- Home
- Conferences
- Conference Proceedings
- Conferences
77th EAGE Conference and Exhibition - Workshops
- Conference date: June 1-4, 2015
- Location: Madrid, Spain
- Published: 01 June 2015
81 - 100 of 128 results
-
-
Near-surface Characterization by Full Elastic Wave-field Inversion Based on the Scattering Integral Equation
Authors G. Rizzuti and A. GisolfSummaryThe near-surface can leave a strong imprint on seismic data. In order to be able to image deeper targets, we propose a full wave-field inversion approach to be followed by redatuming to the specific target level. The inversion scheme presented in this paper estimates iteratively the elastic properties of the model by trying to predict, at each iteration, one order of scattering of the data. We demonstrate the effectiveness of this idea on near-surface synthetic examples.
-
-
-
Semi-discrete Matrix Free Formulation of 3D Full Waveform Elastic Modeling and Inversion
SummaryIn this work we derive the adjoint equations for the SEM model, which enables the incorporation of a generic data misfit functional. The key contributions of the study are a scalable computational strategy and a novel approach for matrix-free evaluation of the adjoint wavefield. The formulation enables computation of the adjoint wavefield, and consequently the gradient (or proximal function of which) of the objective function, while avoiding storage of the forward displacement field in memory. This algorithmic setup grants scalability in both weak and strong scaling tests.
-
-
-
Simultaneous GPR Reconstruction of Electrical Conductivity and Permittivity
Authors H. Pinard, M. Dietrich, S. Garambois, F. Lavoué, L. Métivier and J.M. VirieuxSummaryGround-penetrating radar (GPR) is a non-invasive prospecting technique based on the electromagnetic waves sampling of the near surface. Building quantitative images through these waves requires the reconstruction of both
electrical permittivity and conductivity. This multi-parameter reconstruction is performed through the minimization of a misfit function measuring the discrepancy between observed and synthetic data. The minimization is achieved with a local descent method based on the Newton equation. Both the gradient and the product of the
Hessian matrix with a model vector are necessary for avoiding any trade-off between parameter classes, especially when high contrasts are encountered by electromagnetic waves. This presentation is devoted to the design
of these two key ingredients needed when updating the model, based on efficient first- and second-order adjoint methods. We formulate the problem in the frequency domain and we show that we need two forward modeling for the gradient and two additional forward modeling for the product of the Hessian matrix and a model vector. Our
formulation is such that these quantities are obtained through solution fields, regardless of the numerical scheme used to obtain them.
-
-
-
Application of Crosshole GPR Full-Waveform Inversion to Experimental Data - Overview of Recent Developments and Possible Pitfalls
Authors A. Klotzsche, J. van der Kruk and H. VereeckenSummaryMany theoretical papers show the potential of the crosshole GPR full-waveform inversion, however to apply the full-waveform inversion to experimental data is still challenging. We have applied the crosshole GPR full-waveform inversion to many different datasets in Germany, Switzerland and USA. Here, we will give an overview of all the important steps and developments of the full-waveform inversion during the last years that are necessary for a successful application to experimental GPR data. Thereby, we will discuss important pre-processing steps including time zero correction, source wavelet estimation and different inversion strategies.
-
-
-
Near-surface 2D SH-waveform Inversion - A Case Study from Slovakia
More LessSummaryWe apply elastic time-domain 2D full waveform inversion (FWI) to a near surface SH-wave data set recorded in the village Cachtice in northwestern Slovakia. Aim of the seismic survey was the detection of possible basement structures. The recorded data, consisting of 42 shots and receivers distributed along a 21 m long profile, covers a wide frequency range from 5 Hz to 100 Hz with high signal-to-noise ratio well suited for FWI. In addition to the direct SH/Love-wave a distinct refracted wave is visible. The trend of the first arrivals suggests an approximately 1D velocity gradient in the subsurface. Therefore a 1D initial model for the FWI is estimated by the Wiechert-Herglotz method. For the initial density model a similar gradient medium is assumed. The 2D FWI is based on the global correlation norm as objective function in combination with sequential frequency inversion. The final FWI S-wave velocity and density models reveal a heterogeneous underground with a prominent low-velocity weathering layer and isolated highvelocity anomalies. The structures of the density model correlate well with the velocity model.
-
-
-
Applications of 2-D Elastic Full Waveform Inversion to Shallow Seismic Rayleigh Waves
Authors T. Bohlen, T. Forbriger, L. Groos, M. Schäfer, C. Wolf and M. BinnigSummaryShallow-seismic Rayleigh waves are attractive for geotechnical site investigations. They exhibit a high signal to noise ratio in field data recordings and have a high sensitivity to the S-wave velocity, an important lithological and geotechnical parameter to characterize the very shallow subsurface. Conventional inversion methods assume local 1D models and invert the dispersion curves of Rayleigh waves. In this work we present a workflow for 2-D elastic FWI that allows for the reconstruction of small-scale lateral velocity models. We apply a geometrical correction and an a priori estimation of the quality factors which used as passive modelling parameters during the multi-parameter FWI. We invert for the seismic velocities and densities, but only the S-wave velocitiy seems to be reliable due to its high sensitivity. Different field data examples demonstrate that realistic small-scale lateral S-wave velocities can be inferred by this approach.
-
-
-
Full Waveform Inversion for Shallow Hazard Identification on a Narrow Azimuth Dataset
Authors D. Bright, C.E. Jones, Y. Belhassen, R. Brasil, H. Macintyre and C. FrançaSummarySite survey seismic is usually acquired as 2D seismic lines with specific acquisition configurations for imaging close to the seabed at high frequencies. The resulting images have excellent detail and they are then interpreted and potential drilling hazards identified largely from the stack amplitudes. However we lack other methods by which to constrain the interpretation of high amplitudes, such as being able to map their extent in 3D or use other seismic attributes derived from pre-stack data.
In many prospects around the world, shallow gas is a known issue for drilling and as an experiment we ran a full waveform inversion study to test if we could detect velocity anomalies caused by these accumulations.
To isolate anomalous velocities, the FWI velocity model was smoothed in a horizon-consistent manner and the smoothed model subtracted to give an “anomaly volume”. This was then compared with conventional site survey seismic data. We also verified the FWI velocity model by remigrating the 3D seismic data using the original tomographic model, the smoothed FWI model and the full FWI model.
-
-
-
Application of 2D (visco)-elastic Rayleigh Waveform Inversion to Ultrasonic Data from the Porta Nigra in Trier (Germany)
More LessSummaryBeside geophysical applications from the near-surface to global scale, seismic full waveform inversion (FWI) can be applied to ultrasonic data on the centimeter and decimeter scale for non-destructive testing (NDT) of pavements, facades, plaster, sculptures and load-bearing structures like pillars. Classical NDT approaches are based on the inversion of body-wave travel-times to deduce P-wave velocity models. In contrast, surface waves (Rayleigh or Love waves) are well suited to quantify superficial alterations of material properties, e.g. due to weathering. In this paper we demonstrate the potential of 2D Rayleigh waveform inversion on the ultrasonic scale using a very low coverage acquisition geometry consisting of 1 shot and a few dozen receiver positions. For a 2D elastic FWI with a passive visco-elastic modelling approach the resolution is illustrated using a ultrasonic field data example from the weathered facade of the Porta Nigra, a large Roman city gate from the 2nd century AD, in Trier (Germany).
-
-
-
The Sensitivity of a Seismic Sweep to the Near Surface
By W. KimmanSummarySurface consistent deconvolution compensates for phase perturbations in the near surface. Some of these perturbations are the result of wave propagation near the source or receiver. Using a Born based approach in a highly idealized representation of the near surface I derive analytical sensitivity kernels based on the instantaneous phase of a seismic sweep. This allows predicting changes in instantaneous phase due to velocity perturbations in the very near surface. One interesting observation is the relative large effect that small perturbations near the source/receiver can have on the instantaneous phase. This calls for further study on the relation between surface consistent source/receiver terms and more realistic structures with large velocity contrasts.
-
-
-
Early-arrival FWI - Potential for the Near-surface
More LessSummaryI will discuss an exemplary application of early-arrival FWI to near-surface data from a highly complex environment and the benefit of using the high-resolved FWI model for reflection imaging. To further assess the potential of the method I will summarize some of the experience from early-arrival FWI and derive scaling relationships that allow for constraining the required frequencies. I will discuss the limitations of acoustic FWI and some of the problems and solutions towards fully elastic FWI.
-
-
-
Noise-based Seismic Tomography at the Valhall Oil Field with Using Scholte and Love Waves
Authors G. Tomar, N.M. Shapiro, A. Mordret, S. Singh and J.P. MontagnerSummaryWe present here Scholte and Love wave phase velocity tomography at the Valhall Oil Field using ambient noise recorded by a network of 3D multi-component ocean bottom cable. We have cross-correlated 6.5 hours of continuous recording of noise between vertical-vertical (ZZ), radial-radial (RR), and transverse-transverse (TT) components. After applying an F-K filter, we were able to extract the first overtone of Scholte waves at Valhall from the RR cross-correlations. We then used the filtered overtone waveforms to measure inter-station frequency-dependent phase time delays and constructed 2D phase-velocity maps with the Eikonal tomography method.
Furthermore, we compute average dispersion curves for Scholte and Love waves by combining information from more than 10 millions of individual cross-correlations. We use the Neighbourhood algorithm to invert jointly these dispersion curves and to obtain an average 1D anisotropic model of the Valhall overburden down to depths of ~1 km. We find a significant radial anisotropy at depths below 600 m. This average 1D model is used as a reference model for the 3D inversion.
-
-
-
A Cascaded Approach to Surface Wave Noise Attenuation
By P.J. BilsbySummaryThe removal of source-generated coherent energy propagating in the near-surface continues to be a fundamental step in land surface seismic data processing and one of key importance for subsequent reservoir characterisation. Such noise overlays and obscures a significant portion, in time and space, of the reflection signal and requires effective removal to maximise the value delivered to the reservoir geologist, as well as the perceived value of onshore seismic itself. Common strategies for the attenuation of this complex coherent noise have historically included both acquisition and data processing techniques, with the former using large areal source and receiver arrays and the latter using aggressive velocity filtering, both providing limited effectiveness. We describe a new systematic, cascaded, approach for the progressive attenuation of both the complex direct-arrival and scattered surface wave energy prevalent in land seismic data.
-
-
-
Understanding the Impact of Karst on Seismic Wave Propagation - A Multi-method Geophysical Study
SummaryKarstified areas are known to be difficult ground for seismic exploration. We conducted a combined numerical-modeling and field-experiment study with the objectives to study the impact of karst on seismic wave propagation and to advance geophysical characterization of karst with seismic as well as non-seismic methods (electric and electromagnetic techniques). Finite-difference simulations using models with realistic topography illustrate the pronounced impact of topographic variations in high-velocity carbonate-bedrock environments on the scattered surface/guided waves. Wavefield complexities such as strong lateral changes in the strength of surface/guided waves, which were observed in a Vibroseis gather from a karst terrain in the Middle East, were also evident in our data recorded in Switzerland. In the latter case, amplifications of surface/guided waves could be correlated with low-velocity zones, which are probably due to more intensively karstified zones. Our study demonstrates that because of the strong heterogeneity of karst terrains, dense sampling is required to properly comprehend and disentangle the observed wavefield. Furthermore, we observed in our field study that the electrical-resistivity models correlate more closely with the mapped lithology, whereas karstification seems to more strongly affect the P-wave velocity models.
-
-
-
Quantitative Integration of Measurements in Near Surface Characterization
By M. MantovaniSummaryA valid support to seismic in shallow anomaly detection is offered by additional geophysical measurements such as potential fields or electromagnetic induction soundings. In fact, gravity and electromagnetics (EM) measure complementary properties and, even though they do not sense any quantity directly related to P-velocity, they can be successfully used to improve the statics solution. Since the different geophysical domains are sensitive – to a large extent – to the same geometry structure of subsurface rock formations, there exists a structural link between the properties measured by the various methods, which are supposed to sense consistent anomaly shapes.
A high confidence solution for the near surface is therefore derived through a Simultaneous Joint Inversion (SJI) process ( De Stefano and Colombo, 2007 ). Within this process, seismic and non-seismic data are inverted simultaneously together with the structural (or petrophysical) link established between the corresponding physical properties (velocity, density, resistivity, etc.). The solution explains seismic as well as plenty of logging information other than the simple sonic curve. Moreover, EM and potential field data are also explained by the output model simultaneously with seismic and borehole data. Thus, each of the measurements complements the others, as each is affected by different and uncorrelated noise.
-
-
-
Mitigating Uncertainties in Towed Streamer Acquisition and Imaging by Survey Planning
More LessSummaryUncertainties in seismic images or reservoir characterisation can very often been associated with lack of resolution, illumination problems, or the inability to invert for accurate velocity models. Uncertainties may also be caused by general data quality issues like noise content as well as acquisition and processing footprints. Provided the geophysical challenges are well understood, such uncertainties can be mitigated already in the planning phase of a seismic acquisition project. Geophysical survey planning typically considers requirements for, e.g., temporal and spatial resolution, illumination at target level and also looks into the suppression of incoherent and coherent noise (e.g., multiple energy). Typically, the geophysical objectives and quality requirements have to be traded against survey cost and may be constrained by operational aspects.
Combining state of the art towed streamer acquisition technology with advanced processing and imaging methods enables us to mitigate some of these uncertainties. This paper discusses issues and solutions related to illumination, footprints, and resolution. Also the benefit from utilising seismic wave types in addition to primary reflections and other geophysical methods is addressed.
-
-
-
Understanding and Analysing Seismic Images - Insight through Appropriate Modelling
Authors I. Lecomte, P. Lubrano-Lavadera and D.W. SchmidSummarySeismic images are often provided to interpreters as the only element to work with, besides well logs. However, the possible geological complexity of both overburden and reservoir, as well as survey, wavelet, and processing are all parameters affecting seismic images. An ideal seismic modelling would be using full-waveform methods to generate complete synthetic seismic records, then process the latter in the same manner as the observed data. But it is a complex and time-consuming task, impossible at production and routine levels. Interpreters often use the simplest method of all, i.e., “trace modelling” or “1D convolution”, but this corresponds to strong, often forgotten, model assumptions: homogeneous horizontal layers. Using a 3D-spatial ray-based prestack convolutional method allows instead accounting for many of the above-mentioned effects, at a similar cost than 1D convolution, but in a flexible and more complete manner, permitting sensitivity tests on various parameters. This would help interpreters better understanding what seismic can/cannot see and assessing their results in each specific case.
-
-
-
5D Local Angle Domain Gathers as an Ideal Representation for Directivity Driven Imaging
More LessSummaryIn seismic imaging, 5D recorded seismic data is migrated to the depth domain, forming a much lower dimensionality, the extra dimensions being reduced by integration. As a consequence much information is irretrievably lost, information that could be of use in characterizing with minimal uncertainty the subsurface geological features. In order to retain the full richness of the collected data in a form which can be of use in achieving an unambiguous interpretation of the geology of the subsurface, an approach is presented here that, using migration, maps data from the surface acquisition domain to that of the physically derived subsurface 7D domain. The gathers thus derived can be considered to be complete representations of the scattering event and as such can be processed subsequent to migration, to accentuate any aspect required. We have demonstrated the directionality derived images among many other possible applications.
-
-
-
Seismic Imaging in and around Salt Bodies - Problems and Pitfalls Update
Authors I.F. Jones and I. DavisonSummarySalt movement often results in steeply-dipping complex structures, which pose significant challenges for model building and migration. In recent years, advances in seismic imaging algorithms have permitted imaging of steep structures by exploiting the two-way wave equation via the introduction of reverse time migration (RTM). With such imaging algorithms, double bounces and turning wave reflections can be imaged, thereby enabling the imaging of vertical and overturned salt flanks. However, despite advances in the migration algorithms, the derivation of a suitable earth model incorporating the anisotropic behaviour of the velocity field remains a significant challenge, requiring tight integration of geological interpretation, and geophysical skills.
A major contributing factor to the successful execution of a complex salt imaging project, is the understanding of the many and varied pitfalls involved at every stage of the process. Here we describe and discuss some of these issues, building on the observations made at the predecessor to this workshop which took place at the EAGE in London in 2003.
The various aspects described in this review paper all contribute to the overall uncertainty associated with creating and understanding images of complex structures.
-
-
-
Imaged Depth Versus True Depth
More LessSummaryThe use of the term ‘velocity’ to describe both the true propagation velocity and the modelling “velocity” used in processing (what is here termed ‘pro-velocity’) has given rise to a damaging level of confusion whereby a great majority of geoscientists take the depth image to be a good representation of depth in the subsurface. The relationship between the pro velocity and true velocity is unsystematic. It essentially amounts to a dichotomy between the two quantities. This form of relationship arises as a result of a number of fundamental factors. The depth image could therefore contain considerable structural distortion relative to the actual subsurface. The misconception that anisotropic depth imaging produces true depthing has been behind the significant decline in the use of the solidly sound approach for obtaining an optimal representation of the subsurface, namely: to convert the depth image vertically to time and to carry out a proper vertical depth conversion using true velocities. The application of this approach to practically every PreSDM process is strongly recommended.
-
-
-
Master the Uncertainty with Bayesian Approach -Case Study, Multi-Azimuth Depth Imaging Deep Water, Offshore Brazil
More LessSummaryInterpretation in the depth domain is controversial. Sometimes, the problem comes with constantly changing models. Geologists have found that structures which have been focused come and go along with the model update.
One of these tough imaging challenges occurs when a carbonate feature lies above a salt feature.
Overcoming this carbonate feature is a good challenge to exploration today due to the high variation in velocity and the homogeneity of anisotropic character.
Most of the current depth imaging processes faces uncertainty with an arbitrary and variables. This judgement of adequate volume and pattern of distribution remain controversial. With limited well information in the area, the depth model from the same seismic data could be non-unique. In consequence this uncertainty of the structure causes problems in deep water exploration.
Repsol intelligently designed its proprietary multi-Azimuth acquisition in Campos basin, Offset Brazil. By using an advanced imaging study combining technologies of advanced grid based tomography and inversion tools with structure control and multi-discipline joint inversion processes, we achieved the goal of the imaging the structure with outstanding resolutions.
The imaging results are confirmed by the latest well marker and the accuracy of production structure both approach the limit of resolution by inversion theory. This project sets an outstanding benchmark for computing resource utilization and demonstrates the benefit of advanced migration technologies and well-planned acquisition patterns.
-