Publisher: Society of Exploration Geophysicists   (Total: 3 journals)   [Sort alphabetically]

Showing 1 - 3 of 3 Journals sorted by number of followers
Geophysics     Full-text available via subscription   (Followers: 21, SJR: 1.018, CiteScore: 2)
The Leading Edge     Hybrid Journal   (Followers: 1, SJR: 0.386, CiteScore: 1)
Interpretation     Hybrid Journal   (Followers: 1)
   Most Followed Journals

   Top Publishers

   Top Subjects
Similar Journals
Journal Cover
Geophysics
Journal Prestige (SJR): 1.018
Citation Impact (citeScore): 2
Number of Followers: 21  
 
  Full-text available via subscription Subscription journal
ISSN (Print) 0016-8033 - ISSN (Online) 1942-2156
Published by Society of Exploration Geophysicists Homepage  [3 journals]
  • Advances in mathematical geophysics — Introduction

    • Free pre-print version: Loading...

      PubDate: Tue, 31 Jan 2023 00:00:00 GMT
       
  • To: “Vibroseis sweep sequencing research based on convex optimization
           model,” Zhu Xujiang, Ni Yudong, Wang Jingfu, He Yongqing, Xu Yingpo,
           Wang Wei, Sun Junqing, Li Kanghu, and Wang Mingliang , 2022, Second
           International Meeting for Applied Geoscience & Energy, 103–109, doi:
           10.1190/image2022-3742386.1.

    • Free pre-print version: Loading...

      Abstract: Please note that this SEG/AAPG expanded abstract lacked proper attribution of Figure 2, reproduced below. The caption for this figure should have sourced an image that originally published in a United States patent application, to read:
      PubDate: Thu, 19 Jan 2023 00:00:00 GMT
       
  • To: “Simulation of wave propagation in linear thermoelastic media,”
           Jose M. Carcione, Zhi-Wei Wang, Wenchang Ling, Ettore Salusti, Jing Ba,
           and Li-Yun Fu , 2019, Geophysics , 84 , No. 1, T1–T11, doi:
           10.1190/geo2018-0448.1.

    • Free pre-print version: Loading...

      Abstract: Factor 2 in equation B-1 must be removed.
      PubDate: Thu, 19 Jan 2023 00:00:00 GMT
       
  • Unpaired training: Optimize the seismic data denoising model without
           paired training data

    • Free pre-print version: Loading...

      Abstract: ABSTRACTWith the development of seismic exploration technology, distributed acoustic sensing (DAS) has recently received attention in geophysics. However, owing to the complexity of the layout techniques in the DAS systems, and the unknown or harsh exploration environment, seismic data acquired by this technique usually contain the noise of diverse components, which increases the difficulty in subsequent data analysis and interpretation. This study has (1) trained a deep learning model that effectively suppressed noise with augmented noise data sets to obtain a high signal-to-noise ratio in the DAS vertical seismic profile (VSP) records, (2) introduced an attention module to enhance the extraction and recognition of signal features to recover effective signals under substantial noise interference, and (3) introduced adversarial loss and cycle-consistent loss to replace the commonly used L1 norm or L2 norm to train the network. The obtained hybrid training set containing unpaired synthetic and unpaired field data sets for model pretraining and fine tuning effectively improves the denoising performance of the seismic field data. In summary, this study develops an unpaired training-based DAS seismic data denoising method that transformed noisy DAS VSP data into noise-free data. By analyzing the noise suppression results of other methods, including qualitative and quantitative analyses, we demonstrate that our method successfully suppressed multiple types of noise in the DAS VSP data. The study has indicated clear and continuous signals in the denoising results and improved the denoising performance on the seismic field data.
      PubDate: Mon, 16 Jan 2023 00:00:00 GMT
       
  • Magnetotelluric noise suppression via convolutional neural network

    • Free pre-print version: Loading...

      Abstract: ABSTRACTIt is well known that a magnetotelluric (MT) signal with high signal-to-noise ratio is an important prerequisite for correct interpretation of subsurface structures. However, MT signals collected in the environment of strong cultural noise often are of low data quality due to noise pollution, which seriously affects the accuracy of interpretation. As can be seen from the MT time-domain waveform, the noise is highly energetic, diverse, and random. This means MT denoising methods should have strong applicability to guarantee accurate and effective separation of MT signal from noise data. Therefore, we propose a deep-learning-based data nonlinear mapping method for MT signal-to-noise separation. First, this method focuses on learning the nonlinear mapping relationship between a large amount of noise data and the corresponding noise contour by using the convolutional neural network (CNN) in advance. Then, the mapping transformation of noise data to noise contour in the measured data is realized by CNN model. Specifically, based on the features of MT noise data, we construct a large amount of training data very close to it by mathematical functions. At the same time, we also select some of the measured data to be added to the training set. This not only expands the amount and diversity of the training set but also improves the adaptability of CNN when dealing with complex data. Finally, we evaluate the denoising performance of the proposed method in terms of time-domain waveforms, apparent resistivity-phase curves, and polarization directions before and after denoising. The processing results of the simulated data and the measured data collected in Luzong area have verified the feasibility and effectiveness of the proposed method in MT data denoising.
      PubDate: Mon, 16 Jan 2023 00:00:00 GMT
       
  • Separating seismic diffractions by an improved Cook-distance singular
           value decomposition method

    • Free pre-print version: Loading...

      Abstract: ABSTRACTWhen seismic waves encounter abrupt points of stratum or lithology in the process of propagation, such as fault edges, pinch-out points, or the protrusion of unconformity surfaces, the Snell law will break down and many diffractions will be generated. However, the diffraction information is typically masked by strong reflections; thus, separating diffractions is one key issue for diffraction imaging. The traditional singular value decomposition (SVD) method with the ability for wavefield decomposition and reconstruction is useful in removing strong reflections of large singular values. However, the continuously changing characteristics of the singular value of steep-slope reflections make it difficult to choose a suitable parameter for separating diffractions. To solve the problem of singular value selection in diffractions reconstruction by the SVD method, we have developed a Cook-distance SVD method by mathematically least-squares measuring the contribution of every singular value. The Cook-distance SVD can amplify the difference between continuously changing singular values and easily find the singular values representing diffractions. Synthetic examples and field data applications demonstrate that our Cook-distance SVD method can avoid parameter testing and has good performance in diffraction separation and can highlight reservoir-involved small-scale geologic edges and scatterers.
      PubDate: Mon, 16 Jan 2023 00:00:00 GMT
       
  • Target-oriented waveform inversion based on Marchenko redatumed data

    • Free pre-print version: Loading...

      Abstract: ABSTRACTTarget-oriented inversion (TOI) can resolve subsurface reservoir parameters based on surface waveform data. Moreover, virtual data at the datum level and a precise local forward operator are required to accurately perform inversion in the target area. The Marchenko redatuming, a state-of-the-art seismic redatuming method, is used to obtain the virtual data of the target area. Local forward operators of crosscorrelation and cross-convolution forms are used to generate the local wavefields. A subsurface-domain interferometric objective function and the corresponding inversion algorithm are developed. A simple model is used to demonstrate the accuracy of this approach. This technique is applied to the Chevron 2014 benchmark data set to invert a specific area, indicating that the method can achieve accurate high-resolution inversion results with reduced computational costs. Finally, a time-lapse inversion is performed to exhibit the practicality of the new approach in 4D seismic exploration. These numerical results demonstrate great applications of the TOI based on the redatumed data and local forward operators of crosscorrelation or convolution forms.
      PubDate: Wed, 11 Jan 2023 00:00:00 GMT
       
  • Decoupled wave equation and forward modeling of qP wave in VTI media with
           the new acoustic approximation

    • Free pre-print version: Loading...

      Abstract: ABSTRACTAcoustic approximation has received wide attention in modeling and inversion for the anisotropic wave equation to avoid high computational cost and parameter trade-off in seismic inversion. However, it also is limited by the stability condition, such as the instability for the qP-wave equation in transversely isotropic media with ε<δ. We have developed a new approach that decoupled wave equation and forward modeling of the qP wave in vertical transversely isotropic (VTI) media with the new acoustic approximation. To keep dispersion relations (ωSV) of the qSV wave in each direction equal to zero, we formulate the vertical S-wave velocity (VS0) to be a function of the model parameters and wavenumber component (kx,kz), rather than setting it to zero. Then, the corresponding dispersion relation of pure qP for the new acoustic approximation is derived. According to this decoupled dispersion relation, we obtain the decoupled wave equation of pure qP wave in VTI media by inverse Fourier transform. To solve the wave equation in the space domain efficiently, the operator Sk in the wave equation is characterized in the space domain by its asymptotic approximation operator Sn. From the qP wave equation in the time-space domain, we realize the forward modeling of the pure qP wave in VTI media with the finite-difference method. The dispersion relation analysis and numerical examples find that the decoupled qP-wave equation with the new acoustic approximation does not contain the degenerate qSV wave and is a pure qP-wave equation, which is in good agreement with the simulation results of the elastic wave equation and has high accuracy and is stable in VTI media with ε≥δ or ε<δ.
      PubDate: Wed, 11 Jan 2023 00:00:00 GMT
       
  • Ensemble empirical mode decomposition and stacking model for filtering
           borehole distributed acoustic sensing records

    • Free pre-print version: Loading...

      Abstract: ABSTRACTWe have evaluated the ensemble empirical mode decomposition (EEMD) and stacking model for borehole seismic-data denoising. The borehole records collected by distributed acoustic sensing (DAS) technology have multitype noise contamination, and it is difficult to attenuate these noises while recovering the seismic waves well. We first perform EEMD on the seismic data to obtain the signal-to-noise modal components, then extract the time and frequency information of the decomposed modes using six feature factors, and finally introduce an ensemble learning method to classify the acquired modal features effectively. Stacking is the ensemble learning technique we used in our study. This technique integrates several diverse basic ensemble models using the meta-learning strategy and constructs a highly integrated framework with superior performance and good generalization. In addition, the basic ensemble models consist of many decision tree classifiers following two different ideas of parallelization and serialization. The feature extraction process provides sufficient DAS feature data for the training process of the framework. Synthetic and real experimental results demonstrate that the stacking integration framework effectively separates the signal-to-noise modal features of the borehole DAS records. Furthermore, the EEMD-stacking method performs better than wavelet transform, intrinsic time-scale decomposition, robust principal component analysis, k-means singular value decomposition, and median filtering on the denoising task.
      PubDate: Tue, 10 Jan 2023 00:00:00 GMT
       
  • Three-parameter prestack nonlinear inversion constrained by gradient
           structure similarity

    • Free pre-print version: Loading...

      Abstract: ABSTRACTEstimating the elastic parameters from prestack inversion is of great importance for reservoir characterization. Conventional three-parameter prestack inversion methods rely heavily on well logs, and it is difficult to obtain reliable inversion results in situations with limited numbers of wells. Alternatively, we have developed a joint inversion strategy, integrating the advantages of post- and prestack inversion, to deal with the situation. First, due to the high signal-to-noise ratio of poststack seismic data, the high-precision acoustic impedance (AI) inversion is conducted. Second, the exact Zoeppritz equation is used to establish the objective function of the prestack inversion. To better constrain the elastic parameters (P- and S-wave velocities and density), a new similarity measurement criterion, the gradient structure similarity (GSS), is defined to describe the structural similarity between the prestack inversion results and the inverted AI from the poststack inversion. Third, the LM optimization algorithm is used to solve the nonlinear objective function. Through the model test, we verify the effectiveness of our GSS regularization scheme. Some synthetic and field examples find that our method can provide more stable and accurate inverted results relative to the conventional prestack inversion methods.
      PubDate: Tue, 10 Jan 2023 00:00:00 GMT
       
  • Data-assimilated time-lapse visco-acoustic full-waveform inversion: Theory
           and application for injected CO 2 plume monitoring

    • Free pre-print version: Loading...

      Abstract: ABSTRACTContinuous seismic monitoring for quantifying CO2 plume migration and detection of any potential leakages in the subsurface is essential for the security of long-term anthropogenic carbon dioxide geologic storage. Traditional time-lapse full-waveform inversion (TLFWI) methods aim to map the CO2 distribution by estimating seismic velocity changes, but recent studies find that CO2-induced attenuation is an important complement to seismic velocity for tracking the CO2 plumes and even quantifying the CO2 saturation. We have developed a novel data-assimilated TLFWI method to construct high-resolution time-lapse velocity and attenuation changes from dense time-lapse monitoring data. This method consists of two theoretical developments: visco-acoustic full-waveform inversion (QFWI) and multiparameter hierarchical matrix-powered extended Kalman filter (mHiEKF). The method is capable of (1) posing temporal constraints to retrieve time-lapse information from dense monitoring data by using mHiEKF, (2) accurately recovering high-spatial-resolution velocity and attenuation perturbations using first-order equation system-based QFWI, and (3) providing the model uncertainty by estimating their model standard deviation. With numerical examples, we first find the effectiveness of the new QFWI on estimating accurate velocity and attenuation models simultaneously. Then, a CO2 leakage case and a realistic Frio-II CO2 monitoring case are presented to find the advantages and applicability of our data-assimilated QFWI method for estimating time-lapse changes using dense time-lapse monitoring surveys. By assimilating time-lapse seismic monitoring data over time, our data-assimilated QFWI method can improve the resolution of velocity and attenuation changes and decrease their model uncertainties.
      PubDate: Mon, 09 Jan 2023 00:00:00 GMT
       
  • Fault detection using a convolutional neural network trained with
           point-spread function-convolution-based samples

    • Free pre-print version: Loading...

      Abstract: ABSTRACTAutomatic fault detection in seismic images using deep learning-based methods has attracted great interest in recent years. Detecting faults by deep learning methods is generally considered a supervised learning task, which requires numerous diverse training samples and corresponding accurate labels. Because field data with faults labeled by experienced interpreters are difficult to acquire, training samples are usually generated synthetically, in which the 1D seismic wavelets are convolved with reflectivity models. However, the 1D-convolution-based synthetic seismic images still differ from the real seismic images. Therefore, the network trained by such samples may fail to detect some faults in the field data. Full-wavefield approaches are the optimal seismic modeling methods, but they cannot be widely used at present due to the high cost. In this paper, we present an efficient approach to generate realistic synthetic seismic images by using a point-spread function (PSF)-based convolution. Because the size of a seismic sample used for deep learning training is a small range relative to the field seismic data, assuming the background velocity is close to homogeneous in such a small range, we can efficiently construct the PSF using the analytical Green’s function. With an Intel Core i9-10900K CPU, the PSF approach takes approximately 15 min to generate a seismic image of size 128 × 128 × 128, whereas it takes approximately 20 h to generate a seismic image of the same size using reverse time migration. The examples of one synthetic image and two field seismic images demonstrate that the network trained with the PSF convolution samples can predict more accurate and continuous faults than that trained with the 1D convolution samples.
      PubDate: Fri, 06 Jan 2023 00:00:00 GMT
       
  • Residual learning with feedback for strong random noise attenuation in
           seismic data

    • Free pre-print version: Loading...

      Abstract: ABSTRACTIn random seismic noise attenuation, when the noise energy is higher than or close to a signal, it is difficult to distinguish the signal from the noise. This random noise is relatively strong compared to the signal and is called strong random noise. We have developed a deep learning framework to recover the signal from the strong random noise. The framework is based on a residual learning network and feedback connection and is called the feedback residual network. The residual network (ResNet) suppresses random noise through residual fitting and improves the network’s training efficiency. The feedback connection allows the framework to process data in iterations. In each iteration, the feedback connection proportionally combines the input and output of the ResNet to reconstruct a new input with a lower noise level. This enhances denoising performance by asymptotically decreasing the input noise level and retrieving the remaining signals from the estimated noise, thereby reducing the difficulty of strong random noise attenuation. We terminate the feedback iterations according to the energy change of the estimated noise in each iteration. Synthetic and field examples demonstrate that our network can effectively attenuate the strong random noise.
      PubDate: Fri, 06 Jan 2023 00:00:00 GMT
       
  • Amplitude-variation-with-offset in thermoelastic media

    • Free pre-print version: Loading...

      Abstract: ABSTRACTTemperature is an important factor for evaluating the seismic response of deep reservoirs. We have developed an amplitude-variation-with-offset approximation based on the Lord-Shulman thermoelasticity theory. The model predicts two compressional (P and T) waves (the second is a thermal mode) and a shear (S) wave. The T mode is due to the coupling between the elastic and heat equations. In the thermoelastic case, the approximation is more accurate than in the elastic case. Its accuracy is verified by comparison with the exact equations calculated in terms of potential functions. We examine two reservoir models with high temperatures and compute synthetic seismograms that illustrate the reliability of the approximation. Moreover, we consider real data to build a model and find that the approximate equation not only simplifies the calculations but also is accurate enough and can be used to evaluate the temperature-dependent elastic properties, providing a basis for further application of the thermoelasticity theory, such as geothermal exploration, thermal-enhanced oil recovery, and ultradeep oil and gas resources subject to high temperatures.
      PubDate: Fri, 06 Jan 2023 00:00:00 GMT
       
  • Hierarchical transfer learning for deep learning velocity model building

    • Free pre-print version: Loading...

      Abstract: ABSTRACTDeep learning is a promising approach to velocity model building because it has the potential of processing large seismic surveys with minimal resources. By leveraging large quantities of model-gather pairs, neural networks (NNs) can automatically map data to the model space, directly providing a solution to the inverse problem. Such mapping requires big data, which proves prohibitive for 2D and 3D surveys of realistic size. We have developed a transfer learning (TL) strategy. A network is first trained on a smaller subproblem, which then becomes the starting solution to a larger, more difficult data set, akin to the hierarchical multiscale strategy for full-waveform inversion. We perform TL by having subobjectives that escalate in complexity and by first training an NN at estimating horizontally layered velocity models and then proceeding to train an augmented network at estimating 2D dipping layered models. TL improves convergence and allows using a lesser quantity of 2D models for training. For synthetic tests, the structural similarity index measure of 2D interval velocity models in the time domain is 0.893±0.052 and the root-mean-square (rms) error is (198±91) m/s. We benchmark our algorithm on the Marmousi2 model and observe that our method can apply to velocity models with continuous deformed layers with dips up to 35°. We benchmark our algorithm on 2D marine field data and produce an rms velocity model that leads to coherent stacking and a time interval velocity model that reproduces salient features of the stacked section. TL expedites and regularizes training and data-driven techniques may be applied to field data with minimal preprocessing even though we lack real target velocity models.
      PubDate: Thu, 05 Jan 2023 00:00:00 GMT
       
  • Joint elastic reverse time migration of towed streamer and sparse
           ocean-bottom seismic node hybrid data

    • Free pre-print version: Loading...

      Abstract: ABSTRACTCompared to 1C towed-streamer (TS) seismic exploration, 4C ocean-bottom seismic (OBS) node exploration has great advantages in complex structure imaging, lithology, and fluid identification using elastic waves. However, sparse spatial sampling of OBS node surveys highlights the problems of the imaging acquisition footprint, poor phase continuity, and low signal-to-noise ratio (S/N) with conventional elastic reverse time migration (ERTM) methods. Therefore, a solution of joint ERTM (J-ERTM) is proposed by combining sparse OBS node data with dense TS data. In the J-ERTM of the hybrid data, a novel weighted boundary condition combined with acoustic-elastic coupling equations and a vector-based crosscorrelation imaging condition are presented to perform PP and PS imaging by receiver-side tensorial extrapolation of TS and OBS node hybrid data. A synthetic example demonstrates that our method can effectively process TS and OBS node hybrid data and improve elastic imaging problems caused by OBS node sparse acquisition. J-ERTM techniques also are applied to an active-source OBS data set from the South China Sea. To improve the imaging quality with limited data, some preprocessing procedures for field TS and OBS node hybrid data sets are necessary, such as denoising, OBS node relocation, OBS node orientation correction, OBS node calibration, and others. After preprocessing, the pressure component and velocity components have a more physical energy relationship, more consistent frequency range and wavelet, and a higher S/N. Finally, the preprocessed hybrid data can be used for J-ERTM. The imaging results demonstrate that our method works well with field data.
      PubDate: Thu, 05 Jan 2023 00:00:00 GMT
       
  • Porosity and permeability prediction using a transformer and periodic long
           short-term network

    • Free pre-print version: Loading...

      Abstract: ABSTRACTEffective reservoir parameter prediction is important for subsurface characterization and understanding fluid migration. However, conventional methods for obtaining porosity and permeability are based on either core measurements or mathematical/petrophysical modeling, which are expensive or inefficient. In this study, we develop a reliable and low-cost deep learning (DL) framework for reservoir permeability and porosity prediction from real logging data at different regions. We leverage an advanced learning architecture (i.e., the transformer model) and design a new regression network (RPTransformer) that is sensitive to the depth period change of the logging data. The RPTransformer is composed of 1D convolutional, long short-term memory (LSTM), and transformer layers. First, we use a 1D convolutional layer for the first layer of the network to extract significant features from the logging data. Then, the nonlinear mapping relationships between logging data and reservoir parameters are established using several LSTM layers with a period parameter. Afterward, we use the encoder in the vision transformer with the self-attention mechanism to further extract logging data features. The developed network is a data-driven supervised learning framework and indicates highly accurate and robust prediction results when applied to different geographic regions. To demonstrate the reliable prediction performance of our network, we compare it with several classic machine learning and state-of-the-art DL methods, e.g., random forest, multilayer LSTM, and long short-term time-series network (LSTNet). More importantly, we find the generalization and uncertainty of the network in real-world applications through comprehensive numerical experiments.
      PubDate: Thu, 05 Jan 2023 00:00:00 GMT
       
  • Regularization of anisotropic full-waveform inversion with multiple
           parameters by adversarial neural networks

    • Free pre-print version: Loading...

      Abstract: ABSTRACTThe anisotropic full-waveform inversion (FWI) is a seismic inverse problem for multiple parameters, which aims to simultaneously reconstruct the vertical velocity and the anisotropic parameters of the earth’s subsurface. This multiparameter inverse problem suffers from two issues. First, the objective function of the data fitting is less sensitive to the anisotropic parameters. Second, the crosstalk effect among the different parameters worsens the model update in the iterative inversion. We have developed a method that statistically regularizes the anisotropic FWI using Wasserstein adversarial networks, by penalizing the Wasserstein distance between the distribution of the current model parameters and that of the parameters at the borehole locations. The regularizer can mitigate the issues of anisotropic FWI with multiple parameters and therefore it also can be applied to other inverse problems with multiple parameters.
      PubDate: Thu, 05 Jan 2023 00:00:00 GMT
       
  • Automated hyperparameter optimization for simulating boundary conditions
           for acoustic and elastic wave propagation using deep learning

    • Free pre-print version: Loading...

      Abstract: ABSTRACTWe have developed a deep learning framework to simulate the effect of boundary conditions for wave propagation in anisotropic media. To overcome the challenges associated with the stability of conventional implementation of boundary conditions for strongly anisotropic media, we develop an efficient algorithm using deep neural networks. We incorporate the hyperparameter optimization (HPO) workflow in the deep learning framework to automate the network training process. Hyperparameter selection is a crucial step in model building and has a direct impact on the performance of machine learning models. We implement three different HPO techniques, namely random search, Hyperband, and Bayesian optimization, for simulating boundary conditions and compare the strengths and drawbacks of these techniques. We train the network using a few shot locations and time slices enabling the network to learn how to remove boundary reflections and simulate wave propagation for unbounded media. The automated deep learning framework with HPO improves the performance of deep learning models by achieving the optimal minima and significantly improves the efficiency of the workflow. The benefit of this approach is its simple implementation and significant reduction of reflections at the boundaries, especially in the case of tilted transverse isotropic media. We validate our approach by comparing wave propagation at the boundaries using our algorithm with the output obtained using the unbounded media simulated by padding the model. Tests on different models with acoustic and elastic wave propagation verify the effectiveness of our approach.
      PubDate: Thu, 05 Jan 2023 00:00:00 GMT
       
  • Imaging near-surface S-wave velocity and attenuation models by
           full-waveform inversion with distributed acoustic sensing-recorded surface
           waves

    • Free pre-print version: Loading...

      Abstract: ABSTRACTDistributed acoustic sensing (DAS) technology is, increasingly, the seismic acquisition mode of choice for its high spatial sampling rate, low cost, and nonintrusive deployability. It is being widely evaluated as an enabler of seismic monitoring for CO2 sequestration in building subsurface time-lapse images and in characterizing near-surface environments. To advance this evaluation, field seismic surveys with optical fibers have been conducted at the Containment and Monitoring Institute’s Field Research Station (CaMI.FRS) in Newell County, Alberta, Canada. In comparison to the standard geophones, optical fibers deployed in surface trenches at CaMI.FRS have recorded high-quality surface waves, rich in low frequencies and exhibiting limited spatial aliasing. These benefits have motivated us to apply the full-waveform inversion (FWI) approach to image the S-wave velocity (VS) and attenuation (quality factor QS) models at shallow site using the surface waves recorded by optical fibers. Compared to the conventional surface-wave dispersion approach, FWI can intrinsically incorporate fundamental and high-order modes and produce VS model with high spatial resolution that resolves horizontal variations. The low-frequency components below 10 Hz measured in the DAS recordings are helpful to overcome the cycle-skipping problem of FWI. Following the adjoint-state method, QS sensitivity kernel can be calculated efficiently with memory strain variables. The QS model is iteratively estimated with a new misfit function measuring root-mean-square amplitude differences, which helps to reduce the trade-off artifacts. The synthetic data obtained from the inverted models are consistent with the observed data in amplitude and phase. The inversion results provide valuable information to characterize the near-surface environments at CaMI.FRS and are expected to support seismic imaging in deeper CO2 injection zones.
      PubDate: Wed, 04 Jan 2023 00:00:00 GMT
       
  • Joint inversion with petrophysical constraints using indicator functions
           and the extended alternating direction method of multipliers

    • Free pre-print version: Loading...

      Abstract: ABSTRACTJoint inversions often need to construct an objective function with multiple complex constraining terms, which usually increase the computational cost, take a long time to converge, and often are nondifferentiable. We have developed a joint inversion framework that implements petrophysical constraints by indicator functions and solves the optimization problem using the alternating direction method of multipliers (ADMM). An indicator function can describe arbitrary ranges and relationships of multiple physical property parameters by a feasible set. Objective functions involving multiple nondifferentiable terms are numerically not easy to solve, so we adopt an extended alternating direction method of multipliers (eADMM) that separates the terms as independent subproblems and select the most straightforward and efficient solving technique for each of them. With the eADMM, every objective function term can generate its own model, and the models are forced to converge through equality constraints. Such a mechanism brings further efficiency from parallelization and the flexibility of allowing the final inversion model to not strictly obey the model constraints because of uncertainties in the data and petrophysical information. Our approach is tested on two synthetic examples of joint inversion of gravity and magnetic data. The first example contains two blocks having the same magnetic susceptibility but different densities. Our synthetic inversions verify that the combination of the indicator function and eADMM method effectively recovers exact boundaries and values. The second example uses the data synthesized from a realistic kimberlite exploration project. Our joint inversion can distinguish two kimberlite facies, image their position, and even delineate the boundary between the two facies with much improved accuracy. Based on the same model, the robustness of our method also has been tested with inexact and incorrect petrophysical information. The simplicity, flexibility, effectiveness, and efficiency make our approach a good candidate for a wide range of applications in joint inversion.
      PubDate: Wed, 04 Jan 2023 00:00:00 GMT
       
  • Time-lapse seismic imaging using shot gathers with nonrepeatable source
           wavelets

    • Free pre-print version: Loading...

      Abstract: ABSTRACTIn time-lapse seismic applications, the signal produced by changes in the properties of subsurface rocks is generally obscured by noise associated with imperfect repeatability between surveys. A particularly important obstacle in the formation of time-lapse difference images is variation in the effective source wavelet between baseline and monitoring data sets. However, the partially separable influence of the wavelet within Green’s function model of seismic data permits two frequency-domain matching filters to be designed, which act to reduce source wavelet nonrepeatability. One is based on the spectral ratio of the baseline and monitoring wavelets and can be applied when prior estimates of the wavelets are available; the other is the average spectral ratio of the baseline and monitoring traces and can be applied when prior estimates are unavailable. After balancing the data sets with either of these filters, we further prepare for the imaging step with time-shift corrections, using a published fast local crosscorrelations algorithm, preparing the difference data for use in an imaging algorithm. Reverse time migration is engaged for the imaging task, but we observe that residual repeatability errors tend to be magnified at this stage when source-normalized crosscorrelation imaging conditions are used. Testing indicates that replacing this with a Poynting vector imaging condition strongly suppresses remaining errors in a robust manner. This includes stability within simulated data environments to noise-free data and data with random noise, up to signal-to-noise ratios of roughly one. Furthermore, our method illustrates better performance when compared with the conventional least-squares matching filter and common-depth-point-domain warping. At present, there is no common workflow for seismic imaging directly using time-lapse shot gathers. Our research contribution lies not only in the two matching filters but also in a novel workflow for time-lapse seismic imaging.
      PubDate: Wed, 04 Jan 2023 00:00:00 GMT
       
  • Magnetotelluric data denoising method combining two deep-learning-based
           models

    • Free pre-print version: Loading...

      Abstract: ABSTRACTThe magnetotelluric (MT) data collected in an ore-concentration area are extremely vulnerable to all kinds of noise pollution. However, separating real MT signals from strong noise is still a difficult problem, and the noise in MT data is quite distinct from clean data in morphological features. By performing the signal-noise identification and data prediction, we develop a deep learning method to denoise MT data containing strong noise. First, we use the convolutional neural network (CNN) to learn the feature differences between the samples of massive noise and clean data and use the learned features to realize signal-noise identification of the measured data. Second, we use the measured clean data obtained by CNN identification to train the long short-term memory (LSTM) neural network and perform the prediction denoising of the noisy data. The simulation results clearly demonstrate the following two facts: (1) the predicted data output from LSTM basically matches the time-frequency domain features of the real data and (2) our CNN method performs significantly better than the features parameter classification method in dealing with signal-noise identification. In addition, the validity of our method is verified by the processing results of the measured data.
      PubDate: Wed, 04 Jan 2023 00:00:00 GMT
       
  • Pressure-dependent joint elastic-electrical properties of calcite-cemented
           artificial sandstones

    • Free pre-print version: Loading...

      Abstract: ABSTRACTUnderstanding the correlations between the elastic and electrical properties of various types of rocks is the key to the successful joint interpretation of seismic and electromagnetic survey data to provide petrophysical parameters to better assess the subsurface earth. However, the pressure-dependent joint elastic-electrical properties of calcite-cemented sandstones remain poorly understood, even though such rocks are widely distributed in nature and are all experiencing pressures. To obtain such knowledge, a new method has been developed for the manufacture of calcite-cemented artificial sandstones and investigated comprehensively the effects of porosity and cementation content on the confining pressure-dependent joint elastic-electrical properties of the synthetic samples made using the new recipe. Confining pressure is found to more significantly affect the P- and S-wave velocities and electrical resistivity in the samples with higher and lower porosity, respectively, when their cementation content remains the same. On the other hand, cementation content impacts the pressure-dependent elastic and electrical properties more complexly, and the effects of cementation content can be influenced by the minor fluctuation of porosity in the samples, especially at low confining pressures. More interestingly, P- and S-wave velocities are found to approximately linearly correlate with electrical resistivity as confining pressure varies, and the slopes of the linear joint correlations are demonstrated to vary distinctly with porosity and cementation content. The experimental data are interpreted in terms of the competing effects of porosity and cementation content on the microstructure of the samples. The results have helped to reveal the nature that porosity and cementation content find on affecting the joint elastic-electrical properties with varying pressure and have important practical implications for discriminating the porosity and cementation effects that will pave the way to a more successful interpretation of the joint seismic and electromagnetic survey data.
      PubDate: Wed, 04 Jan 2023 00:00:00 GMT
       
  • Vertical transversely isotropic elastic least-squares reverse time
           migration based on elastic wavefield vector decomposition

    • Free pre-print version: Loading...

      Abstract: ABSTRACTAnisotropic elastic reverse time migration (RTM) is a promising technique for imaging complex oil and gas reservoirs. However, the migrated images often suffer from low spatial resolution, migration artifacts, wave-mode crosstalk, and unbalanced amplitude response. Conventional vertical transversely isotropic elastic least-squares reverse time migration (VTI-elastic LSRTM) defines stiffness parameter perturbations as elastic images, which have different physical meanings from VTI-elastic RTM images. We have developed a VTI-elastic LSRTM method based on elastic wavefield vector decomposition that is a natural extension of VTI-elastic RTM. More specifically, our method applies least-squares inversion to VTI-elastic RTM and defines the compressional- and shear-wave reflectivity as elastic images (PP, PS, SP, and SS images). When computing the elastic images, we decompose the elastic wavefields into compressional and shear wavefields and cross-correlate the corresponding wave modes. We derive the reverse time demigration operator by taking the adjoint of the RTM operator. Using the migration and demigration operators, we formulate the VTI-elastic LSRTM as a linear inverse problem with the least-squares criterion. The conjugate gradient method is used to solve the optimization problem. Three numerical examples are presented to test the feasibility of our method. The VTI-elastic LSRTM images have higher resolution, fewer migration artifacts and wave-mode crosstalk, and improved amplitude response when compared with VTI-elastic RTM images.
      PubDate: Wed, 28 Dec 2022 00:00:00 GMT
       
  • Reflection and diffraction separation in the dip-angle common-image
           gathers using convolutional neural network

    • Free pre-print version: Loading...

      Abstract: ABSTRACTIn exploration seismology, reflections have been extensively used for imaging and inversion to detect hydrocarbon and mine resources, which are generated from subsurface continuous impedance interfaces. When the interface is not continuous and its size reduces to less than half-wavelength, reflected wave becomes diffraction. Reflections and diffractions can be used to image subsurface targets, and the latter is helpful to resolve small-scale discontinuities, such as fault plane, pinch out, Karst caves, and salt edge. However, the amplitudes of diffractions are usually much weaker than that of reflections. This makes it difficult to directly identify and extract diffractions from unmigrated common-shot or common-middle-point gathers. Migrating seismic data into a subsurface location for different reflector dip angles yields a dip-angle-domain common-image gather (DACIG). One DACIG represents the migrated traces at a fixed lateral position for different reflector dips. The reflection and diffraction have different geometric characteristics in DACIG, which provides one opportunity to separate diffractions and reflections. In this study, we present an efficient and accurate diffraction separation and imaging method using a convolutional neural network (CNN). The training data set of DACIGs is generated using one pass of seismic modeling and migration for velocity models with and without artificial scatterers, respectively. Then, a simplified end-to-end CNN is trained to identify and extract reflections from the migrated DACIGs that contain reflections and diffractions. Next, two adaptive subtraction strategies are presented to compute the diffraction DACIGs and stacked images, respectively. Numerical experiments for synthetic and field data demonstrate that the proposed method can produce accurate reflection and diffraction separation results in DACIGs, and the stacked image has a good resolution for subsurface small-scale discontinuities.
      PubDate: Wed, 28 Dec 2022 00:00:00 GMT
       
  • Implications of static low-frequency model on seismic geomechanics
           inversion

    • Free pre-print version: Loading...

      Abstract: ABSTRACTI have developed a novel insight into the differences between static and dynamic moduli and their effects on the performance of seismic geomechanics inversion. This achievement is obtained from triaxial deformation tests and ultrasonic measurements on core plugs and reveals that the static Young’s modulus deviates from the dynamic one in porous media, especially in particular ranges of depth and pressure, although conventional regression relationships suggest the opposite, i.e., similar trends for the static and dynamic Young’s moduli. Next, a novel simple approach is formulated to incorporate laboratory information directly into a seismic low-frequency model (LFM) using an artificial neural network to achieve a static low-frequency model (SLFM). Respecting the critical role of the LFM in the reliability of seismic inversion, any modification to the process of building this model can contribute to higher accuracy of the subsequent seismic geomechanics modeling. For this, LFMs are built using static and dynamic data before proceeding to seismic inversion to derive 3D cubes of static Young’s and bulk moduli. The results are successfully validated using data from known wells as well as a blind well. The modeling outcomes demonstrate that the seismic inversion based on the dynamic low-frequency model (DLFM) would return the same results for static and dynamic bulk moduli. In contrast, the results are erroneous for the static Young’s modulus when the conventional DLFM was adopted. Accordingly, the intelligent approach to static low-frequency modeling is found to be a good interpolation technique for estimating geomechanical parameters, as indicated by the good agreement between the static data and the corresponding inversion results at the well locations. My findings place emphasis on the necessity of reconsidering the relationship between the static and dynamic Young’s moduli and highlight the advantage of using an SLFM to increase the accuracy of geomechanical modeling.
      PubDate: Wed, 28 Dec 2022 00:00:00 GMT
       
  • MEANet: Magnitude estimation via physics-based features time series, an
           attention mechanism, and neural networks

    • Free pre-print version: Loading...

      Abstract: ABSTRACTThe traditional magnitude estimation method, which establishes a linear relationship between a single warning parameter and the magnitude, exhibits considerable scatter and underestimation. In addition, the extraction of features from raw waveforms by a deep learning network is a black box. To provide a more robust magnitude estimation and to construct a deep learning network with an interpretable input, in light of deep learning and earthquake rupture physics, we have established a magnitude estimation network model (MEANet) via the physics-based features time series, an attention mechanism, and neural networks. We use events with 4 ≤ M ≤ 7.5 that occur in Japan and the Sichuan-Yunnan region, China, to train and validate MEANet, and then use MEANet to test additional events. Our results find that MEANet has a more robust magnitude estimation than the traditional τc and Pd methods, with a standard deviation of error of ±0.25 magnitude units at a single station with a 3 s P-wave time window. Within 10 s after the first station is triggered, based on the weighted average of the triggered stations, MEANet provides robust magnitude estimation without underestimation for events with 4 ≤ M ≤ 7.5. Our finding implies that the final magnitude is to some degree deterministic by the combination of deep learning and physics-based features. Meanwhile, MEANet might have potential for earthquake early warning.
      PubDate: Wed, 28 Dec 2022 00:00:00 GMT
       
  • Weighted-average time-lapse seismic full-waveform inversion

    • Free pre-print version: Loading...

      Abstract: ABSTRACTAs seismic data can contain information over a large spatial area and are sensitive to changes in the properties of the subsurface, seismic imaging has become the standard geophysical monitoring method for many applications such as carbon capture and storage and reservoir monitoring. The availability of practical tools such as full-waveform inversion (FWI) makes time-lapse seismic FWI a promising method for monitoring subsurface changes. However, FWI is a highly ill-posed problem that can generate artifacts. Because the changes in the earth’s properties are typically small in terms of magnitude and spatial extent, discriminating the true time-lapse signature from noise can be challenging. Different strategies have been proposed to address these difficulties. In this study, we propose a weighted-average (WA) inversion to better control the effects of artifacts and differentiate them from the true 4D changes. We further compare five related strategies with synthetic tests on clean and noisy data. The effects of seawater velocity variation on different strategies also are studied as one of the main sources of nonrepeatability. We tested different strategies of time-lapse FWI (TL-FWI) using the Marmousi and the SEG Advanced Modeling time-lapse models. The results indicate that the WA method can provide the best compromise between accuracy and computation time. This method also provides a range of possible answers of other TL-FWI strategies.
      PubDate: Tue, 27 Dec 2022 00:00:00 GMT
       
  • A deep learning-enhanced framework for multiphysics joint inversion

    • Free pre-print version: Loading...

      Abstract: ABSTRACTJoint inversion has drawn considerable attention due to the availability of multiple geophysical data sets, ever-increasing computational resources, the development of advanced algorithms, and its ability to reduce inversion uncertainty. A key issue of joint inversion is to develop effective strategies to link different geophysical data in a unified mathematical framework, in which the information obtained from different models can complement each other. We have developed a deep learning-enhanced joint inversion framework to simultaneously reconstruct different physical models by fusing different types of geophysical data. Traditionally, structure similarity constraints are pursued by joint inversion algorithms using manually crafted formulations (e.g., cross gradient). The constraint is constructed by a deep neural network (DNN) during the learning process. The framework is designed to combine the DNN and a traditional independent inversion workflow and improve the joint inversion result iteratively. The network can be easily extended to incorporate multiphysics without structural changes. Numerical experiments on the joint inversion of 2D DC resistivity data and seismic traveltime are used to validate our method. In addition, this learning-based framework demonstrates excellent generalization abilities when tested on data sets using different geologic structures. It also can handle different sensing configurations and nonconforming discretization.
      PubDate: Tue, 27 Dec 2022 00:00:00 GMT
       
  • Improving sparse representation with deep learning: A workflow for
           separating strong background interference

    • Free pre-print version: Loading...

      Abstract: ABSTRACTRevealing hidden reservoirs that are severely shielded by strong background interference (SBI) is critical to subsequent refined interpretation. To enhance the characterization of these reservoirs, current interpretation workflows merge multiple attribute information, necessitating intensive human expertise. As an alternative, we regard SBI suppression as a signal separation problem and develop a workflow to suppress SBI by cascading a sparse representation method and deep learning. SBI has coherent morphological characteristics in seismic sections; reservoir seismic responses, such as channels and karst caves, have a narrow spatial distribution, exhibiting abrupt morphological characteristics. As their morphologies differ, we select two 2D sparse representation dictionaries to identify their individual components. Through the morphological component analysis (MCA) technique, we can obtain adequate SBI separation results. However, the MCA separation is inevitably limited because 2D dictionaries cannot adequately represent 3D structures, but 3D dictionaries are not viable due to computing constraints. As an extension, we use 3D deep learning to improve the separation results based on the 2D MCA results. Specifically, the network is fed with training samples from a region with better SBI suppression results obtained by the MCA method. After learning a direct mapping from noisy data to SBI, the network can improve the separation results and remove more SBI than the previous conventional method. Field data experiments demonstrate that our separation workflow successfully enhances reservoir structures after removing SBI.
      PubDate: Tue, 27 Dec 2022 00:00:00 GMT
       
  • Two-stage broad learning inversion framework for shear-wave velocity
           estimation

    • Free pre-print version: Loading...

      Abstract: ABSTRACTShear-wave (S-wave) velocity is considered an essential parameter for the study of the earth, and Rayleigh wave inversion has been widely accepted and used to determine it. Given high-quality measured dispersion curves, the inversion performance depends on the applied optimization algorithm inside the inversion process. We propose a novel inversion framework to promote efficient and accurate inversion, i.e., a two-stage broad learning inversion framework (TS-BL). The proposed TS-BL not only inherits the powerful mapping capability and simple configured structure of broad learning (BL) network but also makes two significant improvements to better acclimatize itself to Rayleigh wave inversion. First, TS-BL adopts a two-stage inversion strategy to perform optimizing two times. It does not yield the same search space in the two inversion stages. In the first stage, because the inversion aims to find an approximation rather than the accurate value of model parameters, the difficulty in constructing the mapping model is reduced by sacrificing accuracy. Then, an effective BL network can be established using smaller sample sizes. In the second stage, the search space becomes much narrower, commencing with the approximation results obtained in the prior stage. This helps the final BL network to easily and quickly model the actual relationship between measured dispersion curves and unknown model parameters. After that, the forward modeling of measurements rather than the validation data set is exploited for tuning the network’s hyperparameters. The physical model is superior to the validation data set for selecting a suitable network complexity to adapt to the measured dispersion curves because the latter only describes an overall relationship. As a result, accurate S-wave velocities can be efficiently acquired by using the proposed TS-BL with a low cost of training samples. The efficiency and reliability of TS-BL have been demonstrated in numerical and field data examples.
      PubDate: Fri, 23 Dec 2022 00:00:00 GMT
       
  • Deep-learning missing well-log prediction via long short-term memory
           network with attention-period mechanism

    • Free pre-print version: Loading...

      Abstract: ABSTRACTUnderground reservoir information can be obtained through well-log interpretation. However, some logs might be missing due to various reasons, such as instrument failure. A deep-learning-based method that combines a convolutional layer and a long short-term memory (LSTM) layer is proposed to estimate the missing logs without the expensive relogging. The convolutional layer is used to extract the depth-series features initially, which are then input into the LSTM layer. To improve the feature memory and extraction capabilities of the LSTM layer, we construct two LSTM-based components: the first component uses an attention mechanism to optimize the LSTM units by adaptively adjusting network weights, and the second component uses a period-skip mechanism, which enhances the sensitivity of aperiodic changes in the depth series by learning the information of the shallow sequence. In addition, we add an autoregressive component to enhance the linear feature extraction capability while learning the nonlinear relationship between different logs. A total of 13 wells from two different regions are used for experiments, including 11 training and two test wells. We use one well to calculate the uncertainties of four time-series networks, i.e., our proposed network and three benchmark models (recurrent neural network, gated recurrent unit, and LSTM), to demonstrate the stability and robustness of the proposed method. Furthermore, we evaluate the performance of our proposed method in several crossover experiments, e.g., different logging intervals, depths, and input logs. Compared to a state-of-the-art deep learning method and a classic LSTM network, the proposed network has higher reliability, which is reflected in the feature extraction of depth series with a larger span. The experimental results demonstrate that our proposed network can accurately generate sonic and other unknown logs.
      PubDate: Fri, 23 Dec 2022 00:00:00 GMT
       
  • Separation and imaging of seismic diffractions using geometric mode
           decomposition

    • Free pre-print version: Loading...

      Abstract: ABSTRACTSeismic diffractions from small-scale discontinuities or inhomogeneities carry key geologic information and can provide high-resolution images of these objects. Because diffractions characterized by weak energy are easily masked by strong reflections, diffraction-enhancement processing is essential before subwavelength information detection. Therefore, a novel diffraction-separation method is developed that uses the Fourier-based geometric-mode decomposition (GMDF) algorithm to remove reflections and separate diffractions in the common-offset or poststack domain. The key idea of our method is that, in the frequency-wavenumber (f-k) domain, strong reflections concentrate linearly along a certain dip direction, whereas weak diffractions are distributed over a wide range of wavenumbers owing to their variable dips in the time-space domain. The GMDF algorithm can effectively represent reflections with directional and linear geometric features by adaptively decomposing seismic data as a combination of the band-limited modes consisting of linear characteristics in the f-k domain. The alternating direction method of the multipliers algorithm is used to solve the GMDF optimization problem and obtain linear reflections. Because this method considers the energy sparsity property and linear geometric features of reflections in the f-k domain, kinematic and dynamic differences between reflections and diffractions are exploited to separate diffractions. Applied synthetic and field examples demonstrate the good performance of our method in removing strong reflections and separating weak diffractions, providing interpreters with detailed structural and stratigraphic information.
      PubDate: Fri, 23 Dec 2022 00:00:00 GMT
       
  • A two-step singular spectrum analysis method for robust low-rank
           approximation of seismic data

    • Free pre-print version: Loading...

      Abstract: ABSTRACTThe singular spectrum analysis (SSA) method can detect the low-rank structure of data and therefore has become a powerful tool in seismic data processing and analysis. In particular, the SSA method can effectively suppress seismic random noise according to the different behaviors of coherent signal and random noise in the singular spectrum. However, there has been much research and experimentation indicating that the basic SSA method performs poorly when the noise becomes erratic. One of the reasons is that the quadratic misfit adopted in SSA is sensitive to non-Gaussian disturbances. One solution to this problem is to iteratively reweight the low-rank approximation of using the SSA method and finally achieve a particular kind of robust misfit. The low-rank approximation is robustified in this study by a more direct strategy, which consists of two main steps, prediction and elimination, and is called the two-step SSA method. The whole algorithm only runs the SSA filtering twice and hence is more computationally efficient compared to the iteratively reweighted SSA technique. The two-step SSA method adopts two criteria to calculate a weighting matrix and predict the erratic disturbance. This strategy also can be generalized into an iterative reweighting-based technique (e.g., the iteratively reweighted SSA technique) for robust denoising. The performance of the proposed two-step SSA method is tested using simulated and real seismic data. The results demonstrate its feasibility.
      PubDate: Fri, 23 Dec 2022 00:00:00 GMT
       
  • Reduction of normal-moveout stretch using nonstationary scaling
           transformation in time-frequency domain

    • Free pre-print version: Loading...

      Abstract: ABSTRACTThe normal-moveout (NMO) stretch causes decrease in the dominant frequency of seismic wavelet after conventional NMO correction and severely damages the quality of the stacked data for shallower reflectors at far offsets. Muting, which is commonly used to handle this problem, reduces seismic fold and negatively affects results of the amplitude-variation-with-offset analysis within the stretched area. We found a novel approach to reduce the stretching phenomenon through compensating the lost frequencies by increasing the dominant frequency of the seismic wavelet before applying the NMO correction. The added so-called compensated frequencies are defined according to the difference between the dominant frequency of the original seismic wavelet and the assumed stretched wavelet after the NMO correction. The corresponding procedure considers frequency content of each time sample along each trace in the time-frequency domain using the Gabor transform. As such, the dominant frequency of the seismic wavelets is increased in a nonstationary manner. Performance of our method is evaluated by applying it on the synthetic and field data examples. The obtained results suggest that this approach provides common-midpoint (CMP) gathers with reduced stretching effect, with the potential to be considered as another alternative for nonstretch NMO correction. However, it should be noted that the presented method resolves neither the problem of intersecting events nor the multiples in the CMP gather. This method also cannot handle highly contaminated noise data and does not contribute in removing multiples during the frequency compensation process.
      PubDate: Fri, 23 Dec 2022 00:00:00 GMT
       
  • Joint physics-based and data-driven time-lapse seismic inversion:
           Mitigating data scarcity

    • Free pre-print version: Loading...

      Abstract: ABSTRACTIn carbon capture and sequestration, developing rapid and effective imaging techniques is crucial for real-time monitoring of the spatial and temporal dynamics of CO2 propagation during and after injection. With continuing improvements in computational power and data storage, data-driven techniques based on machine learning (ML) have been effectively applied to seismic inverse problems. In particular, ML helps alleviate the ill-posedness and high computational cost of full-waveform inversion (FWI). However, such data-driven inversion techniques require massive high-quality training data sets to ensure prediction accuracy, which hinders their application to time-lapse monitoring of CO2 sequestration. We develop an efficient “hybrid” time-lapse workflow that combines physics-based FWI and data-driven ML inversion. The scarcity of the available training data is addressed by developing a new data-generation technique with physics constraints. The method is validated using a synthetic CO2-sequestration model based on the Kimberlina storage reservoir in California. Our approach is shown to synthesize a large volume of high-quality, physically realistic training data, which is critically important in accurately characterizing the CO2 movement in the reservoir. The developed hybrid methodology can also simultaneously predict the variations in velocity and saturation and achieve high spatial resolution in the presence of realistic noise in the data.
      PubDate: Wed, 21 Dec 2022 00:00:00 GMT
       
  • Simultaneous source separation by shot collocation and strength variation

    • Free pre-print version: Loading...

      Abstract: ABSTRACTSimultaneous shooting offers opportunities for significant cost savings in seismic data acquisitions. The most common strategy uses random delay shots where source separation is achieved during the processing stage, thereby doubling source densities. We have determined that the creation of a collocated source survey, where shots are repeated simultaneously at multiple positions, is a viable alternative strategy, with the additional benefit that it may increase source density even further while keeping the acquisition duration unchanged. Source separation is achieved using overcomplete independent component analysis by first by applying a directional wavelet transform to separate source signals with different slowness, then estimating the mixing matrix, followed by solving an optimization problem with an energy constraint combined with a sparseness inducing prior and obtaining the required waveforms. Synthetic tests find average reconstruction quality on the order of 22.1, 15.4, and 8.4 dB if, respectively, three, four, or five shots are acquired in two mixtures. Examination of the true versus obtained zero-offset sections also demonstrates the robustness of our signal recovery strategy. The advantage of shot collocation over conventional acquisitions is that it may triple or even quadruple the source density for unchanged acquisition durations with superior reconstruction results compared with dithered acquisitions using similar blending factors.
      PubDate: Wed, 21 Dec 2022 00:00:00 GMT
       
  • Handling missing data in well-log curves with a gated graph neural network

    • Free pre-print version: Loading...

      Abstract: ABSTRACTWell logging is a common method that is used to obtain the rock properties of a formation. It is relatively frequent, however, that log information is incomplete due to cost limitations or borehole problems. Existing models predict missing well logs from a fixed combination of other available well logs. However, the missing well logs vary from well to well. We have proposed using a gated graph neural network (GNN) to handle the missing values in well-log curves. It takes sequential data, predicting each missing measurement in the data not only using other available variables measured at the same depth but also available measurements of neighboring observations. Meanwhile, the missing well logs and available well logs could be any possible combinations as long as they are mutually exclusive. This approach has two advantages: (1) the gated GNN does not need to build a specific model for each missing measurement or from every possible combination of available measurements and (2) it can be integrated into the training process of the following predictive model to perform classification tasks. We evaluate the gated GNN model along with two other models: the GRAPE model and the multiple imputation by chained equations (MICE)-gated recurrent unit (GRU) model, on a data set from the North Sea to perform a missing feature imputation task and a lithofacies identification task. The GRAPE model also is a graph-based model, and it predicts values for each missing measurement from available variables measured at the same depth. The MICE-GRU model is a combination of the MICE algorithm and GRU, which handles the feature imputation procedure and the lithofacies identification procedure separately. Our experiments find that the gated GNN model outperforms the MICE algorithm and the GRAPE model on the missing feature imputation task. For the lithofacies identification task, the gated GNN model also provides comparable results to the MICE-GRU model, and they both outperform the GRAPE model.
      PubDate: Tue, 13 Dec 2022 00:00:00 GMT
       
  • Transformations of borehole magnetic data in the frequency domain and
           estimation of the total magnetization direction: A case study from the
           Mengku iron-ore deposit, Northwest China

    • Free pre-print version: Loading...

      Abstract: ABSTRACTBorehole magnetic prospecting measures the three components of the magnetic field and is sensitive to the vertical depth of the magnetic source, which plays an important role in deep mineral exploration. Magnetic field transformations in the frequency domain constitute common and important processing for ground and airborne data but are rarely applied to borehole magnetic data. Here, we deduce component transformation and magnetization-direction transformation formulas for borehole data. It is found that the transformation factors differ from those of ground and airborne data. The new formulas depend on the position of the drillholes. The transformed data are then used to estimate the total magnetization direction by computing the correlation coefficient between the transformed components and the amplitude anomaly. The advantages of this method are that the borehole data yield a local and accurate magnetization direction and the total amplitude anomaly is computed directly from the individual anomalies in the observed components. The frequency-domain transformation is tested on the synthetic data, and the estimated directions are consistent with the true values. The processing procedure is applied to borehole data collected at the Mengku iron-ore deposit in Northwest China, where magnetic surveys from 55 boreholes have been acquired. The estimated magnetization directions yield good approximations of the results from the ground magnetic data and the physical property measurements of rock samples.
      PubDate: Tue, 13 Dec 2022 00:00:00 GMT
       
  • Analysis of the viscoelasticity in coal based on the fractal theory

    • Free pre-print version: Loading...

      Abstract: ABSTRACTCoal is a complex viscoelastic porous medium with fractal characteristics at different scales. To model the macroscale structure of coal, a fractal viscoelastic model is established, and the P-wave velocity dispersion and attenuation characteristics are discussed based on the complex modulus derived from this model. The numerical simulation results indicate that the fractional order α and relaxation time τ greatly affect the P-wave velocity dispersion and attenuation. The fractal viscoelastic model indicates a full-band velocity dispersion between 1 Hz and 104 Hz. Meanwhile, the P-wave velocity has a weaker dispersion with the fractal viscoelastic model than with the Kelvin-Voigt model and Zener model between 1 Hz and 104 Hz for the same relaxation time and elastic modulus, but the velocity at 1 Hz based on the fractal viscoelastic model is higher with the Kelvin-Voigt model and Zener model. Simultaneously, the velocities of five coal samples are tested, and the attenuation factor is calculated using a low-frequency system. The experimental results indicate a strong dispersion in coal in the range of 10–250 Hz. The classic Kelvin-Voigt model and Zener model cannot describe the dispersion characteristics of coal, but the fractal viscoelastic model can describe them well by using the appropriate fractional order and relaxation time.
      PubDate: Tue, 13 Dec 2022 00:00:00 GMT
       
  • Deep physics-aware stochastic seismic inversion

    • Free pre-print version: Loading...

      Abstract: ABSTRACTSeismic inversion allows the prediction of subsurface properties from seismic reflection data and is a key step in reservoir modeling and characterization. With the generalization of machine learning in geophysics, deep learning methods have been proposed as efficient seismic inversion methods. However, most of these methods lack a probabilistic approach to deal with the uncertainties inherent in the seismic inversion problem and/or rely on complete and representative training data, which often is partially or scarcely available. We have explored the ability of deep convolutional neural networks to extract meaningful and complex representations from spatially structured data, combined with geostatistical simulation, to efficiently invert poststack seismic data directly for high-resolution models of acoustic impedance. Our model incorporates physics constraints and sparse direct measurements while leveraging the use of imprecise but widely distributed indirect measurements as represented by the seismic data. The models generated with geostatistical simulation provide additional information with higher spatial resolution than the original seismic data and allow assessing uncertainty in the model predictions by generating multiple realizations of the subsurface properties. Our method can (1) provide an uncertainty assessment of the predictions, (2) model the complex and nonlinear relationship between data and model, (3) extend the seismic bandwidth at low and high ends of the frequency parameters spectrum, and (4) lessen the need for large, annotated training data. Our method is applied to a 1D synthetic example and a real 3D application example from a Brazilian reservoir. The predicted impedance models are compared with those obtained from a full iterative geostatistical seismic inversion. Our method allows retrieving similar models but has the advantage of generating alternative solutions in greater numbers, providing a larger exploration of the model parameter space in less time than the iterative geostatistical seismic inversion.
      PubDate: Tue, 13 Dec 2022 00:00:00 GMT
       
  • Unsupervised deep learning for 3D interpolation of highly incomplete data

    • Free pre-print version: Loading...

      Abstract: ABSTRACTWe propose to denoise and reconstruct the 3D seismic data simultaneously using an unsupervised deep learning (DL) framework, which does not require any prior information about the seismic data and is free of labels. We use an iterative process to reconstruct the 3D highly incomplete seismic data. For each iteration, we use the DL framework to denoise the 3D seismic data and initially reconstruct the missing traces. Then, the projection onto convex sets (POCS) algorithm is used for further enhancement of the seismic data reconstruction. The output of the POCS is considered as the input for the DL network in the next iteration. We use a patching technique to extract 3D seismic patches. Because the proposed DL network consists of several fully connected layers, each extracted patch needs to be converted to a 1D vector. In addition, we use an attention mechanism to enhance the learning capability of the proposed DL network. We evaluate the performance of the proposed framework using several synthetic and field examples and find that the proposed method outperforms all benchmark methods.
      PubDate: Tue, 13 Dec 2022 00:00:00 GMT
       
  • Calculation of the stable Poynting vector using the first-order temporal
           derivative of the seismic wavefield

    • Free pre-print version: Loading...

      Abstract: ABSTRACTThe Poynting vector is a powerful tool for calculating the propagation directions of a seismic wavefield, and it has a wide range of applications in reverse time migration. However, an instability issue commonly arises while calculating the Poynting vector. The Poynting vector is a product of the temporal and spatial derivatives of the wavefield. The two derivatives are equal to zero at the local extrema of the seismic wavefield, so the Poynting vector cannot provide the propagation directions at these points. Stabilizing techniques, such as smoothing, optical flow (OF) methods, and time-shifting methods, can be applied to address this issue. However, each of these three types of methods comes with trade-offs. Smoothing is easy to implement but has a low angular resolution, whereas the OF and time-shift techniques have high angular resolutions but are computationally inefficient. We have developed a new method that achieves high resolution and high computational efficiency. Based on the fact that a seismic wavefield and its first-order temporal derivative have the same direction of propagation, and that the unstable points of their Poynting vectors are at different locations, we use the first-order temporal derivative of the seismic wavefield to stabilize its Poynting vector calculation. Our method is nearly as accurate as the OF and time-shift techniques and is more computationally efficient than the smoothing technique. Finally, we use numerical simulations to verify the effectiveness of our method.
      PubDate: Tue, 13 Dec 2022 00:00:00 GMT
       
  • Uncertainty quantification of geologic model parameters in 3D gravity
           inversion by Hessian-informed Markov chain Monte Carlo

    • Free pre-print version: Loading...

      Abstract: ABSTRACTGeologic modeling has been widely adopted to investigate underground structures. However, modeling processes inevitably have uncertainties due to scarcity of data, measurement errors, and simplification of the modeling method. Recent developments in geomodeling methods have introduced a Bayesian framework to constrain the model uncertainties by considering the additional geophysical data in the modeling procedure. Markov chain Monte Carlo (MCMC) methods are normally used as tools to solve the Bayesian inference problem. To achieve a more efficient posterior exploration, advances in MCMC methods use derivative information. Hence, we introduce an approach to efficiently evaluate second-order derivatives in geologic modeling and adopt a Hessian-informed MCMC method, the generalized preconditioned Crank-Nicolson (gpCN), as a tool to solve the 3D model-based gravity Bayesian inversion problem. The result is compared with two other widely applied MCMC methods, random-walk Metropolis–Hastings and Hamiltonian Monte Carlo, on a synthetic geologic model and a realistic structural model of the Kevitsa deposit. Our experiment demonstrates that superior performance is achieved by the gpCN compared with the other two state-of-the-art sampling methods. This indicates the potential of the proposed method to be generalized to more complex models.
      PubDate: Tue, 06 Dec 2022 00:00:00 GMT
       
  • Wave propagation in thermo-poroelasticity: A finite-element approach

    • Free pre-print version: Loading...

      Abstract: ABSTRACTWe have developed continuous and discrete-time finite-element (FE) methods to solve an initial boundary-value problem for the thermo-poroelasticity wave equation based on the combined Biot/Lord-Shulman (LS) theories to describe the porous and thermal effects, respectively. In particular, the LS model, which includes a Maxwell-Vernotte-Cattaneo relaxation term, leads to a hyperbolic heat equation, thus avoiding infinite signal velocities. The FE methods are formulated on a bounded domain with absorbing boundary conditions at the artificial boundaries. The dynamical equations predict four propagation modes, a fast P (P1) wave, a Biot slow (P2) wave, a thermal (T) wave, and a shear (S) wave. The spatial discretization uses globally continuous bilinear polynomials to represent solid displacements and temperature, whereas the vector part of the Raviart-Thomas-Nedelec of zero order is used to represent fluid displacements. First, a priori optimal error estimates are derived for the continuous-time FE method, and then an explicit conditionally stable discrete-time FE method is defined and analyzed. The explicit FE algorithm is implemented in one dimension to analyze the behavior of the P1, P2, and T waves. The algorithms can be useful for a better understanding of seismic waves in hydrocarbon reservoirs and crustal rocks, whose description is mainly based on the assumption of isothermal wave propagation.
      PubDate: Tue, 06 Dec 2022 00:00:00 GMT
       
  • Reservoir multiparameter prediction method based on deep learning for CO 2
           geologic storage

    • Free pre-print version: Loading...

      Abstract: ABSTRACTTime-lapse seismic difference refers to the comprehensive response of fluid saturation, pore pressure, and porosity. However, the contribution of different parameters to the seismic response is difficult to distinguish. The high-precision prediction of these reservoir parameters is of great significance in CO2 geologic storage and oil and gas development. Therefore, a simultaneous time-lapse reservoir multiparameter prediction method based on a multitask learning network is proposed. Combined with CO2 geologic storage monitoring, the process of generating training data is described, involving numerical simulation, petrophysical models, and seismic forward modeling. Moreover, the Hertz-Mindlin formula, which considers pressure changes, is used to establish the relationship between formation elasticity and physical parameters in CO2 storage. The effects of fluid saturation, pressure, and porosity on P- and S-wave velocities and densities are analyzed, and the amplitude-variation-with-offset response characteristics of fluid saturation, pressure, and porosity changes are discussed. In total, 4700 and 300 sets of reservoir parameters and seismic angle gather data are used for network training and testing, respectively. The prediction results of synthetic and field data find that the time-lapse reservoir multiparameter prediction method based on multitask learning can effectively distinguish changes in each parameter and simultaneously obtain high-precision prediction results of fluid saturation, pressure, and porosity. Once the network is constructed, the prediction will take only a few seconds, which will promote the further development of the CO2 geologic storage theory and technology.
      PubDate: Mon, 05 Dec 2022 00:00:00 GMT
       
  • Improving shallow and deep seismic-while-drilling with a downhole pilot in
           a desert environment

    • Free pre-print version: Loading...

      Abstract: ABSTRACTProcessing seismic data from drillbit-generated vibrations requires a reliable source signature for correlation and deconvolution purposes. Recently, a land field trial has been conducted in a desert environment. A memory-based downhole vibration accelerometer has been used together with a more conventional top-drive sensor to continuously record the pilot signal from 590 to 8600 ft (180–2621 m). Past results indicate that seismic-while-drilling (SWD) data processed using the top-drive accelerometer exhibit good quality in the middle sections of the well but a reduced signal-to-noise ratio for shallow and deep sections. One of the main challenges in using the downhole pilot is a substantial drift of the downhole clock time. To resolve it, a novel automated time-alignment procedure using the GPS-synchronized signal of the top-drive sensor as a reference is applied. The downhole recording provides a source signature of better quality. In shallow sections of the well, it helps to overcome the intense surface-related vibrational noise, whereas, in deeper sections, it provides a cleaner extraction of weaker signals from the polycrystalline diamond compact bits. Processing with the downhole pilot results in better surface seismic data quality than with a conventional top-drive sensor. Therefore, enabling the use of the synchronized downhole pilot signal is of crucial importance for SWD applications. Modern cost-effective near-bit vibrational sensors widely used for different nonseismic applications could be an effective acquisition solution, as shown in this study.
      PubDate: Mon, 05 Dec 2022 00:00:00 GMT
       
  • Numerical modeling of the permeability in Bentheim sandstone under
           confining pressure

    • Free pre-print version: Loading...

      Abstract: ABSTRACTA new model for the determination of the permeability in sandstones under confining pressure is presented. Building on the concepts of digital rock physics, a numerical model is derived from a 3D tomographic scan. The pressure-dependent behavior is mimicked by adding an artificial flow resistance to the pore throats. All permeability simulations are performed using an in-house finite-volume code. In the first step, the proposed model is tested based on a given Bentheim sandstone sample and compared with experimental data. In the second step, the influencing factors of the model are investigated. Mainly discussed are the influences of the tomographic scan and the numerical resolution. Overall, the proposed model is observed to be capable of reproducing general trends of the experimental data, whereby the magnitude of the numerically determined permeabilities can strongly depend on the investigated influencing factors.
      PubDate: Mon, 05 Dec 2022 00:00:00 GMT
       
  • Deep unfolding dictionary learning for seismic denoising

    • Free pre-print version: Loading...

      Abstract: ABSTRACTSeismic denoising is an essential step for seismic data processing. Conventionally, dictionary learning (DL) methods for seismic denoising always assume the representation coefficients to be sparse and the dictionary to be normalized or a tight frame. Current DL methods need to update the dictionary and the coefficients in an alternating iterative process. However, the dictionary obtained from the DL method often needs to be recalculated for different input data. Moreover, the performance of DL for seismic noise removal is related to the parameter selection and the prior constraints of dictionary and representation coefficients. Recently, deep learning demonstrates promising performance in data prediction and classification. Following the architecture of DL algorithms strictly, we have developed a novel and interpretable deep unfolding dictionary learning (DUDL) method for seismic denoising by unfolding the iterative algorithm of DL into a deep neural network (DNN). The proposed architecture of DUDL contains two main parts: the first is to update the dictionary and representation coefficients using least-squares inversion and the second is to apply a DNN to learn the prior representation of dictionary and representation coefficients, respectively. Numerical synthetic and field examples find the effectiveness of our method. More importantly, this method for seismic denoising obtains the dictionary for different seismic data adaptively and is suitable for seismic data with different noise levels.
      PubDate: Thu, 01 Dec 2022 00:00:00 GMT
       
  • On the estimation of reflectivity in reverse time migration:
           Implementational forms of the inverse-scattering imaging condition

    • Free pre-print version: Loading...

      Abstract: ABSTRACTThe inverse-scattering imaging condition (ISIC) for reverse time migration (RTM) aims at recovering amplitudes proportional to seismic reflectivity. It has been derived as the high-frequency asymptotic inverse of Born modeling, which justifies its being called a true-amplitude imaging condition. It involves the temporal and spatial derivatives of the up- and downgoing wavefields, in this way generalizing the conventional crosscorrelation imaging condition. The temporal derivations can be redistributed between different wavefield contributions, in this way deriving a set of different implementational forms of the ISIC. By making use of the wave equation for the up- and downgoing wavefields, one can substitute the time derivatives by the Laplacian operator. This provides a theoretical foundation for a popular filter for reducing the backscattering artifacts in RTM. Using Born data from a simple three-layer model and the Marmousi II model as well as the Sigsbee2b data, we have determined that the theoretical equivalence of the equations leads to similar but not identical images. Our numerical tests indicate that the ISIC versions using spatial derivatives are the most economical approach, and that the images obtained with the second time derivative of the source wavefield indicate slightly improved resolution over the other implementations, making the combination of these two characteristics the best choice.
      PubDate: Thu, 01 Dec 2022 00:00:00 GMT
       
  • High dimensional multistep deblending using supervised training and
           transfer learning

    • Free pre-print version: Loading...

      Abstract: ABSTRACTBlended acquisition significantly improves acquisition efficiency, and deblending algorithms, particularly intelligent deblending methods, continue to be developed to provide separated results for subsequent seismic inversion and imaging. Iterative deblending algorithms improve deblending performance and determining how to evaluate the blending noise level becomes critical. Instead of using a multilevel blending noise strategy to evaluate the blending noise level qualitatively, a supervised multistep deblending algorithm is developed that can evaluate the blending noise level quantitatively in multiple steps. The developed multistep method combines the iterative estimation-subtraction strategy based on sparse inversion and the deep learning strategy. Each deblending step handles a different level of blending noise, ranging from a strong level to a weak level as the deblending steps increase. The first step is to train a U-net to attenuate strong blending noise, and then we can obtain a rough signal estimation for predicting the blending noise to be subtracted. The obtained data, via the blending noise estimation and subtraction following the previous deblending step, are used as the input for the current deblending step, which attenuates weak blending noise and extracts signal leakage in a step-by-step manner. The optimized parameters of the previous deblending step can initialize the current step for efficient fine-tuning based on transfer learning. After sequential blending noise estimation and subtraction, the supervised multistep deblending algorithm with varying input can improve deblending accuracy. A thorough examination of 2D and 3D synthetic blended data demonstrates the validity of our multistep deblending method, particularly when compared with the recently proposed multilevel blending noise strategy. The 3D field blended data processing validates our method in terms of removing blending noise while preserving the signal.
      PubDate: Thu, 01 Dec 2022 00:00:00 GMT
       
  • Depth-domain angle and depth variant seismic wavelets extraction for
           prestack seismic inversion

    • Free pre-print version: Loading...

      Abstract: ABSTRACTMany prestack depth migration methods have been developed and widely used to generate depth-domain seismic images, resulting in a need for depth-domain prestack seismic inversion of the subsurface elastic properties for reservoir characterization. Time-domain inversion techniques often are used after the depth-domain data set is transformed to the time domain. We provide a new technique for directly inverting depth-domain prestack seismic data for subsurface elastic properties: P-impedance (IP), S-impedance (IS), and density (ρ) in the depth domain. The proposed depth-domain workflow eliminates the need for time-depth/depth-time conversion, making it efficient and effective. Using a depth-wavenumber decomposition approach, the suggested workflow first extracts a collection of depth and angle varying wavelets to handle the potential nonstationarity of the depth-domain prestack seismic data. The extracted depth and angle variant wavelets are used in a basis pursuit inversion, which enhances the resolution of the inversion results. The workflow is tested on the Wenan 3D field data set, demonstrating its viability for practical applications.
      PubDate: Thu, 01 Dec 2022 00:00:00 GMT
       
  • Hierarchical wave-mode separation in the poroelastic medium using
           eigenform analysis

    • Free pre-print version: Loading...

      Abstract: ABSTRACTIn the elastic medium, the scalar and vector P- and S-waves decomposition has been extensively studied and some strategies can be extended to the poroelastic medium to extract P- and S-wavefields. However, there are three propagation modes in the poroelastic medium in Biot’s theory, namely, a fast P wave, a slow P wave, and an S wave. Because the propagation characteristic of a slow P wave is different from that of a fast P wave and S wave, the wavefield separation methods in the elastic medium cannot be directly applied to the poroelastic medium to produce a complete wave-mode separation. Based on the eigenform analysis, we have developed a hierarchical wavefield decomposition method to completely separate S waves and fast and slow P waves in the poroelastic medium. Using the Helmholtz decomposition, we first compute scalar and vector potential wavefields to separate P and S waves. Then, a cross-product operator is proposed to decompose fast and slow P waves based on their different polarization directions. To produce correct amplitudes and phases, we apply another cross-product operator and an amplitude correction term to the separated wavefields. Three numerical examples demonstrate that our method can produce accurate fast P-wave, slow P-wave, and S-wave separation results, and the decomposed fast and slow P waves have the same phases and amplitudes as the P-wave potential wavefields.
      PubDate: Thu, 01 Dec 2022 00:00:00 GMT
       
  • 3D generalized spherical multifocusing seismic imaging

    • Free pre-print version: Loading...

      Abstract: ABSTRACTWe introduce a 3D generalized spherical multifocusing (GSMF) algorithm to generate a high-resolution 3D stacked volume that is equivalent to a synthesized 3D zero-offset wavefield for crooked-line/3D seismic data. The proposed algorithm can be applied to arbitrary recording geometry from areas with irregular topography, a complex near the surface, and complex subsurface. The 3D GSMF method simultaneously corrects for elevation statics, nonhyperbolic moveout associated with reflections beneath complex overburden structures, and azimuth-dependent dip-moveout effects. In addition, the formulation is dually generalized for the optical domain and the effective medium. The optical domain and effective medium parameterizations account for heterogeneity either by shifting the reference time to project the problem into the optical image space or by adjusting the velocity of an effective overburden, respectively. We test the performance of our method using 3D synthetic data with 3D and crooked-line surveys. The numerical tests have shown that the accuracy of the new approximation is significant for gently to highly curved interfaces beneath low to relatively high heterogeneous overburden with rugged topography, even at large offsets and midpoint separations. In addition, we rigorously evaluate the method using 3D real seismic data acquired over a complex thrust-belt area with rugged terrain. Compared with conventional 3D stacking, the new formulation yields a high resolution and accurate seismic stacked volume from land seismic data collected with arbitrary 3D geometries.
      PubDate: Wed, 23 Nov 2022 00:00:00 GMT
       
  • A ray perturbation method for the acoustic attenuating transversely
           isotropic eikonal equation

    • Free pre-print version: Loading...

      Abstract: ABSTRACTModeling complex-valued traveltimes is helpful for developing attenuation-associated techniques for seismic data processing. Transverse isotropy can explain the directional variation of velocity and attenuation anisotropy of long-wavelength seismic waves in many sedimentary rocks. The acoustic attenuating transversely isotropic eikonal equation can be used to accurately calculate P-wave complex-valued traveltimes under such a geologic condition. However, no ray-tracing system for this eikonal equation could be found in the literature until now. We have developed a ray perturbation method to solve this eikonal equation. Unlike all existing ray perturbation methods, our newly proposed method does not perturb the exact ray-tracing system but splits the acoustic attenuating transversely isotropic eikonal equation into a nonlinear partial differential equation (PDE) and a first-order PDE. These two PDEs can be solved by the method of characteristics. This gives rise to two sets of ray-tracing equations for the real and imaginary parts of the complex-valued traveltimes, respectively. Both sets of ray-tracing equations share the same raypath, which allows us to merge them into a complete ray-tracing system for complex-valued traveltimes. Numerical examples are used to demonstrate the high accuracy of the newly proposed ray-tracing system, analyze the complex-valued traveltimes of diving P waves, and compare the modeled complex-valued traveltimes with those extracted from constant-Q viscoelastic waveforms.
      PubDate: Wed, 23 Nov 2022 00:00:00 GMT
       
  • Enhancing one-way wave equation-based migration with deep learning

    • Free pre-print version: Loading...

      Abstract: ABSTRACTOne advantage of one-way wave equation-based migration is its low computational cost. However, due to the limited wavefield propagation angle, it is difficult to use one-way wave equation-based migration for high-precision imaging of structures with large inclinations due to issues such as inaccurate amplitudes and migration image artifacts. In addition, when the model has large horizontal velocity differences, it is difficult for the one-way wave propagator to calculate an accurate wavefield phase. Reverse time migration (RTM) based on the two-way wave propagator has a high resolution and avoids the issues associated with one-way wave propagators; however, it has a high computational cost in practical applications. We develop a convolutional neural network (CNN) application mode that improves one migration method by learning from another one and design a CNN with a structure similar to U-net that combines the advantages of both migration methods. The CNN label is the RTM result, and the corresponding input is the result of one-way wave migration with a generalized screen propagator (GSP). The trained CNN model improves the amplitude in the one-way wave migration image and removes the errors caused by large lateral velocity perturbations. Moreover, by maintaining the high migration calculation efficiency, our CNN model allows for a high resolution, few artifacts, and accurate images of steep structures in the one-way wave migration result. With our method, the accuracy of the one-way wave migration result is close to that of the RTM result. The use of GSP-based migration in our CNN model rather than conventional RTM to generate prospecting images can considerably reduce the calculation costs.
      PubDate: Fri, 28 Oct 2022 00:00:00 GMT
       
  • Denoising of distributed acoustic sensing data using supervised deep
           learning

    • Free pre-print version: Loading...

      Abstract: ABSTRACTDistributed acoustic sensing (DAS) is an emerging technology for acquiring seismic data due to its high-density and low-cost advantages. Because of the harsh acquisition environment and other unexpected reasons, the seismic signals acquired in DAS are masked by various types of complex noise, which seriously decreases the signal-to-noise ratio of seismic data. We propose a fully convolutional neural network with dense and residual connections to attenuate complex noise in DAS data. The network is designed to learn features of useful reflection signals recorded from a large number of earthquake and microseismic events, aiming at obtaining an unprecedented generalization ability. First, we generate labels using an integrated framework that attenuates specific types of noise in real DAS data, where the integrated framework includes carefully designed band-pass, structure-oriented median, and dip filters. Then, we use the patching technique to segment the training samples into many small-scale patches to reduce computational cost and improve the extraction of essential features from large-scale passive seismic data. Finally, we use the well-trained network to estimate the heavily polluted hidden signals. Compared with two advanced deep-learning methods and a traditional denoising framework, our proposed method can more effectively attenuate strong and complex noise and recover weak hidden signals in synthetic and real DAS data tests.
      PubDate: Fri, 28 Oct 2022 00:00:00 GMT
       
  • 3D controlled-source electromagnetic inversion in the radio-frequency band

    • Free pre-print version: Loading...

      Abstract: ABSTRACTThe classical radio-magnetotelluric (RMT) method is nowadays routinely applied to various environmental, engineering, and exploration problems. The technique uses radio transmitters broadcasting in the frequency range of 10 kHz to 1 MHz, and the measurements are carried out in the far field. The well-known disadvantages of RMT are a lack of robust radio transmitters in remote areas; the absence of transmitters broadcasting below 10 kHz, which limits the penetration depth; and a possible low signal-to-noise ratio. To overcome these difficulties, controlled sources can be used (controlled-source RMT [CSRMT]). We extend the CSRMT method to perform measurements not only in the far field but also in the transition zone. In CSRMT practice, it often is challenging to maintain far-field conditions for logistical reasons. Therefore, part of the measured data contains signatures of the source field, which cannot be interpreted with magnetotelluric software. In addition, the source placed directly in the survey area allows us to increase the signal-to-noise ratio and resolution. Such CSRMT in the transition zone is, in fact, a controlled-source electromagnetic method but with full impedance tensor and tipper vector transfer functions. We develop new procedures for the 3D modeling and inversion of the tensor radio-frequency data measured in the transition zone of two perpendicular horizontal electric dipole sources. In this case, the geometry of the source must be considered in the forward modeling. The developed modeling and inversion software is tested on a synthetic 3D model. The 3D resistivity models derived from the real data confirm the geologic settings and are consistent with the available borehole information. Therefore, we conclude that the CSRMT approach extended to include the source field is feasible and that the developed procedures are reliable.
      PubDate: Fri, 28 Oct 2022 00:00:00 GMT
       
  • Discontinuous curvilinear collocated grid combined with nonuniform time
           step Runge-Kutta scheme for poroelastic finite-difference modeling

    • Free pre-print version: Loading...

      Abstract: ABSTRACTFor poroelastic media, the existence of a slow P-wave mode, next to the standard fast P and S waves, hinders efficient numerical implementations to propagate poroelastic waves through arbitrary seismic models. The slow P-wave speed can be an order of magnitude smaller than the fast P-wave speed. Hence, a stable and accurate simulation that can capture the slow P wave requires a fine grid and a small time step, which increases the overall computation cost greatly. To decrease the computation cost, we propose a poroelastic finite-difference simulation method that combines a discontinuous curvilinear collocated-grid method with a nonuniform time step Runge-Kutta (NUTS-RK) scheme. The fine grid and small time step are only used for areas near interfaces, where the contribution of the slow P wave is nonnegligible. The NUTS-RK scheme is derived from a Taylor expansion and it can circumvent the need for interpolations or extrapolations otherwise required by communications between different time levels. The accuracy and efficiency of the proposed method are verified by numerical tests. Compared with the curvilinear collocated-grid finite-difference method that uses a globally uniform space grid as well as a uniform time step, the proposed method requires fewer computing resources and can reduce the computing time greatly.
      PubDate: Fri, 28 Oct 2022 00:00:00 GMT
       
  • Seismic data interpolation using nonlocal self-similarity prior

    • Free pre-print version: Loading...

      Abstract: ABSTRACTThe use of a nonlocal self-similarity (NSS) prior, which refers to each reference patch always having many nonlocal similar patches, has demonstrated its effectiveness in seismic data random noise attenuation because of the repetitiveness of textures and structures in their global position. However, NSS-based approaches face challenges when dealing with seismic interpolation. In the presence of missing traces, similar patch matching may be highly unreliable, resulting in a limited performance of interpolation. To solve this problem, a two-stage iterative seismic-interpolation framework based on a rank-reduction (RR) algorithm is developed. In the first stage, preinterpolation seismic data are used to guide the similar patch matching, and the problem of missing trace recovery for the stacked matched patches is converted to the problem of low-rank matrix completion. In the second stage, the similar patches are directly searched on the interpolation result after stage 1 without external help; that is, exploiting its own NSS, which can achieve enhanced interpolation performance. For each iteration, we obtain accurate similarly matched groups and apply a simple and efficient truncated singular value decomposition for RR. Owing to the unique construction method of a low-rank matrix formed by similar patches, our approach can handle irregularly or regularly sampled seismic data. Numerical experiments verify the effectiveness of our method, compared with the curvelet, low-rank matrix fitting, and f-x prediction filtering methods.
      PubDate: Mon, 03 Oct 2022 00:00:00 GMT
       
  • Unsupervised contrastive learning for seismic facies characterization

    • Free pre-print version: Loading...

      Abstract: ABSTRACTSeismic facies characterization plays a key role in hydrocarbon exploration and development. The existing unsupervised methods are mostly waveform-based and involve multiple steps. We have developed a method to leverage unsupervised contrastive learning to automatically analyze seismic facies. To obtain a stable result, we use 3D seismic cubes instead of seismic traces or their variants as inputs of networks to improve lateral consistency. In addition, we treat seismic attributes as geologic constraints and feed them into the network along with the seismic cubes. These different seismic and multiattribute cubes from the same position are regarded as positive pairs and the cubes from a different position are treated as negative pairs. A contrastive learning framework is used to maximize the similarities of positive pairs and minimize the similarities of negative pairs. In this way, we can enforce the samples with similar features to get close while pushing the samples with different features to be separated in the space where we make the seismic facies clustering. This contrastive learning framework is a one-stage, end-to-end, and unsupervised fashion without any manual labels. We have determined the effectiveness of this method by using it to a turbidite channel system in the Canterbury Basin, offshore New Zealand. The obtained facies map is continuous, resulting in a stable and reliable classification.
      PubDate: Mon, 03 Oct 2022 00:00:00 GMT
       
  • Curvature-regularized manifold for seismic data interpolation

    • Free pre-print version: Loading...

      Abstract: ABSTRACTSeismic data can be described in a low-dimensional manifold. Thus, the low dimensionality of the seismic data patch manifold can serve as a good regularizer for seismic data interpolation. However, we have found that having only low-dimensional manifold regularization is not sufficient for interpolating seismic data with large data gaps or spectral aliasing. Therefore, we propose the application of a curvature-regularized low-dimensional manifold method for seismic data interpolation. A windowed version of the method is proposed, which provides adaptability and efficiency for seismic data interpolation. Numerical experiments on the synthetic and field data indicate that the proposed method outperforms the f-x, curvelet, and low-dimensional manifold methods in many missing data cases, especially in the presence of large gaps or spectral aliasing.
      PubDate: Mon, 12 Sep 2022 00:00:00 GMT
       
  • Solving the nonlinear Riccati equation for the 1D plane-wave reflection
           response from a velocity gradient interface by a Heun to hypergeometric
           reduction

    • Free pre-print version: Loading...

      Abstract: ABSTRACTFor 1D acoustic wave propagation in the frequency domain, the complex nonlinear Riccati differential equation describes the dependence of the plane-wave reflection response in depth, defined as the upgoing part of the wavefield divided by the downgoing part of the wavefield at that depth. In the case that the model is a velocity gradient interface expressed in terms of a smooth Heaviside function defined by the Fermi-Dirac function, the Riccati equation is shown to have an analytical solution for the reflection response. The solution accounts for all wave phenomena in the gradient zone. The solution of the Riccati equation for the reflection response is obtained by introducing the Riccati equation for the reciprocal of the reflection response which can be related to a second-order self-adjoint linear differential equation. The two latter equations obey the same radiation condition. Furthermore, a connection between the second-order linear equation and the Heun equation is obtained, both being characterized by having four regular singular points. By a change of variables, the Heun equation can be transformed into Gauss’ hypergeometric equation, whereby the solution is expressed as the sum of two linearly independent functions containing hypergeometric functions. The advantage of applying the Heun-to-hypergeometric reduction to the nonlinear Riccati equation is that solutions to the hypergeometric equation are well known. The two functions allow us to analytically determine the Riccati solution because the ratio of the linearly independent functions at infinity is constant. In contrast to the classic frequency-independent reflection coefficient from two layers in welded contact, the reflection coefficient at the surface associated with the gradient interface depends on frequency. In the limit of zero frequency, the reflection coefficient approaches the classic one. We have implemented the iterative solution of the integral form of the Riccati equation and found that the solution builds up nonlinearly to the correct solution. The model is verified by a numerical example valid for 1D wave propagation.
      PubDate: Mon, 22 Aug 2022 00:00:00 GMT
       
 
JournalTOCs
School of Mathematical and Computer Sciences
Heriot-Watt University
Edinburgh, EH14 4AS, UK
Email: journaltocs@hw.ac.uk
Tel: +00 44 (0)131 4513762
 


Your IP address: 34.232.63.94
 
Home (Search)
API
About JournalTOCs
News (blog, publications)
JournalTOCs on Twitter   JournalTOCs on Facebook

JournalTOCs © 2009-