Layer | Fill | Outline |
---|
Map layers
Theme | Visible | Selectable | Appearance | Zoom Range (now: 0) |
---|
Fill | Stroke |
---|---|
Collaborating Authors
Results
Posterior sampling with convolutional neural network-based plug-and-play regularization with applications to poststack seismic inversion
Izzatullah, Muhammad (King Abdullah University of Science and Technology (KAUST)) | Alkhalifah, Tariq (King Abdullah University of Science and Technology (KAUST)) | Romero, Juan (King Abdullah University of Science and Technology (KAUST)) | Corrales, Miguel (King Abdullah University of Science and Technology (KAUST)) | Luiken, Nick (King Abdullah University of Science and Technology (KAUST)) | Ravasi, Matteo (King Abdullah University of Science and Technology (KAUST))
ABSTRACT Uncertainty quantification is a crucial component in any geophysical inverse problem, as it provides decision makers with valuable information about the inversion results. Seismic inversion is a notoriously ill-posed inverse problem, due to the band-limited and noisy nature of seismic data; as such, quantifying the uncertainties associated with the ill-posed nature of this inversion process is essential for qualifying the subsequent interpretation and decision-making processes. Selecting appropriate prior information is a crucial — yet nontrivial — step in probabilistic inversion because it influences the ability of sampling-based inference algorithms to provide geologically plausible posterior samples. However, the necessity to encapsulate prior knowledge into a probability distribution can greatly limit our ability to define expressive priors. To address this limitation and following in the footsteps of the plug-and-play (PnP) methodology for deterministic inversion, we develop a regularized variational inference framework that performs posterior inference by implicitly regularizing the Kullback-Leibler divergence loss — a measure of the distance between the approximated and target probabilistic distributions — with a convolutional neural network-based denoiser. We call this new algorithm PnP Stein variational gradient descent and determine its ability to produce high-resolution trustworthy samples that realistically represent subsurface structures. Our method is validated on synthetic and field poststack seismic data.
- Geophysics > Seismic Surveying > Seismic Processing > Seismic Migration (1.00)
- Geophysics > Seismic Surveying > Seismic Modeling > Velocity Modeling > Seismic Inversion (1.00)
- Geophysics > Seismic Surveying > Seismic Interpretation (1.00)
- Europe > Norway > North Sea > Central North Sea > South Viking Graben > PL 046 > Block 15/9 > Volve Field > Shetland Group > Åsgard Formation (0.99)
- Europe > Norway > North Sea > Central North Sea > South Viking Graben > PL 046 > Block 15/9 > Volve Field > Shetland Group > Svarte Formation (0.99)
- Europe > Norway > North Sea > Central North Sea > South Viking Graben > PL 046 > Block 15/9 > Volve Field > Shetland Group > Sleipner Formation (0.99)
- (17 more...)
- Information Technology > Artificial Intelligence > Representation & Reasoning > Uncertainty > Bayesian Inference (1.00)
- Information Technology > Artificial Intelligence > Machine Learning > Neural Networks > Deep Learning (1.00)
- Information Technology > Artificial Intelligence > Machine Learning > Learning Graphical Models > Directed Networks > Bayesian Learning (0.93)
ABSTRACT Characterizing the elastic properties in deep-buried reservoirs beneath complex overburden structures remains challenging for seismic inversion. Elastic full-waveform inversion (FWI) is capable of quantitatively estimating the subsurface elastic properties with a reasonably high resolution. However, elastic FWI using high frequencies is computationally expensive because fine discretization is required to stabilize the wavefield simulation. In addition, it is challenging for elastic FWI to obtain the high-resolution inversion results of the deep targets due to complex overburden structures and limited energy illumination of the target of interest. To overcome these limitations, we propose a target-oriented high-resolution elastic FWI scheme by using the estimated elastic data for a virtual survey deployed only above a zone of interest. Specifically, we have developed an elastic redatuming approach to retrieve virtual elastic data by solving an iterative least-squares data-fitting problem. In addition to the elastic multicomponent data recorded at the acquisition surface, an estimate of the overburden model is required. A synthetic data example indicates that an overburden model estimated by elastic FWI with a low-frequency band enables our elastic redatuming approach to reconstruct the virtual elastic data representing the seismic reflection response from the target zone. We then perform regularized elastic FWI by using the redatumed elastic data to estimate the elastic properties in the target zone with reduced computational cost. In the numerical examples, the Marmousi2 synthetic data and 2D Volve field data are used to demonstrate the feasibility and practicality of the proposed method.
- Europe > Norway > North Sea > Central North Sea > South Viking Graben > PL 046 > Block 15/9 > Volve Field > Shetland Group > Åsgard Formation (0.99)
- Europe > Norway > North Sea > Central North Sea > South Viking Graben > PL 046 > Block 15/9 > Volve Field > Shetland Group > Svarte Formation (0.99)
- Europe > Norway > North Sea > Central North Sea > South Viking Graben > PL 046 > Block 15/9 > Volve Field > Shetland Group > Sleipner Formation (0.99)
- (20 more...)
ABSTRACT Time-lapse seismic data acquisition is an essential tool to monitor changes in a reservoir due to fluid injection, such as CO2 injection. By acquiring multiple seismic surveys in the exact same location, the authors can identify the reservoir changes by analyzing the difference in the data. However, such analysis can be skewed by the near-surface seasonal velocity variations, inaccuracy, and repeatability in the acquisition parameters, and other inevitable noise. The common practice (cross equalization) to address this problem uses the part of the data in which changes are not expected to design a matching filter and then apply it to the whole data, including the reservoir area. Like cross equalization, the authors train a recurrent neural network (RNN) on parts of the data excluding the reservoir area and then infer the reservoir-related data. The RNN can learn the time dependency of the data, unlike the matching filter that processes the data based on the local information obtained in the filter window. The authors determine the method of matching the data in various examples and compare it with the conventional matching filter. Specifically, we start by demonstrating the ability of the approach in matching two traces and then test the method on a prestack 2D synthetic data. Then, the authors verify the enhancements of the 4D signal by providing reverse time migration images. The authors measure the repeatability using normalized root-mean-square and predictability metrics and find that, in some cases, our proposed method performed better than the matching filter approach.
- Asia > Middle East > Saudi Arabia (0.68)
- North America (0.68)
- Geophysics > Time-Lapse Surveying > Time-Lapse Seismic Surveying (1.00)
- Geophysics > Seismic Surveying > Seismic Modeling > Velocity Modeling (1.00)
- Geophysics > Seismic Surveying > Seismic Processing > Seismic Migration (0.86)
- Europe > Norway > North Sea > Central North Sea > Central Graben > Block 2/8 > Valhall Field > Tor Formation (0.99)
- Europe > Norway > North Sea > Central North Sea > Central Graben > Block 2/8 > Valhall Field > Hod Formation (0.99)
- Europe > Norway > North Sea > Central North Sea > Central Graben > Block 2/11 > Valhall Field > Tor Formation (0.99)
- Europe > Norway > North Sea > Central North Sea > Central Graben > Block 2/11 > Valhall Field > Hod Formation (0.99)
- Reservoir Description and Dynamics > Reservoir Characterization > Seismic processing and interpretation (1.00)
- Reservoir Description and Dynamics > Reservoir Characterization > Four-dimensional and four-component seismic (1.00)
- Data Science & Engineering Analytics > Information Management and Systems > Neural networks (1.00)
ABSTRACT The ample size of time-lapse data often requires significant event detection and source location efforts, especially in areas such as shale gas exploration regions where a large number of microseismic events are often recorded. In many cases, real-time monitoring and location of these events are essential to production decisions. Conventional methods face considerable drawbacks. For example, traveltime-based methods require traveltime picking of often noisy data, whereas migration and waveform inversion methods require expensive wavefield solutions and event detection. Both tasks require some human intervention, and this becomes a big problem when too many sources need to be located, which is common in microseismic monitoring. Machine learning has recently been used to identify microseismic events or locate their sources once they are identified and picked. We have used a novel artificial neural network framework to directly map seismic data, without any event picking or detection, to their potential source locations. We train two convolutional neural networks (CNNs) on labeled synthetic acoustic data containing simulated microseismic events to fulfill such requirements. One CNN, which has a global average pooling layer to reduce the computational cost while maintaining high-performance levels, aims to classify the number of events in the data. The other network predicts the source locations and other source features such as the source peak frequencies and amplitudes. To reduce the size of the input data to the network, we correlate the recorded traces with a central reference trace to allow the network to focus on the curvature of the input data near the zero-lag region. We train the networks to handle single-, multi-, and no-event segments extracted from the data. Tests on a simple vertical varying model and a more realistic Otway field model demonstrate the approach’s versatility and potential.
- North America > United States > West Virginia > Appalachian Basin > Marcellus Shale Formation (0.99)
- North America > United States > Virginia > Appalachian Basin > Marcellus Shale Formation (0.99)
- North America > United States > Pennsylvania > Appalachian Basin > Marcellus Shale Formation (0.99)
- (3 more...)
High-dimensional wavefield solutions based on neural network functions
Alkhalifah, Tariq (KAUST) | Song, Chao (KAUST) | Huang, Xinquan (KAUST)
Wavefield solutions are critical for applications ranging from imaging to full waveform inversion. These wavefields are often large, especially for 3D media, and multiple point sources, like Green’s functions. A recently introduced framework based on neural networks admitting functional solutions to partial differential equations (PDEs) offers the opportunity to solve the Helmholtz equation with a neural network (NN) model. The input to such an NN is a location in space and the output are the real and imaginary parts of the scattered wavefield at that location, thus, acting like a function. The network is trained on random input points in space and a variance of the Helmholtz equation for the scattered wavefield is used as the loss function to update the network parameters. In spite of the methods flexibility, like handling irregular surfaces and complex media, and its potential for velocity model building, the cost of training the network far exceeds that of numerical solutions. Relying on the network’s ability to learn wavefield features, we extend the dimension of this NN function to learn the wavefield for many sources and frequencies, simultaneously. We show, in this preliminary study, that reasonable wavefield solutions can be predicted using smaller networks. This includes wavefields for frequencies not within the training range. The new NN function has the potential to efficiently represent the wavefield as a function of location in space, as well as source location and frequency.
Target-oriented time-lapse elastic full-waveform inversion assisted by deep learning with prior information
Li, Yuanyuan (King Abdullah University of Science and Technology (KAUST)) | Bakulin, Andrey (EXPEC ARC, Saudi Aramco) | Nivlet, Philippe (EXPEC ARC, Saudi Aramco) | Smith, Robert (EXPEC ARC, Saudi Aramco) | Alkhalifah, Tariq (King Abdullah University of Science and Technology (KAUST))
Time-lapse (TL) monitoring of the elastic property changes in the reservoir of interest is important for optimizing the reservoir interpretation and development plan. Given that elastic full-waveform inversion (EFWI) provides quantitative estimations of the elastic properties (Vp and Vs), its application to time-lapse elastic data is of considerable interest. For practical applications in reservoir monitoring, we need EFWI to provide high-resolution reservoir information at a reasonable cost. Thus, we develop an elastic redatuming technique to provide the required virtual elastic data for a target-oriented inversion, thus improving the computational efficiency by focusing our full-band inversion on the target zone. To improve the inversion resolution, we combine the well information and seismic data in the proposed time-lapse inversion approach using a regularized objective function. To derive the required prior model, we train a deep neural network (DNN) to learn the connection between the seismic estimation and the facies interpreted from well logs. We then apply the trained network to the target inversion domain to predict a prior model. Given the prior model, we perform another time-lapse inversion. We fit the simulated data difference for the virtual survey to the redatumed one from the surface recording and fit the model changes to the predicted prior model. The numerical results demonstrate that the proposed method enables the recovery of the time-lapse changes effectively in the target zone by incorporating the learned model changes from well logs.
Frequency-domain reflection waveform inversion with generalized internal multiple imaging
Wang, Guanchao (China University of Petroleum-Beijing (CUPB), China Railway Design Corporation) | Guo, Qiang (King Abdullah University of Science and Technology, KAUST, Shearwater Geoservices) | Alkhalifah, Tariq (King Abdullah University of Science and Technology, KAUST) | Wang, Shangxu (China University of Petroleum-Beijing (CUPB))
ABSTRACT Full-waveform inversion (FWI) has the potential to provide a high-resolution detailed model of the earth’s subsurface, but it often fails to do so if the starting model differs significantly from the true one. Reflection waveform inversion (RWI) is a popular method for building a sufficiently accurate initial model for FWI. In traditional RWI, the low-wavenumber updates are always computed and captured by smearing the data misfit along the reflection path with the help of migration/demigration. However, the success of RWI relies heavily on accurately reproducing the data in demigration. Thus, we have introduced a new generalized internal multiple imaging-based RWI (GIMI-RWI) implementation, in which we avoid the Born modeling and update the primary reflection kernel directly. In GIMI-RWI, we store one reflection kernel for each source-receiver pair, preserving the unique wavepath for every single source-receiver trace. Subsequently, the convolution between the data residuals and the corresponding reflection kernel can build the tomographic velocity updates. In this situation, the long-wavelength tomographic updates are free of migration footprints and will contribute a smoother background velocity to reduce the cycle-skipping risk and stabilize the followed FWI process. In addition, the GIMI-RWI method is source independent because it entirely relies on the data. Using a synthetic example extracted from the Sigsbee2A model, we find the reliable performance of the GIMI-RWI technique.
- Geophysics > Seismic Surveying > Seismic Processing (1.00)
- Geophysics > Seismic Surveying > Seismic Modeling > Velocity Modeling > Seismic Inversion (1.00)
Mapping full seismic waveforms to vertical velocity profiles by deep learning
Kazei, Vladimir (King Abdullah University of Science and Technology (KAUST), Aramco Services Company) | Ovcharenko, Oleg (King Abdullah University of Science and Technology (KAUST)) | Plotnitskii, Pavel (King Abdullah University of Science and Technology (KAUST)) | Peter, Daniel (King Abdullah University of Science and Technology (KAUST)) | Zhang, Xiangliang (King Abdullah University of Science and Technology (KAUST)) | Alkhalifah, Tariq (King Abdullah University of Science and Technology (KAUST))
ABSTRACT Building realistic and reliable models of the subsurface is the primary goal of seismic imaging. We have constructed an ensemble of convolutional neural networks (CNNs) to build velocity models directly from the data. Most other approaches attempt to map full data into 2D labels. We exploit the regularity of seismic acquisition and train CNNs to map gathers of neighboring common midpoints (CMPs) to vertical 1D velocity logs. This allows us to integrate well-log data into the inversion, simplify the mapping by using the 1D labels, and accommodate larger dips relative to using single CMP inputs. We dynamically generate the training data in parallel with training the CNNs, which reduces overfitting. Data generation and training of CNNs is more computationally expensive than conventional full-waveform inversion (FWI). However, once the network is trained, data sets with similar acquisition parameters can be inverted much faster than with FWI. The multiCMP CNN ensemble is tested on multiple realistic synthetic models, performs well, and was combined with FWI for even better performance.
- North America > United States (0.93)
- Asia > Middle East > Saudi Arabia (0.28)
- Geophysics > Seismic Surveying > Seismic Processing (1.00)
- Geophysics > Seismic Surveying > Seismic Modeling > Velocity Modeling > Seismic Inversion (1.00)
ABSTRACT Full-waveform inversion (FWI) is popularly used to obtain a high-resolution subsurface velocity model. However, it requires either a good initial velocity model or low-frequency data to mitigate the cycle-skipping issue. Reflection-waveform inversion (RWI) uses a migration/demigration process to retrieve a background model that can be used as a good initial velocity in FWI. The drawback of conventional RWI is that it requires the use of least-squares migration, which is often computationally expensive, and it is still prone to cycle skipping at far offsets. To improve the computational efficiency and overcome cycle skipping in the original RWI, we have incorporated it into a recently introduced method called efficient wavefield inversion (EWI) by inverting for the Born-scattered wavefield instead of the wavefield itself. In this case, we use perturbation-related secondary sources in the modified source function. Unlike conventional RWI, the perturbations are calculated naturally as part of the calculation of the scattered wavefield in an efficient way. Because the sources in the reflection-based EWI (REWI) are located in the subsurface, we are able to update the background model along the reflection wavepath. In the background velocity inversion, we calculate the background perturbation by a deconvolution process at each frequency. After obtaining the REWI inverted velocity model, a sequential FWI or EWI is needed to obtain a high-resolution model. We determine the validity of our approach using synthetic data generated from a section of the Sigsbee2A model. To further demonstrate the effectiveness of our approach, we test it on an ocean-bottom cable data set from the North Sea. We find that our methodology leads to improved velocity models as evidenced by flatter angle gathers.
- Europe > Norway > North Sea > Central North Sea > South Viking Graben > PL 046 > Block 15/9 > Volve Field > Shetland Group > Åsgard Formation (0.99)
- Europe > Norway > North Sea > Central North Sea > South Viking Graben > PL 046 > Block 15/9 > Volve Field > Shetland Group > Svarte Formation (0.99)
- Europe > Norway > North Sea > Central North Sea > South Viking Graben > PL 046 > Block 15/9 > Volve Field > Shetland Group > Sleipner Formation (0.99)
- (17 more...)
Target-oriented time-lapse waveform inversion using deep learning-assisted regularization
Li, Yuanyuan (King Abdullah University of Science and Technology) | Alkhalifah, Tariq (King Abdullah University of Science and Technology) | Guo, Qiang (King Abdullah University of Science and Technology, Shearwater Geoservices)
ABSTRACT Detection of the property changes in the reservoir during injection and production is important. However, the detection process is very challenging using surface seismic surveys because these property changes often induce subtle changes in the seismic signals. The quantitative evaluation of the subsurface property obtained by full-waveform inversion allows for better monitoring of these time-lapse changes. However, high-resolution inversion is usually accompanied with a large computational cost. Besides, the resolution of inversion is limited by the bandwidth and aperture of time-lapse seismic data. We have applied a target-oriented strategy through seismic redatuming to reduce the computational cost by focusing our high-resolution delineation on a relatively small zone of interest. The redatuming technique generates time-lapse virtual data for the target-oriented inversion. Considering that the injection and production wells are often present in the target zone, we can incorporate the well velocity information with the time-lapse inversion by using regularization to complement the resolution and illumination at the reservoir. We use a deep neural network to learn the statistical relationship between the inverted model and the facies interpreted from well logs. The trained network is used to map the property changes extracted from the wells to the target inversion domain. We then perform another time-lapse inversion, in which we fit the predicted data difference to the redatumed one from observation, as well as fit the model to the predicted velocity changes. The numerical results demonstrate that our method is capable of inverting for the time-lapse property changes effectively in the target zone by incorporating the learned model information from well logs.
- Research Report > New Finding (0.48)
- Research Report > Experimental Study (0.48)
- Reservoir Description and Dynamics > Reservoir Characterization > Seismic processing and interpretation (1.00)
- Reservoir Description and Dynamics > Formation Evaluation & Management > Seismic (four dimensional) monitoring (1.00)
- Data Science & Engineering Analytics > Information Management and Systems > Neural networks (1.00)