Sixth EAGE Borehole Geophysics Workshop

Seismic imaging of thin stratigraphic features requires high frequency seismic data, accurate estimation of subsurface velocities and high-resolution imaging techniques for proper focusing of seismic energy. Borehole seismic surveys meet all of these requirements and are known to provide higher resolution subsurface images compared to surface seismic mainly because of shorter ray paths and consequently less anelastic attenuation of energy received in a borehole. However, when imaging focus is constrained within a certain depth range of the target reservoir, standard walkaway borehole seismic surveys suffer from large incidence angles at longer source offsets. The reflected seismic waveform at higher incidence angles experiences amplitude and phase variations that are undesirable for imaging. Herein, we present a novel borehole seismic survey geometry – "Moving-source Moving-receiver Vertical Seismic Profile" (MsMr-VSP), where the receiver array is moved up the hole starting at the target depth, while the source is moved away from the well-head. An application of the MsMr-VSP survey is presented from Abu Dhabi, where it was used for high-resolution and improved quality imaging of thin clinoform features in a carbonate reservoir with reasonably low reflection angles and a high reflection point density maintained at the target level along the extent of the clinoform.

Velocity model building is an essential step in seismic exploration, which runs through the whole process of seismic data acquisition, processing, and interpretation. The velocity information is conventionally obtained by iterative optimization methods such as full-waveform inversion or tomography. These traditional methods are computationally expensive, and they require an initial velocity model and human interactions. To simplify the model building problem, we develop a supervised end-to-end conventional neural network to reconstruct the P-wave velocity models directly from raw seismic data. The network takes in one-shot seismic traces simulated with acoustic wave equations in VSP geometry. To train the network, we create 750 2D synthetic seismic images and corresponding labeled images, which are shown to be sufficient for the network to learn the nonlinear relationship between one-shot seismic data and the corresponding velocity model. The numerical examples show that the trained network is capable of predicting accurate layered velocity models from only one-shot seismic data.

Two walkaway DAS-VSP lines, zero offset VSP, offset VSP and two 2D 3C surface seismic lines data were acquired simultaneously using the borehole TF-02 in Southwest Oilfield of CNPC at Southwest China in order to perform the enhanced borehole-driven 3D surface seismic data processing and map the volcanic gas reservoir distribution. An armored optical cable with high temperature single mode fiber, dynamite and vibrator mixed sources were used to acquire the Walkaway DAS-VSP data.

The walkaway DAS-VSP data were used to extract formation velocity, deconvolution operator, absorption, attenuation (Q value), anisotropy parameters (η, δ, ε) as well as enhanced the surface seismic data processing including velocity model calibration and modification, static correction, deconvolution, demultiple processing, high frequency restoration, anisotropic migration, and Q-compensation or Q-migration. In this project, anisotropic Q-PSDM was conducted with the anisotropy parameters (η, δ, ε) data volume and enhanced Q-field data volume obtained from the joint inversion of both the near surface 3D Q-field data volume from uphole data and the mid-deep layer Q-field data volume from all available VSP data in the 3D surface seismic survey area. The anisotropic Q-PSDM results show much sharper and focused faults and clearer subsurface structure.

Shallow gas is a potential target for tight Miocene in some areas of onshore Abu Dhabi, whereas the Gachsaran formation shows a complex thinly bedded intercalation of different lithologies: evaporites, shales and carbonates. Surface seismic methods are key to map and characterize the target layers and collecting enough data for borehole calibration is considered essential.

A large multi-walkaway and walkaround vertical seismic profile surveys were recorded in one onshore exploration well from its total depth up to surface. The survey objectives included seismic amplitude, phase and velocity control, polar and azimuthal anisotropy calibration, higher-resolution walkaway imaging, amplitude versus angle analyses and studying interbed multiples. Together with a complete suite of elastic logs, these measurements supported reservoir geophysics studies and surface seismic imaging.

The walkaround results showed that azimuthal anisotropy was negligible, while the walkaway results indicated significant polar anisotropy at seismic scale. The latter is driven primarily by layering with strong elastic property contrasts. Large offset multiples degrade standard walkaway imaging and best results were obtained with a tau-p domain deconvolution workflow. We review some of the challenges and key results from this project.

In recent years, in order to improve the effect of oilfield development, an important model of large cluster well groups has been established based on high-precision 3D seismic data. However, in the 2D seismic area, the seismic survey lines are sparse and the seismic imaging resolution is very low, which cannot meet the needs of multi-azimuth targeted prediction of horizontal wells. We recognized that the Walkaway-VSP observation method has the characteristics of high imaging resolution and accurate formation information. Walkaway-VSP can play a role in the fine description of well-side reservoirs and horizontal well targeting prediction.

In this study, a multi-directional Walkaway-VSP survey line was deployed, and high-quality in-hole seismic acquisition data was obtained through the observation of a large downhole array three-component digital geophone, in the Ordos Basin. Then, high-quality imaging data are obtained through targeted processing techniques such as weak signal protection deconvolution and time-varying vector wavefield separation. Finally, Walkaway-VSP imaging was used to characterize the reservoir distribution and the development of fault structures in the area around the well. Through comprehensive analysis of seismic interpretation results, geological data and oil test data, recommendations for the deployment of horizontal wells in the study area are drawn.

A case study of Look-ahead VSP from Offshore UAE is presented where it helped to guide drilling exploration well from an island within salt dome for multiple subsurface targets. Salt location and geometry was very uncertain, only few coarse 2D offshore surface seismic (with no line crossing the planned well location) and gravity measurements were available for structural interpretation. Different salt geometry scenarios were considered, source positions were modelled ahead of drilling, taking into account the restrictions of having no offshore sources, to ensure maximum chance of obtaining the location of salt using vibroseis on the island. Multiple Look-ahead VSPs were acquired at four casing points and were rush processed and interpreted to guide drilling subsequent sections. Severe first arrival multi-pathing was observed in the data likely due to geological complexity between salt face and the receivers. This made first arrival picking challenging and caused uncertainty in velocities and model calibration. Interpretative approach to processing was carried out followed by imaging. Salt face location was accurately determined prior to exit. Once out of salt, the well was steered towards salt without re-entering. After reaching TD, data from each section were merged and processed for final sub-salt image.

Time-lapse seismic was performed as part of the Otway Stage 3 Project (Victoria, Australia) to monitor the spread of an injected plume of CO2. A permanent seismic monitoring system based on surface orbital vibrators and DAS VSP technology was used to acquire a new vintage survey every two days. SOV/DAS data was collected from the end of May 2020 to April 2021 (~150 vintages), with the CO2 injection commencing in December 2020. We acquired more than 60 vintages to utilize as a baseline, which provided an opportunity to investigate how a selection of a baseline survey impacts the interpretation of time-lapse changes.

Two baseline vintages from June 2020 and November 2020 were compared to a monitor survey acquired in April 2021. The NRMS was calculated to be 45% and 60% for the June and November baselines, respectively. We suggest that this non-repeatability is created by seasonal variations affecting near-surface conditions. The pairwise analysis of all possible vintage combinations highlights seasonal changes in the NRMS value. Analysis of the time-lapse VSP wavefield suggests that the main source of non-repeatability is a seasonal variation of near-surface conditions. This variation manifests itself in changes in source performance and the pattern of surface-related multiples.

No summary.

We discuss the processing and imaging challenges relating to a Vertical Seismic Profiling (VSP) pilot survey acquired with Distributed Acoustic Sensing (DAS) over the Culzean field, Central North Sea. Acquired during 2019, this survey aimed to provide insight into the potential of DAS technology for 3D imaging and 4D monitoring as well as assessing the limitations of this acquisition. We show how processing this data with a synthetic based cooperative de-noise workflow was effective in attenuating the background noise common in this type of acquisition, significantly improving the signal-to-noise ratio of the data. The data was further improved using a model based de-multiple technique prior to imaging with Reverse Time Migration, achieving a rich, high quality 3D image with improved imaging over the key Lower Cretaceous section of the area. In addition, we highlight the potential of multiple imaging with Reverse Time Migration to achieve an extended image away from the well bore including the near surface.

No summary.

In 2016 ADNOC started injecting CO2 into the two onshore fields to boost oil production and replace the use of hydrogen gas. Carbon capture could assist in enhanced oil recovery and in protecting the environment. At the same time, the time-lapse walkaway VSP technique was considered to monitor fluid displacement in the reservoir where CO2 and water were going to be injected. The data acquisition consisted of a walkaway line recorded with a vibroseis source and with geophone tools anchored behind cemented casing near the reservoir level in a CO2 monitor well. The walkaway VSP line should be repeated over time, when changing the injection cycle from water to CO2 and vice versa.

This study showed how careful survey design, data acquisition and data processing have allowed meeting the required level of repeatability in a time-lapse walkaway survey monitoring project, where CO2 and water have been injected for EOR purposes. The repeatability metrics improved significantly during the successive data processing steps and they provided more reassurance about the final time-lapse differences observed that should allow detecting the injected fluid effects in the final processing stage. Also, the time-lapse corridor stack result at zero-offset VSP was a particularly interesting calibration point.

Accurate well to seismic ties are important during exploratory phases and become critical during reservoir seismic characterization. Vertical Seismic Profiles (VSP) are routinely recorded on wireline or LWD logging for accurate time-depth and velocity control.

In qualitative well tying, VSP corridor stacks are compared against seismic data and synthetic seismograms. Quantitative well to seismic tie workflows go a few steps beyond, providing the synthetic trace best match location, wavelet extraction and error estimates, and enabling a quantitative assessment of the seismic tie around the well. This technique remains one of the most used by the oil & gas industry, being available in mainstream seismic interpretation software.

Ireson et al. (1996) extended White (1980) technique to include additional well tie metrics and made a strong case for using also the VSP corridor stack. A deterministic seismic wavelet can be estimated without well logs and a spectral operator derived to match-filter the seismic data around the well. Unfortunately, the extended technique has not been yet incorporated in most seismic interpretation platforms and few case studies have been published. We revive this important application of VSP corridor stacks in quantitative borehole matching by illustrating their value in a deep exploration well setting.