Sheng Xu

3D angle gathers from reverse time migration

Reverse time migration (RTM), based on directly solving the two-way wave equation, provides a natural way to deal with large lateral velocity variations and imposes no dip limitations on the seismic images. It has become a standard migration tool for subsalt imaging in the Gulf of Mexico. In addition, the adoption of wide-azimuth (WAZ) data acquisition provides better subsurface illumination and more azimuthal information, which significantly aids imaging beneath complex overburdens. Because WAZ data are richer in the propagation directions of the waves recorded, RTM and WAZ acquisition go hand-in-hand.

Antileakage Fourier transform for seismic data regularization in higher dimensions

Wide-azimuth seismic data sets are generally acquired more sparsely than narrow-azimuth seismic data sets. This brings new challenges to seismic data regularization algorithms, which aim to reconstruct seismic data for regularly sampled acquisition geometries from seismic data recorded from irregularly sampled acquisition geometries. The Fourier-basedseismicdataregularizationalgorithmfirstestimates the spatial frequency content on an irregularly sampled input grid. Then, it reconstructs the seismic data on any desired grid. Three main difficulties arise in this process: the “spectral leakage” problem, the accurate estimation of Fourier components, and the effective antialiasing scheme used inside the algorithm. The antileakage Fourier transform algorithm can overcome the spectral leakage problem and handles aliased data. To generalize it to higher dimensions, we propose an area weighting scheme to accurately estimate the Fourier components. However, the computational cost dramatically increases with the sampling dimensions. A windowed Fourier transform reduces the computational cost in high-dimension applications but causes undersampling in wavenumber domain and introduces some artifacts, known as Gibbs phenomena.As a solution, we propose a wavenumber domain oversampling inversion scheme. The robustness and effectiveness of the proposed algorithm are demonstrated with some applications to both synthetic and real data examples.

The benefit of TTI tomography for dual azimuth data in Gulf of Mexico

Over the past decade, the majority of deepwater blocks in the Gulf of Mexico have been covered multiple times with seismic data from narrow-azimuth, towed-streamer acquisition (NAZ). In complex subsalt areas, each NAZ dataset provides unique subsurface illumination benefits. Multiple-azimuth data are now frequently integrated to provide extended subsurface coverage and for better imaging of complex subsalt structures. Multiple-azimuth seismic data, with shot and receiver locations covering a large portion of the two dimensional surface, present a new challenge for deriving a single velocity model that satisfies both datasets. In exploration and development work in the deepwater Gulf of Mexico, there has been an increasing demand to incorporate anisotropy in prestack depth imaging workflows. Incorporating anisotropy improves image quality and well/seismic misties. While most pre-stack depth imaging involves vertical transverse isotropy (VTI) anisotropy, transverse isotropy with tilted symmetry axis (TTI) is generally overlooked. Shale layering near steeply-dipping salt flanks can cause TTI anisotropy issues. In such a case, ignoring the tilted symmetry of salt flank bedding causes image blurring and mispositioning of the salt flank structure. Velocity variation with azimuth is observed in an orthogonal dual-azimuth streamer dataset, as well as wide-azimuth data in the deepwater Gulf of Mexico. The paper presents a study to build a single TTI anisotropy model for pre-stack depth imaging of dual-azimuth data in the deepwater Gulf of Mexico to yield an anisotropy model that flattens gathers for all azimuths, as well as improves focusing and spatial positioning of steeply-dipping salt flanks.

3D Common Image Gathers From Reverse Time Migration

Wide azimuth acquisition (WAZ), together with reverse time migration (RTM), have greatly improved our capability to image the complex structures. To further take advantage of the advanced imaging techniques, we require RTM to output 3D common image gathers (CIGs) for different reflection angles and subsurface azimuth angles. In this paper, we introduce a straightforward method for computing 3D CIGs using the wavefields generated in RTM. We derive the RTM imaging condition for 3D CIGs based on true amplitude migration theory. The method adapts to both isotropic and VTI/TTI anisotropic migrations. The 3D CIGs retain the rich azimuth and reflection angle information carried in WAZ data. They can be used to build velocity models, invert anisotropy parameters and to analyze reservoir attributes. In the numerical section, we demonstrate our method using high resolution 3D CIGs on a synthetic dataset as well as on a WAZ dataset in the Gulf of Mexico.

Enhanced Tomography Resolution By a Fat Ray Technique

Summary We propose a tomography algorithm to enhance the resolution of 3D velocity model building. Our algorithm uses a fat ray instead of the conventional asymptotic ray. Compared to conventional ray tomography, the resulting tomography matrix is far less sparse. Thus, while the new matrix may still be ill-conditioned, much less model regularization is required to produce a well-posed system. In turn this should lead to a higher resolution tomographic update. A subsalt velocity analysis test on the Sigsbee 2a model shows promising result; we also apply the fat ray technique to well-tie tomography that estimates anisotropy parameters.

Anisotropy Estimation For Prestack Depth Imaging - A Tomographic Approach

In areas where well data (sonic logs, check shots et al) are available, we are able to get accurate vertical velocity by calibrating seismic migration velocity with check shot velocity. We present a method to accurately determine anisotropic parameters in transversely isotropic media without the weak anisotropy assumption. We test the algorithm on synthetic data and present a work flow to determine anisotropic parameters for prestack depth imaging in the Gulf of Mexico.

A model-based water-layer demultiple algorithm

This paper focuses on the attenuation of Water-Layer-Related Multiples (WLRMs or peg-leg multiples) which reflect at least once between the water bottom and the water surface. WLRMs are often the most dominant multiples in shallow-water seismic data. We propose a Model-based Water-layer Demultiple (MWD) algorithm to calculate the Greens functions of the Water-Layer Primary Reflections (WLPRs: Greens functions convolved with source signature) based on the known seabed and water-layer velocity model and then convolve them with the recorded data to predict the WLRMs. Combined with adaptive subtraction, MWD can effectively attenuate WLRMs. We apply MWD to field data from the Hibernia oilfield area which has a water depth of 70-90 m. The results show that while Surface-Related Multiple Elimination (SRME) by itself has limited success, MWD is effective in attacking WLRMs. Once the WLRMs have been removed by MWD, successive SRME can then be applied to predict and eliminate other types of surface-related multiples (SRMs).�The combination of MWD and SRME is demonstrated as an effective multiple attenuation package for shallow-water data and results in fewer residual multiples and better preserved pri- maries over tau-p gapped deconvolution. This, in turn, contributes to a more realistic velocity model and higher-quality images. from the auto-correlation, to predict WLRMs. DWD pre- dicts WLRMs with correct amplitudes. However, when the water-bottom is complex (and thus multi-arrivals of WLPRs present) the water-layer model derived from the time-domain

Migration Artifacts And Velocity Analysis

Kirchhoff depth migration is conventionally used for large 3D marine surveys. It has advantages for both fast track imaging and velocity model building. Conventional model building requires a depth-migrated common-image gather (CIG) (Al-Yahya, 1989) in which the reflectors are migrated multiple times using a multi-channel input dataset. If the prestack migration uses the correct velocity the seismic events will focus properly in the CIG regardless of any redundant axes (offset, reflection angle etc.). Otherwise, the seismic event will appear at a different depth. The difference in depths, the so-called residual moveout, provides information to update the velocity model. Generally, the moveouts are automatically picked from coherent seismic events in CIGs and this process provides the basis for an iterative tomography algorithm for complex velocity model building.

Aliasing In RTM 3D Angle Gathers

3D angle domain common image gathers (ADCIGs) from reverse time migration provide a powerful tool for imaging complex geological structures. For wide azimuth (WAZ) data processing, RTM 3D ADCIGs retain localized information indexed by subsurface azimuth and reflection angles, which improve tomographic velocity inversion for WAZ data, provide information on subsurface illumination and are ideal for azimuthal amplitude versus angle (AVA) and azimuthal anisotropy analysis for fractured reservoirs. For conventional WAZ data, the energy is well distributed in the vector (source-receiver) offset domain. However, coarse WAZ shot spacing produces far fewer recorded traces with actual reflection events from shallow depths than from great depths. This ensures that a relatively small number of shallow reflection events will be distributed across a wide range of reflection angles. Theoretical analysis shows that coarse shot spacing produces aliased shallow events, and the problem is more severe at small reflection angles. A second problem, observed at large reflection angles when migration velocity is incorrect, is due to inadequate angular spacing for deeper events. These sampling issues challenge our ability to produce high quality ADCIGs. We first analyze the aliasing problem and derive the angular sampling formula. Then we propose methods to reduce the under-sampling noise for RTM 3D ADCIGs at shallow events. We demonstrate the efficiency and robustness of our noise suppression algorithms using 2D and 3D examples. Finally, we discuss the angular under-sampling effects at large reflection angles.