scholarly journals Horizontal contraction in image domain for velocity inversion

Geophysics ◽  
2015 ◽  
Vol 80 (3) ◽  
pp. R95-R110 ◽  
Author(s):  
Peng Shen ◽  
William W. Symes
Keyword(s):  
Image Domain ◽  
Velocity Inversion

Combining reflection and transmission information in time-lapse velocity inversion: A new hybrid approach

Geophysics ◽  
2019 ◽  
Vol 84 (4) ◽  
pp. R601-R611 ◽  
Author(s):  
Maria Kotsi ◽  
Jonathan Edgar ◽  
Alison Malcolm ◽  
Sjoerd de Ridder
Keyword(s):  
Waveform Inversion ◽  
Hybrid Approach ◽  
Time Lapse ◽  
Velocity Model ◽  
Nonlinear Inversion ◽  
Image Domain ◽  
Velocity Inversion ◽  
Reflected Waves ◽  
Deep Diving ◽  
Velocity Changes

Full-waveform inversion (FWI) uses the information of the full wavefield to deliver high-resolution images of the subsurface. Conventional time-lapse FWI primarily uses the transmitted component (diving waves) of the wavefield to reconstruct the low-wavenumber component of the velocity model. This requires large-offset surveys and low-frequency data. When the target of interest is deep, diving waves cannot reach the target and FWI will be dominated by the reflected component of the wavefield. Consequently, the retrieved model resembles a least-squares migration instead of a velocity model. Image-domain methods, especially image-domain wavefield tomography (IDWT), have been developed to obtain a model of time-lapse velocity changes in deeper targets using reflected waves. The method is able to recover models of deep targets. However, it also tends to obtain smeared time-lapse velocity changes. We have developed a form of time-lapse waveform inversion that we call dual-domain time-lapse waveform inversion (DDWI), whose objective function joins FWI and IDWT, combining information from the diving waves in the data-domain FWI term with information from the reflected waves in the image-domain IDWT term. During the nonlinear inversion, the velocity model is updated using constraints from both terms simultaneously. Similar to sequential time-lapse waveform inversion, we start the time-lapse inversion from a baseline model recovered with FWI. We test DDWI on a variety of synthetic models of increasing complexity and find that it can recover time-lapse velocity changes more accurately than when both methods are used independently or sequentially.

Nonlinear extended images via image-domain interferometry

Geophysics ◽  
2010 ◽  
Vol 75 (6) ◽  
pp. SA105-SA115 ◽  
Author(s):  
Ivan Vasconcelos ◽  
Paul Sava ◽  
Huub Douma
Keyword(s):  
Wave Equation ◽  
Seismic Imaging ◽  
Theory And Practice ◽  
Seismic Interferometry ◽  
Image Domain ◽  
Velocity Inversion ◽  
Nonlinear Imaging ◽  
The One ◽  
Scattered Fields ◽  
And Inversion

Wave-equation, finite-frequency imaging and inversion still face many challenges in addressing the inversion of highly complex velocity models as well as in dealing with nonlinear imaging (e.g., migration of multiples, amplitude-preserving migration). Extended images (EIs) are particularly important for designing image-domain objective functions aimed at addressing standing issues in seismic imaging, such as two-way migration velocity inversion or imaging/inversion using multiples. General one- and two-way representations for scattered wavefields can describe and analyze EIs obtained in wave-equation imaging. We have developed a formulation that explicitly connects the wavefield correlations done in seismic imaging with the theory and practice of seismic interferometry. In light of this connection, we define EIs as locally scattered fields reconstructed by model-dependent, image-domain interferometry. Because they incorporate the same one- and two-way scattering representations usedfor seismic interferometry, the reciprocity-based EIs can in principle account for all possible nonlinear effects in the imaging process, i.e., migration of multiples and amplitude corrections. In this case, the practice of two-way imaging departs considerably from the one-way approach. We have studied the differences between these approaches in the context of nonlinear imaging, analyzing the differences in the wavefield extrapolation steps as well as in imposing the extended imaging conditions. When invoking single-scattering effects and ignoring amplitude effects in generating EIs, the one- and two-way approaches become essentially the same as those used in today’s migration practice, with the straightforward addition of space and time lags in the correlation-based imaging condition. Our formal description of the EIs and the insight that they are scattered fields in the image domain may be useful in further development of imaging and inversion methods in the context of linear, migration-based velocity inversion or in more sophisticated image-domain nonlinear inverse scattering approaches.

Image-domain velocity inversion and event location for microseismic monitoring

Geophysics ◽  
2017 ◽  
Vol 82 (5) ◽  
pp. KS71-KS83 ◽  
Author(s):  
Ben Witten ◽  
Jeffrey Shragge
Keyword(s):  
Oil And Gas ◽  
Random Noise ◽  
Microseismic Monitoring ◽  
Seismic Events ◽  
S Wave ◽  
Origin Time ◽  
Image Domain ◽  
Velocity Inversion ◽  
Velocity Models ◽  
Event Location

Microseismic event locations obtained from seismic monitoring data sets are often a primary means of determining the success of fluid-injection programs, such as hydraulic fracturing for oil and gas extraction, geothermal projects, and wastewater injection. Event locations help the decision makers to evaluate whether operations conform to expectations or parameters need to be changed and may be used to help assess and reduce the risk of induced seismicity. However, obtaining accurate event location estimates requires an accurate velocity model, which is not available at most injection sites. Common velocity updating techniques require picking arrivals on individual seismograms. This can be problematic in microseismic monitoring, particularly for surface acquisition, due to the low signal-to-noise ratio of the arrivals. We have developed a full-wavefield adjoint-state method for locating seismic events while inverting for P- and S-wave velocity models that optimally focus multiple complementary images of recorded seismic events. This method requires neither picking nor initial estimates of event location or origin time. Because the inversion relies on (image domain) residuals that satisfy the differential semblance criterion, there is no requirement that the starting model be close to the true velocity. We determine synthetic results derived from a model with conditions similar to a field-acquisition scenario in terms of the number and spatial sampling of receivers and recorded coherent and random noise levels. The results indicate the effectiveness of the methodology by demonstrating a significantly enhanced focusing of event images and a reduction of 95% in event location error from a reasonable initial model.

Microseismic image-domain velocity inversion: Marcellus Shale case study

Geophysics ◽  
2017 ◽  
Vol 82 (6) ◽  
pp. KS99-KS112 ◽  
Author(s):  
Ben Witten ◽  
Jeffrey Shragge
Keyword(s):  
Marcellus Shale ◽  
Velocity Variation ◽  
Signal To Noise ◽  
Injection Wells ◽  
Hydraulic Stimulation ◽  
Image Domain ◽  
Velocity Inversion ◽  
Velocity Models ◽  
Adjoint State ◽  
Event Location

Seismic monitoring at injection wells relies on generating accurate location estimates of detected (micro-) seismicity. Event location estimates assist in optimizing well and stage spacings, assessing potential hazards, and establishing causation of larger events. The largest impediment to generating accurate location estimates is an accurate velocity model. For surface-based monitoring, the model should capture 3D velocity variation, yet rarely is the laterally heterogeneous nature of the velocity field captured. Another complication for surface monitoring is that the data often suffer from low signal-to-noise levels, making velocity updating with established techniques difficult due to uncertainties in the arrival picks. We use surface-monitored field data to demonstrate that a new method requiring no arrival picking can improve microseismic locations by jointly locating events and updating 3D P- and S-wave velocity models through image-domain adjoint-state tomography. This approach creates a complementary set of images for each chosen event through wave-equation propagation and correlating combinations of P- and S-wavefield energy. The method updates the velocity models to optimize the focal consistency of the images through adjoint-state inversion. We have determined the functionality of the method using a surface array of 192 3C geophones over a hydraulic stimulation in the Marcellus Shale. Applying the proposed joint location and velocity-inversion approach significantly improves the estimated locations. To assess the event location accuracy, we have developed a new measure of inconsistency derived from the complementary images. By this measure, the location inconsistency decreases by 75%. The method has implications for improving the reliability of microseismic interpretation with low signal-to-noise data, which may increase hydrocarbon extraction efficiency and improve risk assessment from injection-related seismicity.

scholarly journals Improving Semi-Supervised Learning for Audio Classification with FixMatch

Electronics ◽  
2021 ◽  
Vol 10 (15) ◽  
pp. 1807
Author(s):  
Sascha Grollmisch ◽  
Estefanía Cano
Keyword(s):  
Neural Networks ◽  
Supervised Learning ◽  
Transfer Learning ◽  
Data Transfer ◽  
State Of The Art ◽  
Training Data ◽  
Audio Classification ◽  
Image Domain ◽  
Full Dataset ◽  
Audio Data

Including unlabeled data in the training process of neural networks using Semi-Supervised Learning (SSL) has shown impressive results in the image domain, where state-of-the-art results were obtained with only a fraction of the labeled data. The commonality between recent SSL methods is that they strongly rely on the augmentation of unannotated data. This is vastly unexplored for audio data. In this work, SSL using the state-of-the-art FixMatch approach is evaluated on three audio classification tasks, including music, industrial sounds, and acoustic scenes. The performance of FixMatch is compared to Convolutional Neural Networks (CNN) trained from scratch, Transfer Learning, and SSL using the Mean Teacher approach. Additionally, a simple yet effective approach for selecting suitable augmentation methods for FixMatch is introduced. FixMatch with the proposed modifications always outperformed Mean Teacher and the CNNs trained from scratch. For the industrial sounds and music datasets, the CNN baseline performance using the full dataset was reached with less than 5% of the initial training data, demonstrating the potential of recent SSL methods for audio data. Transfer Learning outperformed FixMatch only for the most challenging dataset from acoustic scene classification, showing that there is still room for improvement.

The relationship between Vs, Vp, density and depth based on PS-logging data at K-NET and KiK-net sites

2021 ◽  
Author(s):  
Fumiaki Nagashima ◽  
Hiroshi Kawase
Keyword(s):  
Standard Deviation ◽  
Seismic Velocity ◽  
Saturated Soil ◽  
P Wave ◽  
Soil Types ◽  
Depth Range ◽  
Velocity Inversion ◽  
Velocity Models ◽  
Regression Curves ◽  
Ps Logging

Summary P-wave velocity (Vp) is an important parameter for constructing seismic velocity models of the subsurface structures by using microtremors and earthquake ground motions or any other geophysical exploration data. In order to reflect the ground survey information in Japan to the Vp structure, we investigated the relationships among Vs, Vp, and depth by using PS-logging data at all K-NET and KiK-net sites. Vp values are concentrated at around 500 m/s and 1,500 m/s when Vs is lower than 1,000 m/s, where these concentrated areas show two distinctive characteristics of unsaturated and saturated soil, respectively. Many Vp values in the layer shallower than 4 m are around 500 m/s, which suggests the dominance of unsaturated soil, while many Vp values in the layer deeper than 4 m are larger than 1,500 m/s, which suggests the dominance of saturated soil there. We also investigated those relationships for different soil types at K-NET sites. Although each soil type has its own depth range, all soil types show similar relationships among Vs, Vp, and depth. Then, considering the depth profile of Vp, we divided the dataset into two by the depth, which is shallower or deeper than 4 m, and calculated the geometrical mean of Vp and the geometrical standard deviation in every Vs bins of 200 m/s. Finally, we obtained the regression curves for the average and standard deviation of Vp estimated from Vs to get the Vp conversion functions from Vs, which can be applied to a wide Vs range. We also obtained the regression curves for two datasets with Vp lower and higher than 1,200 m/s. These regression curves can be applied when the groundwater level is known. In addition, we obtained the regression curves for density from Vs or Vp. An example of the application for those relationships in the velocity inversion is shown.

Sign in / Sign up

Export Citation Format

Share Document