Method for simulating local prestack depth migrated seismic images

• US 7,376,539 B2
• Filed: 04/16/2003
• Issued: 05/20/2008
• Est. Priority Date: 04/09/2003
• Status: Active Grant

• Current Assignee: Norsar
• Original Assignee: Norsar
• Inventors: Lecomte, Isabelle
• Primary Examiner(s): Rodriguez; Paul
• Assistant Examiner(s): Kim; Eunhee
• Application Number:
  US10/414,972
• Publication Number:
  US 20050088913A1
• Time in Patent Office: 1,861 Days
• Field of Search: 703/3, 703/5, 703/2, 703/10, 703/1, 367 14-116, 367/9, 181101-122, 702/18, 175/1, 33/1.HH
• US Class Current: 703/1
• CPC Class Codes: G01V 1/282 Application of seismic models, synthetic seismograms
• CPC Class Codes: G01V 1/286 Mode conversion
• CPC Class Codes: G01V 2210/512 Pre-stack
• CPC Class Codes: G01V 2210/614 Synthetically generated data

Abstract

A method, article of manufacture, and data set for simulating seismic prestack depth migrated images on the basis of a model of a selected GF-node, without the use of real or synthetic recorded data, provides a very efficient and flexible way to calculate a simulated depth migrated image as a function of parameters such as survey, overburden model, pulse, elastic wavefield, and local reflectivity structure. Important information needed for the method is the scattering wavenumber, calculated for example, by ray methods and other equivalent methods. Complex model geometry can be done in 2-D and 3-D.

Claims

1. Method simulating local prestack depth migrated images on basis of a model of a selected GF-node, comprising:

defining station points from acquisition surfaces;

calcaulating the GF;

extracting at the GF-node scattering wavenumbers K;

creating at the GF-node K-filters in a scattering wavenumber domain;

applying the K-filters to a target model in the scattering wavenumber domain; and

obtaining a simulated local image by transformation from the K-domain to a space domain, wherein said method is conducted without the use of real or synthetic recorded data.

2. Method according to claim 1, wherein the extraction of the K is done with different sub-selections in source/receiver locations, and/or locally in the dip/azimuth/reflection angle domain at the GF-node.

3. Method according to claim 2, wherein re-sampling of the K distribution is done with interpolation to get regular distribution, in order to improve the quality of the simulated local image and speed-up the following mapping process in the K-domain.

4. Method according to claim 2, wherein other GF parameters are extracted in parallel to the K.

5. Method according to claim 4, wherein amplitudes are extracted.

6. Method according to claim 4, wherein traveltimes are extracted.

7. Method according to claim 4, further comprising using amplitudes for creating the K-filters.

8. Method according to claim 1, wherein the K-filters are created by mapping the K on a grid suited for a summation process in terms of efficiency and quality, mapping techniques for which many public and registered techniques existing in the Synthetic Aperture Radar domain, can be used.

9. Method according to claim 8, wherein the mapping of the K on a regular grid is done by the nearest-point mapping.

10. Method according to claim 9, wherein the mapping includes taking care of the Jacobian of the transformation between the original K coordinate system, polar in 2D and spherical in 3D, and the one used for the integration, the simplest being counting the number of K assigned to a sample and dividing the total result by this number.

11. Method according to claim 8, wherein the values assigned to each K sample is a function of factors, selected from the group consisting of propagation amplitudes to simulate non-preserving amplitude PSDM, angle-dependent reflectivity, Fresnel-zone factor to obtain the frequency-dependent “

lateral”

smoothing effect done by PSDM when working with band-limited signals, and scattering-pattern factors.

12. Method according to claim 8, wherein the values assigned to each K sample is multiplied by a pulse to get the band-limited signal effect on the resolution across the reflectors, this pulse being the source-signature if no source correction is to be considered, or the deconvolved pulse if deconvolution is supposed to be done prior to the imaging, or a pulse used to test resolution improvements with the simulated process.

13. Method according to claim 1, wherein target models to study are defined from various types of input selected from the group consisting of parameter grids, interpreted horizons—

either in time and depth—

with attributes, ray-tracing models, finite-differences models, reservoir models, and are created in the same coordinate system than the one used for the K-filters, the target model parameter to map depending on the type of PSDM to simulate.

14. Method according to claim 1, wherein the final transformation from the K-domain to the space-domain is done by a numerical summation method.

15. Method according to claim 1, wherein the numerical summation method is a FFT summation method.

16. A method of producing a data set representing the simulated prestack local images comprising:

(a) extracting at the GF-node scattering wavenumbers K;

(b) creating at the GF-node K-filters in a scattering wavenumber domain;

(c) applying the K-filters to a target model in the scattering wavenumber domain; and

(d) obtaining a simulated local image by transformation from the K-domain to a space domain;

wherein the method is conducted without the use of real or synthetic recorded data.

17. Method simulating seismic traces on basis of a model of a selected Gf-node, comprising

extracting at the GF-node scattering wavenumbers K and traveltime;

creating at the GF-node and along a given K, 1D K-filters, also called 1D K-signals, in a frequency domain;

applying the 1D K-signals to a target model along a given K and in the frequency domain; and

obtaining a simulated seismic trace by transformation from the frequency domain to a time domain, wherein the method is conducted without the use of real or synthetic recorded data.

18. Method according to claim 17, further comprising using amplitudes for creating the 1D K-signals.

19. Method according to claim 17, wherein the K-signals are created by mapping the K along a frequency line suited for a summation process in terms of efficiency and quality.

20. Method according to claim 19, wherein the mapping of the K along a frequency line is done by the nearest-point mapping.

21. Method according to claim 19, wherein the values assigned to each frequency sample of a 1D K-signal is a function of factors selected from the group consisting of propagation amplitudes to simulate non-preserving amplitude PSDM, angle-dependent reflectivity, and other scattering-pattern factors.

22. Method according to claim 19, wherein the values assigned to each frequency sample is multiplied by a pulse to get the band-limited signal effect, this pulse being the source-signature if no source correction is to be considered, or the deconvolved pulse if deconvolution is supposed to be done prior to the imaging, or a pulse used to test resolution improvements with the simulated process.

23. Method according to claim 17, wherein the final transformation from the frequency domain to the time domain is done by a numerical summation method.

24. Method according to claim 23, wherein the numerical summation method is a FFT summation method.

25. A method of producing a data set representing simulated seismic traces, comprising:

(a) extracting at the GF-node scattering wavenumbers K and traveltime;

(b) creating at the GF-node and along a given K, 1D K-filters, also called 1D K-signals, in a frequency domain;

(c) applying the 1D K-signals to a target model along a given K in the frequency domain; and

(d) obtaining a simulated seismic trace by transformation from the frequency domain to a time domain, wherein the method is conducted without the use of real or synthetic recorded data.