EFWI: Multiparameter benchmark datasets for elastic full waveform inversion of geophysical properties
Explore the source record for details and available documents.
Feng, Shihang↗
SEARCH · Engineering Papers
Results for “full waveform inversion”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
ML-Based Rock Properties and Seismic Volume Enhancement
This project aims to improve field-scale Carbon Capture and Storage (CCS) assessments by enhancing petrophysical and geophysical log predictions through machine learning and neural networks. In our work during EY23, we applied Conditional Variational Autoencoders (CVAEs) to predict compressional velocity (Vp) and assess CO2 saturation levels in geological formations at the Illinois Basin Decatur Project (IBDP). In another task, we improved full-waveform inversion (FWI) methods with machine-learning approaches using lithological constraints. Full-waveform inversion (FWI) of seismic data estimates the elastic properties of subsurface rocks with high spatial resolution.
Nathanail, Athanasios↗
Real-time deep-learning inversion of seismic full waveform data for CO 2 saturation and uncertainty in geological carbon storage monitoring
Deep-learning inversion has recently drawn attention in geological carbon storage research due to its potential of imaging and monitoring carbon storage in real time, significantly improving efficiency and safety of carbon storage operations. We present a deep-learning full waveform inversion method that after the neural network has been trained can image CO 2 saturation and its uncertainty in real time. Our deep-learning inversion method is based on the U-Net architecture with the neural network trained on pairs of synthetic seismic data and CO 2 saturation models. Accordingly, our training establishes a mapping relationship between seismic data and CO 2 saturation models and once fully trained directly estimates CO 2 saturation as a function of subsurface location. We further quantify uncertainties of CO 2 saturation estimates using the Monte Carlo dropout method and a bootstrap aggregating method. For this proof-of-concept study, the CO 2 training models and data are derived from the Kimberlina 1.2 model, a hypothetical 3D geological carbon storage model that is constructed based on various geological and hydrological data from the Southern San Joaquin Basin, California. We perform deep-learning inversion experiments using noise-free and noisy training and test data sets and compare the results. Our modelling experiments show that (1) the deep-learning inversion can estimate 2D distributions of CO 2 fairly well even in the presence of Gaussian random noise and (2) both CO 2 saturation imaging and uncertainty quantification can be done in real time. Our results suggest that the deep-learning inversion method can serve as a robust real-time monitoring tool for geological carbon storage and/or other time-varying reservoir/aquifer properties that result from injection, extraction, and/or other subsurface transport phenomena.
58 GEOSCIENCES↗
Joint physics-based and data-driven time-lapse seismic inversion: Mitigating data scarcity
In carbon capture and sequestration (CCS), developing rapid and effective imaging techniques is crucial for real-time monitoring of the spatial and temporal dynamics of CO 2 propagation during/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 CO 2 sequestration. We propose 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 vali dated on a synthetic CO 2 -sequestration model based on the Kimberlina storage reservoir in California. The proposed approach is shown to synthesize a large volume of high-quality, physically realistic training data, which is critically important in accurately characterizing the CO 2 movement in the reservoir. In conclusion, 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.
58 GEOSCIENCES↗
Three-dimensional seismic characterization and imaging of the Soda Lake geothermal field
Accurate characterization of subsurface geophysical properties and detection of the fault system are essential for geothermal energy exploration and production. The Soda Lake geothermal field is in western Nevada with a complex fault system. Previous seismic characterization only produced a low-resolution, smooth velocity model along with a simple, conceptual fault model. Using optimized correlation-based full-waveform inversion, wavefield-separation-based reverse-time migration, and automatic fault detection techniques, we present 3D seismic characterization for the Soda Lake geothermal field using 3D surface seismic data acquired with Vibroseis sources. Here, we obtain 3D high-resolution velocity, density, and acoustic impedance models, 3D seismic images with different grid spacings, and a high-resolution fault system. Consistency check between the constructed faults and currently active injection and production geothermal wells verifies that our seismic inversion and imaging results and detected faults are reliable. These results can provide valuable information for optimizing well placement and geothermal energy production at the Soda Lake geothermal field.
15 GEOTHERMAL ENERGY↗
Making Invisible Visible: Data-Driven Seismic Inversion With Spatio-Temporally Constrained Data Augmentation
Deep learning and data-driven approaches have shown great potential in scientific domains. The promise of data-driven techniques relies on the availability of a large volume of high-quality training datasets. Due to the high cost of obtaining data through expensive physical experiments, instruments, and simulations, data augmentation techniques for scientific applications have emerged as a new direction for obtaining scientific data recently. However, existing data augmentation techniques originating from computer vision yield physically unacceptable data samples that are not helpful for the domain problems that we are interested in. In this article, we develop new data augmentation techniques based on convolutional neural networks. Specifically, our generative models leverage different physics knowledge (such as governing equations, observable perception, and physics phenomena) to improve the quality of the synthetic data. To validate the effectiveness of our data augmentation techniques, we apply them to solve a subsurface seismic full-waveform inversion using simulated CO 2 leakage data. Our interest is to invert for subsurface velocity models associated with very small CO 2 leakage. We validate the performance of our methods using comprehensive numerical tests. Here via comparison and analysis, we show that data-driven seismic imaging can be significantly enhanced by using our data augmentation techniques. Particularly, the imaging quality has been improved by 15% in test scenarios of general-sized leakage and 17% in small-sized leakage when using an augmented training set obtained with our techniques.
58 GEOSCIENCES↗
Seismic Characterization of the Blue Mountain Geothermal Field
Subsurface characterization is crucial for geothermal energy exploration and production. Yet hydrothermal reservoirs usually reside in highly fractured and faulted zones where accurate characterization is very challenging because of low signal-to-noise ratios of land seismic data and lack of coherent reflection signals. We perform an active-source seismic characterization for the Blue Mountain geothermal field in Nevada using active seismic data to reveal the elastic medium property complexity and fault distribution at this field. We first employ an unsupervised machine learning method to attenuate groundroll and near-surface guided-wave noise and enhance coherent reflection and scattering signals from noisy seismic data. We then build a smooth initial P-wave velocity model based on an existing magnetotellurics survey result, and use 3D first-arrival traveltime tomography to refine the initial velocity model. We then derive a set of elastic wave velocities and anisotropic parameters using elastic full-waveform inversion, and obtain PP and PS images using elastic reverse-time migration. We identify major faults by analyzing the variations of seismic velocities and anisotropy parameters, and reveal mid- to small-scale faults by applying a supervised machine learning method to the seismic migration images. Our characterization reveals complex velocity heterogeneities and anisotropies, as well as faults, with a high spatial resolution. These results can provide valuable information for optimal placement of future injection and production wells to increase geothermal energy production at the Blue Mountain geothermal power plant.
58 GEOSCIENCES↗
Toward more-robust, AI-enabled subsurface seismic imaging for geotechnical applications
Non-invasive seismic imaging has the potential to cost-effectively evaluate large volumes of subsurface material to inform geotechnical site investigation. However, seismic imaging using full waveform inversion (FWI) requires significant computational time and is dependent on an initial starting model. As a result, FWI has not yet been widely adopted into geotechnical practice. Previous efforts, on relatively simple two-layered models, indicate that data-driven artificial intelligence (AI) models may be as effective as FWI at predicting 2D images of shear wave velocity (V s ). Furthermore, the AI model predictions can be made almost instantaneously after data acquisition and do not require an initial starting model. We examine the generality of these findings by developing a new AI model for subsurface seismic imaging, whereby we make several notable contributions. First, we architect a multimodal AI model that combines time- and frequency-domain representations of the seismic wavefield to predict a 50 m by 20 m subsurface image of V s . Second, we developed a new diverse dataset of 100,000 images with their corresponding seismic wavefields to train the AI model. Third, we propose four physics-informed data augmentations for data-driven seismic imaging. Fourth, we develop two prediction consistency tests to evaluate the model’s performance when the true subsurface is unknown. Our final model, which has been made publicly available, is capable of predicting a subsurface V s image from a single seismic wavefield with an average, mean absolute percent error (MAPE) of 24 %. The predictive model is applied to a field dataset and shown to be consistent with local geology and shear-wave refraction measurements from the same location.
Artificial intelligence↗
SPiRaL: a multiresolution global tomography model of seismic wave speeds and radial anisotropy variations in the crust and mantle
SUMMARY SPiRaL is a joint global-scale model of wave speeds (P and S) and anisotropy (vertical transverse isotropy, VTI) variations in the crust and mantle. The model is comprised of >2.1 million nodes with five parameters at each node that capture velocity variations for P- and S-waves travelling at arbitrary directions in transversely isotropic media with a vertical symmetry axis (VTI). The crust (including ice, water, sediments and crystalline layers) is directly incorporated into the model. The default node spacing is approximately 2° in the lower mantle and 1° in the crust and upper mantle. The grid is refined with ∼0.25° minimum node spacing in highly sampled regions of the crust and upper mantle throughout North America and Eurasia. The data considered in the construction of SPiRaL includes millions of body wave traveltimes (crustal, regional and teleseismic phases with multiples) and surface wave (Rayleigh and Love) dispersion. A multiresolution inversion approach is employed to capture long-wavelength heterogeneities commonly depicted in global-scale tomography images as well as more localized details that are typically resolved in more focused regional-scale studies. Our previous work has demonstrated that such global-scale models with regional-scale detail can accurately predict both teleseismic and regional body wave traveltimes, which is necessary for more accurate location of small seismic events that may have limited signal at teleseismic distances. SPiRaL was constructed to predict traveltimes for event location and long-period waveform dispersion for seismic source inversion applications in regions without sufficiently tuned models. SPiRaL may also serve as a starting model for full-waveform inversion (FWI) with the goal of fitting waves with periods 10–50 s over multiple broad regions (thousands of kilometres) and potentially the globe. To gain insight to this possibility, we simulated waveforms for a small set of events using SPiRaL and independent waveform-based models for comparison. For the events tested, the performance of the traveltime-based SPiRaL model is shown to be generally on par with regional 3-D waveform-based models in three regions (western United States, Middle East, Korean Peninsula) suggesting SPiRaL may serve as a starting model for FWI over broad regions.
58 GEOSCIENCES↗
Integration of seismic-pressure-petrophysics inversion of continuous active-seismic monitoring data for monitoring and quantifying CO 2 plume (Final Report)
The overall objective of this project is to develop and validate an integrated package of joint seismic-pressure-petrophysics inversion (jSPPI) of continuous active-source seismic monitoring dataset capable of providing real-time monitoring of CO 2 plume during geologic carbon sequestration (GCS). The three specific developments include: (a) the methodologies for fast seismic full waveform inversion of continuous active source seismic monitoring, (CASSM) datasets for simultaneously estimating velocity and attenuation, and with data assimilation; (b) joint Bayesian petrophysical inversion of seismic models and pressure data for providing and updating CO 2 saturation models; (c) the methods using multiple datasets including (Crainfield and Frio-II borehole) synthetic, laboratory, and field CASSM datasets. The outcomes of jSPPI include (a) a workflow for processing CASSM data, (b) Bayesian inversion algorithms using CASSM data and pressure response data, and (c) integration with data assimilation algorithms for continuously updating site-specific models used for prediction and reservoir management. The validation of joint FWI will be conducted using synthetic models based on the Cranfield and Frio experiments as well as field CASSM datasets collected as part of the Frio-II pilot injection. To quantify and map the mass and distribution of CO 2 (saturation), we will jointly invert velocity and attenuation measurements from the FWI with a Bayesian approach using a rock physics model for attenuation (e.g., White’s attenuation model with two selected patch sizes (White, 1976; Dutta and Seriff, 1979)). The Bayesian inversion will be applied to each time step in the CASSM survey in an updating scheme, which integrates with an ensemble of reservoir simulations at each step. A more complete experimental validation dataset will be collected as part of a mesoscale (2-3 m) gas-CO 2 injection experiment utilizing a higher frequency version of the CASSM system developed for laboratory studies; the integrated inversion will be demonstrated using this dataset which will provide both a dense geometry as well as more precise secondary confirmation measurements (e.g. saturation) typically not available in the field. The resulting real-time map of CO 2 saturation is able to provide a deeper scientific understanding of the complex, time-varying dynamics of subsurface fluid flow migration path as well as the rapid detection of CO 2 leakage hazards.
25 ENERGY STORAGE↗
The Collaborative Seismic Earth Model: Generation 2
Geological interpretations, earthquake source inversions and ground motion modeling, among other applications, require models that jointly resolve crustal and mantle structure. With the second generation of the Collaborative Seismic Earth Model (CSEM2), we present a global multi-resolution tomographic Earth model that serves this purpose. The model evolves through successive regional- and global-scale refinements. While the first generation aggregated regional models, with this study, we ensure consistency between all individual submodels, resulting in a model that accurately explains wave propagation across scales. Recent regional tomographic models were incorporated, comprising continental-scale inversions for Asia and Africa, as well as regional inversions for the Western US, Central Andes, Iran, and Southeast Asia. Across all regional refinements, over 793,000 source-receiver pairs contributed. Moreover, the long-wavelength Earth model (LOWE) introduces large-scale structures outside of pre-existing local refinements. A full-waveform inversion for global anisotropic P-and S-wave speed structure over a total of 194 iterations with a minimum period of 50 s on a large data set of 1 hr of waveform data from 2,423 earthquakes and over 6 million source-receiver pairs ensures that regional updates in the crust and uppermost mantle translate into updates of deeper, global-scale structure. To test the performance of CSEM2, we evaluate waveform fits between observed and synthetic seismograms at 50 s for an independent data set on the global scale, and on the regional scale for lower periods. We accurately simulate waveforms within and across regional refinements, maintaining the original resolution of the submodels embedded in the global framework.
58 GEOSCIENCES↗
Paired autoencoders for likelihood-free estimation in inverse problems
Abstract We consider the solution of nonlinear inverse problems where the forward problem is a discretization of a partial differential equation. Such problems are notoriously difficult to solve in practice and require minimizing a combination of a data-fit term and a regularization term. The main computational bottleneck of typical algorithms is the direct estimation of the data misfit. Therefore, likelihood-free approaches have become appealing alternatives. Nonetheless, difficulties in generalization and limitations in accuracy have hindered their broader utility and applicability. In this work, we use a paired autoencoder framework as a likelihood-free estimator (LFE) for inverse problems. We show that the use of such an architecture allows us to construct a solution efficiently and to overcome some known open problems when using LFEs. In particular, our framework can assess the quality of the solution and improve on it if needed. We demonstrate the viability of our approach using examples from full waveform inversion and inverse electromagnetic imaging.
Chung, Matthias (ORCID:0000000178224539)↗
Global centroid moment tensor solutions in a heterogeneous earth: the CMT3D catalogue
SUMMARY For over 40 yr, the global centroid-moment tensor (GCMT) project has determined location and source parameters for globally recorded earthquakes larger than magnitude 5.0. The GCMT database remains a trusted staple for the geophysical community. Its point-source moment-tensor solutions are the result of inversions that model long-period observed seismic waveforms via normal-mode summation for a 1-D reference earth model, augmented by path corrections to capture 3-D variations in surface wave phase speeds, and to account for crustal structure. While this methodology remains essentially unchanged for the ongoing GCMT catalogue, source inversions based on waveform modelling in low-resolution 3-D earth models have revealed small but persistent biases in the standard modelling approach. Keeping pace with the increased capacity and demands of global tomography requires a revised catalogue of centroid-moment tensors (CMT), automatically and reproducibly computed using Green's functions from a state-of-the-art 3-D earth model. In this paper, we modify the current procedure for the full-waveform inversion of seismic traces for the six moment-tensor parameters, centroid latitude, longitude, depth and centroid time of global earthquakes. We take the GCMT solutions as a point of departure but update them to account for the effects of a heterogeneous earth, using the global 3-D wave speed model GLAD-M25. We generate synthetic seismograms from Green's functions computed by the spectral-element method in the 3-D model, select observed seismic data and remove their instrument response, process synthetic and observed data, select segments of observed and synthetic data based on similarity, and invert for new model parameters of the earthquake’s centroid location, time and moment tensor. The events in our new, preliminary database containing 9382 global event solutions, called CMT3D for ‘3-D centroid-moment tensors’, are on average 4 km shallower, about 1 s earlier, about 5 per cent larger in scalar moment, and more double-couple in nature than in the GCMT catalogue. We discuss in detail the geographical and statistical distributions of the updated solutions, and place them in the context of earlier work. We plan to disseminate our CMT3D solutions via the online ShakeMovie platform.
58 GEOSCIENCES↗
Memory-efficient nonsmooth dynamic optimization using adaptive randomized compression
Dynamic optimization problems arise in many applications including flow control, full waveform inversion, and medical imaging. These problems are plagued by significant computational challenges. One such challenge — and the focus of this work — is the memory limitation induced by the size of the underlying dynamical system. In particular, the entire dynamic trajectory is required for derivative computation and therefore must be stored or recomputed using, e.g., checkpointing. Although recent work demonstrated the use of adaptive randomized sketching to overcome the memory challenge, that work only applies to smooth unconstrained problems, prohibiting its use for nonsmooth regularized and constrained problems. The inclusion of nonsmooth regularizers and constraints is critical as they often arise in an attempt to preserve certain physical properties or to promote sparsity. To solve these problems, we introduce a trust-region algorithm for minimizing the sum of a smooth nonconvex function and a nonsmooth convex function that leverages randomized sketching to compress the dynamical system trajectories and adaptively adjust the sketch rank to satisfy a gradient inexactness condition. We prove convergence of this algorithm and demonstrate that it achieves substantial memory reduction on three discretized PDE-constrained optimization applications.
97 MATHEMATICS AND COMPUTING↗
Combining different 3-D global and regional seismic wave propagation solvers towards box tomography in the deep Earth
SUMMARY In previous publications, we presented a general framework, which we called ‘box tomography’, that allows the coupling of any two different numerical seismic wave propagation solvers, respectively outside and inside a target region, or ‘box’. The goal of such hybrid wavefield computations is to reduce the cost of computations in the context of full-waveform inversion for structure within the target region, when sources and/or receivers are located at large distances from the box. Previously, we had demonstrated this approach with sources and receivers outside the target region in a 2-D acoustic spherical earth model, and demonstrated and applied this methodology in the 3-D spherical elastic Earth in a continental scale inversion in which all stations were inside the target region. Here we extend the implementation of the approach to the case of a 3-D global elastic earth model in the case where both sources and stations are outside the box. We couple a global 3-D solver, SPECFEM3D_GLOBE, for the computation of the wavefield and Green’s functions in a reference 3-D model, with a regional 3-D solver, RegSEM, for the computation of the wavefield within the box, by means of time-reversal mirrors. We briefly review key theoretical aspects, showing in particular how only the displacement is needed to be stored at the boundary of the box. We provide details of the practical implementation, including the geometrical design of the mirrors, how we deal with different sizes of meshes in the two solvers, and how we address memory-saving through the use of B-spline compression of the recorded wavefield on the mirror. The proposed approach is numerically efficient but also versatile, since adapting it to other solvers is straightforward and does not require any changes in the solver codes themselves, as long as the displacement can be recovered at any point in time and space. We present benchmarks of the hybrid computations against direct computations of the wavefield between a source and an array of stations in a realistic geometry centred in the Yellowstone region, with and without a hypothetical plume within the ‘box’, and with a 1-D or a 3-D background model, down to a period of 20 s. The ultimate goal of this development is for applications in the context of imaging of remote target regions in the deep mantle, such as, for example, Ultra Low Velocity Zones.
Geochemistry & Geophysics↗
Machine learning for seismic low-frequency extrapolation
The cycle-skipping problem that plagues full waveform inversion (FWI) can be at least partially mitigated if low frequencies (which encode the kinematics of wave propagation in seismic data) are recorded. However, seismic sources and receivers are band-limited, so seismic data does not generally include signals down to 0 Hz. To improve our ability to solve the seismic inverse problem, one can synthesize this missing low-frequency (LF) content from the recorded high-frequency (HF) data using machine learning (ML) models. Deep learning models such as convolutional neural networks (CNNs) demonstrate impressive ability to perform low frequency extrapolation. However, such models require powerful hardware (GPU machines) and careful training. We assess the extrapolation capabilities of three different ML models that do not require GPU machines, namely, random forest, Gaussian process regression and gradient boosting, on both synthetic and real data. Experimental results on two synthetic data sets (generated from a low velocity lens embedded in a homogeneous medium, and the Marmousi model) demonstrate that FWI applied to the extrapolated data consistently improves inversion accuracy relative to FWI applied to the original data sets that do not contain low frequencies. Application of low-frequency extrapolation to real data from the Northwest Shelf of Australia demonstrates that tree-based ML models such as gradient boosting can outperform CNNs in terms of both accuracy and computational cost on non-GPU architectures.
58 GEOSCIENCES↗
Randomized Sketching Algorithms for Low-Memory Dynamic Optimization
This paper develops a novel limited-memory method to solve dynamic optimization problems. The memory requirements for such problems often present a major obstacle, particularly for problems with PDE constraints such as optimal flow control, full waveform inversion, and optical tomography. In these problems, PDE constraints uniquely determine the state of a physical system for a given control; the goal is to find the value of the control that minimizes an objective. While the control is often low dimensional, the state is typically more expensive to store. This paper suggests using randomized matrix approximation to compress the state as it is generated and shows how to use the compressed state to reliably solve the original dynamic optimization problem. Concretely, the compressed state is used to compute approximate gradients and to apply the Hessian to vectors. The approximation error in these quantities is controlled by the target rank of the sketch. This approximate first- and second-order information can readily be used in any optimization algorithm. As an example, we develop a sketched trust-region method that adaptively chooses the target rank using a posteriori error information and provably converges to a stationary point of the original problem. Numerical experiments with the sketched trust-region method show promising performance on challenging problems such as the optimal control of an advection-reaction-diffusion equation and the optimal control of fluid flow past a cylinder.
97 MATHEMATICS AND COMPUTING↗
Decoupled Fréchet kernels based on a fractional viscoacoustic wave equation
We formulate the Fréchet kernel computation using the adjoint-state method based on a fractional viscoacoustic wave equation. We first numerically prove that both the 1/2- and the 3/2-order fractional Laplacian operators are self-adjoint. Using this property, we show that the adjoint wave propagator preserves the dispersion and compensates the amplitude, while the time-reversed adjoint wave propagator behaves identically as the forward propagator with the same dispersion and dissipation characters. Without introducing rheological mechanisms, this formulation adopts an explicit Q parameterization, which avoids the implicit Q in the conventional viscoacoustic/viscoelastic full waveform inversion (Q-FWI). In addition, because of the decoupling of operators in the wave equation, the viscoacoustic Fréchet kernel is separated into three distinct contributions with clear physical meanings: lossless propagation, dispersion, and dissipation. Here, we find that the lossless propagation kernel dominates the velocity kernel, while the dissipation kernel dominates the attenuation kernel over the dispersion kernel.
58 GEOSCIENCES↗