Pub Date : 2024-11-07DOI: 10.1016/j.cageo.2024.105751
Liyuan Feng , Binhong Li , Huailiang Li , Jian He
We present a novel denoising strategy based on empirical curvelet transform (ECT) for noisy microseismic data. Our approach can simultaneously suppress high-frequency, low-frequency, and shared-bandwidth noises and preserve detailed information on the noisy microseismic data. Initially, we design a new threshold estimation method by adding a scale factor for ECT threshold denoising. Subsequently, we construct an adaptive parameter model employing the similarity standard deviation for the non-local means (NLM) algorithm. Then, we divide the coefficients obtained from the ECT decomposition into two sets based on the energy spectrum, subjecting each set to improved adaptive thresholding and improved NLM denoising algorithms. Eventually, we reconstruct the denoised signals using the empirical curvelet inverse transform. Our results demonstrate that under a signal-to-noise ratio (SNR) of 10 dB, the proposed strategy achieves a correlation coefficient of 0.9524, a root mean square error of 0.198, an SNR of 1.36 dB, and reduces the first arrival picking error to 0.00382 s. Furthermore, application on the real microseismic data further confirms that the proposed method can clarify the corresponding first arrival.
我们针对高噪声微地震数据提出了一种基于经验小曲线变换（ECT）的新型去噪策略。我们的方法可以同时抑制高频、低频和共享带宽噪声，并保留噪声微地震数据的详细信息。首先，我们设计了一种新的阈值估计方法，为 ECT 阈值去噪添加了一个比例因子。随后，我们利用非局部均值（NLM）算法的相似性标准偏差构建了一个自适应参数模型。然后，我们根据能谱将 ECT 分解得到的系数分成两组，每组都采用改进的自适应阈值和改进的 NLM 去噪算法。最后，我们使用经验小曲线逆变换重建去噪信号。结果表明，在信噪比（SNR）为 -10 dB 的条件下，所提出的策略实现了 0.9524 的相关系数、0.198 的均方根误差、1.36 dB 的信噪比，并将首次到达的选取误差降低到 0.00382 s。
{"title":"Novel empirical curvelet denoising strategy for suppressing mixed noise of microseismic data","authors":"Liyuan Feng , Binhong Li , Huailiang Li , Jian He","doi":"10.1016/j.cageo.2024.105751","DOIUrl":"10.1016/j.cageo.2024.105751","url":null,"abstract":"<div><div>We present a novel denoising strategy based on empirical curvelet transform (ECT) for noisy microseismic data. Our approach can simultaneously suppress high-frequency, low-frequency, and shared-bandwidth noises and preserve detailed information on the noisy microseismic data. Initially, we design a new threshold estimation method by adding a scale factor for ECT threshold denoising. Subsequently, we construct an adaptive parameter model employing the similarity standard deviation for the non-local means (NLM) algorithm. Then, we divide the coefficients obtained from the ECT decomposition into two sets based on the energy spectrum, subjecting each set to improved adaptive thresholding and improved NLM denoising algorithms. Eventually, we reconstruct the denoised signals using the empirical curvelet inverse transform. Our results demonstrate that under a signal-to-noise ratio (SNR) of <span><math><mo>−</mo></math></span>10 dB, the proposed strategy achieves a correlation coefficient of 0.9524, a root mean square error of 0.198, an SNR of 1.36 dB, and reduces the first arrival picking error to 0.00382 s. Furthermore, application on the real microseismic data further confirms that the proposed method can clarify the corresponding first arrival.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"194 ","pages":"Article 105751"},"PeriodicalIF":4.2,"publicationDate":"2024-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142659016","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 : 2024-11-07DOI: 10.1016/j.cageo.2024.105770
Yu-lian Zhu , Wei-ying Chen , Wan-ting Song , Si-xu Han
The rapid imaging of electrical source transient electromagnetic (TEM) data involves two essential processes: the calculation of apparent resistivity and the conversion of time to depth. Traditionally, the definition of full-time apparent resistivity is defined by considering solely the vertical magnetic field, which is predicated on the monotonic relationship between the resistivity and the electromagnetic field response. Based on the concept of peak time, we have developed distinct methodologies for calculating the apparent resistivity for both the horizontal electric field (ex) and the vertical induced voltage (vz), which demonstrated accuracy across the entire time range examined. We also introduced a formula to address discrepancies in apparent resistivity arising from the non-dipole size effect of the source, thereby ensuring that the algorithm can adapt to any transmitting and receiving configuration. Furthermore, we provided straightforward and precise time-depth conversion equations applicable to both ex and vz, which facilitate the rapid imaging of observational data. Multiple numerical examples were employed to illustrate the effectiveness and robustness of this approach. Finally, we applied this imaging technique to the data processing of actual measured data from a survey area conducted in Ningxia Province, and the imaging results accurately reflected the distribution of the electrical structure of the subsurface strata. The innovative imaging technique presented in this study holds considerable potential for the expedited processing and analysis of ground-based and semi-aerial electrical source transient electromagnetic survey data, which are widely employed in contemporary applications.
电源瞬变电磁（TEM）数据的快速成像涉及两个基本过程：视电阻率的计算和时间到深度的转换。传统上，全时视电阻率的定义仅考虑垂直磁场，其前提是电阻率与电磁场响应之间的单调关系。基于峰值时间的概念，我们开发了不同的方法来计算水平电场（ex）和垂直感应电压（vz）的视电阻率，这些方法在整个考察时间范围内都表现出了准确性。我们还引入了一个公式，以解决源的非偶极子尺寸效应引起的视电阻率差异，从而确保算法能够适应任何发射和接收配置。此外，我们还提供了适用于 ex 和 vz 的直接而精确的时间深度转换方程，这有助于观测数据的快速成像。我们采用了多个数值示例来说明这种方法的有效性和稳健性。最后，我们将该成像技术应用于宁夏某测区实测数据的数据处理，成像结果准确反映了地下地层电性结构的分布。本研究提出的创新成像技术在加快处理和分析当代广泛应用的地面和半航空电源瞬变电磁勘测数据方面具有相当大的潜力。
{"title":"New fast imaging techniques for electrical source transient electromagnetic data: Approaches and application","authors":"Yu-lian Zhu , Wei-ying Chen , Wan-ting Song , Si-xu Han","doi":"10.1016/j.cageo.2024.105770","DOIUrl":"10.1016/j.cageo.2024.105770","url":null,"abstract":"<div><div>The rapid imaging of electrical source transient electromagnetic (TEM) data involves two essential processes: the calculation of apparent resistivity and the conversion of time to depth. Traditionally, the definition of full-time apparent resistivity is defined by considering solely the vertical magnetic field, which is predicated on the monotonic relationship between the resistivity and the electromagnetic field response. Based on the concept of peak time, we have developed distinct methodologies for calculating the apparent resistivity for both the horizontal electric field (<em>e</em><sub>x</sub>) and the vertical induced voltage (<em>v</em><sub>z</sub>), which demonstrated accuracy across the entire time range examined. We also introduced a formula to address discrepancies in apparent resistivity arising from the non-dipole size effect of the source, thereby ensuring that the algorithm can adapt to any transmitting and receiving configuration. Furthermore, we provided straightforward and precise time-depth conversion equations applicable to both <em>e</em><sub><em>x</em></sub> and <em>v</em><sub><em>z</em></sub>, which facilitate the rapid imaging of observational data. Multiple numerical examples were employed to illustrate the effectiveness and robustness of this approach. Finally, we applied this imaging technique to the data processing of actual measured data from a survey area conducted in Ningxia Province, and the imaging results accurately reflected the distribution of the electrical structure of the subsurface strata. The innovative imaging technique presented in this study holds considerable potential for the expedited processing and analysis of ground-based and semi-aerial electrical source transient electromagnetic survey data, which are widely employed in contemporary applications.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"194 ","pages":"Article 105770"},"PeriodicalIF":4.2,"publicationDate":"2024-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142659030","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 : 2024-11-06DOI: 10.1016/j.cageo.2024.105749
Yury Alkhimenkov
An application based on graphical processing units (GPUs) applied to 3-D digital images is described for computing the linear anisotropic elastic properties of heterogeneous materials. The application can also retrieve the property contribution tensors of individual inclusions of any shape. The code can be executed on professional GPUs as well as on a basic laptop or personal computer Nvidia GPUs. The application is extremely fast: a calculation of the effective elastic properties of volumes consisting of about 7 million voxel elements (1913) takes less than 4 s of computational time using a single A100 GPU; 3 min for 100 million voxel elements (4793) using a single A100 GPU; 14 min for 350 million voxel elements (7033) using a single A100 GPU. Several comparisons against analytical solutions are provided. In addition, an evaluation of the anisotropic effective elastic properties of a 3-D digital image of a cracked Carrara marble sample is presented. The software can be downloaded from a permanent repository Zenodo, the link with a doi is given in the manuscript.
{"title":"Digital rock physics: Calculation of effective elastic properties of heterogeneous materials using graphical processing units (GPUs)","authors":"Yury Alkhimenkov","doi":"10.1016/j.cageo.2024.105749","DOIUrl":"10.1016/j.cageo.2024.105749","url":null,"abstract":"<div><div>An application based on graphical processing units (GPUs) applied to 3-D digital images is described for computing the linear anisotropic elastic properties of heterogeneous materials. The application can also retrieve the property contribution tensors of individual inclusions of any shape. The code can be executed on professional GPUs as well as on a basic laptop or personal computer Nvidia GPUs. The application is extremely fast: a calculation of the effective elastic properties of volumes consisting of about 7 million voxel elements (191<sup>3</sup>) takes less than 4 s of computational time using a single A100 GPU; 3 min for 100 million voxel elements (479<sup>3</sup>) using a single A100 GPU; 14 min for 350 million voxel elements (703<sup>3</sup>) using a single A100 GPU. Several comparisons against analytical solutions are provided. In addition, an evaluation of the anisotropic effective elastic properties of a 3-D digital image of a cracked Carrara marble sample is presented. The software can be downloaded from a permanent repository Zenodo, the link with a doi is given in the manuscript.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"194 ","pages":"Article 105749"},"PeriodicalIF":4.2,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142659019","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 : 2024-11-06DOI: 10.1016/j.cageo.2024.105756
Yakin Hajlaoui , Richard Labib , Jean-François Plante , Michel Gamache
The blastability index (BI) is a measure that indicates the resistance of rock to fragmentation when blasting. With novel technologies, miners are now able to collect and calculate BI at different depths while drilling. In this research, we propose an approach to estimate the BI at multiple depths for new areas using only spatial locations and observed BI measurements of previously drilled holes. Spatial interpolation techniques are investigated. This study introduces a novel treatment for Gaussian Processes (GPs) and Inverse Distance Weighting (IDW). Variography is leveraged to ensure an appropriate fit between the data and the spatial component. The parameters controlling anisotropy are constrained to intervals chosen to reflect the observed anisotropy. Gradient descent with back-propagation is used for optimization. The proposed approach improves the performance of GP and IDW at predicting BI. The similarities between the IDW variant proposed and a single-layer neural network are discussed.
可爆性指数（BI）是表示爆破时岩石抗破碎能力的指标。利用新技术，矿工现在能够在钻探时收集和计算不同深度的可爆性指数。在这项研究中，我们提出了一种方法，仅利用空间位置和先前钻孔的观察 BI 测量值来估算新区域多个深度的 BI。研究了空间插值技术。该研究引入了一种新的高斯过程（GPs）和反距离加权（IDW）处理方法。利用变分法确保数据与空间分量之间的适当拟合。控制各向异性的参数受限于所选的区间，以反映观察到的各向异性。采用反向传播梯度下降法进行优化。所提出的方法提高了 GP 和 IDW 预测 BI 的性能。讨论了所提出的 IDW 变体与单层神经网络之间的相似性。
{"title":"Backpropagation-based inference for spatial interpolation to estimate the blastability index in an open pit mine","authors":"Yakin Hajlaoui , Richard Labib , Jean-François Plante , Michel Gamache","doi":"10.1016/j.cageo.2024.105756","DOIUrl":"10.1016/j.cageo.2024.105756","url":null,"abstract":"<div><div>The blastability index (BI) is a measure that indicates the resistance of rock to fragmentation when blasting. With novel technologies, miners are now able to collect and calculate BI at different depths while drilling. In this research, we propose an approach to estimate the BI at multiple depths for new areas using only spatial locations and observed BI measurements of previously drilled holes. Spatial interpolation techniques are investigated. This study introduces a novel treatment for Gaussian Processes (GPs) and Inverse Distance Weighting (IDW). Variography is leveraged to ensure an appropriate fit between the data and the spatial component. The parameters controlling anisotropy are constrained to intervals chosen to reflect the observed anisotropy. Gradient descent with back-propagation is used for optimization. The proposed approach improves the performance of GP and IDW at predicting BI. The similarities between the IDW variant proposed and a single-layer neural network are discussed.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"194 ","pages":"Article 105756"},"PeriodicalIF":4.2,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142659021","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 : 2024-11-05DOI: 10.1016/j.cageo.2024.105754
Sihai Wu , Jiubing Cheng , Jianwei Ma , Tengfei Wang , Xueshan Yong , Yang Ji
Reverse time migration (RTM) plays a crucial role in high-resolution seismic imaging of the Earth’s interior. However, scaling it across millions of cores in parallel to process large-scale seismic datasets poses significant computational challenges, because the conventional storage solutions are insufficient to deal with the I/O and memory bottlenecks. To address this issue, we present a highly scalable 3D RTM algorithm for vertically transverse isotropic (VTI) media, optimized for the Sugon exascale supercomputer, utilizing over 1,024,000 cores with optimal weak-scaling efficiency. Through cache optimizations tailored for the new deep computing unit (DCU) accelerator architecture, our approach achieves a maximum speedup of 6x compared to conventional methods on a single accelerator. Moreover, based on the lossy compression and boundary-saving techniques, we reduce storage requirements by 266 times, which allows for the effective utilization of million-core computing resources and ensures scalability efficiency when handling large-scale datasets for complex geophysical tasks. Finally, when applied to a industrial dataset, the method demonstrates robust scalability and high efficiency, making it well-suited for large-scale geophysical exploration.
{"title":"Million-core scalable 3D anisotropic reverse time migration on the Sugon exascale supercomputer","authors":"Sihai Wu , Jiubing Cheng , Jianwei Ma , Tengfei Wang , Xueshan Yong , Yang Ji","doi":"10.1016/j.cageo.2024.105754","DOIUrl":"10.1016/j.cageo.2024.105754","url":null,"abstract":"<div><div>Reverse time migration (RTM) plays a crucial role in high-resolution seismic imaging of the Earth’s interior. However, scaling it across millions of cores in parallel to process large-scale seismic datasets poses significant computational challenges, because the conventional storage solutions are insufficient to deal with the I/O and memory bottlenecks. To address this issue, we present a highly scalable 3D RTM algorithm for vertically transverse isotropic (VTI) media, optimized for the Sugon exascale supercomputer, utilizing over 1,024,000 cores with optimal weak-scaling efficiency. Through cache optimizations tailored for the new deep computing unit (DCU) accelerator architecture, our approach achieves a maximum speedup of 6x compared to conventional methods on a single accelerator. Moreover, based on the lossy compression and boundary-saving techniques, we reduce storage requirements by 266 times, which allows for the effective utilization of million-core computing resources and ensures scalability efficiency when handling large-scale datasets for complex geophysical tasks. Finally, when applied to a industrial dataset, the method demonstrates robust scalability and high efficiency, making it well-suited for large-scale geophysical exploration.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"194 ","pages":"Article 105754"},"PeriodicalIF":4.2,"publicationDate":"2024-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142659028","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 : 2024-11-05DOI: 10.1016/j.cageo.2024.105752
A. Armandine Les Landes , L. Beaude , D. Castanon Quiroz , L. Jeannin , S. Lopez , F. Smai , T. Guillon , R. Masson
In deep geothermal reservoirs, faults and fractures play a major role, serving as regulators of fluid flow and heat transfer while also providing feed zones for production wells. To accurately model the operation of geothermal fields, it is necessary to explicitly consider objects of varying spatial scales, from the reservoir scale itself, to that of faults and fractures, down to the scale of the injection and production wells.
Our main objective in developing the ComPASS geothermal flow simulator, was to take into account all of these geometric constraints in a flow and heat transfer numerical model using generic unstructured meshes. In its current state, the code provides a parallel implementation of a spatio-temporal discretization of the non-linear equations driving compositional multi-phase thermal flows in porous fractured media on unstructured meshes. It allows an explicit discretization of faults and fractures as 2D hybrid objects, embedded in a 3D matrix. Similarly, wells are modeled as one dimensional graphs discretized by edges of the 3D mesh which allows arbitrary multi-branch wells. The resulting approach is particularly flexible and robust in terms of modeling.
Its practical interest is demonstrated by two case studies in high-energy geothermal contexts.
{"title":"Geothermal modeling in complex geological systems with ComPASS","authors":"A. Armandine Les Landes , L. Beaude , D. Castanon Quiroz , L. Jeannin , S. Lopez , F. Smai , T. Guillon , R. Masson","doi":"10.1016/j.cageo.2024.105752","DOIUrl":"10.1016/j.cageo.2024.105752","url":null,"abstract":"<div><div>In deep geothermal reservoirs, faults and fractures play a major role, serving as regulators of fluid flow and heat transfer while also providing feed zones for production wells. To accurately model the operation of geothermal fields, it is necessary to explicitly consider objects of varying spatial scales, from the reservoir scale itself, to that of faults and fractures, down to the scale of the injection and production wells.</div><div>Our main objective in developing the ComPASS geothermal flow simulator, was to take into account all of these geometric constraints in a flow and heat transfer numerical model using generic unstructured meshes. In its current state, the code provides a parallel implementation of a spatio-temporal discretization of the non-linear equations driving compositional multi-phase thermal flows in porous fractured media on unstructured meshes. It allows an explicit discretization of faults and fractures as 2D hybrid objects, embedded in a 3D matrix. Similarly, wells are modeled as one dimensional graphs discretized by edges of the 3D mesh which allows arbitrary multi-branch wells. The resulting approach is particularly flexible and robust in terms of modeling.</div><div>Its practical interest is demonstrated by two case studies in high-energy geothermal contexts.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"194 ","pages":"Article 105752"},"PeriodicalIF":4.2,"publicationDate":"2024-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142659029","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-02DOI: 10.1016/j.cageo.2024.105753
Hongyu Qi , Zhenwu Fu , Yang Li , Bo Han , Longsuo Li
Full-waveform inversion (FWI) represents an advanced geophysical imaging technique focused on intricately depicting subsurface physical properties by iteratively minimizing the differences between the simulated and observed seismograms. Unfortunately, the conventional FWI utilizing a least-squares loss function suffers from various drawbacks, including the challenge of local minima and the necessity for human intervention in parameter fine-tuning. It is particularly problematic when handling noisy data and inadequate initial models. Recent works have exhibited promising performance in two-dimensional FWI by integrating structural sparse representation to procure adaptive dictionaries. Drawing inspiration from the competitiveness of structural sparse representation, we introduce a paradigm of group sparse residuals that integrates two types of complementary prior information by harnessing both the internal and external subsurface media models. The proposed algorithm is based on an alternate minimization algorithm to guarantee workflow flexibility and efficient optimization capabilities. We experimentally validate our method for two baseline geological models, and a comparison of the results demonstrates that the proposed algorithm faithfully recovers the velocity models and consistently outperforms other traditional or learning-based algorithms. A further benefit from the group sparse coding used in this method is that it reduces the sensitivity to data noise.
{"title":"Regularization by double complementary priors for full waveform inversion","authors":"Hongyu Qi , Zhenwu Fu , Yang Li , Bo Han , Longsuo Li","doi":"10.1016/j.cageo.2024.105753","DOIUrl":"10.1016/j.cageo.2024.105753","url":null,"abstract":"<div><div>Full-waveform inversion (FWI) represents an advanced geophysical imaging technique focused on intricately depicting subsurface physical properties by iteratively minimizing the differences between the simulated and observed seismograms. Unfortunately, the conventional FWI utilizing a least-squares loss function suffers from various drawbacks, including the challenge of local minima and the necessity for human intervention in parameter fine-tuning. It is particularly problematic when handling noisy data and inadequate initial models. Recent works have exhibited promising performance in two-dimensional FWI by integrating structural sparse representation to procure adaptive dictionaries. Drawing inspiration from the competitiveness of structural sparse representation, we introduce a paradigm of group sparse residuals that integrates two types of complementary prior information by harnessing both the internal and external subsurface media models. The proposed algorithm is based on an alternate minimization algorithm to guarantee workflow flexibility and efficient optimization capabilities. We experimentally validate our method for two baseline geological models, and a comparison of the results demonstrates that the proposed algorithm faithfully recovers the velocity models and consistently outperforms other traditional or learning-based algorithms. A further benefit from the group sparse coding used in this method is that it reduces the sensitivity to data noise.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"194 ","pages":"Article 105753"},"PeriodicalIF":4.2,"publicationDate":"2024-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142659017","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 : 2024-11-02DOI: 10.1016/j.cageo.2024.105755
Guido Di Federico, Louis J. Durlofsky
Geological parameterization entails the representation of a geomodel using a small set of latent variables and a mapping from these variables to grid-block properties such as porosity and permeability. Parameterization is useful for data assimilation (history matching), as it maintains geological realism while reducing the number of variables to be determined. Diffusion models are a new class of generative deep-learning procedures that have been shown to outperform previous methods, such as generative adversarial networks, for image generation tasks. Diffusion models are trained to “denoise”, which enables them to generate new geological realizations from input fields characterized by random noise. Latent diffusion models, which are the specific variant considered in this study, provide dimension reduction through use of a low-dimensional latent variable. The model developed in this work includes a variational autoencoder for dimension reduction and a U-net for the denoising process. Our application involves conditional 2D three-facies (channel-levee-mud) systems. The latent diffusion model is shown to provide realizations that are visually consistent with samples from geomodeling software. Quantitative metrics involving spatial and flow-response statistics are evaluated, and general agreement between the diffusion-generated models and reference realizations is observed. Stability tests are performed to assess the smoothness of the parameterization method. The latent diffusion model is then used for ensemble-based data assimilation. Two synthetic “true” models are considered. Significant uncertainty reduction, posterior P10–P90 forecasts that generally bracket observed data, and consistent posterior geomodels, are achieved in both cases.
地质参数化需要使用一小套潜在变量来表示地质模型，并将这些变量映射到孔隙度和渗透率等网格块属性。参数化对于数据同化（历史匹配）非常有用，因为它既能保持地质的真实性，又能减少需要确定的变量数量。扩散模型是一类新的生成式深度学习程序，在图像生成任务中的表现优于以往的方法，如生成式对抗网络。扩散模型经过 "去噪 "训练，能够从随机噪声输入区域生成新的地质现实。潜在扩散模型是本研究中考虑的具体变体，它通过使用低维潜在变量来降低维度。本研究开发的模型包括一个用于降维的变异自动编码器和一个用于去噪的 U 型网络。我们的应用涉及有条件的二维三岩层（通道-岩层-泥浆）系统。结果表明，潜在扩散模型可提供与地理建模软件样本视觉上一致的现实。对涉及空间和流量响应统计的定量指标进行了评估，发现扩散生成的模型与参考现实之间基本一致。还进行了稳定性测试，以评估参数化方法的平稳性。然后将潜在扩散模型用于基于集合的数据同化。考虑了两个合成的 "真实 "模型。在这两种情况下，都能显著减少不确定性，P10-P90 后期预报与观测数据基本保持一致，后期地理模型也保持一致。
{"title":"Latent diffusion models for parameterization of facies-based geomodels and their use in data assimilation","authors":"Guido Di Federico, Louis J. Durlofsky","doi":"10.1016/j.cageo.2024.105755","DOIUrl":"10.1016/j.cageo.2024.105755","url":null,"abstract":"<div><div>Geological parameterization entails the representation of a geomodel using a small set of latent variables and a mapping from these variables to grid-block properties such as porosity and permeability. Parameterization is useful for data assimilation (history matching), as it maintains geological realism while reducing the number of variables to be determined. Diffusion models are a new class of generative deep-learning procedures that have been shown to outperform previous methods, such as generative adversarial networks, for image generation tasks. Diffusion models are trained to “denoise”, which enables them to generate new geological realizations from input fields characterized by random noise. Latent diffusion models, which are the specific variant considered in this study, provide dimension reduction through use of a low-dimensional latent variable. The model developed in this work includes a variational autoencoder for dimension reduction and a U-net for the denoising process. Our application involves conditional 2D three-facies (channel-levee-mud) systems. The latent diffusion model is shown to provide realizations that are visually consistent with samples from geomodeling software. Quantitative metrics involving spatial and flow-response statistics are evaluated, and general agreement between the diffusion-generated models and reference realizations is observed. Stability tests are performed to assess the smoothness of the parameterization method. The latent diffusion model is then used for ensemble-based data assimilation. Two synthetic “true” models are considered. Significant uncertainty reduction, posterior P<sub>10</sub>–P<sub>90</sub> forecasts that generally bracket observed data, and consistent posterior geomodels, are achieved in both cases.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"194 ","pages":"Article 105755"},"PeriodicalIF":4.2,"publicationDate":"2024-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142659015","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 : 2024-11-01DOI: 10.1016/j.cageo.2024.105730
Pedro Ribeiro Mendes Júnior , Soroor Salavati , Oscar Linares , Maiara Moreira Gonçalves , Marcelo Ferreira Zampieri , Vitor Hugo de Sousa Ferreira , Manuel Castro , Rafael de Oliveira Werneck , Renato Moura , Elayne Morais , Ahmed Esmin , Leopoldo Lusquino Filho , Denis José Schiozer , Alexandre Ferreira , Alessandra Davólio , Anderson Rocha
We investigate machine-learning techniques for rock-type classification. A throughout literature review (considering the machine-learning technique, number of classes, rock types, and image types) presents a diversity of datasets employed and a wide range of classification results as well as multiple problem formulations. Throughout the discussion of the literature, we highlight some common machine-learning pitfalls and criticize the decisions taken by some authors on the problem formulation. We present an experimental contribution by evaluating the classification of seven types of rocks found in carbonate reservoirs along with state-of-the-art Convolutional Neural Networks (CNNs) architectures available through a well-known open-source library. For this experimentation, we detail the preparation of the dataset of drill core plugs (DCPs), the experimental setup itself, and the obtained results considering the normalized accuracy and the traditional accuracy as metrics. We performed the manual background segmentation of the employed dataset of DCPs; so the results reported are not influenced by the background of the images. We evaluate top-1, top-2, and top-3 performance for the problem. We apply fusion of multiple CNNs for richer classification decisions. We also contribute by presenting the manual classification — human labeling by looking at the image on the computer screen — of the same seven-class dataset, performed by six non-geologist volunteers. Finally, we present a conclusion for the results obtained with our experiments and share valuable advice for researchers applying machine learning to rock classification.
{"title":"Rock-type classification: A (critical) machine-learning perspective","authors":"Pedro Ribeiro Mendes Júnior , Soroor Salavati , Oscar Linares , Maiara Moreira Gonçalves , Marcelo Ferreira Zampieri , Vitor Hugo de Sousa Ferreira , Manuel Castro , Rafael de Oliveira Werneck , Renato Moura , Elayne Morais , Ahmed Esmin , Leopoldo Lusquino Filho , Denis José Schiozer , Alexandre Ferreira , Alessandra Davólio , Anderson Rocha","doi":"10.1016/j.cageo.2024.105730","DOIUrl":"10.1016/j.cageo.2024.105730","url":null,"abstract":"<div><div>We investigate machine-learning techniques for rock-type classification. A throughout literature review (considering the machine-learning technique, number of classes, rock types, and image types) presents a diversity of datasets employed and a wide range of classification results as well as multiple problem formulations. Throughout the discussion of the literature, we highlight some common machine-learning pitfalls and criticize the decisions taken by some authors on the problem formulation. We present an experimental contribution by evaluating the classification of seven types of rocks found in carbonate reservoirs along with state-of-the-art Convolutional Neural Networks (CNNs) architectures available through a well-known open-source library. For this experimentation, we detail the preparation of the dataset of drill core plugs (DCPs), the experimental setup itself, and the obtained results considering the normalized accuracy and the traditional accuracy as metrics. We performed the manual background segmentation of the employed dataset of DCPs; so the results reported are not influenced by the background of the images. We evaluate top-1, top-2, and top-3 performance for the problem. We apply fusion of multiple CNNs for richer classification decisions. We also contribute by presenting the manual classification — human labeling by looking at the image on the computer screen — of the same seven-class dataset, performed by six non-geologist volunteers. Finally, we present a conclusion for the results obtained with our experiments and share valuable advice for researchers applying machine learning to rock classification.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"193 ","pages":"Article 105730"},"PeriodicalIF":4.2,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142552516","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 : 2024-11-01DOI: 10.1016/j.cageo.2024.105766
Xinmin Ge , Mohmmed Ishag , Haiyan Li , Jundong Liu , Cuixia Qu , Badreldein Mohamed
This study investigates the impact of the drilling mud invasion on the borehole-measured resistivity. The primary objective is to retrieve the true resistivity of the formation, which helps in identifying different fluids in the reservoir. To achieve this goal, We proposed a hybrid inversion approach integrating the Levenberg-Marquardt and Markov Chain Monte Carlo algorithms with a five-parameter formation resistivity model. Synthetic and real-world data are utilized to assess the method's robustness and reliability. The simulated result indicated that the method is reliable when the data noise level is less than 5%.
The method applied to real-world data revealed that the resistivity profile on the water zone showed a slight increase in the inverted resistivity from measured resistivity. Meanwhile, in the oil zone, the calculated resistivity revealed a high deviation from the measured resistivity, indicating the effects of mud invasion. The introduced methods are only applicable when the invasions of mud occur within the range of the logging tool's depth of investigation. Moreover, the method may give no reliable result when the invasion exceeds the tool's investigation depth. It indicates its limitation.
{"title":"A hybrid inversion algorithm to obtain the resistivity of the uninvaded zone based on the array induction log","authors":"Xinmin Ge , Mohmmed Ishag , Haiyan Li , Jundong Liu , Cuixia Qu , Badreldein Mohamed","doi":"10.1016/j.cageo.2024.105766","DOIUrl":"10.1016/j.cageo.2024.105766","url":null,"abstract":"<div><div>This study investigates the impact of the drilling mud invasion on the borehole-measured resistivity. The primary objective is to retrieve the true resistivity of the formation, which helps in identifying different fluids in the reservoir. To achieve this goal, We proposed a hybrid inversion approach integrating the Levenberg-Marquardt and Markov Chain Monte Carlo algorithms with a five-parameter formation resistivity model. Synthetic and real-world data are utilized to assess the method's robustness and reliability. The simulated result indicated that the method is reliable when the data noise level is less than 5%.</div><div>The method applied to real-world data revealed that the resistivity profile on the water zone showed a slight increase in the inverted resistivity from measured resistivity. Meanwhile, in the oil zone, the calculated resistivity revealed a high deviation from the measured resistivity, indicating the effects of mud invasion. The introduced methods are only applicable when the invasions of mud occur within the range of the logging tool's depth of investigation. Moreover, the method may give no reliable result when the invasion exceeds the tool's investigation depth. It indicates its limitation.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"194 ","pages":"Article 105766"},"PeriodicalIF":4.2,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142593369","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}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号