Seismic monitoring of CO 2 geosequestration: CO2CRC Otway case study using full 4D FDTD approach

Glubokovskikh, S..; Pevzner, Roman; Dance, Tess; Caspari, Eva; Popik, Dmitry; Shulakova, Valeriya; Gurevich, Boris

doi:10.1016/j.ijggc.2016.02.022

Cited by 40 publications

(16 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To test our workflow, we compute the vertical component of particle velocity for corresponding offset VSP gathers using 2D finite-difference time-domain approach and invert them. The model we use was designed to replicate the geological conditions of the Otway site (Glubokovskikh et al, 2016), where the field dataset was acquired. The acquisition geometry is similar to the actual field geometry.…”

Section: Synthetic Data Examplementioning

confidence: 99%

Elastic full-waveform inversion of vertical seismic profile data acquired with distributed acoustic sensors

et al. 2018

Self Cite

View full text Add to dashboard Cite

Distributed acoustic sensing (DAS) is a rapidly developing technology particularly useful for the acquisition of vertical seismic profile (VSP) surveys. DAS data are increasingly used for seismic imaging, but not for estimating rock properties. We propose a workflow for estimating elastic properties of the subsurface using full waveform inversion (FWI) of DAS VSP data. Whereas conventional borehole geophones usually measure three components of particle velocity, DAS measures a single quantity, which is an approximation of the strain or strain rate along the fiber. Standard FWI algorithms are developed for particle velocity data, and hence their application to DAS data requires conversion of these data to particle velocity along the fiber. This conversion can be accomplished by a specially designed filter. Field measurements show that the conversion result is close to vertical particle velocity as measured by geophones. Elastic time-domain FWI of a synthetic multi-offset VSP dataset for a vertical well shows that the inversion of the vertical component alone is sufficient to recover elastic properties of the subsurface. Application of the proposed workflow to a multioffset DAS dataset acquired at the CO2CRC Otway Project site in Victoria, Australia reveals salient subhorizontal layering consistent with known geology of the site. The inverted VP model at the well location matches the upscaled VP log with a correlation coefficient of 0.85.

show abstract

Section: Synthetic Data Examplementioning

confidence: 99%

Elastic full-waveform inversion of vertical seismic profile data acquired with distributed acoustic sensors

et al. 2018

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, application of AGC in a workflow is controversial. Our modelling study shows that AGC for high-fold dataset such as the one under consideration can preserve spatial distribution of reflectivity change in stacked domain (Glubokovskikh, et al 2016) while AGC will distort AVO response in the presence of strong remnants of the ground roll present on small offsets. As the goal is to obtain stacked migrated volumes we keep AGC in the workflow.…”

Section: Initial Workflowmentioning

confidence: 83%

Time-lapse surface seismic processing for Stage 2C of CO2CRC Otway Project

Popik¹,

Shulakova²,

Tertyshnikov³

et al. 2018

ASEG Extended Abstracts

Self Cite

View full text Add to dashboard Cite

Stage 2C of the Otway project aims to detect a small injection of CO2-rich mixture into a saline aquifer, verify stabilisation of the injection and evaluate the detectability threshold of CO2 achievable by surface seismic. Over the past three years, we have produced five vintages of high-quality 4D seismic data showing the evolution of the 15,000 tonnes of injected scCO2/CH4 gas mixture into the saline aquifer at 1500 m depth. The time-lapse seismic processing workflow was built based on the findings from a synthetic feasibility study and processing of the baseline dataset. Through this workflow we processed the time-lapse data which allowed monitoring of small incremental injections (about 5 000 tonnes each). Here we elaborate on effect of various processing routines on the quality of 4D image, namely, amplitude restoration, prestack cross-equalization and imaging. We aim at preservation and restoration of time-lapse signal and suppression of time-lapse noise. We evaluate the results of the processing efforts in post-stack domain through estimates of 4D signal-to-noise ratio in vicinity of the injection interval. The usefulness of individual processing routines for improvement of timelapse signal cannot be established independently from other routines in the same workflow. Surprisingly, images with pre-stack AGC applied have consistently higher time-lapse signal-to-noise ratio compared to images produced without it. Improvement of the resolution and appearance of 3D images does not guarantee increase of time-lapse signal-to-noise ratio.

show abstract

“…We tested different hyperparameters and chose the ones that Figure 2: (a) The Otway velocity model (Glubokovskikh et al, 2016) used to generate data for the baseline, (b) the reservoir signal and near-surface noise added to the baseline for the monitor data, and (c) zoom into the random Gaussian 4D noise with 50 m/s mean and 100 m/s standard deviation changes in the first 20 m depth.…”

Section: Trainingmentioning

confidence: 99%

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

Alali

Kazei

Sun

et al. 2020

SEG Technical Program Expanded Abstracts 2020

View full text Add to dashboard Cite

Time-lapse seismic uses repetitive seismic surveys to monitor the fluid in the subsurface. Ideally, the time-lapse data should be identical except for at the target region (i.e., the reservoir), where the fluid changes occur. Unfortunately, it is almost impossible to have identical data for various reasons, such as the static changes in the near-surface or the varying positioning of sources and receivers between surveys. To increase the accuracy of the 4D signal and reduce the noise, we propose to process the time-lapse data using a machine-learning methodology. Specifically, we train a recurrent neural network (RNN) model to map the data from monitor to baseline. The learned RNN model would reveal 4D overburden changes. Therefore, the difference between the predicted baseline and the actual baseline data stets will represent the target signal. We validate the method on synthetic data and show the improvements of the 4D signal by imaging the reservoir and computing the normalized root mean square.

show abstract

Seismic monitoring of CO 2 geosequestration: CO2CRC Otway case study using full 4D FDTD approach

Cited by 40 publications

References 27 publications

Elastic full-waveform inversion of vertical seismic profile data acquired with distributed acoustic sensors

Elastic full-waveform inversion of vertical seismic profile data acquired with distributed acoustic sensors

Time-lapse surface seismic processing for Stage 2C of CO2CRC Otway Project

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

Contact Info

Product

Resources

About