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 Ocean-bottom survey with OBC (left) and OBN (right)
• 그림 OBC survey with single vessel and recording buoy
• 표 Companies offering Ocean-bottom seismic acquisition services.
• 그림 The PP result (left) suffers from problems caused by shallow gas, while the PS result (right) is unaffected by the gas
• 그림 P-wave (left) and PS-wave (middle) data in their native time, with Vp/Vs ratio computed from horizon matching (colored lines).
• 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. Migrated image is constructed by cross-correlation between forward propagation and backward propagation.
• 그림 Reverse time migration algorithm.
• 그림 Comparison of migration speeds (left) using 1 core (right) using 16 cores.
• 표 Results of migration time in the parallel test for 2016 seismic field data
• 그림 Calculation time of reverse time migration using single core computation for seismic field data.
• 그림 Test section using seismic field data applied parallel reverse time migration.
• 그림 Comparison of migration result using commercial software applied by post-stack Kirchhoff migration module (left) and prestack reverse-time migration module (right).
• 그림 Comparison of migration result using commercial software applied by post-stack Kirchhoff migration module (left) and prestack reverse-time migration module (right).
• 그림 Comparison of migration result of Line 16HP-205 using commercial software. Stack section (top) and migrated section (bottom).
• 그림 Comparison of migration result of Line 16HP-102 using commercial software. Stack section (top) and migrated section (bottom).
• 그림 Comparison of migration result of Line 16HP-105A using commercial software. Stack section (top) and migrated section (bottom).
• 그림 Workflow for the simultaneous-source FWI.
• 그림 (a) The Marmousi-2 model and (b) the initial velocity model.
• 그림 Inversion results of (a) the individual-source FWI, (b) the conventional ESSFWI, (c) the correlation-based ESSFWI, and (d) the new ESSFWI using the full-offset observed data.
• 그림 Inversion results of several ESSFWI strategies with (a) every 10th iteration, (b) every 50th iteration, (c) every 100th iteration, and (d) square numbers.
• 표 Comparison of computational speed-up for several FWI strategies.
• 그림 (a) Data misfit and (b) model misfit for the individual-source FWI (solid gray line) and several ESSFWI strategies with every iteration (dotted line), every 10th iteration (dashed line), every 50th iteration (solid line with circles), every 100th iteration (solid line with triangles), and square numbers (solid black line).
• 표 Parameterizations of VTI media in the acoustic approximation
• 그림 Layer model including fault
• 그림 Integrated sensitivity of  according to ray parameter p: (a)  ≤ ∥∥≤  , (b)  ≤ ∥∥≤ , (c)  ≤ ∥∥≤ , and (d) ∥∥≤ .
• 그림 Integrated sensitivity of  according to ray parameter p: (a)  ≤ ∥∥≤  , (b)  ≤ ∥∥≤ , (c)  ≤ ∥∥≤ , and (d) ∥∥≤ .
• 그림 Integrated sensitivity of  according to ray parameter p: (a)  ≤ ∥∥≤  , (b)  ≤ ∥∥≤ , (c)  ≤ ∥∥≤ , and (d) ∥∥≤ .
• 그림 Smoothed initial models
• 그림 (a) Reconstructed vertical velocity,  (b) horizontal velocity,  and (c) using sequential FWI of vertical and horizontal velocities after the 40th iteration.
• 그림 (a) Reconstructed vertical velocity,  (b) horizontal velocity,  and (c) using simultaneous FWI of vertical and horizontal velocities after the 40th iteration.
• 그림 The definition of 3D quadratic shapes expressed as a function of the principal curvatures
• 그림 3D seismic attributes analysis on the collapse features (after Marfurt, 2010; Data courtesy of Devon Energy).
• 그림 Computation of the components of the unit normal(n) from apparent dips, p and q in the inline and crossline direction
• 그림 The results of 3D seismic attributes analysis on Jeju Basin 3D seismic data(time slices at t = 1,624 ms).
• 그림 The results of seismic attributes analysis on Jeju Basin 3D seismic data (time slices at t = 336 ms).
• 그림 Reservoir identification through seismic attributes analysis on a 3D seismic volume acquired in a gas field (time slices at t = 4,220 ms).
• 그림 (a) 1D velocity model. (b) Surface related multiple made by primary reflection from first reflector. (c) Surface related multiple made by primary reflection from second reflector. (d) Surface related multiple that primary reflection from first reflector combines with primary reflection from second reflector.
• 그림 Surface related multiple prediction applied to convolution.
• 그림 1D surface related multiples prediction by convolution of primary reflections.
• 그림 Flowchart of surface related multiple prediction applied to feedback theory.
• 그림 Surface related multiple removal applied to convolution
• 그림 Flowchart of feedback theory applied to source wavelet.
• 그림 Surface related multiple removal of seismic signal convolving with source wavelet applied to deconvolution filter.
• 그림 Surface related multiple removal of seismic signal convolving with source wavelet applied to matching filter.
• 그림 2D surface related multiple prediction
• 그림 Surface related multiple prediction using matrix multiplication.
• 그림 A simple workflow to handle practical issues. The processes are considered to be preprocessing before multiple prediction and elimination.
• 그림 Simple representation of multiple prediction operators
• 그림 Horizontally three-layered velocity model.
• 그림 A common shot gather.
• 그림 The ray paths of reflected wave shown in Fig. 3-2-4-14.
• 그림 A seismic data generated using dipole sources. Since we ignore the surface, the surface related multiples are not synthesized.
• 그림 The Matrix multiplication between a common shot gather(P ) and a common receiver gather(X  ) in frequency domain. Since the X  includes the source wavelet( ), the source has to be removed using deconvolution. That is, predicted multiples(M) are obtained.
• 그림 Predicted surface related multiples.
• 그림 The common shot gathers with (a) and (b) without the surface related multiples.
• 그림 The common shot gather without the surface related multiples (Fig. 3-2-4-19) from 0.3 km to 2.7 km in horizontal distance. The far offset traces still have surface relates multiples due to the head wave so that we attempt to focus on near offset data.
• 그림 Traces extracted from zero offset.
• 그림 The modified Marmousi-2 velocity model.
• 그림 The common shot gather obtained at the center of the velocity model. Its right-hand side only is displayed for more detailed observation.
• 그림 The common shot gather synthesized without free surface and with dipole source signatures to compare with Fig. 3-2-4-23.
• 그림 Predicted surface related multiples.
• 그림 The common shot gather without the surface related multiples. The matching filter was designed based on L2-norm.
• 그림 The common shot gather without the surface related multiples. The matching filter was designed based on L1-norm.
• 그림 The zero offset gathers (a) before and (b, c) after SRME. In order to implement the SRME, (b) L2-norm based and (c) L1-norm based matching filters were chosen. Surface related multiple in (a) is highlighted in red arrows. Noise due to L2-norm based matching filter in (b) is highlighted in yellow circles
• 그림 Predicted surface related multiples at the 2nd iteration
• 그림 The common shot gather without the surface related multiples at the (a) 1st and (b) 2nd iteration.
• 그림 A common shot gather which has missing traces at near offsets.
• 그림 (a) NMO correction result. The common mid-point gathers (b) before and (c) after interpolation.
• 그림 A common shot gather which is interpolated up to the zero offset.
• 그림 The final common shot gather after interpolation using NMO modeling and reciprocity.
• 그림 The comparison of zero offset gathers (a) with and (b) without surface related multiples. Surface related multiple in (a) is highlighted in red arrows.
• 그림 Horizontally 3-layered velocity model which have shallow water layer.
• 그림 The common shot gather obtained at the center of the model.
• 그림 The common shot gather synthesized without free surface and with dipole source signatures to compare with Fig. 3-2-4-37.
• 그림 Predicted surface related multiples.
• 그림 The common shot gather without the surface related multiples. The matching filter was designed based on L1-norm.
• 그림 Comparison result of surface-related multiple elimination.
• 그림 Horizontally 6-layered velocity model which have shallow water layer.
• 그림 A velocity profile extracted from horizontally layered medium.
• 그림 The common shot gather obtained at the center of the model.
• 그림 The common shot gather synthesized without free surface and with dipole source signatures to compare with Fig. 3-2-4-38.
• 그림 Comparison result of surface-related multiple elimination.
• 그림 The seismic data after linear Radon transform.
• 그림 (a) The seismic data after pre-critical reflection muting in radon domain. (b) The seismic data after pre-critical reflection muting in space-time domain.
• 그림 (a) The seismic data after predictive deconvolution in radon domain. (b) The seismic data after predictive deconvolution in space-time domain
• 그림 (a) The seismic data after shallow reflection muting in radon domain. (b) The seismic data after shallow reflection muting in space-time domain.
• 그림 Result of 3-layered model.
• 그림 Result of 6-layered model.
• 그림 A time-domain modeling algorithm using the running integration
• 그림 The SEG/EAGE 3D salt velocity model.
• 그림 A Laplace-domain wavefield example (y = 4 km) when the damping constant is 10   . The surface source location is (x, y) = (4 km, 4 km).
• 표 Time-domain modeling time without Laplace transformation and wavefield transfer.
• 표 Time-domain modeling time with Laplace transformation.
• 표 Time-domain modeling time with Laplace transformation using different block sizes.
• 그림 A simplified Laplace-domain inversion algorithm based on time-domain modeling
• 그림 A diagram showing the data movements between the host and the device.
• 그림 (a) The SEG/EAGE 3D salt velocity model and (b) the initial velocity model used for velocity model building.
• 그림 Velocity update directions from (a) the 1st iteration and (b) 100th iterations.
• 그림 (a) An inversion result after 200 iterations and (b) the error history of the inversion.
• 그림 Parallel program structure.
• 표 Time-domain modeling time with Laplace transformation for four-shot experiments.
• 표 Velocity modeling time using CPUs and GPUs.
• 그림 KIGAM High Performance Computing (HPC) system.
• 그림 Workflow chart for data processing to remove multiples.
• 그림 A variety of seismic events represented by their raypaths.
• 그림 SRME processing procedure in shot gather and stack section after applying SRME.
• 그림 Radon filter processing procedure in CDP gather and stack section after applying radon filter.
• 그림 Stack section after applying SRME and radon filter sequently in deep sea.
• 그림 Comparison between stack section before and after applying combination of SRME, SRWEMR and radon filter.
• 그림 Seismic survey line in Yeosu, South Sea of Korea.
• 그림 2D marine seismic survey equipment and conceptual diagram
• 그림 Mimetic diagram on air-gun array configuration.
• 그림 Offset layout diagram of seismic survey equipment used in seismic survey 2016.
• 표 The 4 kinds of seismic survey and main data acquisition parameters.
• 표 Seismic exploration data acquisition parameters using Sercel streamer survey.
• 표 Geodetic parameters used in data acquisition.
• 표 Information of seismic survey line on research area.
• 그림 Line shoot log of 16HP survey.
• 그림 Screens of module inputing seismic rawdata into seismic processing software; left) SEG-D input and right) SEG-Y input.
• 그림 ‘Binning sequence’ window and ‘trace QC’ window in making geometry data.
• 그림 Common shot gather and its trace header display of (a) before matching shot gather with geometry and (b) after matching shot gather with geometry.
• 그림 (a) shot gather before applying the bandpass filter and (b) shot gather after applying the bandpass filter
• 그림 shot gather (a) before applying the true amplitude recovery and (b) after applying the true amplitude recovery.
• 그림 Velocity analysis panel made by velocity analysis precompute module.
• 그림 The stack section of (a) 8 channel streamer data, (b) GeoEel streamer data and (c) Sercel streamer data.
• 그림 Seismic section acquired by the three types of seismic equipment
• 그림 SeaSPY marine magnetometer.
• 표 Data format of SeaSPY marine magnetometer.
• 그림 Editing raw magnetic data.
• 그림 Total magnetic anomaly.
• 그림 Filtering total magnetic anomaly by moving average method.
• 그림 Filtered total magnetic anomaly.
• 그림 Each track data.
• 그림 Deleted data by quality control of each track data (red line).
• 그림 CCOP total magnetic anomaly (top) and bathymetry (bottom).
• 그림 Crossover error before crossover error correction (rms=20.69 nT).
• 그림 Crossover error after crossover error correction (rms=0.47 nT).
• 그림 Crossover error corrected total magnetic anomaly (C.I.=50 nT).
• 그림 Each track data for final quality control.
• 그림 Available other data set in the study area.
• 그림 Crossover error before crossover error correction (rms=28.83 nT).
• 그림 Crossover error after crossover error correction (rms=0.30 nT).
• 그림 Final total magnetic anomaly (C.I.=50 nT).
• 그림 OBC equipment deployment
• 그림 Tamhae II OBC survey.
• 그림 Tamhae II air gun shooting
• 그림 OBC seismic data recording.
• 표 Main data acquisition parameters for OBC seismic survey.
• 그림 Data acquisition and monitoring system for OBC seismic survey.
• 그림 KIGAM seismic instruments rack and Geores display.
• 그림 OBC seismic survey data display.
• 그림 Geophone orientation and rotation of ocean-bottom cable
• 그림 OBC receiver gather for each OBC component (OBC node1).
• 그림 Remarkable characteristics of hydrophone and inline component data
• 그림 Survey track of 16HP OBC data acquisition (left). Right figure shows the 1st and 2nd OBC position with respect to each shot point number and survey directions of each shooting
• 그림 (a) Basemap of 16HP-105C survey. Black dots indicate source positions and white dots indicate receiver (OBC) positions. (b) CMP positions of the survey, (c) binning is carried out using 3D land geometry method and (d) enlarged grid of the red box in (c), which shows the mid-point gathers in 6.25 x 100 m size of each grid. All x-axes are exaggerated for clarity.
• 그림 2D Foldmap for 2 survey lines.
• 그림 Dip noises are removed from Radon dip filter after top muting.
• 그림 Low frequency noises are removed after FX deconvolution, Frequency median filtering and Spectral whitening
• 그림 Semblance velocity analysis for OBS P wave data.
• 그림 Final PP stacked section of 2 survey lines
• 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 Various Vp/Vs test for ACP binning. Vp/Vs=2.2 shows best fold map without fluctuation.
• 그림 CCP gather for 16HP-105E PS wave after ACP binning.
• 그림 Before and after polarity reversal correction on PS receiver gather.
• 그림 Coherent noise attenuation for low frequency linear noise removal.
• 그림 Application of FX decon. and time variant spectral whitening.
• 그림 Velocity analysis for PS wave.
• 그림 Final PS velocity structure.
• 그림 Final PS stacked section of 2 survey lines.
• 그림 Combined PP and PS stacked section fo 16HP-105C and 16HP-105E survey line. Total 1.6 km’s stack section was derived. Two images are correlated with Vp/Vs=2.2.
• 그림 Comparison between streamer P and OBC section.
• 그림 The snapshots of the displacement wavefield in homogeneous media
• 그림 The P-wavefield separated traces by using
• 그림 Modified Marmousi-2 model
• 그림 Initial velocity models for waveform inversion
• 그림 Inversion results with 830 receivers
• 그림 Inversion results with 41 receivers
• 그림 Inversion results with 200 receivers in center area
• 그림 Inversion results with 20 receivers in center area
• 그림 Inversion results with 200 receivers in left gas area
• 그림 Inversion results with 20 receivers in left gas area
• 그림 Inversion results with the Multi-scale method
• 그림 Initial models by using the poisson's ratio
• 그림 Inversion results by using both hydrophone data and OBC data sets
• 그림 Case of broadband seismic survey using variable-depth streamer in China
• 그림 Seismograms and its amplitude spectra of dual-sensor streamer. left) pressure components, middle) velocity components and right)deghosted up-going pressure components
• 그림 Deghosted stack section by using variable depth streamer. left) conventional stack section and right) variable depth streamer stack section