|
Date
|
Group
|
Speakers
|
Summary
|
|
1-17
|
S-Team
|
Filippo and Kanu
|
Filippo: Connection of scattering principles: a visual and mathematical tour
Abstract:
Inverse scattering, Green's function reconstruction, focusing, imaging, and the optical theorem are subjects usually studied as separate problems in different research areas. We speculate that a physical connection exists between them because the equations that rule these scattering principles have a similar functional form. We first lead the reader through a visual explanation of the relationship between these principles, and then we present the mathematics that illustrates the link between the governing equations of these principles. Throughout this work, we describe the importance of the interaction between the causal and anti-causal Green's functions, and we provide maximum clarity and physical insight to emphasize the connection between these principles.
Kanu: I will briefly show seismic signals from repeating microseismic events from the Paradox Valley, CO. This is to illustrate the attributes of the signals especially the coda section of the signals from which we can make deterministic inference of the subsurface.
|
|
1-24
|
C-Team
|
Yong
|
Projected Hessian matrix for quasi-Newton waveform inversion
Abstract:
A Hessian matrix is difficult to directly compute because of its high computational cost and especially huge memory request. Therefore, applying Newton methods is barely possible in a FWI problem with a realistic size. Here, we modify the BFGS method and propose a projected Hessian matrix, which is constructed to iteratively approximate the true Hessian. The projected Hessian matrix has advantages of computational efficiency and a much smaller size, thereby making a quasi-Newton solution to FWI potentially promising.
|
|
1-31 (Shifted)
|
I-Team
|
Jeff
|
Automatic microseism event location by full-waveform triangulation
Abstract: Locating and characterizing microseismic events induced by hydraulic fracturing is quickly becoming a large subfield of geophysics because of increased interest in producing hydrocarbons from unconventional plays. Most methods for locating microseisms are direct extensions of concepts from earthquake seismology, and rely on methods that may not be appropriate for microseismic datasets due to their low signal-to-noise ratios. In this talk, we explore some alternative methods to locate microseisms using time-reversed acoustics and by exploiting the spatio-temporal autocorrelation of microseismic events. The methods that we develop are conceptually similar to those already in widespread use within industry, but do not rely on arrival time picking, and may allow us to locate microseismic sources that are well below the noise threshold in our data. Additionally, it appears that we may be able to perform automated velocity analysis for microseismic sources by exploiting wavefield semblance.
|
|
2-4
|
|
Dan Burns, ERL Executive Director
MIT Earth Resources Lab
|
Fractured reservoir characterization
and passive production monitoring – an overview of two reservoir
geophysics projects at MIT’s Earth Resources Lab
Abstract:
The ERL at MIT is an applied geophysics lab focused on integrated reservoir applications. Founded in 1982, the lab currently has about 25 PhD students and involves faculty from multiple departments. Areas of research include reservoir characterization and imaging, reservoir monitoring, and borehole science and rock physics. Integration across disciplines is the primary focus. Two particular projects will be highlighted. The first is the use of scattered seismic energy to characterize fractured reservoirs. Several methods, based on numerical modeling studies and field data testing, have been developed to estimate fracture orientation and density. Results are promising and are consistent with both borehole measurements (e.g. FMI and borehole breakout) and other seismic analysis methods (e.g., AVOA and shear wave splitting). The second project is a passive monitoring project of a producing field in the Middle East. GPS, InSAR, and passive seismic data were acquired over a period of almost 10 years. The data provide insight into the compaction of the reservoir and the reactivation of faults in the field, information that is helpful for reservoir management.
Venue: GC 281
Time: 2:00 – 3:00 PM
|
|
2-7
|
I-Team
|
Jeff
|
Automatic microseism event location by full-waveform triangulation
Abstract: Locating and characterizing microseismic events induced by hydraulic fracturing is quickly becoming a large subfield of geophysics because of increased interest in producing hydrocarbons from unconventional plays. Most methods for locating microseisms are direct extensions of concepts from earthquake seismology, and rely on methods that may not be appropriate for microseismic datasets due to their low signal-to-noise ratios. In this talk, we explore some alternative methods to locate microseisms using time-reversed acoustics and by exploiting the spatio-temporal autocorrelation of microseismic events. The methods that we develop are conceptually similar to those already in widespread use within industry, but do not rely on arrival time picking, and may allow us to locate microseismic sources that are well below the noise threshold in our data. Additionally, it appears that we may be able to perform automated velocity analysis for microseismic sources by exploiting wavefield semblance.
|
|
2-14
|
A-Team
|
Mamoru
|
Velocity analysis for TI media with quadratic velocity variation
Abstract: Lateral velocity variation with a scale smaller than spreadlength can significantly distort moveout parameters and stacked sections. To efficiently handle such lateral heterogeneity, I employ a so-called factorized-block model with quadratic velocity variation. I
first show that the error in the lateral quadratic variation amplifies residual moveout of the events at large depths. Synthetic tests confirm that the MVA algorithm reconstructs velocity and anisotropy parameters for TI media with quadratic velocity variations.
|
|
2-21
|
Presidents' Day
|
|
|
|
2-28
|
S-Team
|
Nori
|
Estimating shear wave velocity of a shallow zone in Japan by using seismic interferometry and KiK-net
Abstract: In Japan, there is a seismometer network (KiK-net), and it has about 700 stations. Each station has two receivers: one is on the ground surface and another is in the bottom of a borehole. The depth of the borehole is 100 m or more. These stations have recorded big earthquakes for more than 10 years. When I apply seismic interferometry to each station and each earthquake, I can estimate the velocity between two receivers. In this study, I use horizontal component, so I estimate the shear wave velocity. I apply the same procedure to all stations, and I obtain the velocity between two receivers all of Japan. By using 10 years data, I also obtain time-lapse change of shear-wave velocity.
|
|
3-7
|
C-Team
|
Simon
|
Seismic image flattening using vector shift fields
Abstract: Many flattening methods are limited to vertical shearing and
stretching of traces in an image. Because of this limitation, these methods may have difficulty flattening more complicated seismic images without significantly distorting features in the image. I propose a new image flattening method that flattens by solving for a vector shift field, rather than a scalar field of vertical shifts. The method
can still vertically shear traces, but it can also rotate portions of an image in order to flatten. Because it is not limited to vertical shearing, the method can flatten in ways that are more consistent with the actual geologic deformation that occurred, and it can flatten with less distortion of features in the image..
|
|
3-14
|
Spring Break
|
|
|
|
3-21
|
I-Team
|
Tongning
|
Image-domain waveform tomography with extended common-image-point gathers
Abstract: I will introduce an image-domain waveform tomography approach. The objective function is defined in such a way that the image inaccuracy is penalized. Therefore, minimizing the objective function optimizes the velocity model and improve the image quality simultaneously. The gradient is computed using the adjoint-state method, in a way similar to the gradient calculation in full waveform inversion. The input for the image-domain waveform tomography can be various type of common-image gathers, and I will focus on the implementation of using extended common-image-point gathers (CIPs). A synthetic steep salt flank example will be used to illustrate the workflow and the advantages of using CIPs for velocity model building in complex subsurface environments.
|
|
3-28
|
A-Team
|
Xiaoxiang
|
Ray-based gridded TTI tomography
Abstract I will present a 2D ray-based tomographic algorithm designed to iteratively update TTI parameters on square grids. The symmetry axis is set perpendicular to the interfaces that may be dipping or curved. To construct the Frechet matrix, which links the model update and the data misfit, the traveltime derivatives with respect to the parameters at each grid point are computed analytically along the trajectory. Finally, I will talk about a synthetic test on a model that contains a TTI syncline.
|
|
4-4
|
C-Team
|
Dave
|
Simple readable slides
Abstract The need for graphics in both printed manuscripts and oral presentations raises a question. Can we use a single graphic for both? For graphics with text, the answer is no. Text in graphics for manuscripts is always too small for presentations. A simple algorithm can be used to compute the correct font size when generating graphics and other material for presentations.
|
|
4-11
|
|
Francesco
|
CWP Meeting Rehearsal
|
|
4-18
|
|
|
|
|
4-25
|
|
|
|
|
5-2
|
|
Karl Schleicher, University of Texas
|
Processing open-access seismic data using open-source geophysical software
Abstract I have started to build a library of open-access seismic data and scripts that process the data using open source geophysical software. Research staff, especially at universities, will be able to recreate my processing or modify my scripts to test their ideas. This library will better organize data that is already publicly available.
I will present the processing of a 2D land line from Alaska using Seismic Unix through geometry initialization, initial data quality control, shot record velocity filtering, velocity interpretation, residual statics, and stack and compare the results to the commercial results produced in 1981.
|