- Home
- Conferences
- Conference Proceedings
- Conferences
70th EAGE Conference and Exhibition - Workshops and Fieldtrips
- Conference date: 09 Jun 2008 - 12 Jun 2008
- Location: Rome, Italy
- ISBN: 978-94-6282-104-0
- Published: 09 June 2008
61 - 80 of 91 results
-
-
Shallow versus deep thermal circulations at Bagni di S. Filippo (M.te Amiata, Tuscany, Italy)
Authors A. Baietto, G. Giudetti, S. Governi, L. Fusani and E. SalvaticiThe M.te Amiata sector constitutes a volcano-geothermal area of the southern Tuscany,.
This area has been affected by extensional tectonics from the Early-Middle Miocene onwards
(Carmignani et al., 1994), that led, in Pliocene, to the emplacement of a deep seated
intrusive body and to the eruption of dacitic-rhyodacitic lavas through a NW-SE-trending
fissure (Ferrari et al., 1996). Currently, this area is characterized by a high heat flow (up to
600 mW/m2; Baldi et al., 1995) that feeds important geothermal fields. Main geothermal
reservoirs are located at depths of several hundreds of meters within Triassic evaporitic
horizons, constituting the base of the Tuscan Unit.
-
-
-
Resistivity reduction in the vapour-dominated field of Travale (Italy)
Authors C. Giolito, G. Ruggieri, A. Manzella and G. GianelliThe aim of this multidisciplinary work is to find out what can account the significant reduction
in resistivity (from 103 to 100 m) observed at the depth of the geothermal reservoirs in the
Travale area (SE of Larderello, Italy). Since the exploited fluid is supersaturated steam and
thus resistive, its presence and localisation can not explain the observed resistivity
reductions. The observed reduction in resistivity could be related: 1) to the lithology and
heterogeneities of reservoir rocks and the alteration affecting them (i.e. presence-abundance
of conductive minerals), and/or 2) to the presence of brines within a fracture net sufficiently
interconnected to produce electrolytic conduction.
-
-
-
Waveform tomography - Successes, Cautionary tales, and future directions
By R. G. PrattI would like to highlight some of the issues in evaluating the success (or otherwise) of
waveform tomography. For synthetic data it is always tempting to commit inverse crimes, for
real data a scrutiny of the detailed data fit is "sine qua non". I'll go through some new
animations that illustrate the benefits of the frequency domain, and I'll finish with some topics
that we are working on - real data examples of Q-inversion being a current interest.
-
-
-
Applications of Waveform Inversion
More LessWaveform inversion techniques aim to fit the entire seismic wavefield including those phases
that conventional processing and migration seek to remove. Such methods have the potential
to image the subsurface with significantly improved spatial resolution. The inversion in this
paper uses a frequency-domain, finite-difference modeling method to solve the full acoustic
wave equation, so high-order effects such as diffractions and multiple scattering are
accounted for automatically. It is a local descent algorithm that refines a starting model
iteratively to reduce the waveform misfit between observed and modeled data.
Waveform inversion is applied to a number of synthetic models, including the complicated
BP EAGE model. The results have demonstrated that waveform inversion has the potential to
reconstruct high-resolution velocity structure. Some practical strategies were found to be
critical in the application to real data. These strategies include: using a diving wave
tomography model as a starting model; starting the inversion with the lowest available
frequency; using complex-valued velocity to take care of undesired amplitude discrepancies;
using complex-valued frequencies to simulate the damping of late arrivals; and handling
surface-related multiples effectively.
-
-
-
Effects of Surface Scattering in Waveform Inversion
Authors F. Bleibinhaus and S. RondenayFor seismic waveform inversion of body waves, the free surface is usually neglected on
grounds of computational efficiency. Subsurface parameters are simply extended in the air,
and sources and receivers are embedded in the model at their true locations. Many synthetic
studies have shown that this is an acceptable approximation if the surface is flat and,
naturally, surface waves are excluded from the inversion. In this study, we investigate the
effects of ignoring P wave scattering from irregular topography in acoustic waveform
inversion.
-
-
-
Full Elastic Waveform Inversion: Future of Quantitative Seismic Imaging
Authors S. Singh, T. Sears, M. Roberts, A. Gosselet, G. Royle and P. Batoneismic reflection data are acquired at a very high cost. Conventional processing (stacking
and migration) provides very high-quality image of the sub-surface, but does not provide
quantitative measure of the physical properties of the sub-surface. Amplitude versus offset
(AVO) analyses can be used to estimate P and S-wave impedances. Since the method is local,
i.e. assumes 1D media, linear approximation to the reflection coefficient, and ignores
interference effects, the results are very approximative. Over the last ten years, we have
developed a suite of 1D and 2D full waveform inversion. We have particularly focused on the
use of wide-aperture data, containing near- and post-critical angle reflections, which helped to
constrain medium scale features of the velocity model, allowing convergence towards the
global minimum. The algorithm has been applied to surface seismic reflection, ocean bottom
cable and walk-away VSP data. Both vertical (Vz) and horizontal (Vx) particle velocity
records aree used, allowing fine-scale estimation of both P- and S-waves velocities.
The 2D elastic waveform inversion scheme is based on Shipp and Singh (2002) and
Freudenreich et al. (2002), utilizing a finite-difference solution to the 2D elastic wave
equation (Levander, 1988) operating in the time-distance domain. The aim of the scheme is
to model shot gathers accurately, and to use the residual between observed and modeled
wavefields to update the velocity model appropriately using a conjugate gradient method.
Both Vp and Vs may be inverted for, whilst density is coupled empirically with Vp.
-
-
-
Full-waveform inversion results when using acoustic approximation instead of elastic medium
Authors C. Barnes and M. ChararaSeismic marine data inversion is a very heavy process, especially for the 3D seismic case. Often approximations are made to limit the number of physical parameters or to speed up the forward modeling. Because the data are often dominated by uncoverted P waves, one popular approximation is to consider the earth as purely acoustic: no shear modulus; even sometimes with constant density. Nonlinear waveform seismic inversion consists in iteratively minimizing the misfit between the amplitudes of the measured and the modeled data.
-
-
-
3D wavefield tomography: Problems, opportunities and future directions
More LessWavefield tomography, otherwise known as full-waveform inversion, of two-dimensional
seismic data, has become a well-established technique over the past decade, with impressive
recovery of realistically complex synthetic models being reported by several groups.
However, despite its proven potential, its uptake to tackle real-world exploration and
production problems has been rather limited. In our view, this has been principally because
the increased spatial resolution, accuracy, and other benefits that the method brings are only
genuinely realised for field data when the method is extended to deal with three-dimensional
velocity structure, three-dimensional reflection geometry, and a three-dimensional array of
sources and receivers. Since the real world is always three-dimensional, very-accurate twodimensional
solutions to three-dimensional problems are nearly always illusory – the higher is
the spatial resolution of the method, and the more accurate is the physics of wave propagation
that is employed, then the more significant will be the errors that are introduced by neglect of
the third dimension. In essence, there is little utility to be gained from a model that is highly
resolved in two dimensions, but that is not at all resolved in the third, and where structure
from the missing third dimension is mapped incorrectly onto the 2D plane.
-
-
-
3D Full Waveform Inversion: a Complex Recipe for Success
By L. SirgueWaveform inversion has been an established technique for more than two decades
(Lailly 1983; Tarantola, 1984). Numerous 2D applications on synthetic and real data
have been published (Mora, 1988; Pratt et al., 1996, Operto et al., 2004, Sirgue and
Pratt, 2004). In particular, studies related to the influence of subsurface angle
illumination (Jannane et al., 1989, Sun and McMechan, 1992; Pratt et al., 1996;
Sirgue and Pratt , 2004) have demonstrated the importance of wide-angle/large offset
surface seismic data.
On the other hand, the highly non-linear nature of waveform inversion begs for the
need of low frequencies in the seismic data. Multi-scale strategies in either time or
frequency domains (Bunks et al, 1995; Forgues et al., 1998) have shown that inverting
initially for the low-end of the frequency spectrum may be an efficient approach for
the mitigation of non-linearity.
This need for low frequencies however may not be dissociated from the accuracy of
the starting model. As a result, inaccuracy of the starting model will results in more
demanding requirements in terms of low frequencies (Sirgue and Pratt, 2002). The
combination of wide-angle illumination, low frequencies and starting model hence
constitute the primary ingredients of a successful inversion (Sirgue, 2006). Each of
these ingredients plays a key role and interacts with one another in a complex
relationship that will depend on the geophysical problem that one is trying to solve.
More recently, the first examples of 3D waveform inversion were shown (Sirgue et
al., 2007; Ben-Hadj-Ali et al., 2007). While the extension of waveform inversion to
3D problems will not fundamentally change the importance of wide-angle data,
starting model and low frequencies, additional aspects will need to be assessed such
as the impact of the azimuthal coverage.
-
-
-
Velocity inversion based on one-way migration and semblance maximization versus fullwaveform inversion
Authors R. Soubaras and B. GratacosIn this paper, we provide a theoretical comparison between twomethods of velocity estimation:
full-waveform inversion based on the minimization between the recorded data and the reconstructed
data based on a reflectivity model and a velocity model, and stack energy maximization
methods where the energy of a one-way migration is maximized. We first analyze the process of
migration and least-squares migration, then show that minimizing the data misfit is equivalent
to first order to energy maximization of a migration, when an appropriated weighting matrix is
used. We then obtain, by using 1D migration, a first order approximation of this weighting matrix.
Synthetic examples illustrates this derivation. The influence of using one-way propagation
instead of two-way is then discussed.
-
-
-
Velocity analysis in the data domain – overview and prospects
Authors T. van Leeuwen and W. A. MulderAutomatic versions of migration velocity analysis provide velocity background models that
can serve as a starting point for migration or least-squares inversion. When dealing with
primary reflections, current methods perform well. They tend to break down, however, when
strong multiples are present. This is due to the single scattering approximation that underlies
most migration algorithms. When multiples are interpreted as primaries, they may end up at
the wrong depth with the wrong apparent velocity. In the data-domain, it should in principle
be possible to match observed with predicted multiples and use them for velocity estimation.
We present a method for performing the velocity analysis in the data domain and include tests
on multiple-free synthetic data and on real data, using the convolutional model to model the
data. We also outline ideas on how the method is expected to behave in the presence of
multiples.
-
-
-
On the gradient generated wave-paths in differential semblance velocity analysis
By P. ShenDifferential semblance velocity analysis (“DSVA”, (Symes, 1986)) estimates velocity models
from waveform data, by means of prestack migration and its linearized adjoint state process.
Several authors have presented DSVA in its explicit differential forms for laterally
heterogeneous velocity models using various methods of prestack migration (Symes and
Versteeg, 1993; Kern and Symes, 1994; Chauris and Nobel, 2001;Mulder and ten Kroode,
2002). Shen et al. (2003) presented a version DSVA based on an objective of focusing the
image in subsurface offset. In this approach the wrong velocity is penalized by simply
multiplication of subsurface offset to the image volume. The name of “differential
semblance” is justified through its subsurface angle substitute (Shen, 2004), where an explicit
differential with respect to angle is formulated.
-
-
-
High-resolution imaging of basin-bounding normal faults in the Southern Apennines seismic belt (Italy) by traveltime and frequency-domain full-waveform tomography
Authors L. Improta, S. Operto, C. Piromallo and L. ValorosoWe apply a two-step seismic imaging flow by combined first-arrival traveltime and
frequency-domain waveform tomographies to dense wide aperture data collected in the Val
d’Agri basin (southern Italy). A large wavelength Vp model determined by first-arrival
traveltime tomography is used as a starting model for waveform tomography. The multiscale
waveform tomography consisting of successive inversion of increasing frequencies allows to
progressively reconstruct the short wavelengths of the velocity model, providing valuable
information on the Quaternary basin and on range-bounding normal-faulting systems.
-
-
-
Quantitative imaging of the Permo-Mesozoic complex and its basement by frequency domain waveform tomography of wide-aperture seismic data from the Polish Basin
Authors M. Malinowski and S. OpertoRecently we observe an increasing interest in acquisition of global-offset seismic data for
commercial prospecting in geologically complex areas, eg. in areas of basalt flows or thrust
belts (Operto et al. 2004, Colombo, 2005). The broad range of recorded offsets provides a
sufficient ray coverage for traveltime tomography and enhances the depth-migrated images by
using more energetic wide-angle reflections. Such data are also well suited for frequencydomain
full waveform inversion (FWI) – a method which was recently used for imaging
complex structures (Ravaut et al., 2004).
In this study we present the workflow and results of 2-D frequency domain waveform
tomography (WT) applied to the global-offset seismic data acquired in central Poland along a
50-km long profile during GRUNDY 2003 experiment (Malinowski et al. 2007). The WT
method allows full exploitation of the wide-aperture content of these data and produces in a
semi-automatic way both the detailed P-wave velocity model and the structural image (i.e.,
perturbations in respect to the starting model).
-
-
-
2D Full wave form inversion in time-lapse mode: CO2 quantification at Sleipner
Authors A. Gosselet and S. SinghConventional processing of seismic time-lapse data is very valuable in gaining qualitative
insights into reservoir history. For example, this allows characterization of fluid front
displacements, identification of fluid migration pathway or detection of flow barriers and
compartments. Quantitative analyses are generally based on Amplitude Versus Offset (AVO)
and attribute generation techniques. However, AVO is fundamentally a 1D approach and is
generally applied using linear approximations. Integrating time-lapse seismic into historymatching
is another way to obtain quantitative conclusions. Nevertheless, in this case, seismic
forward modeling is often based on a simple 1D convolution model and a too large
computational load precludes any proper optimization loop. To overcome such limitations, we
propose to apply 2D elastic full waveform inversion to time-lapse seismic data. The method is
computer intensive but allows modeling the different propagation modes (reflections, wide
angles, multiples, converted) to achieve a rigorous non-linear inversion of the seismic data.
The approach is applied to reflection time-lapse data from monitoring surveys of the Sleipner
CO2 injection site, North Sea. CO2 separated from methane is injected into the Utsira Sands,
a deep saline aquifer. Inverted P-wave velocity variations are related to CO2 saturation using
Gassmann theory. Since gas injection into a water bearing formation is a drainage process,
saturation is likely to be patchy. Consequently, we also used the patchy Vp-saturation
relationship to determine the maximum possible saturations. Investigations about the level of
patchiness, with respect to the frequency bandwidth used for the inversion, is required to
determine the most likely CO2 saturation.
-
-
-
What initial velocity model do we need for full waveform inversion?
Authors H. Chauris, M. Noble and C. TaillandierIn the context of velocity model building, we examine if the velocity models obtained after
first-arrival traveltime tomography are accurate enough for subsequent full waveform
inversion of reflected energy. For that purpose, we test the quality of the velocity model
obtained by first-arrival traveltime tomography on the BP salt dome model. Several 1-D
inversions are conducted in two different zones. In the simplest zone corresponding to smooth
velocity models, the tomographic models are good enough for waveform inversion with
realistic frequency contents. In the complex part going through a salt body, one need very low
frequencies (starting at around 1 Hz) or a further refinement of the tomographic model.
-
-
-
Influence of acquisition parameters for 2D acoustic frequency-domain full-waveform inversion.
Authors C. Ravaut, M. Alerini, J. A. Haugen and B. ArntsenIn this paper, we studied and illustrated the influence of the acquisition parameters on the results of 2D
acoustic frequency domain full-waveform inversion. We considered two synthetic geological models: a
tilted layered blocks model and a complex salt dome model. In the first case, the inverse problem is quite
linear up to 5Hz frequency and full-waveform inversion gives well constrained models for more
industrial seismic acquisitions. In the salt dome context, the inverse problem is strongly non linear and
very low frequencies and large offsets are necessary for full-waveform inversion to reconstruct properly
the true velocity model. In this very complex case, dedicated acquisitions, like wide-offsets and low
frequency sources, need to be designed to ensure the success of the method
-
-
-
Comparison of acoustic full waveform tomography in the time - and frequency - domain
Authors A. Kurzmann, D. Köhn and T. BohlenFor better parameter estimation, both in active source and earthquakes seismology, we need to exploit the richness of full seismic waveforms. Full waveform tomography (FWT) is a powerful method to reach this goal. Although first implementations in the 1980's were conducted in the time-domain by Tarantola, the frequency-domain version of FWT developed in the 1990's by G. Pratt and coworkers has now emerged as an efficient imaging tool. The main advantage of the frequency-domain approach is the possibility of starting the inversion at low frequencies (large scale structures) and then moving to higher frequency compounds (smaller scale structures), thereby realizing a multiscale approach. The main advantage of the time-domain method is the efficient parallelization by domain decomposition leading to a significant speedup on parallel computers. In this study, we demonstrate the performance of our parallel acoustic time-domain code. We present the results for a very complex example - a random medium model. Last but not least, we compare our time-domain inversion results with the frequency-domain results calculated using the FULLWV code by G. Pratt et al.
-
-
-
Frequency-domain acoustic wave modeling using a hybrid direct-iterative solver based on a parallel domain decomposition method: a tool for 3D Full Waveform Inversion?
Authors F. Sourbier, A. Haidar, L. Giraud, S. Operto and J. VirieuxFrequency-domain full-waveform tomography has been extensively developed during last decade to build
high-resolution velocity models (Pratt, 2004). One advantage of the frequency domain is that inversion
of few frequencies are enough to build velocity models from wide-aperture acquisitions. Multi-source
frequency-domain wave modeling requires resolution of a large sparse system of linear equations with
multiple right-hand side (RHS). In 2D, the method of choice for solving this system relies on direct solver
because multi-RHS solutions can be efficiently computed once the matrix was LU factorized. In 3D or
for very large 2D problems, the memory complexity of direct solvers precludes applications involving
hundred millions of unknowns. To overcome this limitation, we investigate a domain decomposition
method based on a Schur complement approach for 2D/3D frequency-domain acoustic wave modeling.
The method relies on a hybrid direct-iterative solver. Direct solver is applied to sparse matrices assembled
on each sub-domain, hence, mitigating the memory complexity of the overall simulation. Iterative solver
based on a preconditioned Krylov method is used to solve the interface nodes between adjacent domains.
Drawback of the hybrid approach is that the time complexity of the iterative part linearly increases with
the number of RHS. In the following, we introduce the domain decomposition method before illustrating
its potentialities with 2D and 3D simulations.
-
-
-
Efficient iterative solution of the 3D acoustic wave equation
By A. UmplebyOutlines the iterative solver that we have developed for the 3D visco-acoustic wave equation
in the frequency-domain, explains how the system of equations is preconditioned and how the
method is parallelised for tomography, and presents hardware benchmarks for the method.
-