Mapping of seismic and lithologic facies from 3D reflection data plays a key role in depositional environment analysis reservoir characterization during hydrocarbon exploration development. Although variety machine-learning methods have been developed to speed up interpretation improve prediction accuracy, there still exist significant challenges multiclass classification practice. Some these limitations include complex representation, limited training with labels, imbalanced class...

Seismic facies analysis interprets depositional environment and types from the reflection seismic data, an important step in exploration reservoir characterization. While machine learning methods, especially deep models such as convolutional neural networks (CNNs) have been applied to assist interpretation salt identification, significant challenges still remain for 3D multi-class classification: complex data representation, limited labeled training, imbalanced class distribution lack of...

Onshore seismic monitoring for CO 2 injection in carbonate reservoirs the Middle East is a major challenge many reasons. The 4D signal generally much smaller than that of clastic or chalk due to high bulk moduli rocks and relatively small fluid effect. In addition, data are characterized by low signal-to-noise ratio (S/N) poor penetration below high-contrast near-surface layers, poorly consolidated materials at surface, conversion source energy into surface waves trapped modes, scattering...

We describe here methods of estimating interval velocities based on two nonlinear optimization methods; very fast simulated annealing (VFSA) and a genetic algorithm (GA). The objective function is defined using prestack seismic data after depth migration. This inverse problem involves optimizing the lateral consistency reflectors between adjacent migrated shot records. In effect, normal moveout correction in velocity analysis replaced by When least‐squared difference each pair shots at...

We present results from a first of its kind permanent seismic monitoring system in desert environment. This new consists sensors buried at 70 m depth and surface vibroseis sources. describe processing challenges associated with this single-sensor single-source dataset initial solutions that allowed us to obtain robust 3D reservoir images. The has achieved remarkable repeatability mean NRMS 4-5 % across high fold area between closely spaced repeat surveys. Presentation Date: Wednesday,...

ABSTRACT The orientation of three‐component borehole geophones used for recording during a microseismic monitoring experiment is estimated. standard technology deploying multi‐component in deep wireline‐based, which the azimuthal rotation geophone string cannot be controlled. Each receiver can have different angle that compensated by particle motion analysis direct P‐wave arrivals, picked from walk‐around VSP carried out proximity well. Knowing receivers critical, as inaccuracies lead to...

A seismic field acquisition test was conducted in an onshore Saudi Arabia. The effects of near-surface complexity (in the form sand, karsts, topography), environmental noise as well large surface temperature variations are illustrated and quantified by 4D attribute analysis using permanent sources buried geophones. We show that burying receivers dramatically improves wavelet amplitude stability naturally reduces effect man-made noise. layers comprise sand from 3 m to more than 20 thick this...

To evaluate repeatability of emerging seismic technologies for future 4D reservoir monitoring studies on Middle East carbonate reservoirs, a field acquisition test was conducted over an onshore in Saudi Arabia. The effects near-surface complexity (in the form sand and karsts) as well large surface temperature variations are illustrated quantified by attribute analysis using permanent piezoelectric sources. Even though measured does not reach values observed non-desert environments, we show...

Research Article| September 01, 2013 Enhancement of Passive Microseismic Events Using Seismic Interferometry Abdullatif Al‐Shuhail; Al‐Shuhail aEarth Sciences Department, KFUPM, Dhahran, 31261, Saudi Arabiaashuhail@kfupm.edu.saskaka@kfupm.edu.sa Search for other works by this author on: GSW Google Scholar SanLinn I. Kaka; Kaka Michael Jervis bGeophysics Technology Team, EXPEC Advanced Center, Aramco, 31311, Arabiamichael.jervis@aramco.com Author and Article Information Publisher:...

We have proposed a deep neural network-based framework for seismic facies classification. implement two different networks based on the architectures of DeepLabv3+ and generative adversarial network segmentation compare mapping results from reflection data to lithologic facies. predictions sharper boundaries between predicted whereas output has better continuity incorporate uncertainty analysis into workflow using Bayesian framework. The approach consisting joint multiple along with in...

Despite the current easing in demand for increased oil production linked to global downturn crude prices, energy continuously increases and long-term will require maximizing productivity of reservoirs a search into exploitation new resources increasingly challenging environments. In this study, we present results from monitoring very first multistage stimulation experiment at shale gas reservoir Saudi Arabia, presenting an analysis microseismicity induced during treatment. Our aim was...

PreviousNext No AccessSEG Technical Program Expanded Abstracts 1990Nonlinear inversion of seismograms: State the artAuthors: Albert TarantolaEdward CraseMichael JervisZvi KonenJonas LindgrenKlaus MosegaardMark NobleAlbert TarantolaInstitut de Physique du Globe Paris (IPGP), France, Edward CraseMontana Tech University, Michael JervisIPGP, Zvi KonenELF Aquitaine, Jonas LindgrenIPGP, Klaus MosegaardUniversity Copenhagen, Denmark, and Mark NobleIPGP, Francehttps://doi.org/10.1190/1.1889948...

Summary Onshore time-lapse seismic is a major challenge, particularly in the harsh desert environments encountered Saudi Arabia. Injection of CO2 into stiff carbonate reservoir results weak 4D signal, which can easily be contaminated by noise generated changing near surface conditions. Using hybrid acquisition system, consisting permanent buried receivers and vibroseis sources, seasonal variations data repeatability are clearly identified. Data acquired under same conditions (e.g., both...

A new imaging application is presented using sonic waveform data for ranging to locate a nearby borehole. The challenge of locating well from borehole commonly addressed with electromagnetic (EM) or passive magnetic methods, which can suffer poor resolution and penetration require the presence conductive casing. In addition, deeper reading active EM methods do not work when two boreholes are orthogonal. contrast, acoustic requires only an impedance contrast between target object surrounding...

We address the problem of velocity estimation in heterogeneous media using a combination nonlinear inversion and migration analysis. In estimation, travel time information seismic reflection data are nonlinearly related to perturbations subsurface. By taking functional traces, migrated themselves, we define misfit criterion which greatly reduces oscillatory nature objective function. Migration is inherently smoothing process; it collapses diffractions, focuses reflected energy suppresses...

Wavelet domain hidden Markov Models (HMM) have been proven to be useful tools in signal processing for image denoising and interpolation. Here the complex wavelet transform is used with a HMM applied problem of noise reduction on post‐stack seismic data volumes. aim capture statistical structure smooth (data) singular (edgy) parts signal. Experiments real show improved separation after filtering compared regular methods principle component filtering.

We present a new approach to perform Bayesian linearized AVO inversion directly in the depth domain using non-stationary wavelets. has gained immense popularity past decade because of its superior computational speed and ability estimate uncertainties inverted parameters. is typically performed on time-domain seismic data; therefore, depth-imaged cannot be this method, would require depth-to-time conversion before inversion. Subsequently, time-to-depth volumes required for reservoir modeling...

We propose a deep neural network based framework for seismic facies classification. The proposed utilizes generative adversarial segmentation to learn mapping from reflection data lithological facies. incorporate uncertainty analysis into the workflow using Bayesian framework. approach accelerates interpretation process by reducing need human intervention, and also lessens individual biases that an interpreter may bring. demonstrate effectiveness of algorithm testing on field examples, show...

A comprehensive onshore seismic experiment was conducted in Saudi Arabia to evaluate various configurations for monitoring a desert environment. line of 80 receiver stations deployed comprised both geophones and hydrophones permanently cemented the boreholes at four different depth levels from 0 30 m. This study focuses on analysis data acquired using surface vibrator source. Due challenging near-surface conditions poor signal-to-noise ratio observed pre-stack data, dense source grid with...

Deep learning is increasingly being applied in many aspects of seismic processing and interpretation. Here, we look at a deep convolutional neural network approach to multiclass lithofacies characterization using well logs data. In particular, focus on performance hyperparameter tuning. Several tuning approaches are compared, including true directed random search methods such as very fast simulated annealing Bayesian optimization. The results show that improvements predictive capability...

A series of field tests were conducted to quantify repeatability land seismic data acquired with a surface vibrator in desert environment. Tests included six repeated 2D surveys as well daily and hourly sweep using buried geophones buried, cemented geophones. The results indicate that contrast marine data, even small source geometry changes (<2 m) may seriously degrade survey repeatability. In addition, for fixed geometry, there is significant non-repeatability caused by the interaction...

The analysis of microtremors has been claimed to be useful as a possible direct hydrocarbon indicator, but so far only surface measurements were presented support the claim. However, recent results obtained while processing microseismic data from an oil‐producing reservoir, which show no significant correlation in average polarization between and downhole data.

Abstract In 2015, Saudi Aramco started a CO2 Water-Alternating-Gas (WAG) EOR pilot project in an onshore carbonate reservoir. To monitor lateral expansion of the plume, area was instrumented with hybrid surface/downhole permanent seismic monitoring system. This system consists over 1000 buried sensors at depth around 70 m, below expected weathering layer to mitigate time-lapse noise. Despite receiver burial, data still suffers from numerous challenges including: significant amounts...

In this paper we propose a new method to locate bed boundaries by carrying out 1-D nonlinear inversion of electromagnetic (EM) logging data. We first solve for resistivity structure in which the earth is modeled using layers constant thickness. This thickness determined based upon tool resolution and desired from user. use general measures data misfit model construct piecewise‐constant models through iteratively reweighted least‐squares (IRLS) procedure, minimize generic global objective...

A large passive seismic array has been installed over a field in Saudi Arabia. This comprises both downhole and surface 3C sensors is designed to record microseismic events from the reservoir zone about 2km deep an area of 9 sq km. The main objective demonstrate that which occur are detectable they can be used for mapping fluid flow fracture pathways.