보고서 내 다른 이미지

• 그림 The scope of this project.
• 그림 Multiscale strategy for elastic FWI with application to the overthrust model.
• 그림 Laplace-Fourier-domain waveform inversion.
• 그림 (a) The extended Marmousi2 velocity model and (b) the starting velocity model for the waveform inversions. The inverted velocity models using the objective function of (c) the conventional least-square norm and (d) the global correlation for EMSFWI of the marine streamer data.
• 그림 OBC survey image with 3 vessels.
• 그림 3D volume comparison between towed streamer data (left) and OBC data (right) from Lomond structure (BP, 1998).
• 그림 Process diagram for attaining the final conclusion.
• 그림 Real data example of a salt dome in the Gulf of Mexico.
• 그림 Shear wave analysis of OBS data.
• 그림 Schematic diagram of Reverse Time Migration.
• 그림 Velocity model of inclined 4 layer including isolated sediment at the seafloor.
• 그림 Synthetic seismogram which made by forward modeling using velocity model presented at the Fig. 3-2-1-2.
• 그림 Reverse-time migrated section of velocity model presented at the Fig. 3-2-1-2.
• 그림 Marmousi velocity model.
• 그림 Synthetic seismogram of Marmousi velocity model.
• 그림 Second derivative wavefield derived by back propagation of synthetic seismogram presented at the Fig. 3-2-1-6.
• 그림 Comparison of migration result using commercial software applied by post-stack Kirchhoff migration module (left) and prestack reverse-time migration module (right).
• 그림 A wavefield generated from isotropy acoustic wave equation at the 0.25 second.
• 그림 A wavefield generated from VTI acoustic wave equation at the 0.25 second.
• 그림 Input velocity and Thomsen
• 그림 Shot gather generated by FDM for VTI media given in Fig. 3-2-1-11.
• 그림 Virtual source calculation at different positions showing in Fig. 3-2-1-11.
• 그림 Partial derivative wavefield calculation at different locations showing in Fig. 3-2-1-11.
• 그림 Result of zero lag crosscorrelation between partial derivative wavefields respect to P wave velocity showing examples in Fig. 3-2-1-13 and shot gather data showing in Fig. 3-2-1-14.
• 그림 Result of zero lag crosscorrelation of partial derivative wavefields and shot gathers data.
• 그림 Result of zero lag crosscorrelation of virtual source and back propagating wavefields.
• 그림 Input velocities and Thomsen
• 그림 Examples of a shot gather from forward modeling as input for back propagating wavefield.
• 그림 Single image gathers as the result of inner product of virtual source and back propagation wavefields.
• 그림 Result of VTI RTM as after summation of 240 single image gathers.
• 그림 Workflow for the simultaneous-source FWI.
• 그림 (a) The Marmousi-2 model and (b) the initial velocity model.
• 그림 The residual seismograms calculated by simple subtraction
• 그림 Illustration of the problem of missing traces for the ESSFWI using the marine-streamer acquisition.
• 그림 The residual seismograms calculated by the correlation-based approach.
• 그림 (a) The new full-offset observed data obtained by combining the original observed data with the modeled data and (b) the residual seismograms calculated with the new algorithm.
• 표 Comparison of computational speed-up of the FWI with the number of cores.
• 그림 (a) Conventional node-based and (b) cell-based grid sets. Material properties are defined within cells for the cell-based grid.
• 그림 Wavefield snapshots at 0.8 s for a homogeneous anisotropic model, calculated by different wave equations
• 그림 Seismograms calculated by (a) the elastic wave equation and (b) the pure-acoustic wave equation when the receiver is located on the white line of Fig. 3-2-2-7.
• 그림 Wavefield snapshots according to different anisotropic parameter sets.
• 그림 Plane wavefield shot gather.
• 그림 Multitrace analysis windows commonly used in coherence calculations for (a) the crosscorrelation algorithm and (b) the semblance and eigenstructure algorithm.
• 그림 A schematic diagram showing crosscorrelation between two traces.
• 그림 A schematic diagram showing the steps used in semblance estimation of coherence.
• 그림 A schematic diagram showing the steps used in eigenstructure estimation of coherence.
• 그림 Sign convention for structural curvature in seismic data.
• 그림 (a) A quadratic surface with the normal, n, defined at point P (after Mai et al., 2009). The circle tangent to the surface whose radius, R, is minimum defines the magnitude of the maximum curvature, |kmax|≡1/Rmin (in blue). The planeperpendicular to that containing the previously defined blue circle will contain one whose radius is maximum, which defines the magnitude of the minimum curvature, |kmin|≡1/Rmax (in red). (b) Concept of  and  most principal curvatures.
• 표 List of 3D curvature attributes.
• 그림 The definition of 3D quadratic shapes expressed as a function of the principal curvatures.
• 그림 Shale resources play map of the North America.
• 그림 Schematic cross section illustrating platform-to-basin relationships.
• 그림 Synthetic seismogram using sonic and density log from Well #A in the Horn River Basin, Canada.
• 그림 Conventional 3D seismic amplitude volume acquired in Horn River Basin, Canada.
• 그림 Variance attribute (Semblance Coherence) of 3D seismic data from Horn River Basin, Canada.
• 그림 Schematic diagram showing calculation of reflector dip and azimuth using AASPI dip3d module.
• 그림 Time slices of 3D seismic volume from Horn River Basin (t = 1400 ms).
• 그림 Fracture lineament analysis using Petrel Ant-tracking algorithm.
• 그림 Fracture lineament analysis in Horn River Basin using AASPI generate roses module.
• 그림 Optimization of shale resources development in Horn River Basin, Canada.
• 그림 (a) First reflector-related internal multiple, (b) water bottom-related internal multiple, (c) surface-related multiple, (d) water layer multiples, (e) receiver side water layer multiple, (f) source and receiver side water layer multiple(peg-legs), (g) other surface-related multiples.
• 그림 (a) Long-period multiple, (b) short-period multiple.
• 그림 The ray path of multiple (yellow solid line) and primary reflection (green dashed line) on a linearly increasing velocity model.
• 그림 Seismic data containing various primaries and multiples, and (b) NMO-corrected data.
• 그림 The relationship between an event in the space-time and the corresponding event in the tau-p domain.
• 그림 (a) Three reflected events (a) in the space-time domain, and (b) those transformed by linear Radon transform.
• 그림 Water layer reverberations
• 그림 A signal containing a primary and multiples, (b) predicted multiples, and (c) the signal after removing the multiples shown in (a).
• 그림 (a) An ideal predictive deconvolution filter, and (b) practical predictive deconvolution filter.
• 그림 A velocity profile extracted from horizontally layered medium.
• 그림 The seismic data obtained from Fig. 3-2-4-10.
• 그림 The auto-correlation of the seismic data in the space-time domain.
• 그림 The seismic data after removing water layer reverberations by the space-time-domain predictive deconvolution.
• 그림 The seismic data after linear Radon transform
• 그림 The auto-correlation of the seismic data in the tau-p domain.
• 그림 The seismic data after removing multiples by the tau-p-domain predictive deconvolution.
• 그림 The muted seismic data.
• 그림 Multiple-free seismic data after inverse Radon transform.
• 그림 A common shot gather (CSG) of Jeju basin data.
• 그림 Workflows for data processing to remove multiples of Jeju basin data including predictive deconvolution in the space-time domain (red) and tau-p domain (blue), respectively.
• 그림 The band-pass filtered seismic data.
• 그림 A gain function.
• 그림 The seismic data after gain recovery.
• 그림 The auto-correlation of the seismic data in the space-time domain.
• 그림 Multiple removal using predictive deconvolution.
• 그림 A stack section.
• 그림 Linear Radon transformed data.
• 그림 The auto-correlation of the seismic data in the tau-p domain.
• 그림 The predictive deconvolved data in the tau-p domain and (b) the data after muting.
• 그림 A stack section.
• 그림 Point As are on the wavefield (a) of closed space and (b) under the surface S
• 그림 Ray paths (top) and seismograms (bottom).
• 그림 The comparisons of traces extracted from modeled (red dashed line) and extrapolated data (black solid line).
• 그림 Ray paths (top) and seismograms (bottom).
• 그림 The comparisons of traces extracted from modeled (red dashed line) and extrapolated data (black solid line).
• 그림 Two layered horizontal velocity having 1500 m/s water layer and ray paths.
• 그림 The synthesized observed data.
• 그림 The ray paths indicating the forward extrapolation and (b) the results of the forward extrapolation.
• 그림 (a) The ray paths indicating the second forward extrapolation, and (b) the results of the second forward extrapolation.
• 그림 The workflow of designing the Wiener filter.
• 그림 A matching filter to match the amplitudes of a trace on zero offset.
• 그림 The multiple-removed observed data.
• 그림 The comparisons of traces extracted from (a) the observed data, (b) the predicted multiples, and (c) the multiple-free data.
• 그림 (a) The ray paths indicating the inverse extrapolation, and (b) the results of the inverse extrapolation.
• 그림 The comparisons of traces extracted from (a) the inverse-extrapolation data (b) the forward-extrapolation data and (c) the multiple-free data.
• 그림 The multiple-removed observed data.
• 그림 Structure of a GPU card.
• 그림 Grid, Block, and Thread.
• 그림 Tesla K40.
• 표 Tesla K40.
• 그림 GeForce GTX Titan Black.
• 표 GeForce GTX Titan Black.
• 그림 Firepro D700.
• 표 Firepro D700.
• 그림 Firepro W9000.
• 표 Firepro W9000.
• 그림 CUDA Hardware Architecture.
• 그림 CUDA Memory Architecture.
• 표 OpenCL procedure & API.
• 표 CUDA & OpenCL terms.
• 그림 Kernel space.
• 그림 OpenCL device structure.
• 그림 CUDA vector addition.
• 그림 Kepler Streaming Multiprocessor (SMX).
• 그림 Kepler GPU memory structure.
• 그림 GPU memory optimization.
• 그림 Data pre-loading.
• 그림 Wave propagation modeling algorithm.
• 그림 3D finite difference stencil.
• 그림 Speed ups obtained by GPUs.
• 그림 Shared memory model.
• 그림 Speed ups obtained by the shared memory.
• 그림 The pageable memory and the page-locked (pinned) memory.
• 그림 Speed ups obtained by the shared memory and the pinned memory.
• 그림 Overlapping processes using CUDA Streams.
• 그림 Speed ups obtained by the CUDA Stream.
• 그림 Speed ups obtained by the shared memory without memory copy.
• 그림 Cluster system.
• 그림 Seismic survey line in Jeju basin, northeast sea of Korea.
• 그림 2D marine seismic survey equipment and conceptual diagram.
• 그림 Seismic data quality control in eSQC-Pro system
• 그림 Mimetic diagram on air-gun array configuration.
• 그림 Offset layout diagram of geophysical survey equipment installed in Tamhae-2.
• 그림 Marine seismic data acquisition diagram using the Tamhae-2.
• 표 Seismic exploration data acquisition parameters.
• 표 The kinds of seismic survey and main data acquisition parameters.
• 표 Geodetic parameters used in data acquisition.
• 그림 Source-side ghost and receiver-side ghost.
• 그림 Bird depth values used for broadband seismic survey.
• 그림 Streamer configuration layout applied to broadband seismic survey.
• 그림 Survey line of research area.
• 그림 Seismic survey track chart in Jeju basin, northeast sea of Korea.
• 그림 Flow chart for data processing standardization.
• 그림 Entering the seismic recording data into data processing system.
• 그림 Entering geometric information and defining formation.
• 그림 Entering acquisition parameters into ‘Setup’ and ‘Source’ window.
• 그림 Setting the offset bin increment as survey equipment array.
• 그림 (a) Checking geometry information entered in ‘TraceQC’ window, (b) The graph on CDP fold and statistical analysis.
• 그림 Seismic shot gather and trace header information before and after entering geometry information database.
• 그림 (a) Amplitude spectrum analysis in shot gather, (b) Amplitude spectrum analysis after removing low frequency noise, (c) Shot gather before and after applying bandpass filter.
• 그림 (a) Trace spikes attribute graph (b) Trace average energy attribute graph (c) Shot gather containing faulty traces.
• 그림 Shot gather before and after applying true amplitude correction.
• 그림 (a) Predictive deconvolution parameter test (b) Shot gather before and after applying predictive deconvolution (c) Stack section before and after applying predictive deconvolution.
• 그림 Seismic survey track chart in research area.
• 표 Data acquisition information on each survey line.
• 그림 Seismic data processing flow.
• 그림 Entering geometry information into ‘TraceQC’ window.
• 그림 The graphs on CDP fold and statistical analysis.
• 그림 Near-trace gather before and after applying shot delay correction.
• 그림 (a) Amplitude spectrum analysis for bandpass filter (b) Shot gather before and after applying bandpass filter.
• 그림 (a) The graphs of race average energy and standard deviation attribute on average frequency, (b) Shot gather before and after removing faulty traces.
• 그림 Shot gather before and after applying true amplitude correction.
• 그림 (a) Predictive deconvolution parameter test, (b) Shot gather before and after applying predictive deconvolution.
• 그림 (a) Velocity analysis panel (b) Velocity distribution in stack section.
• 그림 Stack section after completing data processing.
• 그림 Microg-LaCoste shipboard gravitymeter.
• 표 Calibration information for S-118 shipboard gravitymeter.
• 표 Data format of S-118 shipboard gravitymeter.
• 그림 Flow chart for freeair anomaly.
• 그림 Editing raw gravity data.
• 그림 Heights for meter drift correction.
• 그림 Preliminary freeair anomaly.
• 그림 Filtering preliminary freeair anomaly using moving average method.
• 그림 Filtered freeair anomaly.
• 그림 Data of each track for QC.s.
• 그림 Edited data by each track line QC (red lines).
• 그림 Altimeter satellite freeair anomaly (up) and bathymetry (down).
• 그림 Cross over errors before correction (rms=1.166 mGal).
• 그림 Cross over errors after correction (rms=0.000 mGal).
• 그림 The final freeair anomlay (C.I.=1 mGal).
• 그림 Each track map for QC.
• 그림 Flow chart for complete Bouguer anomaly.
• 그림 Available freeair anomaly inside the study area.
• 그림 Cross over errors correction for complied freeair anomaly.
• 그림 The complete Bouguer anomaly.
• 표 Results of poll of shear-wave specialists at a workshop held by the SEG in 2000.
• 그림 OBC acquisition system.
• 표 Specifications of key modules of OBC system.
• 그림 Geological model for synthetic data generation.
• 표 Acquisition Parameters.
• 그림 Synthetic multicomponent dataset acquired by modeling.
• 그림 Example of a linearly polarized direct arrival showing the recorded and rotated seismic traces of multicomponent data.
• 그림 Rotated Radial component (lower left S’x section) could be regarded as SV (PS in 2D) energy section.
• 그림 Original geophone data with shot point 20, 40, 60, 80 and 100, respectively (left), and rotated P and S wavefield (right).
• 그림 Static correction parameters applied for synthetic OBC data.
• 그림 PP fold map.
• 그림 Velocity analysis.
• 그림 (left) Stack section for PP after elevation static and (right) depth migrated section for PP data.
• 그림 (a) CCP offsets (red lines) for mode converted data are different from conventional CMP method. (b) CCP offsets varies with Vp/Vs ratio and source-receiver offset.
• 그림 ACP(Asymptotic conversion point) binning test with variable Vp/Vs value. Vp/Vs=2.1 is best for binning.
• 그림 CCP gather for PS wave after ACP binning.
• 그림 Initial Vs model derived using Vp/Vs and PP RMS velocity.
• 그림 NMO using initial Vs.
• 그림 Polarity reversal correction. (left) before and (right) after.
• 그림 Velocity analysis for PS wave.
• 그림 Final RMS velocity of PS wave.
• 그림 (a) PP and PS interpretation for correlation and (b) calculation of time shift of PS event to derive updated Vp/Vs.
• 그림 Depth variant stack of PS after elevation static.
• 그림 Implicit FD depth migration result of PS data.
• 그림 Heuristic illustration of the multi-scale technique.
• 그림 Marmousi true velocity model.
• 그림 Initial velocity model.
• 표 Inversion parameters of multi-scale FWI.
• 그림 Results of time-domain multi-scale FWI using.
• 그림 Comparison between signals in (a) time domain and (b) frequency domain.
• 그림 The spectrogram obtained by shot-time Fourier transform.
• 그림 (a) Decomposed signals from the spectrogram and (b) spectra of extracted signals.
• 그림 Two velocity models for the accuracy test of the partial derivative wavefield.
• 그림 The results of the computed partial derivative wavefield using (a) the raw; the corresponding results using the transformed data with the following decomposed componenets: (b) 5 Hz.
• 그림 The velocity models used for spectrogram inversion.
• 그림 (a) The extracted 35th raw trace from shot gather obtained at the horizontal position at 1.7km, and (b) the amplitude spectrum of the trace.
• 그림 (a) The Transformed traces, and (b) the frequency information of transformed trace with the decomposed 4 Hz, 7 Hz, 9 Hz, and 11 Hz component.
• 그림 The results of the spectrogram inversion using transformed data with (a) 7 Hz, and (b) 9 Hz components.
• 그림 Seismic profiles of the normal acquisition(a), slanted#1(b) and curved#1(c).
• 그림 Seismic profiles of 15HP-309.
• 그림 Starting menu for using Xover program.
• 그림 Diagram showing cross-over positions of the data acquired from different institutes along the given profile.
• 그림 Diagram showing cross-over errors caused by the position shifting. In this case, cross-over errors are in general corrected using the horizontal shifting method, which unreliable data or profile move to the reliabledata. Reliable data can be selected comparing the neighboring data.
• 그림 High-performance data processing services provided by the major exploration company.
• 그림 (Left) Data point map of original survey and (right) its 5D interpolated map
• 그림 FWI example from Schlumberger website. Left shows the velocity from standard PSDM tomography and right figure shows the result from FWI
• 그림 Comparison between conventional TTI RTM result(left) and Orthorhombic RTM result(right) on Perdido fold belt in GOM area(CGG).
• 표 Companies offering OBC acquisition services.
• 그림 OBC buoy array of ION (left) and OBC VSO system (right).
• 그림 PP PSDM image (left) shows the poor continuity of the event due to the shallow gas and PS PSDM image (right) shows better continuity without the effect of shallow gas (CGG).
• 그림 Seismic depth imaging methods and HPC evolution.
• 그림 BP's supercomputing center in Houston built in 2012
• 그림 Total's supercomputing center in Pau built in 2013