Cross-equalization of time-lapse seismic data using recurrent neural networks

Alali, Abdullah A.; Kazei, Vladimir; Sun, Bingbing; Smith, Robert; Nivlet, Philippe; Bakulin, Andrey; Alkalifah, Tariq

doi:10.1190/segam2020-3424773.1

Cited by 5 publications

(2 citation statements)

References 21 publications

(17 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A successful inversion of land data requires building of an accurate model of the near-surface (Baeten et al, 2013) because inaccuracies accumulated in the shallow subsurface dramatically magnify at depth. The data-driven methods found broad range of applications in geophysics (Alali et al, 2020;Sun and Alkhalifah, 2020;Song et al, 2021). Here, we focus on initial velocity model building, which might be approached in data and model domain.…”

Section: Introductionmentioning

confidence: 99%

Dual-band generative learning for low-frequency extrapolation in seismic land data

Ovcharenko

Kazei

Peter

et al. 2021

First International Meeting for Applied Geoscience &Amp; Energy Expanded Abstracts

Self Cite

View full text Add to dashboard Cite

The presence of low-frequency energy in seismic data is one of the desirable prerequisite for a successful full-waveform inversion. However, it requires costly machinery and operations to acquire such data. Extrapolation for the missing lowfrequency content in field data might be addressed in a datadriven framework. In particular, the deep learning models are commonly trained for this purpose on synthetic data and then used for inference on field data. However, it is always a challenge to switch the application domains. We propose the concept of generative dual-band learning to facilitate the knowledge transfer between synthetic and field seismic data applications of low-frequency data extrapolation. We first explain the two-step procedure for training of a generative adversarial network (GAN) that extrapolates low frequencies. Then we describe the workflow for synthetic dataset generation. Finally, we explore the feasibility of the dual-band learning concept on real near-surface land data acquired in Saudi Arabia.

show abstract

Section: Introductionmentioning

confidence: 99%

Dual-band generative learning for low-frequency extrapolation in seismic land data

Ovcharenko

Kazei

Peter

et al. 2021

First International Meeting for Applied Geoscience &Amp; Energy Expanded Abstracts

Self Cite

View full text Add to dashboard Cite

show abstract

“…In data processing: Ovcharenko et al (2019) extrapolated low frequencies from high ones using deep learning; Slang et al (2019) used deep learning for denoising and deblending. For time-lapse processing: Alali et al (2020a) correct for the time shift in the data using a fully-connected layer in the latent space of an autoencoder; Duan et al (2020) showed that a trained network can outperform a conventional cross-correlation method in estimating the time-shift; Alali et al (2020b) suggested to use recurrent neural networks to better account for time dependency in the data.…”

Section: Introductionmentioning

confidence: 99%

Time-lapse seismic cross-equalization using temporal convolutional networks

Alali

Smith

Nivlet

et al. 2021

First International Meeting for Applied Geoscience &Amp; Energy Expanded Abstracts

Self Cite

View full text Add to dashboard Cite

Monitoring carbon dioxide sequestered in geological reservoirs using 4D seismic faces many challenges. The repeatability between different surveys needs to be optimal in which changes are only present in the target zone, which will require having the exact acquisition parameters and no near-surface variations, like those caused by seasonal changes. It is not possible to meet these requirements without data processing and matching techniques. We propose to align the data trace by trace using deep learning. We utilize the sequential nature of the data to train a temporal convolutional network (TCN). The network will learn to map the monitor to the base data in the overburden region. Therefore, the difference between the predicted base and the actual base will eliminate the overburden effect and improve the 4D signal. We validate the method on a synthetic time-lapse zero-offset data and show improvements in the repeatability. We also perform an initial test on real 4D land data and show the potential of the method for real applications.

show abstract

Repeatability enhancement of time-lapse seismic data via a convolutional autoencoder

Jun

Cho

2021

Geophysical Journal International

View full text Add to dashboard Cite

Summary In an ideal case, the time-lapse differences in 4D seismic data should only reflect the changes of the subsurface geology. Practically, however, undesirable discrepancies are generated because of various reasons. Therefore, proper time-lapse processing techniques are required to improve the repeatability of time-lapse seismic data and to capture accurate seismic information to analyze target changes. In this study, we propose a machine learning-based time-lapse seismic data processing method improving repeatability. A training data construction method, training strategy, and machine learning network architecture based on a convolutional autoencoder are proposed. Uniform manifold approximation and projection are applied to the training and target data to analyze the features corresponding to each data point. When the feature distribution of the training data is different from the target data, we implement data augmentation to enhance the diversity of the training data. The method is verified through numerical experiments using both synthetic and field time-lapse seismic data, and the results are analyzed with several methods, including a comparison of repeatability metrics. From the results of the numerical experiments, we can conclude that the proposed convolutional autoencoder can enhance the repeatability of the time-lapse seismic data and increase the accuracy of observed variations in seismic signals generated from target changes.

show abstract

Cross-equalization of time-lapse seismic data using recurrent neural networks

Cited by 5 publications

References 21 publications

Dual-band generative learning for low-frequency extrapolation in seismic land data

Dual-band generative learning for low-frequency extrapolation in seismic land data

Time-lapse seismic cross-equalization using temporal convolutional networks

Repeatability enhancement of time-lapse seismic data via a convolutional autoencoder

Contact Info

Product

Resources

About