Seismic reservoir characterization is of great interest for sweet spot identification, reservoir quality assessment, and geological model building. The sparsity of the labeled samples often limit the application of supervised machine learning for seismic reservoir characterization. Unsupervised learning methods, on the other hand, explore the internal structure of data and extract low-dimensional features of geologic interest from seismic data without the need for labels. We compare various unsupervised learning approaches, including the linear method PCA, manifold learning methods T-distributed Stochastic Neighbor Embedding (t-SNE) and Uniform Manifold Approximation and Projection (UMAP), and the Convolutional Autoencoder (CAE), on both the 3D synthetic and field seismic data of a deep carbonate reservoir, SW China. On the synthetic data, the low-dimensional features extracted by UMAP and CAE provide a better indication of porosity and gas saturation than traditional seismic attributes. In particular, UMAP better preserves the global structure of geological features, and show the potential of decoupling the gas saturation and porosity effects from seismic responses. We demonstrate that by joint use of multi type of seismic attributes instead of using single type of seismic attributes can better delineate the reservoir structures using unsupervised ML. On the field seismic data, UMAP can effectively characterize sedimentary facies distribution, which is consistent with the geological understanding. Nevertheless, the porosity and saturation can not be reliably identified from field seismic data using unsupervised ML, which is likely to be caused by the complex pore structures in carbonates complicating the mapping relationship between seismic responses and reservoir parameters.
{"title":"Deep carbonate reservoir characterization using multi seismic attributes:A comparison of unsupervised machine learning approaches","authors":"Luanxiao Zhao, Xuanying Zhu, Xiangyuan Zhao, Yuchun You, Minghui Xu, Tengfei Wang, Jianhua Geng","doi":"10.1190/geo2023-0199.1","DOIUrl":"https://doi.org/10.1190/geo2023-0199.1","url":null,"abstract":"Seismic reservoir characterization is of great interest for sweet spot identification, reservoir quality assessment, and geological model building. The sparsity of the labeled samples often limit the application of supervised machine learning for seismic reservoir characterization. Unsupervised learning methods, on the other hand, explore the internal structure of data and extract low-dimensional features of geologic interest from seismic data without the need for labels. We compare various unsupervised learning approaches, including the linear method PCA, manifold learning methods T-distributed Stochastic Neighbor Embedding (t-SNE) and Uniform Manifold Approximation and Projection (UMAP), and the Convolutional Autoencoder (CAE), on both the 3D synthetic and field seismic data of a deep carbonate reservoir, SW China. On the synthetic data, the low-dimensional features extracted by UMAP and CAE provide a better indication of porosity and gas saturation than traditional seismic attributes. In particular, UMAP better preserves the global structure of geological features, and show the potential of decoupling the gas saturation and porosity effects from seismic responses. We demonstrate that by joint use of multi type of seismic attributes instead of using single type of seismic attributes can better delineate the reservoir structures using unsupervised ML. On the field seismic data, UMAP can effectively characterize sedimentary facies distribution, which is consistent with the geological understanding. Nevertheless, the porosity and saturation can not be reliably identified from field seismic data using unsupervised ML, which is likely to be caused by the complex pore structures in carbonates complicating the mapping relationship between seismic responses and reservoir parameters.","PeriodicalId":55102,"journal":{"name":"Geophysics","volume":"181 2","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135679549","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Organic matter (OM) maturity is closely related to organic pores in shales. Quantitative characterization of organic and inorganic pores in shale is crucial for rock physics modeling and reservoir porosity and permeability evaluation. Focused ion beam-scanning electron microscopy (FIB-SEM) can capture high-precision three-dimensional (3D) images and directly describe the types, shapes, and spatial distribution of pores in shale gas reservoirs. However, due to the high scanning cost, wide 3D view field, and complex microstructure of FIB-SEM, more efficient segmentation for the FIB-SEM images is required. For this purpose, a multiphase segmentation workflow in conjunction with a U-Net is proposed to segment pores from the matrix and distinguish organic pores from inorganic pores simultaneously in the entire 3D image stack. The workflow is repeated for FIB-SEM datasets of seventeen organic-rich shales with various characteristics. The analysis focuses on improving the efficiency and relevance of the workflow, that is, quantifying the minimum number of training slices while ensuring accuracy and further combining the Fractal Dimension (FD) and Lacunarity (La) to study a simple and objective way of selection. Meanwhile, the computational efficiency, accuracy, and robustness to noise of the 2D U-Net model are discussed. The intersection over union (IoU) of automatic segmentation can amount to 8095% in all datasets with manual labels as ground truth. In addition, calculated by the FIB-SEM multiphase segmentation, the organic porosity (OP) is used to quantitatively evaluate the OM decomposition level. Deep learning-based segmentation shows great potential for characterizing shale pore structures and quantifying OM maturity.
{"title":"Quantitative Characterization of Organic and Inorganic Pores in Shale Based on Deep Learning","authors":"Bohong Yan, Langqiu Sun, Jianguo Zhao, Zixiong Cao, Mingxuan Li, KC Shiba, Xinze Liu, Chuang Li","doi":"10.1190/geo2023-0352.1","DOIUrl":"https://doi.org/10.1190/geo2023-0352.1","url":null,"abstract":"Organic matter (OM) maturity is closely related to organic pores in shales. Quantitative characterization of organic and inorganic pores in shale is crucial for rock physics modeling and reservoir porosity and permeability evaluation. Focused ion beam-scanning electron microscopy (FIB-SEM) can capture high-precision three-dimensional (3D) images and directly describe the types, shapes, and spatial distribution of pores in shale gas reservoirs. However, due to the high scanning cost, wide 3D view field, and complex microstructure of FIB-SEM, more efficient segmentation for the FIB-SEM images is required. For this purpose, a multiphase segmentation workflow in conjunction with a U-Net is proposed to segment pores from the matrix and distinguish organic pores from inorganic pores simultaneously in the entire 3D image stack. The workflow is repeated for FIB-SEM datasets of seventeen organic-rich shales with various characteristics. The analysis focuses on improving the efficiency and relevance of the workflow, that is, quantifying the minimum number of training slices while ensuring accuracy and further combining the Fractal Dimension (FD) and Lacunarity (La) to study a simple and objective way of selection. Meanwhile, the computational efficiency, accuracy, and robustness to noise of the 2D U-Net model are discussed. The intersection over union (IoU) of automatic segmentation can amount to 8095% in all datasets with manual labels as ground truth. In addition, calculated by the FIB-SEM multiphase segmentation, the organic porosity (OP) is used to quantitatively evaluate the OM decomposition level. Deep learning-based segmentation shows great potential for characterizing shale pore structures and quantifying OM maturity.","PeriodicalId":55102,"journal":{"name":"Geophysics","volume":"167 2","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135679400","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Acoustic single-well imaging (SWI) of logging-while-drilling (LWD) is an advanced logging method in reservoir exploration, which uses reflected waves to detect the around-borehole geological structures and quickly determines the drilling direction for enhancing the drilling-encounter ratio and reducing the drilling risk. Forward acoustic modelling is a fundamental problem for SWI in LWD. Due to the complex structures, it is a challenge to simulate the wave propagation and investigate wavefield characteristics based on the forward model. Numerical modeling is a commonly used method for calculating wavefields, however it is too computationally expensive. In this study, we propose a fast method for calculating the full reflected pressure and displacement waves (i.e., P-P, SV-SV, SH-SH, and P-SV/SV-P) in SWI of LWD considering different types of sources including arcuate, monopole and dipole transmitters. The analytical algorithm is proposed by applying the reciprocity relation between the virtual force (displacement) sources located at the receiver position and the outside-borehole virtual forces which are equivalent to the reflections from the formation interfaces. Numerical experiments show that the analytical solutions agree well with the reference solutions from 3D finite-difference time-domain method, demonstrating the accuracy and high efficiency of the analytical method. Based on the analytical solutions, we find that LWD reflected waves are much more sensitive to the azimuth than those in the wireline case, showing that the availability of LWD is important for identifying the reflector azimuth. Furthermore, to enhance the reception efficiency of reflected waves, we present the optimized LWD parameters: For slow formations, we suggest using a dipole source with dominant excitation-frequency band being from 1 kHz to 3 kHz; For fast formations, a dipole with wider excitation-frequency band from 1 kHz to 5 kHz is recommended; For all formations, recording pressure signals shows much higher reception efficiency than the displacement signals.
{"title":"A fast algorithm for simulation and analysis of wavefields in acoustic single-well imaging of logging-while-drilling considering arbitrary types of sources","authors":"Jiaqi Xu, Hengshan Hu, Bo Han","doi":"10.1190/geo2023-0177.1","DOIUrl":"https://doi.org/10.1190/geo2023-0177.1","url":null,"abstract":"Acoustic single-well imaging (SWI) of logging-while-drilling (LWD) is an advanced logging method in reservoir exploration, which uses reflected waves to detect the around-borehole geological structures and quickly determines the drilling direction for enhancing the drilling-encounter ratio and reducing the drilling risk. Forward acoustic modelling is a fundamental problem for SWI in LWD. Due to the complex structures, it is a challenge to simulate the wave propagation and investigate wavefield characteristics based on the forward model. Numerical modeling is a commonly used method for calculating wavefields, however it is too computationally expensive. In this study, we propose a fast method for calculating the full reflected pressure and displacement waves (i.e., P-P, SV-SV, SH-SH, and P-SV/SV-P) in SWI of LWD considering different types of sources including arcuate, monopole and dipole transmitters. The analytical algorithm is proposed by applying the reciprocity relation between the virtual force (displacement) sources located at the receiver position and the outside-borehole virtual forces which are equivalent to the reflections from the formation interfaces. Numerical experiments show that the analytical solutions agree well with the reference solutions from 3D finite-difference time-domain method, demonstrating the accuracy and high efficiency of the analytical method. Based on the analytical solutions, we find that LWD reflected waves are much more sensitive to the azimuth than those in the wireline case, showing that the availability of LWD is important for identifying the reflector azimuth. Furthermore, to enhance the reception efficiency of reflected waves, we present the optimized LWD parameters: For slow formations, we suggest using a dipole source with dominant excitation-frequency band being from 1 kHz to 3 kHz; For fast formations, a dipole with wider excitation-frequency band from 1 kHz to 5 kHz is recommended; For all formations, recording pressure signals shows much higher reception efficiency than the displacement signals.","PeriodicalId":55102,"journal":{"name":"Geophysics","volume":"307 1‐2","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135679082","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-11-01DOI: 10.1190/geo2023-1017-tiogeo.1
In this article, the Editor of Geophysics provides an overview of all technical articles in this issue of the journal.
在这篇文章中，地球物理学编辑提供了本刊所有技术文章的概述。
{"title":"This issue of G<scp>eophysics</scp>","authors":"","doi":"10.1190/geo2023-1017-tiogeo.1","DOIUrl":"https://doi.org/10.1190/geo2023-1017-tiogeo.1","url":null,"abstract":"In this article, the Editor of Geophysics provides an overview of all technical articles in this issue of the journal.","PeriodicalId":55102,"journal":{"name":"Geophysics","volume":"19 2","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135112336","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Himanshu Barthwal, Matthew van den Berghe, Robert Shcherbakov
Microseismic event locations and moment tensors in underground mines can provide insights into the subsurface deformation and the current state of stress. However, reliable estimation of these source parameters is rather challenging due to the high-frequency waveforms and low signal-to-noise ratio for negative magnitude events. We study microseismicity in an underground potash mine in Saskatchewan, Canada, recorded between March 1 and June 30, 2021, by a network of broadband seismometers. The active mining is carried out in low-velocity evaporites at depths of approximately 1 km below the ground level. The theoretical dispersion curves show that guided waves in the form of leaky P and P-SV/SH normal modes can exist in a 1D velocity model representing the mine geology. These guided waves are detected as high-energy dispersive arrivals on the seismograms recorded at the underground receivers. We locate the events using the arrival times of the guided waves and their mean group velocities. Most (∼80%) of the detected events cluster around the mine layout between depths of 0.95 to 1.05 km. Next, we compute moment tensors for 92 events using waveforms of guided phases. The moment tensors show non-double couple components with only 28 events having double-couple percentages greater than 50%. These events occur near the mined-out cavities with source mechanisms corresponding to layer delamination in the roof and floor or pillar yield related to the closure of cavities. No abnormal microseismicity is detected away from the mine levels in the more competent carbonate rocks above or below the evaporite formations. Thus, guided waves enable the detection of microseismic events up to large distances and can provide high-resolution event locations and moment tensor inversion. These can be interpreted in the context of local geology and mining activities to identify the dominant factors affecting microseismicity.
{"title":"Microseismic event locations and source mechanisms using dominant guided waves recorded in an underground potash mine","authors":"Himanshu Barthwal, Matthew van den Berghe, Robert Shcherbakov","doi":"10.1190/geo2023-0359.1","DOIUrl":"https://doi.org/10.1190/geo2023-0359.1","url":null,"abstract":"Microseismic event locations and moment tensors in underground mines can provide insights into the subsurface deformation and the current state of stress. However, reliable estimation of these source parameters is rather challenging due to the high-frequency waveforms and low signal-to-noise ratio for negative magnitude events. We study microseismicity in an underground potash mine in Saskatchewan, Canada, recorded between March 1 and June 30, 2021, by a network of broadband seismometers. The active mining is carried out in low-velocity evaporites at depths of approximately 1 km below the ground level. The theoretical dispersion curves show that guided waves in the form of leaky P and P-SV/SH normal modes can exist in a 1D velocity model representing the mine geology. These guided waves are detected as high-energy dispersive arrivals on the seismograms recorded at the underground receivers. We locate the events using the arrival times of the guided waves and their mean group velocities. Most (∼80%) of the detected events cluster around the mine layout between depths of 0.95 to 1.05 km. Next, we compute moment tensors for 92 events using waveforms of guided phases. The moment tensors show non-double couple components with only 28 events having double-couple percentages greater than 50%. These events occur near the mined-out cavities with source mechanisms corresponding to layer delamination in the roof and floor or pillar yield related to the closure of cavities. No abnormal microseismicity is detected away from the mine levels in the more competent carbonate rocks above or below the evaporite formations. Thus, guided waves enable the detection of microseismic events up to large distances and can provide high-resolution event locations and moment tensor inversion. These can be interpreted in the context of local geology and mining activities to identify the dominant factors affecting microseismicity.","PeriodicalId":55102,"journal":{"name":"Geophysics","volume":"655 ","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135869490","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yuxiao Ren, Jiansen Wang, Qingyang Wang, Senlin Yang
Tunnel and underground engineering construction often encounter unfavorable geology, leading to disasters such as water and mud inrushes, landslides, etc. In order to prevent geological hazards, it is important to look ahead and predict the location and distribution of adverse geology ahead of the tunnel face. This process is known as seismic forward-prospecting in tunnels, and it typically requires an accurate calculation of velocity. Seismic waveform inversion methods based on deep learning have demonstrated potential in estimating velocity from synthetic seismic data. However, the superiority of these methods over traditional ones on field data is still an area of active research. Here, we use the Pearl River Delta Water Resources Allocation Project in China as an example to develop a self-supervised learning waveform inversion method for building a reliable velocity distribution in front of the tunnel. By introducing the background velocity as large-scale information and implementing multi-scale loss functions, the previous self-supervised learning inversion method on synthetic data is improved. Additionally, the corresponding network-based workflow for field data is proposed. To demonstrate the effectiveness of the proposed method, we conducted a comparison with practical tunneling exposure, where the low-velocity zone corresponds with the fault-fractured zones and the water-flowing zones. This indicates that the results obtained from our proposed method can be used as geological guidance for safe tunneling practices. In the end, the applicability and disadvantages of the proposed deep-learning inversion method for seismic forward-prospecting in tunnels are discussed.
{"title":"Self-supervised learning waveform inversion for seismic forward-prospecting in tunnels: A case study in Pearl River Delta Water Resources Allocation Project in China","authors":"Yuxiao Ren, Jiansen Wang, Qingyang Wang, Senlin Yang","doi":"10.1190/geo2023-0113.1","DOIUrl":"https://doi.org/10.1190/geo2023-0113.1","url":null,"abstract":"Tunnel and underground engineering construction often encounter unfavorable geology, leading to disasters such as water and mud inrushes, landslides, etc. In order to prevent geological hazards, it is important to look ahead and predict the location and distribution of adverse geology ahead of the tunnel face. This process is known as seismic forward-prospecting in tunnels, and it typically requires an accurate calculation of velocity. Seismic waveform inversion methods based on deep learning have demonstrated potential in estimating velocity from synthetic seismic data. However, the superiority of these methods over traditional ones on field data is still an area of active research. Here, we use the Pearl River Delta Water Resources Allocation Project in China as an example to develop a self-supervised learning waveform inversion method for building a reliable velocity distribution in front of the tunnel. By introducing the background velocity as large-scale information and implementing multi-scale loss functions, the previous self-supervised learning inversion method on synthetic data is improved. Additionally, the corresponding network-based workflow for field data is proposed. To demonstrate the effectiveness of the proposed method, we conducted a comparison with practical tunneling exposure, where the low-velocity zone corresponds with the fault-fractured zones and the water-flowing zones. This indicates that the results obtained from our proposed method can be used as geological guidance for safe tunneling practices. In the end, the applicability and disadvantages of the proposed deep-learning inversion method for seismic forward-prospecting in tunnels are discussed.","PeriodicalId":55102,"journal":{"name":"Geophysics","volume":"22 5","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135871692","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Seismic interpretation of gas hydrates requires the assistance of rock physics. Changes in gas hydrate saturation can alter the elastic properties of formations, and this relationship can be considerably influenced by the occurrence state of gas hydrates. Pore-filling, load-bearing, and cementing types are three single gas hydrate occurrence states commonly considered in rock-physics investigations. However, many gas hydrate-bearing formations are observed to have mixed occurrence states, and their rock-physics properties do not fully conform to models of single occurrence states. We present a generalized rock-physics model for gas hydrate-bearing formations with three mixed occurrence states observed in the field or laboratory experiments: coexisting pore-filling-type and matrix-forming-type gas hydrate (case 1); pore-filling type when S h (gas hydrate saturation) < S c (critical saturation) and pore-filling + matrix-forming type when S h > S c (case 2); and matrix-forming type when S h < S c and matrix-forming + pore-filling type when S h > S c (case 3). Instead of initial porosity, the apparent porosity (the volume fraction of an effective pore filler) φ as represents the influence of occurrence states on the pore space. These three mixed occurrence states can be modeled using a unified workflow, in which the volume fractions of various gas hydrate types are expressed in general forms in terms of the apparent porosity. In addition, the model considers the effect of a pore filler on shear modulus. The proposed model is validated through calibration with real well-log data and published experimental data corresponding to five gas hydrate-bearing formations. The model effectively interprets the influences of gas hydrate saturation and occurrence state on these formations. Thus, the generalized model provides a theoretical basis for the analysis of sensitive elastic parameters and quantitative interpretation for gas hydrate reservoirs.
天然气水合物的地震解释需要岩石物理学的帮助。天然气水合物饱和度的变化会改变地层的弹性性质，这种关系受天然气水合物赋存状态的影响较大。孔隙充填型、承载型和胶结型是岩石物理研究中通常考虑的三种单一天然气水合物赋存状态。然而，许多含天然气水合物地层具有混合赋存状态，其岩石物理性质并不完全符合单一赋存状态模型。本文提出了一种具有三种混合赋存状态的含天然气水合物地层的广义岩石物理模型:孔隙填充型和基质形成型天然气水合物共存(案例1);S h(天然气水合物饱和度)<时的孔隙填充类型;S c(临界饱和度)和S h >时孔隙填充+基质形成型;S c(情况2);S h <时为基体形成型;S h >时S c与成基+充孔型;表观孔隙率(有效孔隙填料的体积分数)φ as表示赋存状态对孔隙空间的影响，而不是初始孔隙率。这三种混产状态可以使用统一的工作流程进行建模，其中各种天然气水合物类型的体积分数用表观孔隙度的一般形式表示。此外，该模型还考虑了孔隙填料对剪切模量的影响。通过实际测井资料和已发表的5个含天然气水合物地层的实验数据进行标定，验证了该模型的有效性。该模型有效地解释了水合物饱和度和赋存状态对这些地层的影响。为天然气水合物储层弹性敏感参数分析和定量解释提供了理论依据。
{"title":"Rock physics model of gas hydrate reservoir with mixed occurrence states","authors":"Cun-Zhi Wu, Feng Zhang, Pin-Bo Ding, Peng-Yuan Sun, Zhi-Guang Cai, Bang-Rang Di","doi":"10.1190/geo2023-0211.1","DOIUrl":"https://doi.org/10.1190/geo2023-0211.1","url":null,"abstract":"Seismic interpretation of gas hydrates requires the assistance of rock physics. Changes in gas hydrate saturation can alter the elastic properties of formations, and this relationship can be considerably influenced by the occurrence state of gas hydrates. Pore-filling, load-bearing, and cementing types are three single gas hydrate occurrence states commonly considered in rock-physics investigations. However, many gas hydrate-bearing formations are observed to have mixed occurrence states, and their rock-physics properties do not fully conform to models of single occurrence states. We present a generalized rock-physics model for gas hydrate-bearing formations with three mixed occurrence states observed in the field or laboratory experiments: coexisting pore-filling-type and matrix-forming-type gas hydrate (case 1); pore-filling type when S h (gas hydrate saturation) < S c (critical saturation) and pore-filling + matrix-forming type when S h > S c (case 2); and matrix-forming type when S h < S c and matrix-forming + pore-filling type when S h > S c (case 3). Instead of initial porosity, the apparent porosity (the volume fraction of an effective pore filler) φ as represents the influence of occurrence states on the pore space. These three mixed occurrence states can be modeled using a unified workflow, in which the volume fractions of various gas hydrate types are expressed in general forms in terms of the apparent porosity. In addition, the model considers the effect of a pore filler on shear modulus. The proposed model is validated through calibration with real well-log data and published experimental data corresponding to five gas hydrate-bearing formations. The model effectively interprets the influences of gas hydrate saturation and occurrence state on these formations. Thus, the generalized model provides a theoretical basis for the analysis of sensitive elastic parameters and quantitative interpretation for gas hydrate reservoirs.","PeriodicalId":55102,"journal":{"name":"Geophysics","volume":"49 ","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135872708","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Seismic data interpolation is essential in a seismic data processing workflow, recovering data from sparse sampling. Traditional and deep learning based methods have been widely used in the seismic data interpolation field and have achieved remarkable results. In this paper, we propose a seismic data interpolation method through the novel application of diffusion probabilistic models (DPM). DPM transform the complex end-to-end mapping problem into a progressive denoising problem, enhancing the ability to reconstruct complex situations of missing data, such as large proportions and large-gap missing data. The inter polation process begins with a standard Gaussian distribution and seismic data with missing traces, then removes noise iteratively with a Unet trained for different noise levels. Our#xD;proposed DPM-based interpolation method allows interpolation for various missing cases, including regularly missing, irregularly missing, consecutively missing, noisy missing, and different ratios of missing cases. The generalization ability to different seismic datasets is also discussed in this article. Numerical results of synthetic and field data show satisfactory interpolation performance of the DPM-based interpolation method in comparison with the f- x prediction filtering method, the curvelet transform method, the low dimensional mani fold method (LDMM) and the coordinate attention (CA)-based Unet method, particularly in cases with large proportions and large-gap missing data. Diffusion is all we need for seismic data interpolation.
{"title":"Generative interpolation via diffusion probabilistic model","authors":"Qi Liu, Jianwei Ma","doi":"10.1190/geo2023-0182.1","DOIUrl":"https://doi.org/10.1190/geo2023-0182.1","url":null,"abstract":"Seismic data interpolation is essential in a seismic data processing workflow, recovering data from sparse sampling. Traditional and deep learning based methods have been widely used in the seismic data interpolation field and have achieved remarkable results. In this paper, we propose a seismic data interpolation method through the novel application of diffusion probabilistic models (DPM). DPM transform the complex end-to-end mapping problem into a progressive denoising problem, enhancing the ability to reconstruct complex situations of missing data, such as large proportions and large-gap missing data. The inter polation process begins with a standard Gaussian distribution and seismic data with missing traces, then removes noise iteratively with a Unet trained for different noise levels. Our#xD;proposed DPM-based interpolation method allows interpolation for various missing cases, including regularly missing, irregularly missing, consecutively missing, noisy missing, and different ratios of missing cases. The generalization ability to different seismic datasets is also discussed in this article. Numerical results of synthetic and field data show satisfactory interpolation performance of the DPM-based interpolation method in comparison with the f- x prediction filtering method, the curvelet transform method, the low dimensional mani fold method (LDMM) and the coordinate attention (CA)-based Unet method, particularly in cases with large proportions and large-gap missing data. Diffusion is all we need for seismic data interpolation.","PeriodicalId":55102,"journal":{"name":"Geophysics","volume":"23 8","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135871373","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Surface wave exploration technology has been extensively employed in the inspection of construction engineering quality and shallow surface surveys. In order to enhance the efficiency of surface wave exploration field acquisition and achieve high precision and high-density surface wave profile imaging, a wireless distributed seismic surface wave signal acquisition system has been developed based on the principles of active source transient surface wave signal acquisition and dispersion curve calculation methods. For the purpose of achieving rapid multiple coverage signal acquisition and enhancing field work efficiency, a method for rapidly configuring Common Midpoint Signal-Couples (CMC) for multiple coverage common-shot signal acquisition has been devised, and a high-precision visualization method for dispersion curve calculation based on the CMC array has been formulated. When compared with the Multichannel Analysis of Surface Wave (MASW) method under identical conditions, the CMC array can effectively enhance surface wave dispersion curve survey station density and lateral resolution, thereby enabling High-density Analysis of Surface Wave (HASW) profile imaging. Through model analysis and field examples related to construction quality detection, including foundation compactness and earth and rock mixture compactness, it has been demonstrated that this method offers significant advantages in terms of high accuracy, high density, and a wide application range. These advantages greatly enhance the efficiency of surface wave exploration and the accuracy of profile imaging for construction engineering projects.
{"title":"High-density analysis of surface wave (HASW) profile imaging based on a multiple coverage common midpoint signal-couple array","authors":"Chen Li, Yunsheng Wang, Guozhong Gao","doi":"10.1190/geo2023-0100.1","DOIUrl":"https://doi.org/10.1190/geo2023-0100.1","url":null,"abstract":"Surface wave exploration technology has been extensively employed in the inspection of construction engineering quality and shallow surface surveys. In order to enhance the efficiency of surface wave exploration field acquisition and achieve high precision and high-density surface wave profile imaging, a wireless distributed seismic surface wave signal acquisition system has been developed based on the principles of active source transient surface wave signal acquisition and dispersion curve calculation methods. For the purpose of achieving rapid multiple coverage signal acquisition and enhancing field work efficiency, a method for rapidly configuring Common Midpoint Signal-Couples (CMC) for multiple coverage common-shot signal acquisition has been devised, and a high-precision visualization method for dispersion curve calculation based on the CMC array has been formulated. When compared with the Multichannel Analysis of Surface Wave (MASW) method under identical conditions, the CMC array can effectively enhance surface wave dispersion curve survey station density and lateral resolution, thereby enabling High-density Analysis of Surface Wave (HASW) profile imaging. Through model analysis and field examples related to construction quality detection, including foundation compactness and earth and rock mixture compactness, it has been demonstrated that this method offers significant advantages in terms of high accuracy, high density, and a wide application range. These advantages greatly enhance the efficiency of surface wave exploration and the accuracy of profile imaging for construction engineering projects.","PeriodicalId":55102,"journal":{"name":"Geophysics","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135871412","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Borehole images are measured by logging tools in a well, providing a microresistivity map of the rock properties surrounding the borehole. These images contain valuable information related to changes in mineralogy, porosity, and fluid content, making them essential for petrophysical analysis. However, due to the special design of borehole imaging tools, vertical strips of gaps occur in borehole images. We propose an effective approach to fill these gaps using a convolutional neural network with partial convolution layers. To overcome the challenge of missing training labels, we introduce a self-supervised learning strategy. Specifically, we replicate the gaps found in borehole images by randomly creating vertical blank strips that mask out certain known areas in the original images. We then employ the original images as label data to train the network to recover the known areas masked out by the defined gaps. To ensure that the missing data does not impact the training process we incorporate partial convolutions, which exclude the null-data areas from convolutional computations during both forward and backward propagation of updating the network parameters. Our network, trained by this way, can then be used to reasonably fill the gaps originally appearing in the borehole images and obtain full images without any noticeable artifacts. Through the analysis of multiple real examples, we demonstrate the effectiveness of our method by comparing it to three alternative approaches. Our method outperforms the others significantly, as demonstrated by various quantitative evaluation metrics. The filled fullbore images obtained through our approach enable enhanced texture analysis and automated feature recognition.
{"title":"Filling Borehole Image Gaps with Partial Convolution Neural Network","authors":"Lei Jiang, Xu Si, Xinming Wu","doi":"10.1190/geo2022-0344.1","DOIUrl":"https://doi.org/10.1190/geo2022-0344.1","url":null,"abstract":"Borehole images are measured by logging tools in a well, providing a microresistivity map of the rock properties surrounding the borehole. These images contain valuable information related to changes in mineralogy, porosity, and fluid content, making them essential for petrophysical analysis. However, due to the special design of borehole imaging tools, vertical strips of gaps occur in borehole images. We propose an effective approach to fill these gaps using a convolutional neural network with partial convolution layers. To overcome the challenge of missing training labels, we introduce a self-supervised learning strategy. Specifically, we replicate the gaps found in borehole images by randomly creating vertical blank strips that mask out certain known areas in the original images. We then employ the original images as label data to train the network to recover the known areas masked out by the defined gaps. To ensure that the missing data does not impact the training process we incorporate partial convolutions, which exclude the null-data areas from convolutional computations during both forward and backward propagation of updating the network parameters. Our network, trained by this way, can then be used to reasonably fill the gaps originally appearing in the borehole images and obtain full images without any noticeable artifacts. Through the analysis of multiple real examples, we demonstrate the effectiveness of our method by comparing it to three alternative approaches. Our method outperforms the others significantly, as demonstrated by various quantitative evaluation metrics. The filled fullbore images obtained through our approach enable enhanced texture analysis and automated feature recognition.","PeriodicalId":55102,"journal":{"name":"Geophysics","volume":"66 ","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135872348","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：