**Peer Reviewed**

**Source**

**Journal**

**Conference**

**Publisher**

**Theme**

**Author**

- Aravkin, Aleksandr Y. (1)
- Asnaashari, Amir (1)
- Ata, Elias Z. (1)
- Audebert, François (1)
- Benetazzo, Alvise (1)
- Bernardelli, P. (1)
- Bonar, David (1)
- Brewer, Joel D. (1)
- Brossier, Romain (1)
- Brouwer, Wouter (1)
- Bube, Kenneth (1)
- Burdette, Jason (1)
- Chang, Dar-Lon (1)
- Christie, Philip (1)
- Coraggio, F. (1)
- Dragoset, William H. (1)
- Eggenberger, Kurt (3)
- Eick, Peter M. (1)
- Farquharson, Colin (1)
- Fedele, Francesco (1)
- Foks, Leon (1)
- Frost, Elton (1)
- Gabbriellini, G. (1)
- Gallego, Guillermo (1)
- Garambois, Stéphane (1)
- Griffiths, Roger (1)
- Grønaas, Halvor (1)
- Habashy, Tarek (1)
- Halliday, David (1)
- Herrmann, Felix J. (1)
- Huang, Hao (1)
- Janiszewski, Frank D. (1)
- Kaplan, Sam T. (3)
- Karmonik, Christof (1)
- Kemal Özdemir, Ahmet (1)
- Keskula, Erik (1)
- Keys, Robert G. (1)
- Kjellesvig, Bent (1)
- Krahenbuhl, Richard (1)
- Kreimer, Nadia (2)
- Laws, Robert (1)
- Li, Chengbo (3)
- Li, Yaoguo (1)
- Liu, Kuang-Hung (1)
- Long, Ted Alan (1)
- Martin, James (1)
- McDuff, Darren (1)
- Miles, Jeffrey (1)
- Moghaddam, Peyman Poor (1)
- Morley, Larry C. (1)
- Morriss, Chris (1)
- Mosher, Charles C. (3)
- Mulder, Wim A. (1)
- Naghizadeh, Mostafa (1)
- Olson, Robert A. (1)
- Omeragic, Dzevat (1)
- Quinn, Terrence (1)
- Rasmus, John (1)
- Rentsch, Susanne (1)
- Robertsson, Johan (1)
- Robertsson, Johan O.A. (1)
- Sacchi, Mauricio (1)
- Sacchi, Mauricio D. (1)
- Shetty, Sushil (1)
- Skauge, A. (1)
- Sood, Sanjay (1)
- Sorbie, K.S. (1)
- Stanton, Aaron (1)
- Thore, Pierre (1)
- Tsingas, Constantine (1)
- Valsecchi, Pietro (1)
- van Leeuwen, Tristan (1)
- van Manen, Dirk-Jan (4)
- Vasconcelos, Ivan (1)
- Vassallo, Massimiliano (3)
- Verschuur, Eric (1)
- Virieux, Jean (1)
- Vrolijk, Jan Willem (1)
- Yezzi, Anthony (1)
- Zeroug, Smaine (1)
- Zheglova, Polina (1)
- Özbek, Ali (4)
- Özdemir, Kemal (2)

**Concept Tag**

- acidizing (1)
- acquisition (5)
- adaptive subtraction (1)
- adjoint (1)
- airgun (1)
- algorithm (6)
- amplitude (1)
- analysis-based optimization model (1)
- angle well (1)
- annual meeting (3)
- anomaly (1)
- application (5)
- artifact (1)
- Artificial Intelligence (14)
- attenuation (1)
- azimuth (2)
- azimuthal variation (1)
- background (1)
- Blunt (1)
- borehole (3)
- boundary (3)
- CMP gather (1)
- crossline (3)
- CSI (1)
- data reconstruction (5)
- Engineering (1)
- equation (2)
- experiment (3)
- flow in porous media (2)
- Fluid Dynamics (2)
- Fourier (1)
- frequency (4)
- FWI (1)
- geometry (2)
- geophysics (7)
- GMP (2)
- Gravity Gradiometry (1)
- Herrmann (1)
- Horizontal Gradient (1)
- ICA (1)
- IEEE Transaction (1)
- Imaging (1)
- InfoMax (1)
- information (5)
- Information Theory (1)
- interpolation (5)
- interpretation (2)
- inverse problem (2)
- inversion (6)
- log analysis (2)
- machine learning (3)
- matrix (2)
- McDougall (1)
- Modeling & Simulation (1)
- Monopole (1)
- MRI (1)
- Mulder (1)
- Multicomponent (3)
- Multiscale (1)
- Multiscale noise attenuation (1)
- MWNI (1)
- Network Modeling (1)
- noise (2)
- Noise Attenuation Algorithm (1)
- objective function (3)
- optimization problem (5)
- particle velocity (2)
- POC (1)
- prediction (2)
- pressure wavefield (3)
- Radon transform (1)
- receiver (3)
**reconstruction (23)**- Reservoir Characterization (17)
- resolution (2)
- Robertsson (1)
- Sacchi (2)
- Scenario (1)
- SEG (4)
- SeisADM (1)
- seismic data (4)
- seismic data reconstruction (3)
- separation (2)
- SIAM Journal (1)
- Signature (1)
- Sorbie (1)
- spatial (3)
- streamer (4)
- Symposium (1)
- synthetic data (2)
- TCOM (1)
- Upstream Oil & Gas (22)
- Variational (1)
- Vassallo (1)
- Verschuur (1)
- Visualization (1)
- wavefield (4)
- Wavefield Reconstruction (2)
- Waveform Inversion (1)
- well logging (2)

**File Type**

ABSTRACT

Over the last two decades there has been an increase in activity on the pore-scale modeling of multiphase flow in porous media. Excellent progress has been made in many areas of pore-scale modeling, particularly in (1) the representation of the rock itself and (2) our description of the pore-scale displacement physics (in model pore geome-tries). Three-dimensional voxelized images of actual rocks can be generated either numerically (e.g. from 2D thin sections) or from micro-CT imaging. A simplified network involving more idealized nodes and bonds can then be extracted from this numerical rock model and this can be used in modeling pore-scale displacement processes. Much progress has also been made in understanding these pore-scale processes (i.e. piston-like displacement, snap-off events, layer formation/collapse, pore-body filling draining). These processes can be mathematically modeled accurately for pores of non uniform wettability, if the geometry of the pore is sufficiently simple. In fact, in recent years these various pore-level processes in mixed and fractionally wet systems have been classified as "events" in an entire capillary-dominated "phase space" which can be defined in a thermodynamically consistent manner. Advances in our understanding and ability to compute several two- (and three-) phase properties a priori have been impressive and the entire flooding cycle of primary drainage (PD), aging/wetting change, and imbibition can be simulated.

In this paper, we review the successes of pore-scale network modeling and explain how it can be of great use in understanding and explaining many phenomena in flow through porous media. However, we also critically examine the issue of how predictive network modeling is in practice. Indeed, one of our conclusions on pore-scale modeling in mixed-wet systems is that we cannot predict two-phase functions reliably in "blind" tests. Interestingly, we make this statement not because we do not understand the pore-scale physics of the process, but because we do understand the physics. It is hoped that our comments will stimulate a more critical debate on the role of pore-scale modeling and its use in core analysis.

Artificial Intelligence, Blunt, drainage, Engineering, enhanced recovery, flow in porous media, Fluid Dynamics, McDougall, modeling, network model, Network Modeling, physics, pore, pore-scale network, prediction, reconstruction, relative permeability, Sorbie, two-phase flow, Upstream Oil & Gas, van Dijke, wettability

SPE Disciplines:

Technology:

Asnaashari, Amir (Université Joseph Fourier Grenoble) | Brossier, Romain (Université Joseph Fourier Grenoble) | Garambois, Stéphane (Université Joseph Fourier Grenoble) | Virieux, Jean (Université Joseph Fourier Grenoble) | Audebert, François (TOTAL E&P) | Thore, Pierre (TOTAL E&P)

Full Waveform Inversion (FWI) is an appealing technique for time-lapse imaging, especially when the prior model information is included into the workflow. After baseline reconstruction, several strategies such as: differential, parallel difference, and sequential difference can be used to assess the physical parameter changes. Using the synthetic Marmousi data-sets, we study which strategy could be more robust and give more accurate time-lapse velocity changes in the presence of noise. We illustrate that the sequential difference method, starting from a reconstructed baseline model and inverting the monitor data-set, can give a better result in the case of random ambient noise. However, the differential approach could also be interesting if the time-lapse response can be preserved from the noise level.

application, Artificial Intelligence, baseline, baseline model, difference method, differential, differential data, FWI, information, inversion, noise, noisy, optimization problem, parallel difference, reconstruction, regularized fwi, Reservoir Characterization, robustness, sequential difference, time-lapse imaging, Upstream Oil & Gas, weighting

SPE Disciplines: Reservoir Description and Dynamics > Reservoir Characterization > Seismic processing and interpretation (1.00)

Technology: Information Technology > Artificial Intelligence > Representation & Reasoning > Optimization (0.46)

In this paper, we present a direct and reliable approach to the adaptive down-sampling of potential-field data for large inversion problems. In contrast to traditional down-sampling methods, the approach significantly reduces the number of data parameters in relatively smooth/quiet regions of the data, while preserving the signal anomalies that contain the relevant target information. This allows for a simple and effective approach for compressive inversion of large datasets, without the need for large computing power, while maintaining the resolution of the recovered structures. The formulation has the flexibility to decimate large data sets for an optimal balance between data number and signal shape, or data can be decimated more conservatively if desired. We first present the data adaptive down-sampling technique, and then demonstrate the approach, applied to both synthetic and field data.

Oilfield Places:

- North America > United States > Texas > West Gulf Coast Tertiary Basin > Mineral Field (0.98)
- North America > Canada > Nova Scotia > Minas Basin (0.97)
- South America > Brazil (0.95)

SPE Disciplines:

Full waveform inversion with the classic least-squares cost functional suffers from spurious local minima caused by cycle skipping in the absence of low frequencies in the seismic data. We present an alternative cost functional that is less sensitive to cycle skipping. It has the property of an annihilator, just like the functional used for velocity analysis with extended images based on subsurface shifts. For 2-D models with a line acquisition, the proposed functional applies a singular-value decomposition on the observed data and uses the eigenvectors to build a data panel that should be diagonal in the correct velocity model but has significant off-diagonal entries in the wrong model. By penalizing off-diagonal entries or maximizing values close to the main diagonal, the correct model should be found. We therefore call it the diagonalator.

A convexity test demonstrates the superiority of the proposed functional over the classic least-squares approach. We present initial synthetic data tests on a subset of a North Sea velocity model. The diagonalator performs better than least-squares data fitting in terms of resolving deeper events with full-bandwidth data. It also converges to an acceptable velocity model in the absence of low frequencies, when least-squares minimization fails.

Artificial Intelligence, cost functional, diagonalator, frequency, full waveform inversion, geophysics, inversion, least-square cost functional, least-square inversion, machine learning, migration, minimization, Mulder, objective function, reconstruction, Reservoir Characterization, SEG, Upstream Oil & Gas, van Leeuwen, velocity analysis, velocity model, Waveform Inversion

SPE Disciplines: Reservoir Description and Dynamics > Reservoir Characterization > Seismic processing and interpretation (1.00)

Summary We present a fast and effective method to detect and eliminate seismic interference from 3D marine data measured by four-component (4C) streamers. The interference elimination method we propose acts on each shot record independently from the others, relying on the pressure wavefield being reconstructed (and simultaneously deghosted) on a 2D grid, densely sampled in both the inline and the crossline directions. Such reconstruction is enabled by matching-pursuit-based signal processing techniques proposed recently in the literature that have the capability to explicit the information in the multicomponent measurements. Without these measurements, the reconstruction capability is seriously compromised by the strong crossline aliasing.

acquisition, annual meeting, azimuth, crossline, crossline direction, crossline view, elimination, inline, interference, marine acquisition, Multicomponent, particle velocity, pressure wavefield, reconstruction, Reservoir Characterization, SEG, seismic interference, seismic interference elimination, streamer, Upstream Oil & Gas, wavefield

SPE Disciplines: Reservoir Description and Dynamics > Reservoir Characterization > Seismic processing and interpretation (1.00)

There are many features in the Earth’s crust that involve a jump in physical property across a sharp boundary, for example, an ore deposit in host rocks. Such well defined boundaries are often of interest to geophysicists, however traditional minimum-structure inversion methods produce blurred images of the subsurface. In this paper we explore the application of a level set inversion method to recover a sharp boundary between two slowness values, one characterizing an inclusion, e.g. an ore body, the other characterizing a background, e.g. host rock, from first arrival travel time data. The scenario considered is that of cross-borehole tomography in two dimensions.

SPE Disciplines:

A comparison is made between three 5D reconstruction methods– Projection Onto Convex Sets (POCS), Tensor Completion (TCOM), and Minimum Weighted Norm Interpolation (MWNI). A method to measure of the quality of synthetic data reconstructions is defined and applied under various scenarios. Two different measures of performance in the case of real data reconstructions are also provided and applied to a real data example taken from a land dataset acquired in the Western Canadian Sedimentary Basin. We find that TCOM and POCS are better able to reconstruct data in the presence of low SNR. We also find that TCOM provides superior results in most synthetic data scenarios, but in the case of real data reconstruction all three methods have similar performance, with POCS giving slightly better preservation of amplitudes.

amplitude, annual meeting, Artificial Intelligence, data reconstruction, frequency, geophysics, input data, interpolation, midpoint, MWNI, POC, real data, reconstruction, reconstruction method, reconstruction result, Reservoir Characterization, Sacchi, Scenario, spatial, synthetic data, TCOM, Upstream Oil & Gas

Oilfield Places:

- North America > Canada > Saskatchewan > Western Canada Sedimentary Basin > Alberta Basin (0.99)
- North America > Canada > Northwest Territories > Western Canada Sedimentary Basin > Alberta Basin (0.99)
- North America > Canada > Manitoba > Western Canada Sedimentary Basin > Alberta Basin (0.99)
- (2 more...)

Mosher, Charles C. (ConocoPhillips) | Keskula, Erik (ConocoPhillips) | Kaplan, Sam T. (ConocoPhillips) | Keys, Robert G. (ConocoPhillips) | Li, Chengbo (ConocoPhillips) | Ata, Elias Z. (ConocoPhillips) | Morley, Larry C. (ConocoPhillips) | Brewer, Joel D. (ConocoPhillips) | Janiszewski, Frank D. (ConocoPhillips) | Eick, Peter M. (ConocoPhillips) | Olson, Robert A. (ConocoPhillips) | Sood, Sanjay (ConocoPhillips)

Ideas from the field of compressive sensing are rapidly making their way into the geophysical realm. We believe that these concepts will motivate major changes in the way that our industry acquires, processes, and images seismic data. In preparation for these changes, we have undertaken an initiative to build a consistent framework for learning, investigating, and applying compressive sensing concepts to the full range of technologies used in seismic acquisition, processing, and imaging. We refer to this framework as Compressive Seismic Imaging (CSI). The components of our CSI framework include compressive sensing theory, acquisition design, processing and imaging algorithms, and the work flows that link these components into a complete system. A key element of our CSI program is the use of field trials to expose algorithms, processes, and people to the realities of deploying new technology in our industry. Before going to the field, we use extensive computer modeling to identify CSI concepts that are either ready for deployment, or require testing in the field to advance the technology. A number of 2D and 3D field trials were undertaken by ConocoPhillips in 2011 to test compressive sensing design ideas for seismic data acquisition. To date, we have acquired test datasets for validating CSI concepts for land, marine, and ocean bottom recording configurations. The key compressive sensing concepts we have tested so far include non-uniform sampling for sources and receivers, data reconstruction, simultaneous shooting, and source encoding. Initial results from these trials show that compressive sensing concepts have the potential to significantly improve acquisition efficiency. Use of the CSI framework has allowed us to quickly focus our attention on the most relevant problems for compressive sensing technology deployment, resulting in rapid progress in our understanding.

Oilfield Places:

- Oceania > Australia > Western Australia > North West Shelf > Carnarvon Basin > Sinbad Field (0.98)
- North America > United States > Texas > Rice Field (0.94)

Recent research indicates that compressive sensing (CS) can be successfully applied to seismic data reconstruction. It also provides a powerful tool that reduces the acquisition cost, and allows for the exploration of new seismic acquisition designs. Most seismic data reconstruction methods require a predefined nominal grid for reconstruction, and the seismic survey must contain observations that fall on the corresponding nominal grid points. However, the optimal nominal grid depends on many factors, such as bandwidth of the seismic data, geology of the survey area, and noise level of the acquired data. It is understandably difficult to design an optimal nominal grid when sufficient prior information is not available. In addition, it may be that the acquired data contain positioning errors with respect to the planned nominal grid. We propose an interpolated compressive sensing method which is capable of reconstructing the observed data on an irregular grid to any specified nominal grid, provided that the principles of CS are satisfied. We first describe the theory and implementation of this interpolated CS method. Then we illustrate this approach using synthetic and real data examples, and make comparisons to the traditional CS method. We show that the interpolated CS method provides an improved data reconstruction compared to results obtained from the traditional CS method.

annual meeting, application, data reconstruction, Fourier, geophysics, Herrmann, IEEE Transaction, information, Information Theory, interpolation, main menu, reconstruction, reference list, Reservoir Characterization, scientific computing, SEG, seismic data, seismic data reconstruction, SIAM Journal, Upstream Oil & Gas

The particle velocity vector enables us to calculate the 3D upgoing wavefield at any desired position within the aperture of the seismic spread and this allows improving not only the temporal bandwidth by removing ghost notches, but also the spatial bandwidth. However, particle motion sensors measure streamer-borne noise with amplitudes typically several orders of magnitude stronger than the corresponding noise recorded by hydrophones at frequencies below about 20-Hz. Therefore, stronger noise attenuation for particle velocity data is needed at these frequencies. In this paper, we introduce a multiscale noise attenuation algorithm that provides a high-fidelity particle motion measurement at frequencies down to 3 Hz. We also show that the new technique provides improved noise attenuation on pressure data. In this paper, we present the characteristics of the noise recorded by the particle motion sensors in the multicomponent (4C) towed streamer and introduce the multiscale noise attenuation (MSNA) algorithm that is used to attenuate noise on both pressure and particle motion data. We show that the MSNA algorithm provides increased noise attenuation on hydrophone data compared to conventional methods and a high-fidelity particle motion measurement at frequencies down to 3-Hz. Introduction Robertsson et al. (2008) introduced the concept of a multicomponent (4C) towed streamer that acquires both pressure and the full particle velocity vector with inline, crossline, and vertical components.

Thank you!