ABSTRACTS

IdentificationofKarstcavesinseismicdatabasedondeepconvolutionalneuralnetwork.YANXingyu1,2,LIZongjie3,GUHanming1,2,CHENBenchi4,DENGGuangxiao3,andLIUJun3.OilGeophysicalProspecting,2022,57(1):1-11.

Karst cave identification is significant for the exploration and development of fracture-cavity oil and gas reservoirs. Conventional identification methods are multi-solution and inefficient. Therefore, a deep learning method with strong feature learning and generalization capabilities is introduced into Karst cave identification. However, it is still a challenging task to identify Karst caves by deep learning due to the complex response characteristics of Karst caves to the seismic wavefield, the small sizes of anomalies, and the difficulties in obtaining training samples. Faced with this pro-blem, we propose a “two-step” deep learning me-thod for identifying Karst caves in seismic data. Spe-cifically, the U-Net model is used to identify the “bead-shaped” anomalous reflection on the seismic section. Then, according to the identification results of the “bead-shaped” anomalies, seismic data are cropped into small patches and input into the deep residual network to implement the prediction of the actual Karst cave profile. Considering the difficulties in obtaining training data for actual Karst cave prediction, we propose implementing wave equation forward modeling to generate seismic Karst cave data with accurate labels. The application of field seismic data shows that the me-thod is accurate in Karst cave identification, has strong noise resistance, and can greatly save the cost of manual interpretation.

Keywords： fracture-cavity oil and gas reservoir, Karst cave identification, deep learning, U-Net model, deep residual network

1. Institute of Geophysics & Geomatics, China University of Geosciences (Wuhan), Wuhan, Hubei 430074, China

2. Hubei Subsurface Multiscale Image Key Laboratory, Wuhan, Hubei 430074, China

3. Research Institute of Exploration and Development, Northwest Oilfield Branch Co., SINOPEC, Urumqi, Xinjiang 830011, China

4. Oil Field Department of Science and Technology Ministry, SINOPEC, Beijing 100728, China

Robustseismicdatadenoisingbasedondeeplear-ning.ZHANGYan1，LIXinyue1，WANGBin1，LIJie1，WANGHongtao2，andDONGHongli3.OilGeophysicalProspecting,2022,57(1):12-25.

Noise in seismic data is complicated, and the traditional modeling methods based on prior knowledge cannot describe the noise distribution accurately. In the denoising methods based on deep learning, a multi-layer convolutional neural network is employed to automatically extract the deep features of seismic data, and its nonlinear approximation ability is used for adaptive learning, which yields a complex denoising model and thus brings a new idea for the denoising of seismic data. How-ever, poor generalization ability is found in the cur-rent denoising methods based on deep learning in the case of insufficient sample coverage, greatly reducing the denoising effect. Therefore, this paper proposes a robust deep learning algorithm for denoising. The model is composed of two sub-networks, which realize the estimation of noise distribution and noise suppression of noisy seismic data respectively. The sub-network for estimating noise distribution is a multi-layer convolutional neural network. The sub-network for denoising introduces a strategy of feature fusion, which comprehensively considers the global and local information of seismic data, and a residual learning strategy is utilized to extract noise features. L1norm loss is taken as the loss function for the two sub-networks to enhance the generalization ability of the model. Experiments show that the method proposed in this paper has a higher generalization ability than similar algorithms. Data processing results indicate that it better preserves event features and has a higher signal-to-noise ratio.

Keywords: seismic data denoising，deep learning，robustness，L1loss function，feature fusion，residual network

1. School of Computer & Information Technology，Northeast Petroleum University，Daqing，Heilongjiang 163318，China

2. Daqing Information Technology Research Center，Daqing，Heilongjiang 163318，China

3. Artificial Intelligence Energy Research Institute，Northeast Petroleum University，Daqing，Heilongjiang 163318，China

S-wavevelocitypredictionmethodforsand-shaleformationbasedonquadraticoptimizationnetwork.SHANBo1,ZHANGFanchang1,andDINGJicai2.OilGeophysicalProspecting,2022,57(1):26-33.

Shear wave (S-wave) velocity is an important parameter that reflects the lithological characteristics of a reservoir. However, it is often absent in actual logging data. In this paper, an end-to-end quadratic optimization network is constructed according to the relationships of S-wave velocity with other parameters. Instead of solving intermediate parameters, it uses the gamma ray, porosity, and compressional wave (P-wave) velocity to predict the S-wave velocity directly. The quadratic optimization algorithm is applied in the network training process to replace the Adam algorithm and achieves higher accuracy and efficiency. In addition, the orthogonal experiment is used to analyze the influences of different parameters and training strategies (including optimization algorithm, number of network layers, and number of training wells) on the prediction of the S-wave velocity. The results show that the optimization algorithm has the greatest impact on the prediction. The quadratic optimization algorithm has a better prediction effect and higher efficiency than those of the Adam optimization algorithm. A suitable activation function has a positive effect on the prediction. According to the experiment results, the optimal network parameters and training strategy are selected for S-wave velocity prediction. The prediction results on the test set show that the method can predict the S-wave velocity accurately and effectively.

Keywords：S-wave velocity prediction, neural network, quadratic optimization, orthogonal experiment

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

2. CNOOC Research Institute Co., Ltd., Beijing 100027, China

Griddingandfilteringmethodofgravityandmagneticdatabasedonself-attentiondeeplearning.MAGuoqing1,WANGZekun1,andLILili1.OilGeophysicalProspecting,2022,57(1):34-42.

The gridding and filtering of gravity and magnetic data directly influence the result of data processing. This paper designs a more rational deep learning model to improve the accuracy of gridding and filtering. The gridding method based on self-attention deep learning is constructed, and the self-attention mechanism layer is utilized to process the two-dimensional position embeddings. In this way, the position embeddings vector is obtained with global and local information integrated. Then the position information and anomaly information are fused to output node anomaly, thus reducing the distortion. For the random and stripe noises of gravity and magnetic data, a convolutional neural network is first employed to classify noise. The stripe noise is filtered by self-attention convolutional neural network and the random noise by convolutional autoencoder to get high-quality basic data. Model experiment shows that the gridded structure of deep learning is closer to the real result than that of the traditional method. The proposed filtering method can remove various noises, providing more accurate basic data for the follo-wing inversion. The gridding and filtering method based on deep learning applied to practical magne-tic data achieves good results, proving that it has strong feasibility and practicability.

Keywords：gravity and magnetic data, gridding, filtering, deep learning, self-attention

1. College of Geoexploration Science and Techno-logy, Jilin University, Changchun, Jilin 130026, China

Seismicrandomnoiseattenuationbasedonstationarywavelettransformanddeepresidualneuralnetwork.WUGuoning1,YUMengmeng1,WANGJunxian1,andLIUGuochang2.OilGeophysicalProspecting,2022,57(1):43-51.

There are many conventional denoising me-thods, but each method is limited by certain assump-tions or conditions. In addition, multiple local extrema may cause the denoising algorithm to converge to a local optimal solution instead of the global one. For this reason, a random noise suppression method based on the stationary wavelet transform and deep residual neural network (WaveResNet) is proposed. It combined the topology structure of the residual neural network (ResNet) with the stationary wavelet transform. The residual module effectively avoids the vanishing gradient or computational consumption caused by the deep network but loss function saturation. In addition, the wavelet transform is an efficient feature extraction method. It can obtain the low-frequency and high-frequency feature information in different directions and learn the characteristics of signal or noise in different regions. First, each picture in the Train400 dataset is rotated by different angles to increase the amount of data in the training set, after which Gaussian noise is added. Then, the one-level stationary Haar wavelet decomposition is performed on each picture to gain a training dataset. The high- and low-frequency information in the wavelet transform domain is extracted through training. On this basis, the wavelet decomposition of the learned noise is subtracted from that of the noisy data, thus achieving the wavelet decomposition of the denoised signal through the direct channel. Finally, the denoised signal is obtained through the inverse stationary wavelet transform. Experiments of synthetic signals and field seismic data show that the proposed method can suppress seismic random noise well, and the signal-to-noise ratio and its peak of the denoised signal are higher than those of conventional methods.

Keywords：random noise, noise attenuation, stationary wavelet transform, residual neural network, signal-to-noise ratio

1. College of Science, China University of Petro-leum (Beijing), Beijing 102249, China

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

AttributefusionmethodbasedonNSST—PAPCNNforfractureprediction.TANGWei1,2,LIJingye1,2,WANGJianhua3,BOXin1,2,GENGWeiheng1,2,YEWei4.OilGeophysicalProspecting,2022,57(1):52-61.

The commonly used post-stack seismic attri-butes mainly include the coherent body (describing the similarity of waveform), curvature body (describing the bending degree of formation caused by tectonic stress), dip-angle body (describing the structural changes of formation). However, it is difficult to accurately predict the distribution of underground fractures only by a single attribute. Therefore, this paper proposes a comprehensive attribute fusion method based on nonsubsampled shear wave transform-parameter adaptive pulse coupled neural network (NSST—PAPCNN) for fracture prediction. This method relies on the NSST decomposition algorithm to decompose the multiple attribute data into high- and low-frequency sub-bands. After fusion, the multi-scale and multi-direction high- and low-frequency sub-bands were refactored. From the final multi-attribute fusion result, the contour and detail information of fractures can be further extracted. The detailed steps are as follows: ①Multiple seismic attributes (coherence, curvature and dip-angle attributes) of the same-scale fractures were extracted and decomposed into high- and low-frequency sub-bands by NSST. The high-frequency sub-band contains more fracture information, and the low-frequency sub-band can better describe the fracture contour and has rich energy information. ②PA-PCNN model was used for the fusion of high-frequency sub-band without manual parameter setting, which generated more comprehensive high-frequency data. The weighted sum of eight-neighbor modified Laplacian (WSEML) and the weighted local energy (WLE) were combined to fuse the low-frequency sub-bands, enriching the low-frequency data by retaining more details and energy information. ③The inverse NSST method was applied to predict the fracture effectively based on attribute fusion. The proposed method was used to test the attri-bute data in M Zone, and the fracture prediction results of different methods are compared. It proves that the attribute fusion based on NSST—PAPCNN can predict fractures more effectively.

Keywords：NSST, PCNN, parameter adaptive, multi-attribute fusion, fracture prediction

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. National Engineering Laboratory for Offshore Oil Exploration, Beijing 100028, China

4. Research Institute of Exploration & Development, Huabei Oilfield Company, PetroChina, Renqiu, Hebei 062552, China

ApplicationandeffectofkeyseismicacquisitiontechnologiesforPermianvolcanicrocksinwesternSichuanBasin.LIShuqin1,WANGXiaoyang1,ZHANGMeng1,ZHAOXiaohong2,WANGXuemei1,andZHOUXiaoji1.OilGeophysicalProspecting,2022,57(1):62-73.

The volcanic rocks developed in Sichuan Basin are well distributed in sources, reservoirs, and cap rocks with favorable conditions for hydrocarbon accumulation. High-quality volcanic porous reservoirs have been successfully drilled in Well YT1 in 2018, and thus the high-yield industrial gas flow was obtained. This proves that Permian volcanic rocks in Sichuan Basin have great oil and gas exploration potential. However, due to the low signal-to-noise ratio and imaging accuracy of previous seismic data, it is difficult to meet the needs of fine characterization for favorable faces of volcanic rocks and well location deployment. High-precision key acquisition technologies have been successfully applied to seismic exploration in volcanic rocks, including high-precision three-dimensional geometry design, shot and vibroseis joint acquisition operation, well depth design of dynamic stimu-lation, and low-frequency geophone receiving technology. As a result, the signal-to-noise ratio and resolution of the seismic data of the target layer have been greatly improved, and the seismic identification mode of volcanic rock effusive faces has been formed by using new data. The plane distribution of effusive and overflow faces of Permian volcanic rocks in this area has been effectively identified, which has greatly promoted the process of natural gas exploration in new areas and new fields of Sichuan Basin.

Keywords：volcanic rock, high-precision seismic acquisition, shot and vibroseis joint acquisition ope-ration, optimal geometry design, dynamic well depth design, low-frequency geophone

1. Southwest Branch, BGP Inc., CNPC, Chengdu, Sichuan 610000, China

2. Exploration Department, Southwest Oil & Gasfield Company, PetroChina, Chengdu, Sichuan 610000, China

Acquisitionandapplicationresultanalysisofthree-componentwalkawayVSPdataofcomplexpiedmontstructuresatthesouthernmarginofJunggarBasin.CHENGZhiguo1,CHENYong1,WANGXiaotao1,CHENPeng2,andCAIZhidong3.OilGeophysicalProspecting，2022,57(1):74-81.

The piedmont structures at the southern margin of the Junggar Basin are complex, with great oil-gas exploration potential. Well GT1 has achieved high-yield industrial oil-gas flows. In contrast, the target formations of the Gaoquan Anticline are deeply buried, the reservoir is thin, and small faults have developed. Consequently, surface seismic data has a limited resolution. In response, efforts are made to achieve technical breakthroughs in walkaway vertical seismic profile (W-VSP). Three-component W-VSP data are collected. Comparative data analysis shows that the resolution of W-VSP data is significantly higher than that of the surface seismic data and small faults and thin re-servoirs are thereby clearly identified. This research not only provides a basis for analyzing the distribution of main oil-gas target formations and the types of oil-gas reservoirs of the Gaoquan Anticline but also offers experience for W-VSP application at Xinjiang Oilfield.

Keywords：walkaway VSP, three-component, thin reservoir, small fault

1. Institute of Geophysics, Research Institute of Exploration and Development, Xinjiang Oilfield Company, PetroChina, Urumqi, Xinjiang 830013, China

2. Urumqi Branch, GRI, BGP Inc., CNPC, Urumqi, Xinjiang 830016, China

3. New Resources Geophysical Exploration Division, BGP Inc., CNPC, Zhuozhou, Hebei 072750, China

QualityfactorQestimationbasedonStransformandvariationalmethod.XULi’nan1,GAOJinghuai1,YANGYang1,GAOZhaoqi1,andWANGQian2.OilGeophysicalProspecting，2022,57(1):82-90.

Quality factorQis an important parameter to quantitatively describe the viscoelastic attenuation. AccurateQestimation is beneficial to reservoir identification and hydrocarbon detection. It can also be used for inverseQfiltering to improve the resolution of seismic data. The traditionalQestimation methods include logarithmic spectrum ratio (LSR), center frequency shift (CFS), and peak frequency shift (PFS), etc. LSR has poor noise immunity. Both CFS and PFS depend on the type of seismic wavelet. In response to these problems, this paper proposes a robustQestimation method based on the S transform and variational method. Firstly, by studying the nonstationary convolution model, we derived the approximate representation of nonstationary seismic data in the S domain. Seco-ndly, on the basis of approximate representation, the objective function of quality factorQand seismic wavelet is established and minimized based on the variational method, thereby obtaining the expression ofQestimation. Finally, we designed an adaptive selection scheme of an integral interval to improve the accuracy and noise resistance of the method. This scheme can automatically calculate the frequency parameters of the integration area based on the time-frequency spectrum of seismic data. Model examples demonstrate that the proposed method does not rely on the wavelet type and the length of the window function and shows good robustness to noise. The real data further verifies the effectiveness of the method inQestimation.

Keywords：seismic wave attenuation, S transform, variational method, adaptive parameter selection,Qestimation

1. School of Information and Communications Engineering, Xi’an Jiaotong University, Xi’an, Shaanxi 710049, China

2. School of Mathematics and Statistics, Hubei University of Arts and Science, Xiangyang, Hubei 441053, China

Amulti-iterativemethodfordeghostingbasedonGreen’stheorem.SONGJianguo1,2，MAAn1,2，HUANGSheng3，LIUJiong4，andCHENFeixu5.OilGeophysicalProspecting,2022,57(1):91-100.

As one of the marine broadband seismic data processing techniques，deghosting is the key to broadening the frequency band and improving the resolution of seismic data. The commonly used deghosting method is removing the predicted ghost from the whole wave field. Evidently，the ghost is hard to predict accurately. According to Green’s theorem，we proposed the iterative techniques and flow to predict the ghost based on its ray path. The technology is then combined with the deve-loped matching subtraction techniques in the curve-let domain，effectively suppressing ghost. The complexity of predicted ghost tended to cause da-mage to primary waves or incomplete deghosting during matching subtraction. Therefore，multiple iterations are introduced. The primary wave field after matching subtraction is taken as input to predict the ghost again by Green’s theorem followed by subtraction. After several iterations，we can improve the accuracy of ghost prediction and get a better primary wave. The test of model data and real seismic data manifests that the proposed method can preserve more primary waves during the deghosting.

Keywords：deghosting，Green’s theorem，multi-iteration，curvelet transform，marine broadband seismic

1. Shangdong Provincial Key Laboratory of Deep Oil & gas, Qingdao, Shangdong 266580, China

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

3. Hainan Company，China National Offshore Oil Corporation，Haikou，Hainan 570311，China

4. Petroleum Exploration and Production Research Institute，SINOPEC, Beijing 100083，China

5. Exploration and Production Research Institute，Tarim Oilfield Company, PetroChina, Korla，Xinjiang 841000，China

Forwardmodelingofscalarwaveequationwithge-neralizedfinitedifferencemethod.JIAZongfeng1，WUGuochen1，LIQingyang1，YANGLingyun1，andWUYou1.OilGeophysicalProspecting,2022,57(1):101-110.

The forward modeling of seismic wavefields with the conventional finite difference method is constrained by fixed mesh spacing，and thus the mesh subdivision is inevitably inconsistent with the actual speed interface，which brings about problems such as stair diffraction in undulating interface and inaccurate travel time of reflection waves. The generalized finite difference method is a meshless algorithm based on Taylor function expansion and weighted least squares fitting. In this method，the partial derivative of unknown parameters in the differential equation is expressed as a linear combination of the function values of adjacent nodes，and suitable distribution forms of field nodes can be established according to different geological body models. Thus，it overcomes the problem of mesh dependence that the traditional finite difference method faces. In this paper，the non-uniform subdivision method which sets nodes along the layer is applied to make the generated nodes conform to the undulating surface or boundary. The test results of different models show that the generalized finite difference method can effectively solve the pro-blems of the spurious reflection and inaccurate travel time of reflection waves，with stability.

Keywords：scalar wave equation，forward modeling，generalized finite difference method，non-uniform subdivision，undulating interface

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

RapidcalculationofreflectedwavetraveltimeinlayeredTImediawithdippinginterfaces.ZHANGJianzhong1，ANQuan1，YUJianming1,andCHENLong2.OilGeophysicalProspecting,2022,57(1):111-117.

On the basis of calculating effective normal moveout (NMO) velocity of layered media with dipping interfaces using the derived analytic solution of NMO-velocity surface of the reflected wave in single-layer TI media with a dipping interface, rapidly calculating the travel time of reflected waves from common middle point (CMP) and common shot point (CSP) gathers is realized with the travel time-distance relation of CMP under conventional spread (the ratio of the maximum spread length to the reflection depth is close to 1). This method only needs to trace one zero-offset ray in line with the normal incidence theory and achieve an effective NMO velocity ellipse (encompassing all-around NMO velocities) with the Dix average algorithm proposed by Grechka and the analytic solution of NMO-velocity surface. Thus, it has a calculation efficiency of several orders of magnitude higher than multi-azimuth, multi-offset ray tracing, suitable for travel time inversion in TI media. Numerical tests show that the relative error between the effective NMO velocity and the NMO velocity when the travel time is calculated by ray tracing of optimal fitting under the conventional spread is in the magnitude of 10-3, while the relative error between the travel time calculated from the travel time-distance relation of CMP with effective NMO velocity and the travel time obtained by ray tracing is in the magnitude of 10-4～10-3. In short, the method has high calculation accuracy and is suitable for the media without too complex structures and conventional spread. It can be extended to layered anisotropic models with smooth curved interfaces.

Keywords：TI media，conventional spread，normal moveout (NMO) velocity surface，effective NMO velocity，travel time-distance relation of reflection

1. Inner Mongolia Autonomous Region Seismolo-gical Bureau，Hohhot，Inner Mongolia 010080，China

2. New Resources Geophysical Exploration Division, BGP Inc., CNPC, Zhuozhou, Hebei 072751, China

Comparisonoffouroptimizationmethodsinelasticfull-waveforminversion.LIUYuhang1,HUANGJianping1,2,YANGJidong1,2,LIZhenchun1,2,KONGLinghang1,andDINGZhaoyuan3.OilGeophysicalProspecting，2022,57(1):118-128.

Elastic full-waveform inversion (EFWI) is a high-precision imaging method. Since it is a highly nonlinear problem in nature, local optimization algorithms are often used to solve it, but the inversion results of different optimization algorithms are very different. On the basis of the commonly used conjugate gradient (CG) method and limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) method, the pseudo-Hessian matrix is adopted as the gradient preconditioning operator to devise a preconditioning CG (P-CG) method and a preconditioning L-BFGS (P-L-BFGS) method. The principles and implementation processes of these four optimization algorithms are expounded. Then, the diffraction model and the Marmousi Ⅱ model are used to test the four algorithms. The following conclusions can be drawn from the results: ①The pseudo-Hessian preconditioning operator can compensate for the deep energy and accelerate the convergence of inversion; ②The CG method and the P-CG method are easy to implement, but they cannot suppress the multi-parameter coupling effect because they only use first-order gradients. However, the P-CG method can deliver inversion results that are slightly inferior to those of the L-BFGS method for the complex Marmousi Ⅱ mo-del. ③The implementation of the L-BFGS method and the P-L-BFGS method is more complicated. Nevertheless, because the pseudo-Hessian matrix is calculated in the inversion process, the multi-parameter coupling effect is suppressed to a certain extent. ④As for the Marmousi Ⅱ model, both the L-BFGS method and the P-L-BFGS method can obtain compressional wave (P-wave) and shear wave (S-wave) velocity models with high inversion accuracy, but density inversion will be subject to the overfitting phenomenon.

Keywords：elastic full-waveform inversion, optimization algorithm, conjugate gradient method, L-BFGS method, preconditioning operator

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

2. Pilot National Laboratory for Marine Science and Technology (Qingdao), Qingdao, Shandong 266071, China

3. Tarim Oilfield Company, PetroChina, Korla, Xinjiang 841000, China

Comparisonofregularizationmethodsforfull-waveforminversion.LIXinjie1，WANGWeihong1，GUOXuebao1，andZHANGTingjun1.OilGeophysicalProspecting,2022,57(1):129-139.

Regularization is an important way to alleviate the ill-posedness of inversion and the characteristics of constrained solutions. Tikhonov regularization and total variation (TV) regularization are two regularization methods commonly used in full-waveform inversion. They can suppress high wavenumbers and protect the formation edge respectively. Two-parameter shaping regularization, hybrid two-parameter regularization, and sparse structure constraint regularization are developed on the basis of the former two and have their advantages. To systematically demonstrate the characte-ristics of different regularization methods, this paper makes a comparative analysis of the full-waveform inversion methods constrained by the five regularization methods. The anticline-overlap model and Marmousi model tests show that diffe-rent regularization methods all improve the inversion results to various degrees. Two-parameter shaping regularization combines the advantages of Tikhonov regularization and TV regularization, which improves the deep accuracy. Hybrid two-parameter regularization further improves the accuracy of shallow inversion. Compared with other methods, sparse structure constraint regularization has obvious advantages in both stratigraphic continuity and the description of edge structure.

Keywords：full-waveform inversion，Tikhonov regularization，total variation (TV) regularization，two-parameter regularization， sparse structure constraint regularization

1. School of Earth Sciences，Northeast Petroleum University，Daqing，Heilongjiang 163318，China

Asimplified2Dpetrophysicalmodelforregularpo-lygonpores.LIUZhishui1，BAOQianzong1，LIUJunzhou2，andSHILei2.OilGeophysicalProspecting，2022,57(1):140-148.

It is common for petrophysical modeling to make pores in rock equivalent to two-dimensional (2D) ellipses or three-dimensional (3D) ellipsoids，while the modeling of other pore types is rarely studied. The Kachanove 2D model is a classical petrophysical model for regular polygon pores，which involves multiple pore shape factors and can only characterize a few pore shapes. It can hardly be combined with mathematical algorithms such as adaptive algorithms. Considering this，we introduce a single pore shape factorgto act as the equivalent of multiple shape factors in the above model. In this way，a simplified 2D petrophysical model for regular polygon pores is obtained，and the theoretical value range ofgis given. The numerical forward modeling of sandy mudstone illustrates seve-ral conclusions：Whengis greater， the elastic mo-dulus is smaller，and thus whengis closer to 1， the elastic modulus is greater. Moreover，the decline rate of the elastic modulus decreases with the growth ofg，and the change rate of the elastic mo-dulus becomes smaller as the porosity declines. In practical applications，the range ofgdoes not reach infinity. The proposed model is applied in laboratory tests on sandy mudstone and actual logging data，and the results show that the proposed model in this paper can achieve good performance in applications.

Keywords：petrophysical model，Kachanove 2D model，2D regular polygon pores，pore shape factor，sandy mudstone

1. College of Geological Engineering and Geoma-tics，Chang’an University，Xi’an，Shaanxi 710054，China

2. Research Institute of Petroleum Exploration and Development，SINOPEC，Beijing 100083，China

Apetrophysicalmodelofdual-porositymediumconsideringdiageneticconsolidationanditsapplication.GAOQiang1，LIHongbing1，andPANHaojie2.OilGeophysicalProspecting,2022,57(1):149-158.

Diagenetic consolidation，pore structure，and pore structure type are important factors affecting the elastic properties of tight oil and gas reservoirs. However，most of the existing petrophysical models，focusing on a single pore type or only considering diagenesis，fail to accurately describe the characteristics of tight oil and gas reservoirs such as low porosity and low permeability，uneven fluid distribution，and complex pore structures and pore structure types. Therefore，in this paper，the differential equivalent medium (DEM) theory and the Pride model are fully used to deduce a petrophysical model of a dual-porosity medium considering diagenetic consolidation. Consolidation parameters and the aspect ratios and volume percentages of soft and hard pores are introduced for the joint characterization of diagenesis and microscopic pore structures. The effects of porosity，consolidation parameters，pore aspect ratio，and volume ratio of soft pores on compressional wave velocity，shear wave velocity，and elastic modulus are quantitatively analyzed with this model. The validity and applicability of the model are verified by the experimental data of sandstone and mudstone under different pressures and the experimental data of the Sulige area. Finally，the model is applied to the prediction of shear wave velocity in Sulige tight sandstone gas reservoirs. Compared with the prediction results of other petrophysical models，the shear wave velocity predicted by this model has the smallest error，confirming that the proposed petrophysical model has high prediction accuracy of shear wave velocity in tight sandstone reservoirs.

Keywords：shear wave velocity prediction，dual-porosity structure，petrophysics，differential equi-valent medium (DEM) model，Pride model

1. Department of Geophysical Exploration Technology，Research Institute of Petroleum Exploration & Development，PetroChina, Beijing 100083，China

2. College of Geophysics and Petroleum Resources，Yangtze University，Wuhan，Hubei 430110，China

Recognitiontechnologyintegratingloggingandseismicdataforthinsandreservoirinnarrowchannelanditsapplication：TakingtheAGLareainwesternDaqingplacanticlineasanexample.YANGChun-sheng1,2，JIANGYan2，SONGBaoquan2，WANGGaowen2，andZHANGXiuli2.OilGeophysicalProspecting，2022,57(1):159-167.

A small-scale sand body reservoir develops in the subaqueous distributary channel in front of the delta in the Saling oil group of the AGL area in the western Daqing placanticline. The delta front witnesses the thin interbedded deposition of sand shale. The channel is narrow, and sand bodies feature thin thickness and fast sedimentary facies transition. These make reservoir prediction challenging. In response, this paper proposes a recognition technology integrating logging and seismic data for the thin sand reservoir in the narrow channel of the study area. First, three seismic response models are established for the sand bodies in the narrow channel through the analysis of the forward model close to the underground geologic structure with seismic data under the control of a fine stratigraphic framework. Then according to the sequence from point to plane and then to volume, we calibrate the reflection characteristics of sand bodies in the narrow channel with logging data, trace the boundary and trend of the channel with the seismic attribute in the plane, analyze the sedimentary period and evolution law of the narrow channel on the three-dimensional (3D) seismic data volume, and qualitatively describe the distribution characteristics of the narrow channel. Finally, depending on the known seismic reflection waveform characteristics in wells and sandstone thickness, we perform the quantitative thickness prediction regarding the thin sand reservoir in the channel based on waveform pattern recognition with the correlation analysis method. This technology is applied to deploy 9 development wells in TA2 block without well control in the AGL area in western Daqing placanticline. After drilling, the consistency of channel prediction reaches 100%. The relative error of sand body thickness is 9.6%, and the maximum daily oil production of a single well reaches 4.1t. This research is capable of guiding the hydrocarbon exploration and development in the surrounding areas of old oilfields and achieves the goal of increasing reserves and production.

Keywords：delta front sub-facies，sand bodies in narrow channel，recognition technology integrating logging and seismic data，meticulous depiction of sand body，Daqing placanticline

1. School of Earth Sciences, Northeast Petroleum University，Daqing，Heilongjiang 163318，China

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

Anewhigh-resolutiontime-frequencyanalysisme-thodbasedonWigner-VilledistributionandChrip-Ztransform.LISiyuan1andXUTianji1.OilGeophy-sicalProspecting,2022,57(1):168-175,211.

The time-frequency analysis method has been widely applied in the processing and interpretation of seismic signals for a long time. At present，the oil and gas exploration is exposed to problems such as the small trap size，thin reservoir thickness，and scattered distribution of rich resources. In addition，with the narrow frequency band and low resolution of effective signals in seismic data，high-resolution time-frequency analysis methods are especially required to improve the identification accu-racy of small-scale oil and gas targets. The bilinear time-frequency analysis method of smoothed pseudo Wigner-Ville distribution (SPWVD) has good time-frequency focusing，and the spiral sampling interpolation characteristic of Chrip-Z transform (CZT) can highlight the local details of 3D space. By capitalizing on the advantages of the two，we present a new method，i.e.，the SPWVD-CZT method，to improve the time-frequency resolution of seismic signals. The interpolation calculations of signals are achieved by sampling to raise the number of effective frequency division points，improve the local details of time-frequency distribution，and realize the refinement of spectrum information in the time-frequency domain. The application of si-mulated signal calculation and channel microfacies identification of actual seismic signals shows that the method can effectively improve the time-frequency resolution of seismic signals and develop a new theory for the discovery of small-scale geological bodies and oil-gas development.

Keywords： time-frequency analysis，small-scale geological body，seismic，high resolution，smoothed pseudo Wigner-Ville distribution (SPWVD)，Chrip-Z transform (CZT)，SPWVD-CZT

1. School of Resources and Environment，University of Electronic Science and Technology of China，Chengdu，Sichuan 611731，China

Studyoftime-varyingreservoirpermeabilitybasedontime-lapseseismicdata.GUOQi1,2，ZHUANGTianlin3，HEShumei3,LIZhen1，WEIPengpeng3，andLIULijie1.OilGeophysicalProspecting,2022,57(1):176-183.

In the high-water-cut development stage, the prediction of reservoir permeability parameters is the key to tapping the potential of remaining oil. The conventional permeability data is obtained by logging interpretation. However, due to the influence of reservoir development and water injection, the logging data of production wells in the late development period cannot truly reflect the permeability of reservoirs in their original state. With the time-lapse seismic data of different water cut stages, seismic attributes are extracted and screened. As a result, five time-lapse seismic attributes are selected for the establishment of a data set, i.e., root-mean-square amplitude, reflection intensity, instantaneous frequency, amplitude envelope and dominant frequency. The relationship between time-lapse seismic data and logging permeability data is built by a partial least squares regression model, and then the permeability of production wells in low-, high- and ultra-high-water-cut development stages is further predicted, which yields a three-stage permeability model for the whole area. The distribution of permeability change multiples is verified by tracer data. The results show that the permeability change trend in the well group is consistent with tracer data. The established reservoir permeability prediction method can objectively reflect the permeability changes caused by water injection. The new method provides a basis for reservoir parameter prediction in different development stages.

Keywords：time-lapse seismic, permeability model, time-varying reservoir parameters, partial least squares regression, attribute optimization

1. Exploration and Development Research Institute, Shengli Oilfield Company, SINOPEC, Dong-ying, Shandong 257015, China

2. Working Station for Postdoctoral Scientific Research, Shengli Oilfield Company, SINOPEC, Dongying, Shandong 257000, China

3. Exploration and Development Research Institute, Dagang Oilfield Company, PetroChina, Tianjin 300280, China

Identificationofmicrofaultopeninginsweet-spotmemberofLucaogouFormationinJimusarSagofJunggarBasinbymaximumpositiveandnegativecurvature.LIWei1,CHENGang1,2,WANGDong-xue3,HANBao1,WANGZhenlin1,2,andQIHongyan1.OilGeophysicalProspecting,2022,57(1):184-193.

Micro faults are richly developed in the sweet-spot member of Lucaogou Formation in the Jimusar Sag, which seriously affects the development effect. In the past, only conventional logging, imaging logging, and field outcrops were used to identify micro faults, which was limited to single-well fracture identification, and the plane distribution patterns of micro faults could not be effectively obtained. At present, the geometric characteristics of waveforms calculated by pre-stack/post-stack seismic data are mainly employed to identify faults, but this method has high requirements for the signal-to-noise ratio of seismic data and can hardly predict the opening of faults. Therefore, this paper proposes to obtain the maximum positive and negative curvature attributes by the Kalman filter technology based on horizontal and vertical combinations. Meanwhile, the fault opening is studied in combination with the information of real drilling lost circulation, fracturing disturbance, and formation dip angles. The specific process is as follows: Firstly, the Kalman filter technology based on horizontal and vertical combinations is applied to filter the original seismic data. Due to the weak fault anisotropy, the filtered seismic data is not processed in different directions to ensure that the data has a high signal-to-noise ra-tio, and the maximum positive and negative curvature volumes are directly obtained. Secondly, the plane curvature attribute of the sweet-spot member is extracted by using structural interpretation layers to identify faults and fault plane combinations. Thirdly, according to the identified fault, the well trajectory data near the breakpoint such as drilling lost circulation, fracturing disturbance, and formation dip angles, are counted. Finally, the relationships between the curvature type, curvature value, and fault strikes and drilling lost circulation, fracturing disturbance, and formation dip angles are calculated. The following conclusions are drawn: ①Most of the NE-SW trending faults with the maximum positive curvature of greater than 500ft-1are open faults, and most of the NW-SE and NE-SW trending faults with the maximum positive curvature of about 300ft-1are semi-open faults; most of the near NS and NE-SW trending faults with the maximum negative curvature absolute value of greater than 300ft-1are closed faults. ②The formation dip angle near the fault changes greatly, and a greater curvature value leads to a greater angle change.

Keywords：curvature attribute, signal-to-noise ratio, Kalman filter, fault opening, formation dip angle

1. Research Institute of Exploration and Development, Xinjiang Oilfield Company, PetroChina, Karamay, Xinjiang 834000, China

2. Unconventional Oil and Gas Science and Technology Research Institute, China University of Petroleum (Beijing), Beijing 102249, China

3. Department of Reservoir Evaluation, Xinjiang Oilfield Company, PetroChina, Karamay, Xinjiang 834000, China

Quantitativepredictionofdeepeffectivesandstonereservoiringeomorphologicandtectoniccomplicatedarea:AnexamplefromBashenjiqikeFormationofMiddleQiulitageAreainTarimBasin.XUZhaohui1,XUZhenping2,ZHANGRonghu3,WANGLu1,HUZaiyuan4,andQINLianbin5.OilGeophy-sicalProspecting,2022,57(1):194-205.

The high and steep “blade mountain” topography in the Middle Qiulitage Area makes it difficult to acquire and process seismic data. The complex subsurface structure of the foreland thrust belt leads to poor seismic imaging and low signal-to-noise ratios. The above-mentioned "dual complexity" situation has caused a strong ambiguity in seismic interpretation and many interference factors in the prediction of sedimentary reservoirs. To this end, on the basis of the newly processed post-stack depth-domain 3D seismic and drilling/logging data, we use seismic sedimentology to qualitatively restore the distribution and evolution of lithofacies in three fault walls, respectively. The principal component analysis is utilized to convert seismic attributes into principal components. Through the fitting of the accumulated thickness of the effective reservoir (porosity>6%) measured at well points, the effective reservoir thickness of the target interval is calculated. Then main controlling factors of effective reservoir distribution are discussed. The following conclusions were obtained: ①The target interval is dominated by medium-thick sandstone, showing low frequency on the blended profile and high frequency in a small amount of thin sandstone or mudstone. ②On the large-area distribution of sandstone, the lateral distribution of effective reservoirs is heterogeneous, in which the hanging wall has good continuity and large thickness (50～65m); the middle wall has poor continuity with thickness of 45～75m; the effective reservoir of the foot wall has medium continuity and small thickness (50～60m). ③Lithofacies is the main factor that controls the lateral continuity and thickness of effective reservoirs, and the northeast-southwest trending faults also regulate the distribution of effective reservoirs. In this paper, the application scope of seismic sedimentology is extended from the shallow layers in areas with simple structures to deep layers in the “dual complexity”, namely surface and subsurface areas. The paper also provides a reference for effective reservoir prediction in areas with similar geologic backgrounds.

Keywords：Tarim Basin, Middle Qiulitage area, Bashenjiqike Formation, principal component analysis, quantitative prediction, seismic sedimentology

1. Research Institute of Petroleum Exploration & Development, PetroChina, Beijing 100083, China

2. Tarim Oilfield Company, PetroChina, Korla, Xinjiang 841000, China

3. PetroChina Hangzhou Research Institute of Geology, Hangzhou, Zhejiang 310023, China

4. Northwest Branch, Research Institute of Petroleum Exploration & Development, Lanzhou, Pe-troChina, Gansu 730020, China

5. Xinjiang Tadong Oil and Gas Exploration and Development Co., Ltd, Daqing Oilfield Company, PetroChina, Korla, Xinjiang 841000, China

Applicationofscatteringimaginginsmall-scalefracturecavityrecognition：AcasestudyofDengyingFormationincentralSichuanpaleo-uplift.JIANGXiaoyu1，SONGTao1，GANLideng1，DAIXiaofeng1，DINGQian1，andZHOUXiaoyue1.OilGeophysicalProspecting,2022,57(1):206-211.

The response characteristics of small-scale Karst fracture-cavity units in Dengying Formation of the central Sichuan paleo-uplift are not distinct on the seismic profile, which makes the seismic characterization difficult for fracture cavities in Dengying Formation. To strengthen the seismic reflection characteristics and predict the distribution of small-scale fracture cavities, we perform the scattering imaging on the full-azimuth common dip-angle gather with the full-azimuth local angle-domain imaging method. This intensifies the seismic reflection characteristics of small-scale fracture cavities. According to the strong energy spectrum characteristics of the small-scale fracture cavities in the scattering imaging, the energy body attributes are extracted from scattering imaging data for the description of the fracture cavity development cha-racteristics. The following conclusions are obtained: ①The full-azimuth local angle-domain imaging method is more effective than the conventional Kirchhoff migration imaging method; ②The scattering imaging data volume can enhance the seismic reflection characteristics of discontinuous geolo-gical bodies and reflect the geological phenomena such as small-scale fracture cavities; ③With the reservoir in Dengying Formation in the central Sichuan paleo-uplift as an example, the energy body attribute features extracted from the scattering imaging data volume are better than those extracted from the conventional post-stack seismic data in the reflection of the distribution of discontinuous geological bodies such as fracture cavities.

Keywords：angle-domain imaging，scattering imaging，Dengying Formation in central Sichuan，energy body，small-scale fracture cavity

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

LoggingcharacteristicsofCretaceoussequencestratigraphyandthegenesisanddistributionoflow-resistivityoilandgaslayersinBongorBasin,Chad.MAOZhiqiang1,JIANGZhihao2,LIChangwen3,4,LINGHUSong3,4,andZHANGLili3,4.OilGeophysicalProspecting,2022,57(1):212-221.

After multi-stage structural evolution of faults and depressions in Bongor Basin, differences occur in oil-bearing properties of different fault depressions, and there are also unconventional oil and gas reservoirs. The resistivity of oil and gas reservoirs varies widely, and it is difficult to identify oil and gas reservoirs with low resistivity and low contrast. Therefore, according to the data of core, logging, oil test and geochemical analysis, this paper analyzed the logging characteristics of Cretaceous sequence stratigraphy, as well as the genesis and distribution of low-resistivity oil and gas reservoirs, with Baobab Block and Daniela Block in Bongor Basin as the main research objects. The results show that: ① The logging responses of R-K, M and P Formations are obviously different. The mudstone of M-P Formation develops source rocks with thin organic matter layers, and its logging characteristics are typical and easy to distinguish. M Formation in Baobab, Daniela and other blocks exhibit rapid sedimentary and overpressure logging. ② Self-generation and self-storage P Formation is the main oil-bearing formation of Cretaceous system. ③ There are two types of low-resistivity oil and gas reservoirs in M-P Formation (The first type is the thin sand layer or thin argillaceous sand layer, and the second type is the unconventional oil and gas reservoir.) and one type of low-contrast oil and gas reservoir. The low-resistivity oil and gas reservoirs are mainly developed at the bottom of M Formation or the top of P Formation, showing inside source or near source accumulation and possessing extremely high natural productivity under formation overpressure. The low-contrast oil and gas reservoirs are mainly developed in the middle and lower part of the main oil layer of P Formation. The water layer of the corresponding formation is marked with flooding, and the resistivity is generally high, close to or even higher than the that of adjacent oil layer.

Keywords：Bongor Basin, sequence stratigraphy, logging response characteristics, low-resistance oil and gas layers, low-contrast oil and gas reservoirs

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

2. School of Earth Sciences and Engineering, Xi’an Shiyou University, Xi’an, Shaanxi 710065, China

3. CNPC Logging International Division, Beijing 102206, China

4. Well Logging Key Laboratory, CNPC, Xi’an, Shaanxi 710077, China

Influenceofundulatingterrainonthree-dimensionalcontrolled-sourceelectromagneticresponse.SHANGXiaorong1,2,YUEMingxin1,2,3,YANGXiaodong1,2,WUXiaoping1,2,andLIYong3.OilGeophysicalProspecting,2022,57(1):222-236.

The controlled-source electromagnetic (CSEM) method is one of the important means for oil, gas, and mineral resource exploration. At present, the conventional interpretation of CSEM data is usually based on the assumption of flat terrain, which leads to the distortion of the positions and shapes of inversion anomalies and even false anomalies. To explore the influence of undulating terrain on the propagation of the three-dimensional (3D) CSEM field and enhance data processing and interpretation, this paper applies the vector finite element algorithm based on unstructured grids to carry out numerical simulation of the 3D CSEM field under undulating terrain and discusses the influence of terrain on each component of the CSEM field. For this purpose, a layered medium model and a simple anomaly model are constructed for digital simulation, and the results prove the correctness and timeliness of the algorithm. Then, the effects of undulating terrain on each component of the CSEM field are discussed in detail with multiple 3D undulating terrain models. Finally, the electromagnetic response characteristics of the metal mine model under actual complex terrain are analyzed, which proves the practicability of the proposed algorithm.

Keywords：controlled-source electromagnetic method, terrain influence, electromagnetic component, vector finite element, 3D forward

1.School of Earth and Space Sciences, University of Science and Technology of China, Hefei, Anhui 230026, China

2.Anhui Mengcheng National Geophysical Observatory, University of Science and Technology of China, Mengcheng, Anhui 233500, China

3.Key Laboratory of Geophysical Electromagnetic Probing Technologies of Ministry of Natural Resources, Institute of Geophysical and Geochemical Exploration, Chinese Academy of Geological Sciences, Langfang, Hebei 065000, China

Simulationofdeep-seareservoirdevelopmentmonitoringusingmarinecontrolled-sourceelectromagne-ticmethod.LIUYong1,LIWenbin1,DENGFangshun1,CHENHangyu2,DINGXuezhen2,andLIUYing2,3.OilGeophysicalProspecting,2022,57(1):237-244.

The marine controlled-source electromagnetic (MCSEM) method has application values in monitoring the development process of reservoirs. On the basis of two-dimensional (2D) marine geoelectric models, we use a 2D forward modeling algorithm of marine controlled sources to calculate the electromagnetic response of different geoelectric models. Simulation analysis is carried out for the influencing factors of reservoir development monitoring by MCSEM under different environment and working modes. In addition, we compare three different observation methods, namely, emitting and receiving at the sea botttom, emitting at the sea bottom and receiving in the well, and emitting and receiving in the well. The results show that the emission frequency, reservoir thickness, burial depth of the reservoir, and depth of seawater all affect the monitoring effect. The observation method that places the receivers in the well has better performance in reservoir development monitoring, which can highlight the electromagnetic response sourced from the high-resistivity reservoir.

Keywords：marine controlled-source electromagnetic method, reservoir monitoring, 2D forward mode-ling, simulation

1. Laboratory of Low Frequency Electromagnetic Communication Technology with the 722 Research Institute, CSSC, Wuhan, Hubei 430079, China

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

3. Key Lab of Submarine Geosciences and Prospecting Techniques of Ministry of Education, Qing-dao, Shandong 266100, China