- Home
- Conferences
- Conference Proceedings
- Conferences
75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013
- Conference date: 10 Jun 2013 - 13 Jun 2013
- Location: London, UK
- ISBN: 978-90-73834-48-4
- Published: 10 June 2013
41 - 60 of 1113 results
-
-
Analysis of Seismic Attenuation and Multiple Scattering with a Version of the Shannon Entropy
By K.A. InnanenMultiple scattering and intrinsic friction both contribute to the attenuation of seismic waves, and are difficult to distinguish in seismic experiments. In this paper we will examine attenuation using theoretical tools that put scattering and friction on more or less an equal footing. Multiple scattering is, after all, a process that increases the disorder of mechanical motions in a continuum. If it occurs well below the resolution of an experiment, the energy it carries is, in a sense, classifiable as thermal. The Shannon entropy, defined on snapshots of a propagating wave, is a measure of disorder in something like the above sense. Its rate of monotonic increase, as a wave experiences either or both of multiple scattering and attenuation, may be a useful means to quantify intrinsic and extrinsic Q. Synthetic experiments and a field VSP data set appear to support this idea.
-
-
-
Continuous Mapping of P Wave Velocity Dispersion - A Useful Tool for Reservoir Characterization
Authors L.F. Sun, A. Campbell and B. MilkereitFor reservoir rocks, P wave velocity dispersion may provide a potential link to rock physical properties. This requires continuous measurement of P wave velocity in a broad frequency band to determine the critical frequency of the attenuation mechanism. In this study, we measure P wave velocity dispersion continuously in sonic frequency range using broadband full-waveform multi-channel sonic logging data with the beam-forming and cross-correlation techniques. Results of the field data from 5L-38 Mallik gas hydrate research well have demonstrated the robustness of this method. The profile of P wave velocity dispersion matches very well with geological settings. The gas hydrate zones show very strong P wave velocity dispersion and a pronounced critical frequency around 15 kHz; weakly-laminated sediments also have strong P wave velocity dispersion with critical frequency lower than 5 kHz; P wave velocity dispersion in consolidated sediments is mild and gradual. In addition, the total magnitude of P wave velocity dispersion is positively correlated to resistivity and gas hydrate saturation. Therefore, continuous P wave velocity dispersion mapping can be a promising tool for reservoir characterization.
-
-
-
Inversion by Trace Matching
Authors P.A. Connolly and M.J. HughesWe propose a method to estimate reservoir properties by matching seismic to very large numbers of pseudo-wells. Each pseudo-well consists of a full suite of well-log curves based on a comprehensive rock physics model. The pseudo-wells are calibrated to real wells and constrained by the seismic interpretation with stochastic micro-layering. The input seismic data are colour inverted angle stacks (optionally multiple angle stacks). The inversion algorithm is 1D, Monte-Carlo. All pseudo-wells are generated completely independently of any previous result and no spatial correlation is enforced. No low-frequency model is required. The procedure also provides estimates of uncertainties allowing the results to be integrated with other data types in a Bayesian framework. We outline the process and illustrate showing application to a deep marine turbidite field.
-
-
-
Amplitude Inversion of Depth-imaged Seismic Data from Areas with Complex Geology
Authors S.H. Archer, X. Du and R.P. FletcherConventional amplitude inversion assumes that the input migrated image has preserved relative amplitude information and is free from the effects of illumination. Under this assumption, stretching a depth migrated image back to time and applying inversion based on 1D convolutional modeling can produce reasonable results. However, illumination effects in complex geological settings (such as shadow zones in subsalt imaging) pose a challenge to even the most advanced imaging algorithms such as reverse-time migration (RTM). Traditional approaches to compensate for illumination effects in migrated images are difficult to regularize in areas of very poor illumination. We address this problem by using the modelled response of the acquisition and imaging process, defined by Point Spread Functions (PSFs), to include these effects in forward modeling for inversion directly in the depth domain. We demonstrate this approach for poststack inversion of synthetic, subsalt data and also apply it to field data from the Gulf of Mexico.
-
-
-
Data Driven Versus Model Based Inversion - When and Why?
By P. ThoreI present a classification of Seismic Inversions in two categories; Data Driven and Model Based. Based on examples I show that no inversion is purely data driven and that the user has to provide prior information which constrains the solution to a certain behavior. On the other hand, Model Based Inversion can provide very appealing results (an example breaks the traditional view of seismic resolution) provided that the underlying model is sound. I conclude by proposing a workflow which tries to get the best from both approaches.
-
-
-
Model-Based Inversion of Low-Frequency Seismic Data
Authors P. Gavotti, D.C. Lawton, G. Margrave and J.H. IsaacThe Hussar experiment was carried out in central Alberta, Canada, in September 2011 with the purpose of acquiring low frequency seismic data to be used in inversion methods. Three wells located close to the seismic line and a dynamite-source dataset, acquired with three-component 10 Hz geophones, were used for a post-stack inversion test using commercial software. Several low-frequency cut-off filters applied to the data were tested with the 3-5 Hz model being selected as the optimum. The resultant impedance reflects lateral changes that were not present in the initial model and therefore are derived from the seismic reflections. Impedance changes in the target zone shows the general trend and relative variations, which would allow changes in the reservoir to be monitored as variations in the rock properties occur. A final inversion was performed to simulate traditional approaches when the low-frequency component is absent in the seismic data. Filtered seismic-data (10-15-60-85 Hz) and an initial model with a 10-15 Hz cut-off were used for this test. The results at the well locations show a good match but the lateral variation and character of the events resemble more the initial model character.
-
-
-
The Accuracy of Spectral Decomposition Methods
More LessIn the context of exploration seismology the resolution of spectral decomposition techniques has been extensively reviewed but the accuracy of these methods has only been discussed qualitatively in regards to the resolution. Using a marginal condition the fundamental accuracy of the matching pursuit decomposition, Stockwell transform, Wigner-Ville distribution and reduced interference distribution are quantified. A cut-off for what can be deemed an acceptable decomposition method is set at 0.3 normalized residual energy, the normalized residual energy for the matching pursuit decomposition is 0.59. Used in conjunction with a PSQI method, the Stockwell transform provided the most accurate determination of the quality factor, Q, of a synthetic shot gather, recovering a Q value of 50.4±3 from a synthetic shot gather with an attenuative layer of quality factor 50. It is recommended that the best spectral decomposition method for a specific task be identified through rigorous testing of spectral decomposition methods with synthetic data.
-
-
-
Spectral Decomposition by Synchrosqueezing Transform
Authors J. Han, R.H. Herrera and M. van der BaanThe synchrosqueezing transform (SST) is a wavelet-based time-frequency reassignment method, which has a grounded mathematical foundation. It produces a well defined time-frequency representation allowing the identification of instantaneous frequencies to highlight individual components. The field data examples demonstrate the high time-frequency resolution feature of SST, therefore render this technique is promising for seismic processing and interpretation.
-
-
-
Gabor Deconvolution Based on Hyperbolic Smoothing in Log Spectra
More LessThe choice of t-f smoothing method is viewed as a significant procedure of Gabor deconvolution. Inspired by homomorphic deconvolution, this paper extends Gabor deconvolution into the log spectra, wherein the explicit linear form of nonstationary convolution model brings great facilities for studying its reverse process. More pertinently, we propose a hyperbolic smoothing method there, in which the linear analysis is iteratively employed to approach the factorization task. Since divisions are transformed to subtractions in log spectra, the usage of the pre-whitening factors in the iteration is avoided, thus reducing human intervention and boosting calculation accuracy. Moreover, a simple piecewise linearization is also introduced to enhance this technique's practical value. And our experiences on reef reservoir and carbonate reservoir show that nonstationary deconvolution based on this new hyperbolic smoothing can acquire higher resolution than ordinary hyperbolic smoothing, thus providing more detailed information and revealing more subtle geological phenomena, which enable better reservoir characterizations and definitions.
-
-
-
Gabor Deconvolution - Hyperbolic Smoothing with Variable-step Sampling
More LessEstimation of nonstationary wavelet from seismic trace is the core problem of the nonstationary deconvolution. The accuracy of estimation wavelet affects the deconvolution result. There are several methods to smooth the Gabor magnitude spectrum of seismic trace, and each of these leads to a distinct method of nonstationary deconvolution. Among all of the methods, hyperbolic smoothing has proven most robust. However, uniform sampling in conventional hyperbolic smoothing cause subsampling in the observable area and oversampling in the noise area. Thus, this lead to an inaccurate wavelet estimation and unreliable deconvolution result. In this paper, we propose a new hyperbolic smoothing method with variable-step sampling (VSS), which can lead to an effective division in the time-frequency plane. Tests on synthetic and real data have shown that more reliable results can be obtained using the new method; moreover, it can also reduce the calculation time.
-
-
-
The Application of Robust Principal Component Analysis for Weak Seismic Signal Enhancement
Authors X. Gao, W.K. Lu, F.Y. Li and X.D. JiangFor weak seismic signal enhancement, a new application of a signal decomposition method, known as Robust Principal Component Analysis (RPCA) is introduced. The motivation of this work stems from the observation that the interested weak seismic signals are always interfered by strong ones besides noises, causing the loss of some detailed seismic information. Principal Component Analysis (PCA), based on second order statistics, however, requires the data to be white and Gaussian, which seismic data may not satisfy. Being an extension of traditional PCA, RPCA utilizes L1-norm instead as the error function and the iterative algorithm obtains the optimal projections one by one with a greedy strategy. The synthetic data experiment indicates that RPCA outperforms PCA in seismic data processing as RPCA forms less artificial horizontal events. Moreover, an example of a 3-D field data is considered on which there are two wells close to each other. The seismic events are continuous across the wells, whereas the oil and gas production of the wells is distinct. The results demonstrate that RPCA is effective for weak seismic signal enhancement and helpful to improve the reliability of oil and gas detection.
-
-
-
Seismic Reflectivity Inversion in a Multichannel Manner
More LessIn reflectivity inversion problems, due to complicated structures and low signal-to-noise ratio (SNR) of input seismic traces, the conventional Cauchy constrained deconvolution method is normally not able to provide the results that can clearly characterize the geology structure. One of the reasons is that the inversion process is performed on a trace-by-trace basis, and as a result the continuity along reflectors in seismic images may be deteriorated. In this paper, we develop a multichannel algorithm to perform this inversion process, where a multichannel precondition filter is incorporated into the conventional Cauchy constrained deconvolution method and the information of adjacent traces is applied during the inversion process. Numerical experiments have verified the validity and feasibility of this method by field data, showing that the reflectivity profiles obtained using the proposed method can have an improved lateral continuity and clearer structure.
-
-
-
2D Spiking Deconvolution Approach to Resolution Enhancement of Prestack Depth Migrated Seismic Images
Authors A.K. Takahata, L.J. Gelius, R.R. Lopes, M. Tygel and I. LecomteComplex velocity models, limitations in acquisition geometry and frequency bandwidth, give rise to distortions in prestack depth migrated (PSDM) images. Such distortions can be modelled as the 2D convolution between the actual reflectivity and a resolution function. In the case of Born scattering, the resolution function is referred to as point spread function (PSF). The PSFs can be calculated with relatively low computational effort by ray tracing. In this work, we review the basic idea of the PSF and its relationship with seismic images generated by PSDM. With the help of the PSF concept, we propose the use of 2D spiking deconvolution with the aim of minimizing these image distortions. Finally, the potential and limitations of the proposed method are explored with applications on controlled synthetic data.
-
-
-
Matching Pursuit with Enhanced Performance and Adaptability
More LessOne of the advantages which matching pursuit (MP) method has over conventional time-frequency decomposition is its ability to get capture of time-frequency fingerprint with extremely high resolution, which also makes itself quite sensitive to data noises and results in the non-uniqueness of its solution. In order to further diminish the influence of noises and stabilize the performance, two targeted modifications are incorporated in this work. Specifically, the lateral continuity of the underground layers is exploited as a constraint and as a reference for aligning the signal traces in the decomposition process. As a consequence, both the number of decomposed wavelets and the computation time can be significantly reduced. Moreover, choosing the best wavelet in each iteration follows a minimal residual energy criterion, which further results in a smaller number of resultant wavelets and thus makes the decomposition more robust.
-
-
-
Hybrid Algorithm for Simulation of Seismic Wave Propagation in Complex Media - Anisotropy, Attenuation, Multi-scale
Authors V.V. Lisitsa, V.I. Kostin, G.V. Reshetova, D.M. Vishnevsky and V.A. TcheverdaThis paper presents an original algorithm for simulation of seismic waves in models containing geological formations with complex properties such as anisotropy, attenuation, and small-scale inhomogeneities. Each of these structures require special treatment either small gridding or computationally intense models and algorithms. Meanwhile, these formations typically take as few as 25% of the model, thus computationally expensive approaches can be used locally, while efficient algorithm can be applied elsewhere. Proper combination of several techniques allows improving the performance and efficiency of the parallel algorithm, in particular the speed-up of the hybrid algorithm is up to 10 if the formations with the complex properties take about 25 % of the model.
-
-
-
Illumination/Reliability in the Local Angle Domain
Authors R. Levy, I. Ravve, L. Korkidi and Z. KorenWe demonstrate how we solve the two-point ray tracing problem in the presence of complex geological structures. For a point diffractor (PD) single-ray ray fan, we define surface-to-subsurface mapping (and vice versa) which maps the subsurface angles (take-off dips and azimuths) of individual traced rays to their surface locations. This mapping provides us with an interpolated ray for any subsurface angle or any surface location. We also introduce the reliability factor, which is based on a high-order interpolation technique in which we integrate several physical parameters calculated along the rays. This factor serves as a way of measuring the quality of the surface-to-subsurface mapping, and allows us to obtain a full ray (wave) field representation at any location, similar to wave equation methods. The same methodology is used for Common Reflection Point (CRP) specular ray pairs which are traced from a given subsurface location, with various opening angles and azimuths and a given reflector directivity, to the surface.
-
-
-
Efficient Traveltime Solutions of the TI Acoustic Eikonal Equation
Authors U. Waheed and T. AlkhalifahNumerical solutions of the eikonal (Hamilton-Jacobi) equation for transversely isotropic (TI) media are essential for integral imaging and traveltime tomography applications. Such solutions, however, suffer from the inherent higher-order nonlinearity of the TI eikonal equation, which requires solving a quartic polynomial at each computational step. Using perturbation theory, we approximate the first-order discretized form of the TI eikonal equation with a series of simpler equations for the coefficients of a polynomial expansion of the eikonal solution in terms of the anellipticity anisotropy parameter. Such perturbation, applied to the discretized form of the eikonal equation, does not impose any restrictions on the complexity of the perturbed parameter field. Therefore, it provides accurate traveltime solutions even for the anisotropic Marmousi model, with complex distribution of velocity and anellipticity anisotropy parameter. The formulation allows tremendous cost reduction compared to using the exact TI eikonal solver. Furthermore, comparative tests with previously developed approximations illustrate remarkable gain in accuracy of the proposed approximation, without any addition to the computational cost.
-
-
-
Low-frequency Layer-induced Anisotropy
Authors A. Stovas, Y. Roganov, K. Duffaut and A. CarterWe develop a low-frequency dynamic effective medium for a horizontally layered transversely isotropic medium with vertical symmetry axis (VTI), and approximate this effective medium by a homogeneous VTI medium. This results in frequency-dependent anisotropy parameters valid for low frequencies and quasi-vertical propagation.
-
-
-
Renormalized Integral-equation Method for 3D Acoustic Full-waveform Modeling
Authors A. Abubakar and T.M. HabashyWe present a frequency-domain renormalized integral-equation formulation for solving a three-dimensional visco-acoustic medium using an iterative solver. Upon applying this special renormalization, the resulting integral-equation operator can be proven to have a contraction property. Hence, solving the linear-system of equations using a Krylov optimization method, will result in a good convergence rate. Furthermore since the matrix-vector multiplication can be done using a Fast-Fourier transform (FFT) technique, its operation is of the order of N Log N, where N is the size of the discretization grid. This technique also allows us to use matrix-free implementation. Hence, the memory usage is about N. Numerical tests show that the computational time and memory usage of this renormalized integral-equation approach can be quite competitive with the frequency-domain finite-difference iterative solver. Further, the numerical examples demonstrate that it is possible to solve a problem with over 100 million unknowns using an integral-equation approach.
-
-
-
The Optimizied Expansion Method for Wavefield Extrapolation
Authors Z. Wu and T. AlkhalifahSpectral methods are fast becoming an indispensable tool for wave-field extrapolation, especially in anisotropic media, because of its dispersion and artifact free, as well as highly accurate, solutions of the wave equation. However, for inhomogeneous media, we face difficulties in dealing with the mixed space-wavenumber domain operator.In this abstract, we propose an optimized expansion method that can approximate this operator with its low rank representation. The rank defines the number of inverse FFT required per time extrapolation step, and thus, a lower rank admits faster extrapolations. The method uses optimization instead of matrix decomposition to find the optimal wavenumbers and velocities needed to approximate the full operator with its low rank representation.Thus,we obtain more accurate wave-fields using lower rank representation, and thus cheaper extrapolations. The optimization operation to define the low rank representation depends only on the velocity model, and this is done only once, and valid for a full reverse time migration (many shots) or one iteration of full waveform inversion. Applications on the BP model yielded superior results than those obtained using the decomposition approach. For transversely isotopic media, the solutions were free of the shear wave artifacts, and does not require that eta>0.
-