Imaging technologies, such as reverse-time migration (RTM), generate migration images for complex structures. However, the migrations can be distorted due to undersampled acquisition geometry, limited recording aperture and illumination effects. In order to partially correct the distortions, we present multi-iteration image-domain least-squares migration (LSM) by approximating the Hessian through point spread functions (PSFs). Since information in a migration image is insufficient to uniquely determine a reasonable reflectivity model and LSM itself can cause artifacts, we include an L p -norm regularization of the difference between the migration image and the reflec-tivity model and a total variation (TV) regularization of the reflectivity model. The L p -norm regularization reduces artifacts and makes the reflectivity model to be similar to the migration image, while the TV regularization helps to maintain structural continuity. We illustrate the LSM through RTM. Given a migration velocity model, synthetic data for a grid of scatterer points is generated through Born modeling operators and, then, is migrated to calculate PSFs. A reflectivity model is convolved with the PSFs to match a migration image. We use a nonlinear conjugate gradient method to seek an optimal reflectivity model. Preliminary results from numerical examples show that the scheme is helpful to improve resolution and reduce artifacts in image domain and broadens spectrum in wavenumber domain.
Summary Correctly injecting recorded multicomponent seismic data to generate receiver-side wavefield is a major step in elastic reverse time migration (RTM). In land reflection survey, the data are often recorded at the free surface where the traction free condition is naturally maintained. Based on the representation theorem, we propose a numerical approach to inject receiver-side signals from a free surface to reconstruct the receive-side wavefield. By injecting the traction-free time-reversed multicomponent seismic records from the surface, the method can generate correct down-going wavefields, which can further be used in the elastic RTM. Numerical examples are used to validate the injection and wavefield reconstruction procedure. Synthetic dataset is calculated in a model with a free surface boundary condition, and then re-injected to the model with a traction-free boundary condition. The reconstructed wavefields are carefully compared to the original forward propagated waves to evaluate the accuracy of the boundary treatment. By combining the regenerated receiver-side wavefield, the source-side wavefield and a vectorized image condition, we also tested the multicomponent elastic RTM. The PP, PS, SP and SS images are calculated.
In the last decade, important oil and gas deposits were discovered in complex mountainous areas in Tarim basin, western China, but many of them were accompanied by serious HSE (health, safety and environment) risks due to poor topographic conditions there. Extremely complex terrains limited ground access and made deployment of sources and receivers difficult, in particular for the dynamite-source seismic exploration, where transportation of drilling rigs was a tremendous challenge and raised serious safety concerns. It is long expected to reduce the exposure of field personnel to hazardous situations and environments by significantly reducing the fieldwork load and reasonably planning implementation route, both reducing the construction risk and improving the survey efficiency. With the application of high precision remote sensing information, especially the DEM (digital elevation model), many important results were achieved. In this paper, we present a terrain risk classification method based on the DEM, and introduce its applications, such as optimized deployment of source and receivers, planning survey routes and lift locations of helicopters, in mountainous areas of western China.
PreviousNext No AccessSEG Technical Program Expanded Abstracts 2020Predicting mineralogy using a deep neural network and fancy PCAAuthors: Dokyeong KimJunhwan ChoiDowan KimJoongmoo ByunDokyeong KimHanyang UniversitySearch for more papers by this author, Junhwan ChoiHanyang UniversitySearch for more papers by this author, Dowan KimHanyang UniversitySearch for more papers by this author, and Joongmoo ByunHanyang UniversitySearch for more papers by this authorhttps://doi.org/10.1190/segam2020-3426151.1 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractMineralogy is strongly related to the rock properties of reservoir formations. To evaluate mineralogy, the core analysis is generally conducted. Core data can be directly measured. However, it is uneconomical to acquire cores continuously for all depth intervals. On the other hand, the additional logs give continuous measurement to estimate the mineralogy. However, it is not easy to discriminate the various mineral compositions with these logs. A deep neural network (DNN), which is one of machine learning methods, has actively been implemented to geophysical problems. It can establish relationships among multiple nonlinear features. In this study, we propose a DNN model to predict the weight fractions of minerals from conventional log data and X-ray diffraction results analyzed using core samples. To prevent overfitting from limited training data, Fancy principal component analysis was adopted to augment training data before training the DNN model. Blind test was carried out to verify the effectiveness of the trained DNN model. The trained DNN model is reliable and cost effective, demonstrating applicability to the prediction of mineralogy. Presentation Date: Tuesday, October 13, 2020Session Start Time: 9:20 AMPresentation Time: 11:00 AMLocation: Poster Station 9Presentation Type: PosterKeywords: core-log integration, machine learning, reservoir characterizationPermalink: https://doi.org/10.1190/segam2020-3426151.1FiguresReferencesRelatedDetailsCited byThe Estimation of Magnetite Prospective Resources Based on Aeromagnetic Data: A Case Study of Qihe Area, Shandong Province, China23 March 2021 | Remote Sensing, Vol. 13, No. 6 SEG Technical Program Expanded Abstracts 2020ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2020 Pages: 3887 publication data© 2020 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 30 Sep 2020 CITATION INFORMATION Dokyeong Kim, Junhwan Choi, Dowan Kim, and Joongmoo Byun, (2020), "Predicting mineralogy using a deep neural network and fancy PCA," SEG Technical Program Expanded Abstracts : 2315-2319. https://doi.org/10.1190/segam2020-3426151.1 Plain-Language Summary Keywordscore-log integrationmachine learningreservoir characterizationPDF DownloadLoading ...
PreviousNext No AccessSEG Technical Program Expanded Abstracts 2020Hessian-based multiparameter fractional viscoacoustic full-waveform inversionAuthors: Guangchi XingTieyuan ZhuGuangchi XingPennsylvania State UniversitySearch for more papers by this author and Tieyuan ZhuPennsylvania State UniversitySearch for more papers by this authorhttps://doi.org/10.1190/segam2020-3426699.1 SectionsSupplemental MaterialAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractRecent progress on fractional modeling enables incorporatingseismic attenuation into wavefield simulation in an accurate and effcient way. But its inverse problem, i.e., the multiparameter viscoacoustic full waveform inversion (FWI), still suffers from various issues, especially the crosstalk between velocity and attenuation. In this study, we integrate the Hessian information via the Newton-CG framework and develop the multiparameter fractional viscoacoustic FWI algorithm. It significantly mitigates the crosstalk problems and sheds light upon simultaneous inversion for both velocity and Q models.Presentation Date: Wednesday, October 14, 2020Session Start Time: 1:50 PMPresentation Time: 3:05 PMLocation: Poster Station 3Presentation Type: PosterKeywords: full-waveform inversion, viscoelastic, attenuation, time-domain, crosswellPermalink: https://doi.org/10.1190/segam2020-3426699.1FiguresReferencesRelatedDetailsCited byPropagating Seismic Waves in VTI Attenuating Media Using Fractional Viscoelastic Wave Equation16 April 2022 | Journal of Geophysical Research: Solid Earth, Vol. 127, No. 4Decoupled Fréchet kernels based on a fractional viscoacoustic wave equationGuangchi Xing and Tieyuan Zhu6 December 2021 | GEOPHYSICS, Vol. 87, No. 1Crosswell Seismic Imaging Using Q-Compensated Viscoelastic Reverse Time Migration With Explicit StabilizationIEEE Transactions on Geoscience and Remote Sensing, Vol. 60Data assimilated time-lapse viscoacoustic full-waveform inversion for monitoring CO2 geological storageChao Huang and Tieyuan Zhu1 September 2021 SEG Technical Program Expanded Abstracts 2020ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2020 Pages: 3887 publication data© 2020 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 30 Sep 2020 CITATION INFORMATION Guangchi Xing and Tieyuan Zhu, (2020), "Hessian-based multiparameter fractional viscoacoustic full-waveform inversion," SEG Technical Program Expanded Abstracts : 895-899. https://doi.org/10.1190/segam2020-3426699.1 Plain-Language Summary Keywordsfull-waveform inversionviscoelasticattenuationtime-domaincrosswellPDF DownloadLoading ...
PreviousNext No AccessSEG Technical Program Expanded Abstracts 2020Component analysis of 3D elastic 9C full-waveform inversion: Ettlingen Line case studyAuthors: T. M. IrnakaR. BrossierL. MetivierT. BohlenY. PanT. M. IrnakaUniversity of Grenoble Alpes, ISTerre, and Universitas Gadjah MadaSearch for more papers by this author, R. BrossierUniversity of Grenoble Alpes, ISTerreSearch for more papers by this author, L. MetivierUniversity of Grenoble Alpes, CNRS, LJK, ISTerreSearch for more papers by this author, T. BohlenKarlsruhe Institute of TechnologySearch for more papers by this author, and Y. PanKarlsruhe Institute of TechnologySearch for more papers by this authorhttps://doi.org/10.1190/segam2020-3426088.1 SectionsSupplemental MaterialAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractFull Waveform Inversion (FWI) is one of the most popular seismic imaging techniques. In the exploration scale, FWI has become one of the industrial standards and proven to be accurate. Following that trend, FWI in shallow seismic scale also starts to gain attraction in the past decade. Several publications have demonstrated and proposed workflow to tackle the challenges in shallow seismic scales, such as sparse and limited acquisition, weak signal to noise ratio, high complexity propagation due to the strong elastic effect, and strong attenuation. In this research, we focus on the analysis of the effect of multicomponent data in shallow seismic scale towards 3D elastic FWI. The experiment’s target is the Ettlingen Line (EL), a defensive trench-line which was built by the German Troop in 1707, located at Rheinstetten, Germany. We perform both synthetic (using cartesian direction’s source) and field data (using Galperin source) in order to demonstrate the effect of multi-component data on FWI. Sixteen component combinations are analyzed for each case. By doing so, we find out that incorporating multi-component data generally has a positive impact on FWI in terms of model and data misfit, especially if the horizontal components are taken into account. The importance of each component can be used for future shallow seismic acquisition design for FWI with similar conditions.Presentation Date: Tuesday, October 13, 2020Session Start Time: 9:20 AMPresentation Time: 9:45 AMLocation: Poster Station 3Presentation Type: PosterKeywords: 3D, full-waveform inversion, multicomponent, archaeology, shallowPermalink: https://doi.org/10.1190/segam2020-3426088.1FiguresReferencesRelatedDetails SEG Technical Program Expanded Abstracts 2020ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2020 Pages: 3887 publication data© 2020 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 30 Sep 2020 CITATION INFORMATION T. M. Irnaka, R. Brossier, L.Metivier, T. Bohlen, and Y. Pan, (2020), "Component analysis of 3D elastic 9C full-waveform inversion: Ettlingen Line case study," SEG Technical Program Expanded Abstracts : 795-799. https://doi.org/10.1190/segam2020-3426088.1 Plain-Language Summary Keywords3Dfull-waveform inversionmulticomponentarchaeologyshallowPDF DownloadLoading ...
PreviousNext No AccessSEG Technical Program Expanded Abstracts 2020Using principal component analysis to decouple seismic diffractions from specular reflectionsAuthors: Raanan DafniRina SchwartzRonit LevyZvi KorenRaanan DafniEmersonSearch for more papers by this author, Rina SchwartzEmersonSearch for more papers by this author, Ronit LevyEmersonSearch for more papers by this author, and Zvi KorenEmersonSearch for more papers by this authorhttps://doi.org/10.1190/segam2020-3421255.1 SectionsSupplemental MaterialAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractMethods developed for diffraction imaging are commonly based on a separation technique between continuous and discontinuous structural elements, based on their seismic response. Seismic imaging in the dip-angle domain decomposes direction-dependent images where the response of different structural elements is clearly distinguished. The subsurface structural geometry dictates a local dip-angle seismic signature of preferable scattering directions, accordingly. Assuming these signatures are uncorrelated, we propose a practical workflow for dip-angle domain principal component analysis as a comprehensive structural feature separator. Once the dip-angle images are transformed into their principal components, a back projection yields structural-specific images of the subsurface main building blocks. Our proposal emphasizes the strength of principal component analysis as a producer of probable geologic features from pre-stack dip-angle data. These should be incorporated as structural attributes to enhance seismic interpretation, reduce uncertainty and enable automation.Presentation Date: Wednesday, October 14, 2020Session Start Time: 1:50 PMPresentation Time: 3:55 PMLocation: 362DPresentation Type: OralKeywords: diffraction, common angle, interpretation, prestack, seismic attributesPermalink: https://doi.org/10.1190/segam2020-3421255.1FiguresReferencesRelatedDetailsCited bySeismic lineaments characterization using independent component analysisAbdulmohsen AlAli, Yazeed Altowairqi, Constantinos Tsingas, and Ali AlSultan15 August 2022 SEG Technical Program Expanded Abstracts 2020ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2020 Pages: 3887 publication data© 2020 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 30 Sep 2020 CITATION INFORMATION Raanan Dafni, Rina Schwartz, Ronit Levy, and Zvi Koren, (2020), "Using principal component analysis to decouple seismic diffractions from specular reflections," SEG Technical Program Expanded Abstracts : 2958-2962. https://doi.org/10.1190/segam2020-3421255.1 Plain-Language Summary Keywordsdiffractioncommon angleinterpretationprestackseismic attributesPDF DownloadLoading ...
PreviousNext No AccessSEG Technical Program Expanded Abstracts 2020Wavefield reconstruction inversion via machine learned functionsAuthors: Chao SongTariq AlkhalifahChao SongKing Abdullah University of Science and TechnologySearch for more papers by this author and Tariq AlkhalifahKing Abdullah University of Science and TechnologySearch for more papers by this authorhttps://doi.org/10.1190/segam2020-3427351.1 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractWavefield reconstruction inversion (WRI) is a PDE constrained optimization problem that aims to mitigate cycle skipping in full-waveform inversion (FWI) among other potential features. WRI is often implemented in the frequency domain, and thus, requires expensive matrix inversions to reconstruct the wavefield. A recently introduced machine learning (ML) framework, called physics-informed neural networks (NNs), is used to predict PDE solutions by setting the physical laws as loss functions. These NNs have shown their effectiveness in solving the Helmholtz equation specifically for the scattered wavefield. By including the recorded data at the sensors’ locations as a data constraint, the NNs can predict the wavefields which simultaneously fit the recorded data and satisfy the Helmholtz equation for a given initial velocity model. Using the predicted wavefields, we build another independent NN to predict the velocity that fits the wavefield. In this new NN, we use spatial coordinates as input to the network, and use the scattered Helmholtz wave to define the loss function. After we train this deep neural network, we are able to predict the velocity in the domain of interest. We demonstrate the potential of the proposed method using a square anomaly model and a simple layered model, and the initial results considering single frequency data show that the ML-based WRI is able to invert for reasonable velocity models.Presentation Date: Wednesday, October 14, 2020Session Start Time: 9:20 AMPresentation Time: 11:25 AMLocation: Poster Station 1Presentation Type: PosterKeywords: machine learning, full-waveform inversion, frequency-domain, acoustic, velocity analysisPermalink: https://doi.org/10.1190/segam2020-3427351.1FiguresReferencesRelatedDetailsCited byA deep learning-enhanced framework for multiphysics joint inversionYanyan Hu, Xiaolong Wei, Xuqing Wu, Jiajia Sun, Jiuping Chen, Yueqin Huang, and Jiefu Chen27 December 2022 | GEOPHYSICS, Vol. 88, No. 1Small-data-driven fast seismic simulations for complex media using physics-informed Fourier neural operatorsWei Wei and Li-Yun Fu24 November 2022 | GEOPHYSICS, Vol. 87, No. 6Waveform inversion of seismic first arrivals acquired on irregular surfaceXiang Li, Gang Yao, Fenglin Niu, Di Wu, and Nengchao Liu4 April 2022 | GEOPHYSICS, Vol. 87, No. 3A versatile framework to solve the Helmholtz equation using physics-informed neural networks23 October 2021 | Geophysical Journal International, Vol. 228, No. 3High-dimensional wavefield solutions based on neural network functionsTariq Alkhalifah, Chao Song, and Xinquan Huang1 September 2021A modified physics-informed neural network with positional encodingXinquan Huang, Tariq Alkhalifah, and Chao Song1 September 2021Solving the frequency-domain acoustic VTI wave equation using physics-informed neural networks11 January 2021 | Geophysical Journal International, Vol. 225, No. 2 SEG Technical Program Expanded Abstracts 2020ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2020 Pages: 3887 publication data© 2020 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 30 Sep 2020 CITATION INFORMATION Chao Song and Tariq Alkhalifah, (2020), "Wavefield reconstruction inversion via machine learned functions," SEG Technical Program Expanded Abstracts : 1710-1714. https://doi.org/10.1190/segam2020-3427351.1 Plain-Language Summary Keywordsmachine learningfull-waveform inversionfrequency-domainacousticvelocity analysisPDF DownloadLoading ...
PreviousNext No AccessSEG Technical Program Expanded Abstracts 2020High-resolution velocity model building and least-squares imaging offshore Canada: A deep-water Orphan Basin exampleAuthors: Tiago AlcantaraEric FrugierBruno VirlouvetTiago AlcantaraPGSSearch for more papers by this author, Eric FrugierPGSSearch for more papers by this author, and Bruno VirlouvetPGSSearch for more papers by this authorhttps://doi.org/10.1190/segam2020-3419351.1 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractA 500 km2 proof-of-concept project in deep-water Canada demonstrated the benefits of combining recently acquired multisensor streamer data and new imaging technologies to produce a detailed velocity model and a robust seismic image that can minimize uncertainties prior to drilling. We describe the full waveform inversion (FWI) and leastsquares migration (LSM) workflow that not only resulted in high-resolution seismic sections, with improved fault definition and signal to noise (S/N) ratio in the Cretaceous and Jurassic sections, but also produced AVO-compliant image gathers that can benefit quantitative interpretation.Presentation Date: Monday, October 12, 2020Session Start Time: 1:50 PMPresentation Time: 2:15 PMLocation: Poster Station 11Presentation Type: PosterKeywords: imaging, full-waveform inversion, least-squares migrationPermalink: https://doi.org/10.1190/segam2020-3419351.1FiguresReferencesRelatedDetailsCited byImproving seismic images for frontier exploration in East Coast Canada: Lessons learnedJun Cai, Jin Tan, Xiaojing Liu, Pengfei Dong, Timmy Dy, and Weiping Gou1 September 2021From a regional AVA screening to focused elastic attributes estimation: An integrated solution from seismic acquisition to reservoir characterizationCyrille Reiser, Elena Polyaeva, and Scott Opdyke25 September 2020 SEG Technical Program Expanded Abstracts 2020ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2020 Pages: 3887 publication data© 2020 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 30 Sep 2020 CITATION INFORMATION Tiago Alcantara, Eric Frugier, and Bruno Virlouvet, (2020), "High-resolution velocity model building and least-squares imaging offshore Canada: A deep-water Orphan Basin example," SEG Technical Program Expanded Abstracts : 2810-2814. https://doi.org/10.1190/segam2020-3419351.1 Plain-Language Summary Keywordsimagingfull-waveform inversionleast-squares migrationPDF DownloadLoading ...
PreviousNext No AccessSEG Technical Program Expanded Abstracts 2020An integrated approach to understand the failure mechanism in cement and formationAuthors: Arpita P. BathijaRoland MartinezArpita P. BathijaAramco AmericasSearch for more papers by this author and Roland MartinezAramco AmericasSearch for more papers by this authorhttps://doi.org/10.1190/segam2020-3408911.1 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractThe interaction between cement and formation is a significant concern in the oil and gas industry. This interaction can lead to failure to provide zonal isolation that can be a well control hazard. The cement-formation interface can fail by the stresses incurred during drilling, completion, stimulation or production phases. For inventing improved cements for effective zonal isolation, it is important to evaluate the bond strength of the cement-formation interface. Instead of just providing a number to evaluate cement bond strength, our method gives more detail on the failure mechanism in cement, sandstone and cement-sandstone composite specimen utilizing acoustic emission (AE), velocity and strain data at downhole stress condition. The AE count, strain and velocity help us understand the fracture mechanism that caused the failure of a specimen. The same three dominant events namely, compaction, multicracking and sliding are present in all specimen studied, yet there are differences related to the basic structure of each specimen.Presentation Date: Wednesday, October 14, 2020Session Start Time: 1:50 PMPresentation Time: 1:50 PMLocation: Poster Station 6Presentation Type: PosterKeywords: acoustic, rock core laboratory measurements, rock physics, ultrasonic, sandstonePermalink: https://doi.org/10.1190/segam2020-3408911.1FiguresReferencesRelatedDetailsCited byMethod of evaluating cement-to-formation bond strength with computed tomography image analysisArpita P. Bathija, Madhumita Sengupta, and Shannon L. Eichmann1 September 2022 | The Leading Edge, Vol. 41, No. 9Cement to formation bond strength evaluation using image analysisArpita P. Bathija, Madhumita Sengupta, and Shannon L. Eichmann1 September 2021 SEG Technical Program Expanded Abstracts 2020ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2020 Pages: 3887 publication data© 2020 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 30 Sep 2020 CITATION INFORMATION Arpita P. Bathija and Roland Martinez, (2020), "An integrated approach to understand the failure mechanism in cement and formation," SEG Technical Program Expanded Abstracts : 2535-2539. https://doi.org/10.1190/segam2020-3408911.1 Plain-Language Summary Keywordsacousticrock core laboratory measurementsrock physicsultrasonicsandstonePDF DownloadLoading ...
PreviousNext No AccessSEG Technical Program Expanded Abstracts 2020Frequency-dependent AVO analysis in fracture porous mediaAuthors: Tingting ZhangTingting ZhangDaqing Oilfield Co. Ltd. Search for more papers by this authorhttps://doi.org/10.1190/segam2020-3425401.1 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractSeismic wave induced fluid flow between mesoscale fractures and background pores is an important reason for seismic wave dispersion and attenuation in fracture reservoirs. Velocity dispersion and attenuation will have an important impact on the AVO (amplitude versus offset variations) characteristics of reflected waves in fracture reservoirs. In this paper, the Chapman model is used to describe fracture reservoir rocks, and the frequencydependent reflection coefficients of seismic wave are calculated based on the generalized propagation matrix theory. The results show that when the fractures are connected with the pores, the AVO characteristics and phase angles show complex variations related to frequency. The combinations of different elastic parameters between fracture reservoir and overburden will lead to different frequency-dependent AVO characteristics of reflected waves, and will show different amplitudes varying with frequency on the spectral decomposition profiles of poststack seismic data. This study lays a theoretical foundation for the identification of fracture reservoirs based on frequency-dependent AVO analysis and spectral decomposition technique.Note: This paper was accepted into the Technical Program but was not presented at the 2020 SEG Annual Meeting.Keywords: AVO/AVA, anisotropy, dispersion, frequency-domain, fracturesPermalink: https://doi.org/10.1190/segam2020-3425401.1FiguresReferencesRelatedDetailsCited byApplication of AVO to thickness prediction of thin sandJournal of Computational Methods in Sciences and Engineering, Vol. 22, No. 3Numerical modeling of seismic responses from fractured reservoirs in 4D monitoring — Part 1: Seismic responses from fractured reservoirs in carbonate and shale formationsVladimir Leviant, Naum Marmalevsky, Igor Kvasov, Polina Stognii, and Igor Petrov25 October 2021 | GEOPHYSICS, Vol. 86, No. 6 SEG Technical Program Expanded Abstracts 2020ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2020 Pages: 3887 publication data© 2020 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 30 Sep 2020 CITATION INFORMATION Tingting Zhang, (2020), "Frequency-dependent AVO analysis in fracture porous media," SEG Technical Program Expanded Abstracts : 365-369. https://doi.org/10.1190/segam2020-3425401.1 Plain-Language Summary KeywordsAVO/AVAanisotropydispersionfrequency-domainfracturesPDF DownloadLoading ...
Machine learning models have been widely used by geoscientists to accelerate their interpretation and highlight hidden patterns in their data. However, as the complexity of the model increases, the interpretation of the results can become quite challenging. The SHAP technique provides a measure of the importance of each of the input seismic attributes on the model’s output. We illustrate the value of the SHAP technique using a tree-based machine learning implementation trained to distinguish between Mass Transport Deposits (MTDs) and salt seismic facies in a Gulf of Mexico survey.
PreviousNext No AccessSEG Technical Program Expanded Abstracts 2020Automated identification and quantification of rock types from drill cuttingsAuthors: Youssef TamaazoustiMatthias FrançoisJosselin KherroubiYoussef TamaazoustiSchlumbergerSearch for more papers by this author, Matthias FrançoisGeoservicesSearch for more papers by this author, and Josselin KherroubiSchlumbergerSearch for more papers by this authorhttps://doi.org/10.1190/segam2020-3426273.1 SectionsSupplemental MaterialAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractIdentifying and quantifying rock types from drill cuttings is a key step in characterizing the reservoir. Today, because it is manually done, this step is subjective and time-consuming. Automating this task is crucial to gain efficiency and objectivity. Some work has been proposed in the literature, but all consider a classification approach that does not allow for quantification. Here, we propose formalizing the problem as a segmentation task and solve it with a convolutional neural network in a transfer-learning scenario. Extensive experiments have been conducted and very promising results were obtained on single lithology samples, mixed ones and even wet cuttings.Presentation Date: Wednesday, October 14, 2020Session Start Time: 1:50 PMPresentation Time: 4:45 PMLocation: 351FPresentation Type: OralKeywords: rock physics, well-log interpretation, reservoir characterizationPermalink: https://doi.org/10.1190/segam2020-3426273.1FiguresReferencesRelatedDetailsCited byAutomation in Cuttings Analysis: Futuristic Preview of Digital Enablement for Geology 10118 March 2022 SEG Technical Program Expanded Abstracts 2020ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2020 Pages: 3887 publication data© 2020 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 30 Sep 2020 CITATION INFORMATION Youssef Tamaazousti, Matthias François, and Josselin Kherroubi, (2020), "Automated identification and quantification of rock types from drill cuttings," SEG Technical Program Expanded Abstracts : 1591-1595. https://doi.org/10.1190/segam2020-3426273.1 Plain-Language Summary Keywordsrock physicswell-log interpretationreservoir characterizationPDF DownloadLoading ...
PreviousNext No AccessSEG Technical Program Expanded Abstracts 2020Multispectral aberrancyAuthors: Bin LyuJie QiFangyu LiKurt J. MarfurtBin LyuUniversity of OklahomaSearch for more papers by this author, Jie QiUniversity of OklahomaSearch for more papers by this author, Fangyu LiUniversity of GeorgiaSearch for more papers by this author, and Kurt J. MarfurtUniversity of OklahomaSearch for more papers by this authorhttps://doi.org/10.1190/segam2020-3426483.1 SectionsSupplemental MaterialAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractThe seismic aberrancy attribute measures the lateral change of curvature along a surface, which complements other seismic geometric attributes such as coherence and curvature. Aberrancy provides a means to map the trace of faults that appear as flexures that may have insufficient offset to delineate with coherence. Generally, we first compute the vector dip using the full-bandwidth seismic amplitude volume. Aberrancy is then computed as the second derivative of the vector dip in the three different directions. Due to either seismic data quality or to the underlying geology, certain spectral components of the seismic amplitude volume often appear higher quality than others, which further result in higher quality geometric attributes. We develop a multispectral aberrancy method to further improve the imaging quality of small-scale geologic features. We first decompose the full-bandwidth seismic data after data-conditioning into different spectral voices, to build the multispectral covariance matrix. Next, we compute the eigenvectors and eigenvalues from the multispectral covariance matrix, followed by the generation of inline and crossline dip volumes. Finally, we provide the multispectral aberrancy by computing the second derivative of the vector dip and rotating the coordinate system. We evaluate the proposed multispectral aberrancy method using a 3D seismic dataset imaging complex channel reservoirs.Presentation Date: Tuesday, October 13, 2020Session Start Time: 1:50 PMPresentation Time: 4:45 PMLocation: 361APresentation Type: OralKeywords: seismic attributes, spectral analysis, 3DPermalink: https://doi.org/10.1190/segam2020-3426483.1FiguresReferencesRelatedDetailsCited byMultiscale fracture prediction technique via deep learning, seismic gradient disorder, and aberrance: Applied to tight sandstone reservoirs in the Hutubi block, southern Junggar BasinZhiguo Cheng, Long Bian, Haidong Chen, Xiaotao Wang, Di Ye, and Luming He11 August 2022 | Interpretation, Vol. 10, No. 4 SEG Technical Program Expanded Abstracts 2020ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2020 Pages: 3887 publication data© 2020 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 30 Sep 2020 CITATION INFORMATION Bin Lyu, Jie Qi, Fangyu Li, and Kurt J. Marfurt, (2020), "Multispectral aberrancy," SEG Technical Program Expanded Abstracts : 1120-1124. https://doi.org/10.1190/segam2020-3426483.1 Plain-Language Summary Keywordsseismic attributesspectral analysis3DPDF DownloadLoading ...
Ground roll attenuation of land seismic data is still an outstanding and challenging problem. Deep learning is a powerful tool for separating signal from noise. Recently, supervised deeplearning-based methods have been applied to ground roll attenuation. However, they require a large set of corresponding clean seismic datasets as labels. Constructing realistic training samples for network training is an unsolved problem. To circumvent it, we proposed an unsupervised deep learning method for attenuating ground roll where no training labels are utilized. The generator network first learns self-similar features before any learning. Therefore, if the reflections are selfsimilar in the time-space domain, but the ground roll is not, the network can extract the reflections before the ground roll. To make reflections look more self-similar than the ground roll, we apply the normal moveout (NMO) correction to flatten the reflections. Access to NMO correction makes the method also model-driven. The combinations of data-driven deep learning and a model-driven procedure are critical to the success of the proposed method. We use both synthetic and field shot data to illustrate the fidelity and validity of the proposed methods. The field data example shows that our proposed method can attenuate strong scattered ground roll. Presentation Date: Tuesday, October 13, 2020 Session Start Time: 1:50 PM Presentation Time: 2:15 PM Location: Poster Station 13 Presentation Type: Poster
