas concentrate in the structural highs.

Keywords: Ying’er sag,Changshaling transfer zone,tectonic evolution,episode of faulting,fault-horst reservoir

1.School of Geology Engineering and Geomatics,Chang’an University,Xi’an,Shaanxi 710064,China

2.BGP Geological Research Center,Zhuozhou,Hebei 072751,China

3.Exploration and Development Research Institute,PetroChina Yumen Oilfield Company,Jiuquan,Gansu 735200,China

Stochasticgeologicalmodelingconstrainedbywelltestdata.FENGGuoqing1,HEYujun1,LIUHonglin2，CHENYan3,ZHANGPing1,4，andXUEFangfang5.OilGeophysicalProspecting，2020，55(2)：435-441.

Production performance data have seldom been involved in stochastic simulation; thus,it is difficult to describe such production performance as pressure and yield using the geologic model.We present a method for well test data constrained stochastic modeling.The initial model,which is built through stochastic simulation,is updated using simulated annealing algorithm until the model matches well test data.In this process,bottom hole pressure is calculated using analytic solution combined with numerical solution; this may save more than 90% of computation time.A case study shows that the final model has been improved and matches well test data.

Keywords: stochastic simulation，well data constraint，simulated annealing，objective function，geological modeling

1.State Key Laboratory of Reservoir Geology and Development Engineering,Southwest Petroleum University,Chendu，Sichuan 610500,China

2.The Second Oil Production Plant,Xinjiang Oilfield Company,CNPC,Karamay,Xinjiang 834008,China

3.The Second Oil Production Plant,Henan Oilfield Company,SINOPEC,Nanyang,Henan 473100,China

4.Chongqing Fuling Shale Gas Exploration and Development Company,SINOPEC,Chongqing 408014,China

5.Oil and Gas Technology Research Institute,Changqing Oilfield Company,CNPC,Xi’an,Shaanxi 710018,China

Edgerecognitionusingjointcorrelationcoefficientofgravitygradient.HOUZhenlong1,2,YUANYuan3,WANGEnde1,2,FUJianfei1,2,andZHENGYujun1.OilGeophysicalProspecting，2020，55(2)：442-453.

To improve the quality of gravity gradient ima-ging and resolution of identifying subsurface structure boundaries,we use the joint correlation coefficient of total horizontal derivative inxandydirection with gravity gradient for edge recognition.Correlation coefficients calculated at multi-directional windows are integrated to enhance image.Binarization and thresholding are also utilized to highlight edges.Model tests show that model boundaries,especially the corners,could be clearly imaged.We apply this method to the aerial data observed at the Bay St.George’s,Canada and identify major subsurface structure boundaries.The tests demonstrate the feasibility of the method.We also discuss the impacts of window size,observation point spacing,and line spacing on imaging.Compared with other methods,the method is theoretically simple and can yield better image.

Keywords: gravity gradient,joint correlation coefficient,multi-directional window,directional total horizontal derivative,edge recognition,Bay St.George’s

1.Key Laboratory of Ministry of Education on Safe Mining of Deep Metal Mines,Northeastern University,Shenyang,Liaoning 110819,China

2.College of Resources and Civil Engineering,Northeastern University,Shenyang,Liaoning 110819,China

3.Key Laboratory of Submarine Geosciences,Se-cond Institute of Oceanography,Ministry of Natural Resources,Hangzhou,Zhejiang 310012,China

Balancededgedetectionbasedoneigenvaluesoffullmagneticgradienttensor.ZHENGQiang1,GUOHua2,ZHANGGuibin1,HANSong2,WANGMing2,andLIUHaojun2.OilGeophysicalProspecting，2020，55(2)：454-464.

Full gravimetric-magnetic gradient tensor data,which feature high precision,high resolution,and multi-parameter,is a focus in potential field data processing and interpretation.For accurate and robust edge detection,we present a balanced edge detection method based on the maximum eigenvalue of full magnetic gradient tensor.Model tests show that the boundaries of the deep and shallow models could be identified simultaneously with high precision and robustness; It also has some resistance to oblique magnetization.To test its feasibility,we apply this method to field data acquired in a prospect in north China and obtain the clear image of structure boundaries.

Keywords: full magnetic gradient tensor,eigenva-lue,balanced edge detection

1.School of Geophysics and Information Technology,China University of Geosciences (Beijing),Beijing 100083,China

2.Airborne Geophysical Prospecting and Remote Sensing Center,Ministry of Natural Resources,Beijing 100083,China

Anoverviewofjointelectromagnetic-seismicinversionanditsfuturedevelopment.PENGGuomin1andLIUZhan1.OilGeophysicalProspecting，2020，55(2)：465-474.

Different geophysical data reflect rock properties from different aspects,and joint utilization of such complemental information may better image subsurface geology.Electromagnetic prospecting and seismic prospecting are two ways of geophysical exploration with different mechanisms.Seismic prediction suffers from the issue of non-uniqueness,and electromagnetic prospecting may render useful information which is absent in seismic data.Thus,joint utilization of electromagnetic and seismic data can provide more credible support for exploration and development decision making.Joint electromagnetic-seismic inversion,as a method of quantitative interpretation,may mitigate the non-uniqueness of single-data inversion and improve the credibility of geophysical interpretation.We give an overview of the classification,progress,and opportunities of joint inversion in China and around the world,as well as its future development.

Keywords: joint inversion,electromagnetic,seismic,rock physics,cross gradient

1.School of Geosciences,China University of Petroleum(East China),Qingdao,Shandong 266580,China