An inversion algorithm is commonly used to estimate the elastic properties, such as P-wave velocity ([Formula: see text]), S-wave and density text]) of earth’s subsurface. Generally, seismic problem solved using one traditional optimization algorithms. These algorithms start with a given model update at each iteration, following physics-based rule. The applied common depth point (CDP) independently parameters. Here, we have developed technique convolutional neural network (CNN) solve same...

Prestack or angle stack gathers are inverted to estimate pseudologs at every surface location for building reservoir models. Recently, several methods have been proposed increase the resolution of All these methods, however, require that total number model parameters be fixed a priori. We investigated an alternate approach in which we allow data themselves choose parameterization. In other words, addition layer properties, layers is also treated as variable our formulation. Such...

Seismic data are used to generate high resolution subsurface images, which require detailed velocity models. Full Waveform Inversion (FWI), has recently gathered immense popularity in inverting for the elastic wave velocities from seismic data. FWI is a non-linear and non-unique inverse problem that uses complete time amplitude information estimating properties. Typically performed using local optimization methods model described by large number of grids. The parameters determined priori. In...

We describe a new method based on the Machine Learning (ML) technique for normal moveout correction (NMO) and estimation of stacking velocity. A Recurrent Neural Network (RNN) is used to calculate velocity directly from seismic data. Finally, this NMO Adam optimization algorithm train network neurons estimate batch gathers. This then compared with correct update weight. The training minimizes cost function defined as mean squared error between estimated velocities. trained rest Here we...

Seismic inverse problem of estimating P- and S-wave reflectivity from seismic traces has recently been revisited using a basis pursuit denoising inversion (BPI) approach. The BPI uses wedge dictionary to define model constraints, which successful in resolving thin beds. Here we address two fundamental problems associated with BPI, namely, the uniqueness estimate choice regularization weight λ be used norm. We investigated these very fast simulated re-annealing (VFSR) gradient projection...

ABSTRACT We present a methodology for seismic inversion that generates high‐resolution models of facies and elastic properties from pre‐stack data. Our algorithm uses transdimensional approach where, in addition to the layer properties, number layers is treated as unknown. In other words, data itself determine correct model parameterization, is, layers. The reversible jump Markov Chain Monte Carlo method an effective tool solve such problems it reservoir along with uncertainty estimates....

Summary We present a new scheme to solve seismic Amplitude Vs. Angle (AVA) inversion problem in probabilistic framework. have used recently developed Stein Variational Gradient Descent (SVGD) method our proposed Bayesian Integrated Reservoir Characterization (BIRCh) workflow. BIRCh is based on framework inference algorithm that seeks set of points (or particles) approximate the target distribution using iterative gradient-based updates. It surpasses several disadvantages traditional...

A pre-stack inversion is performed to estimate elastic properties like VP, VS, r of the earth's subsurface. Pre-stack inversions are generally solved employing a global or local optimization technique and on each CDPs (common-depth-point) separately properties. These algorithms starting with an initial solution, updates model following rule till it meets set error tolerance level when numbers models have been sampled. We develop novel using one increasingly popular neural networks. use...

Machine learning (ML) has recently gained immense popularity because of its successful application in complex problems. It develops an abstract relation between the input and output. We have evaluated ML to most basic seismic processing normal moveout (NMO) correction. The arrival times reflection events a common midpoint (CMP) gather follow hyperbolic trajectory; thus, they require correction term flatten CMP before stacking. This depends on rms velocity, also referred as NMO velocity. In...

Receiver functions (RFs) are derived by deconvolution of the horizontal (radial or transverse) component ground motion from vertical component, which segregates PS phases.Many methods have been proposed to employ in frequency as well time domain.These vary their approaches impose regularization that addresses stability problem.Here, we present application a new time-domain technique called basis pursuit (BPD) has recently applied seismic exploration data.Unlike conventional methods, BPD uses...

Commercial and low gas saturation (also called paleo-residual gas, PRG) both appear as bright amplitude on P-wave seismic data. This is a common exploration risk in Columbus basin. Since density has strong inverse correlation with saturation, it should effectively differentiate zone of full from PRG. However, extracting amplitude-variations-with-offset (AVO) inversion can be challenging, due to weak sensitivity reflection density, lack good quality AVO With the recent improvements...

PreviousNext No AccessThird International Meeting for Applied Geoscience & Energy Expanded AbstractsBayesian discriminative classification with kernel density estimation rock and fluid property characterization of seismic elastic inversion resultsAuthors: Kevin WolfJingfeng ZhangMatt WalkerPedro ParamoJeff WinterbourneReetam BiswasAtish RoyCarole DecalfKevin WolfBP, Jingfeng ZhangBP, Matt WalkerBP, Pedro ParamoBP, Jeff WinterbourneBP, Reetam BiswasBP, Atish RoyBP, Carole...

PreviousNext No AccessThird International Meeting for Applied Geoscience & Energy Expanded AbstractsSpecialized processing of a massive distributed acoustic sensing (DAS) 3D VSP in producing wellAuthors: Ali SayedFrancesca TwynamRan BachrachPhilippe CaprioliMarcus HollowayQingsong LiReetam BiswasDaniel TeboSam BuistAli SayedSLB, Francesca TwynamSLB, Ran BachrachSLB, Philippe CaprioliSLB, Marcus HollowaySLB, Qingsong LiBP, Reetam BiswasBP, Daniel TeboBP, and Sam...

Summary Pre-stack or angle stack gathers are inverted to estimate pseudo-logs at every surface location for building reservoir models. Recently several methods have been proposed increase the resolution of All these methods, however, require that total number model parameters be fixed a priori. Here, we investigate an alternate approach in which allow data choose parameterization. In other words, addition layer properties, layers is also treated as variable our formulation. Such...

Full-Waveform Inversion (FWI) has now become a widely accepted tool to obtain high-resolution velocity models from seismic data. Typically, the model in its discrete form is represented on rectangular grid, and we solve for elastic properties at these grid points. FWI mostly solved employing local optimization method, where one obtains update by minimizing misfit between observed calculated seismograms. Note also that highly non-linear problem which known be prone non-uniqueness. The...

Full Waveform Inversion (FWI) has become a powerful tool to generate high-resolution subsurface velocity models. FWI attempts solve non-linear and non-unique inverse problem, is traditionally based on local optimization technique. As result, it can easily get stuck in minimum. To mitigate this deleterious effect, requires good starting model, which should be close enough the optimal model properly converge global Here, we investigate two-step approach for solving problem. In first step, FWI,...

Uncertainty estimation in subsurface characterization workflows is an important input to decision-making earth science-related problems. We present three methods characterize seismic-related uncertainty, each of which includes a real data case study. Two these are designed depth uncertainty positioning migrated reflectors. Such estimates may be used derisking well prognoses, analysis misties, and for estimating ranges on resource volume. The first method derives rapid robust around existing...

PreviousNext No AccessThird International Meeting for Applied Geoscience & Energy Expanded AbstractsHigh-resolution imaging using least-squares migration in the image domain (LSMi) a marine distributed acoustic sensing (DAS) 3D-VSPAuthors: Ali SayedFrancesca TwynamRan BachrachMaud CavalcaLeo LeonMaria ShadrinaQingsong LiReetam BiswasDaniel TeboSam BuistAli SayedSLB, Francesca TwynamSLB, Ran BachrachSLB, Maud CavalcaSLB, Leo LeonSLB, Maria ShadrinaSLB, Qingsong LiBP, Reetam BiswasBP, Daniel...

High quality seismic impedance volumes directly benefit a wide range of workflows in reservoir characterization. The conventional method calculating impedances, however, has several limitations which can result suboptimal products. We propose an alternative way obtaining volumes. First, we utilize pre-stack Bayesian inversion, honoring Zoeppritz equation, to obtain Vp/Vs/Rho volumes, and then analytically calculate This inversion was made possible by combining the Stein Variational Gradient...

Summary Sleipner gas field in the North Sea is world's first industrial scale CO2 injection project designed specifically to reduce emission of greenhouse gas. Here separated from natural produced at injected into Utsira sand, which a major saline aquifer basin. In time-lapse three-dimensional seismic data (4D), plume imaged as number bright sub-horizontal reflections within reservoir. Correlation log with indicates that accumulates series interbedded sandstones and mudstones beneath thick...

Summary Seismic inverse problem of estimating P- and S-wave reflectivity from seismic traces has recently been revisited using a basis pursuit inversion (BPI) approach. The BPI uses wedge dictionary to define model constraints, which successful in resolving thin beds. Here we have addressed mainly estimation the regularization weight BPI. It plays an important role as it acts weighting factor decides how much importance is given sparsity solution over data fitting. This by developing...

Commercial and low-saturation gas (also called paleoresidual [PRG]) show similar strong amplitude signatures on P-wave seismic data. This poses an exploration risk in reservoir regions. However, density correlates inversely with saturation can differentiate a zone of full from PRG. improve the chances success terms predrill prediction saturation. Amplitude-variation-with-offset (AVO) inversion using prestack data is most commonly used technique that estimate elastic parameters such as...

Raw seismic data are inverted to estimate the earth properties at every surface location for building reservoir models. Three elastic parameters needed describing model: density, acoustic impedance and shear impedance. Various inversion methods have been developed these parameters. Out of those, in recent times, Full Waveform Inversion (FWI) is a popular method, which employs both time amplitude information data, properties. However, due very high computational complexities required process...