 Full waveform inversion for reflected seismic data
EAGE, 2012
Sheng Xu | Feng Chen | Yu Zhang | Gilles Lambare
Summary
Full waveform inversion has been successful to build the shallow high-resolution velocity models. To achieve this requires the refraction information or the low frequencies in the reflection data. To relax the dependence of low frequency reflections, we revisit the gradient function in full waveform inversion and avoid the Born?s approximation which has been widely used in seismic inversion. We propose a new inversion method which is based on the accurate expression of the scattering waves. We provide some primitive applications with a synthetic example and a 2D real data example in the Gulf of Mexico, and demonstrate that the proposed method can invert long wavelength components with reflection information contained in conventional stream data.
|
FWI for streamer data in GOM
EAGE, 2012
Sabaresan Mothi-Venkatesan | Hongbo Bi
Summary
In this paper, we show that 3D time domain anisotropic (VTI) FWI may be used on a 3D Narrow Azimuth Towed Streamer (NATS) GOM deep water dataset. We observe a shallow gas pocket accompanied by radial faulting that could not be resolved using a high res tomography. This complex geology warrants more focused treatment using FWI to be able to accurately resolve the underlying velocity model. We demonstrate that the waveform-based tomography improves not only the resolution of the underlying velocity model, but also the position and quality of the complex structures present in the seismic section.
|
Geomechanical approach to resolve severe velocity variations and remove image distortions below rugose seafloor.
PGCE, 2012
Sergey Birdus | Alexey Artyomov | Li Li | Stephen Kefeng Xin
Summary
Seismic imaging and depth-velocity modeling below a seafloor with complex topography are always challenging. The problems are caused by two types of lateral velocity variations: (1) contrast between the water and sediments and (2) velocity variations within shallow sediments created by the variable water depth. Geomechanical effects contribute to velocity variations in shallow sediments. Variable water depth always creates some variations in overburden pressure and variable pressure changes seismic velocities. We propose and apply to synthetic and real data a new version of tomographic inversion that is focused on these types of geomechanical velocity anomalies associated with rugose seafloor.
|
Geomechanical approach to resolve severe velocity variations and remove image distortions below rugose seafloor.
SEG/EAGE Research Workshop, 2012
Sergey Birdus | Alexey Artyomov | Li Li | Stephen Kefeng Xin
Summary
Seismic imaging and depth-velocity modelling below a seafloor with complex topography are always challenging. Geomechanical effects contribute to velocity variations in shallow sediments. Variable water depth always creates some variations in overburden pressure and variable pressure changes seismic velocities. We propose a new version of tomographic inversion that is focused (constrained) to recover these types of geomechanical (compaction driven) velocity anomalies associated with rugose seafloor. Applied together with geomechanical modelling and standard seismic tomography, it helps to solve the strongest and most complicated water bottom related velocity anomalies. Synthetic and real data examples illustrate how this combination works.
|
High definition tomography brings velocities to light
SEG, 2012
Saverio Sioni | Patrice Guillaume | Gilles Lambare | Anthony Prescott | Xiaoming Zhang | Gregory Culianez | Jean-Philippe Montel
Summary
Velocity model building remains a crucial step in seismic depth imaging. A general drawback of conventional tomographic approaches is that the estimated velocity models do not conform enough to the structures. We present several applications of an innovative high-resolution tomography that inverts densely picked dip and residual move-out data to reveal detailed structurally conformable velocities. The application to the synthetic 2D Marmousi II dataset offers the possibility to carefully assess the method. It demonstrates its ability to produce structurally conformable velocity models with a level of detail that promotes velocity attributes as an aid to geological interpretation. As such it can offer an alternative to full waveform inversion for the interpretation of reflected waves. Finally we show an application to 3D marine dataset where obtained higher resolution velocity model results in improved focusing of migrated images and improved match to well velocities.
|
High-resolution tomography for paleo-canyons: a case study in Para-Maranhao Basin, Brazil
SEG, 2012
Guang Chen | Hao Shen
Summary
The shallow geological complexities in Para-Maranhao Basin, offshore Brazil, characterized by rugose topography and complex V-shaped paleo-canyons directly below the water bottom, brings challenges in velocity model building for prestack depth imaging, especially when using a global grid-based tomography. As a result, image distortions, such as non-geological depth undulation and amplitude variations, are often present after global tomography updates. These distortions can affect the fidelity of reservoir interpretation. Here we propose to use a high-resolution layer-constrained tomography workflow with structure-guided weighting scheme to build a velocity model with better accuracy. The migrated common image gathers are thus flatter, and the image distortions below complex near-surface overburdens are attenuated.
|
How Long Should the Sweep Be?
SEG, 2012
Julien Meunier | Thomas Bianchi
Summary
A sweep, a sinusoid with a continuously variable frequency, can be defined by its amplitude A(f) and its sweep rate Sr(f). Provided the sweep is long enough (longer than 5 or 6 seconds), the amplitude spectrum of the sweep at frequency f is proportional to A(f) and to the square root of Sr(f). Target-oriented sweep design consists in defining A(f) and Sr(f) to obtain the desired signal-to-noise ratio of the target reflection.
|
How to do 4D processing between Broadseis and conventional data
EAGE, 2012
Pierre Charrier | Thomas Rebert | Ziqin Yu
Summary
Processing the variable-depth streamer acquisition has recently become possible, with a new advanced algorithm, called joint deconvolution (Soubaras, 2010). In this particular acquisition, the receiver depth regularly increases with offset, which allows a wide diversity of receiver ghosts and so increases dramatically the possible frequency bandwidth, in both low & high frequencies side, from 2.5hz to source notch. Compared to conventional flat streamer data, processing the variable depth streamer data implies a major change: the receiver ghosts are rigorously taken into account, because they cannot be included in a wavelet like in conventional flat streamer processing. This breaks implicit assumptions made in various processes such as 4D cross-equalization. While most acquisitions in the future will certainly be realized with broadband techniques, the question of 4D matching between conventional and Broadseis data will remain during the intermediate period.
The variable receiver?s depth of Broadseis results in asymmetrical ray paths which are solved by the imaging process and by the proper summation of the up-going and down-going wave fields thanks to the joint deconvolution. Cross-equalization in a 4D process aims at solving issues related to differences in the vintages acquisition (time and amplitude de-striping?s) and positioning (4D binning), that is to say in the non migrated domain. This antagonism will be discussed here, together with the specific topics related to 4D processing: wavelet processing, time and amplitude de-striping?s, 4D binning, regularization, imaging and final matching. The dataset used for this comparison is a dual recording done by Shell in the deep offshore Gabon.
|
