3-D prestack/poststack multiple prediction

• US 6,735,527 B1
• Filed: 02/26/2003
• Issued: 05/11/2004
• Est. Priority Date: 02/26/2003
• Status: Expired due to Fees

• Current Assignee: Landmark Graphics Corporation (Halliburton Company)
• Original Assignee: Landmark Graphics Corporation (Halliburton Company)
• Inventors: Levin, Stewart A.
• Primary Examiner(s): Barlow, John
• Assistant Examiner(s): Taylor, Victor J.
• Application Number:
  US10/374,368
• Time in Patent Office: 440 Days
• Field of Search: 702/14, 702/16, 702/10, 703/5, 703/10, 367/73, 367/83, 367/53
• US Class Current: 702/14
• CPC Class Codes: G01V 1/36 Effecting static or dynamic corrections on records, e.g. correcting spread; Correlating seismic signals; Eliminating effects of unwanted energy
• CPC Class Codes: G01V 2210/52 Move-out correction
• CPC Class Codes: G01V 2210/56 De-ghosting; Reverberation compensation

Abstract

System and method for analyzing seismic data from a formation. Stacked seismic data are provided, including a plurality of stack traces, e.g., by collecting seismic data from source and receiver locations and stacking the collected seismic data to produce the stacked seismic data. 3-dimensional (3-D) prestack traces are generated from the plurality of stack traces, e.g., by performing inverse moveout of stack traces, e.g., in a specified neighborhood, at common-depth-points, e.g., by inverse normal moveout, ray tracing, spike synthesis, etc. The inverse moveout corrected traces are convolved to compute predicted multiples which are useable in analyzing the formation. The multiples may be adaptively subtracted from the stacked seismic data, or optionally, from prestack data, to generate processed seismic data useable in analyzing the formation, e.g., for petroleum production potential. Dip moveout (DMO) corrected seismic data may be used, where DMO velocities are adjusted by dividing by cosine of the dip angle.

Claims

1. A method for processing seismic data from a formation, the method comprising:

receiving stacked seismic data, wherein the stacked seismic data comprise a plurality of stack traces;

generating 3-dimensional (3-D) prestack traces from said plurality of stack traces by performing inverse moveout of each stack trace to each of a plurality of locations to compute a plurality of corresponding inverse moveout corrected traces comprising the 3-D prestack traces; and

convolving the generated 3-D prestack traces to compute predicted multiples, wherein the predicted multiples are usable in analyzing the formation.

2. The method of claim 1, further comprising:

receiving a seismic data set acquired over said formation;

applying a transform to the predicted multiples to generate transformed predicted multiples; and

applying said transform to the seismic data set to generate a transformed seismic data set;

wherein said transformed predicted multiples and said transformed seismic data set are in a format suitable for comparison.

3. The method of claim 2, wherein said transformed predicted multiples comprise a multiples template, the method further comprising:

overlaying the transformed predicted multiples on the transformed seismic data set to generate a multiples template;

wherein said multiples template is usable in analyzing the formation.

4. The method of claim 2, further comprising:

adaptively subtracting the transformed-predicted multiples from the transformed seismic data set.

5. The method of claim 1, further comprising:

receiving a seismic data set acquired over said formation;

adaptively subtracting the predicted multiples from the seismic data set.

6. The method of claim 5, wherein the received seismic data set comprises said stacked seismic data.

7. The method of claim 5, wherein the received seismic data set comprises prestack seismic data.

8. The method of claim 1, wherein said performing inverse moveout of each stack trace comprises performing inverse normal moveout (NMO) of each stack trace.

9. The method of claim 1, wherein said convolving the generated 3-D prestack traces to compute predicted multiples comprises:

convolving each inverse moveout corrected trace with itself to generate a corresponding autoconvolved trace; and

summing the autoconvolved traces at each of the plurality of locations to compute said predicted multiples.

10. The method of claim 1, further comprising:

generating a plurality of spike traces corresponding to said inverse moveout corrected traces.

11. The method of claim 10, wherein said convolving the generated 3-D prestack traces to compute predicted multiples comprises:

convolving each inverse moveout corrected trace with a corresponding spike trace to generate a corresponding convolved trace; and

summing the convolved traces at each of the plurality of locations to compute said predicted multiples.

12. The method of claim 1, wherein each said stacked trace has a stacked trace location, and wherein said plurality of locations comprises a user specified neighborhood around each stacked trace location.

13. The method of claim 1, wherein said performing inverse moveout of each stack trace comprises:

using a reflector and velocity model to ray trace arrivals of reflectors for inverse moveout parameterization, thereby generating corresponding inverse moveout corrected traces at each of the plurality of locations.

14. The method of claim 1,

wherein each stack trace has a plurality of associated prestack source and receiver location pairs;

wherein said generating 3-D prestack traces from said stack traces comprises;

for each stack trace, for each source and receiver location pair associated with said stack trace, selecting an aperture, comprising a neighborhood of stack trace locations around the location of the stack trace;

for each stack trace location in the aperture, applying inverse moveout to a stack trace from the source location of said source and receiver location pair to said stack trace location in the aperture, thereby producing a first inverse moveout corrected trace; and

applying inverse moveout to a stack trace from the receiver location of said source and receiver location pair to said stack trace location in the aperture, thereby producing a second inverse moveout corrected trace;

wherein said first inverse moveout corrected traces and said second inverse moveout corrected traces comprise the generated 3-D prestack traces.

15. The method of claim 14, wherein said convolving the generated 3-D prestack traces to compute predicted multiples comprises:

for each source and receiver location pair of said stack trace,

for each stack trace location in the aperture, convolving the first inverse moveout corrected trace and the second inverse moveout corrected trace together to produce a convolved trace; and

summing the convolved trace into a predicted multiple trace corresponding to the source and receiver location pair;

wherein the predicted multiple traces compose prestack predicted multiple gathers.

16. The method of claim 1,

wherein each stack trace has a plurality of associated prestack source and receiver location pairs;

wherein said generating 3-D prestack traces from said stack traces comprises;

for each stack trace, for each source and receiver pair associated with said stack trace, selecting an aperture, comprising a neighborhood of stack trace locations around the location of the stack trace;

for each stack trace location in the aperture, applying inverse moveout to a spike trace from the source location of said source and receiver location pair to said stack trace location in the aperture, thereby producing a first inverse moveout corrected trace; and

applying inverse moveout to a stack trace from the receiver location of said source and receiver location pair to said stack trace location in the aperture, thereby producing a second inverse moveout corrected trace;

wherein said first inverse moveout corrected traces and said second inverse moveout corrected traces comprise the generated 3-D prestack traces.

17. The method of claim 16, wherein said convolving the generated 3-D prestack traces to compute predicted multiples comprises:

for each source and receiver location pair of said stack trace,

for each stack trace location in the aperture, convolving the first inverse moveout corrected trace and the second inverse moveout corrected trace together to produce a convolved trace; and

summing the convolved trace into a predicted multiple trace corresponding to the source and receiver location pair;

wherein the predicted multiple traces compose prestack predicted multiple gathers.

18. The method of claim 1,

wherein each stack trace has a plurality of associated prestack source and receiver location pairs;

wherein said generating 3-D prestack traces from said stack traces comprises;

for each stack trace, for each source and receiver pair associated with said stack trace, selecting an aperture, comprising a neighborhood of stack trace locations around the location of the stack trace;

for each stack trace location in the aperture, applying inverse moveout to a spike trace from the source location of said source and receiver location pair to said stack trace location in the aperture, thereby producing a first inverse moveout corrected trace; and

applying inverse moveout to a spike trace from the receiver location of said source and receiver location pair to said stack trace location in the aperture, thereby producing a second inverse moveout corrected trace;

wherein said first inverse moveout corrected traces and said second inverse moveout corrected traces comprise the generated 3-D prestack traces.

19. The method of claim 18, wherein said convolving the generated 3-D prestack traces to compute predicted multiples comprises:

for each source and receiver location pair of said stack trace,

for each stack trace location in the aperture, convolving the first inverse moveout corrected trace and the second inverse moveout corrected trace together to produce a convolved trace; and

summing the convolved trace into a predicted multiple trace corresponding to the source and receiver location pair;

wherein the predicted multiple traces compose prestack predicted multiple gathers.

20. The method of claim 1, wherein said stacked seismic data comprise dip moveout (DMO) corrected data, including DMO velocities, and wherein said generating 3-dimensional (3-D) prestack traces from said plurality of stack traces comprises:

adjusting the DMO velocities for the stacked seismic data; and

using the adjusted DMO velocities to generate said 3-D prestack traces from said plurality of stack traces.

21. The method of claim 20, wherein said adjusting DMO velocity comprises:

dividing the DMO velocities by the cosine of a nominal dip angle.

22. The method of claim 1, further comprising:

collecting seismic data from a plurality of source and receiver locations; and

stacking the collected seismic data to produce said stacked seismic data.

23. The method of claim 1, wherein said generating 3-dimensional (3-D) prestack traces from the plurality of stack traces comprises:

for each stack seismic trace location of a plurality of specified stack seismic trace locations;

creating an initial all-zero predicted multiple trace;

performing inverse moveout of each stack trace in a specified neighborhood to the selected stack seismic trace location using a stacking velocity, thereby generating corresponding inverse moveout corrected traces at the stack seismic trace location; and

wherein said convolving the generated 3-D prestack traces to compute predicted multiples comprises;

for each stack seismic trace location of said plurality of specified stack seismic trace locations;

autoconvolving each inverse moveout corrected trace; and

summing the autoconvolved trace into the predicted multiple trace at the stack seismic trace location.

24. The method of claim 23, further comprising:

adaptively subtracting the predicted multiple traces from the stack traces.

25. The method of claim 23, further comprising:

receiving a stacked seismic data set acquired over said formation;

transforming the seismic data set and the predicted multiple traces to generate a transformed seismic data set and transformed predicted multiples,

overlaying the transformed predicted multiples on the transformed seismic data set to generate a multiples template;

wherein said multiples template is usable in analyzing the formation.

26. The method of claim 1, wherein said generating 3-dimensional (3-D) prestack traces from the plurality of stack traces comprises:

for each stack seismic trace location of a plurality of specified stack seismic trace locations;

creating an initial all-zero predicted multiple trace;

for each stack trace in a specified neighborhood, wherein each stack trace has a trace location;

synthesizing a spike trace at the trace location of the stack trace;

performing inverse moveout to the selected stack trace and the spike trace, thereby generating corresponding inverse moveout corrected traces, wherein said synthesizing and said inverse moveout are performed using stacking velocity and reflector times; and

wherein said convolving the generated 3-D prestack traces to compute predicted multiples comprises;

for each stack seismic trace location of said plurality of specified stack seismic trace locations;

convolving the inverse moveout corrected traces to generate a convolved trace; and

summing the convolved trace into the predicted multiple trace at the stack seismic trace location.

27. The method of claim 26, further comprising:

adaptively subtracting the predicted multiple traces from the stack traces.

28. The method of claim 26, further comprising:

receiving a stacked seismic data set acquired over said formation;

transforming the seismic data set and the predicted multiple traces to generate a transformed seismic data set and transformed predicted multiples,

overlaying the transformed predicted multiples on the transformed seismic data set to generate a multiples template;

wherein said multiples template is usable in analyzing the formation.

29. The method of claim 1, wherein said generating 3-dimensional (3-D) prestack traces from the plurality of stack traces comprises:

for each stack seismic trace location of a plurality of specified stack seismic trace locations;

creating an initial all-zero predicted multiple trace;

for each stack trace in a specified neighborhood, ray tracing arrivals of reflectors for inverse moveout parameterization using a reflector and velocity model, thereby generating corresponding inverse moveout corrected traces at the stack seismic trace location; and

wherein said convolving the generated 3-D prestack traces to compute predicted multiples comprises;

for each stack seismic trace location of said plurality of specified stack seismic trace locations;

autoconvolving each inverse moveout corrected trace; and

summing the autoconvolved trace into the predicted multiple trace at the stack seismic trace location.

30. The method of claim 29, further comprising:

adaptively subtracting the predicted multiple traces from the stack traces.

31. The method of claim 29, further comprising:

receiving a stacked seismic data set acquired over said formation;

transforming the seismic data set and the predicted multiple traces to generate a transformed seismic data set and transformed predicted multiples,

overlaying the transformed predicted multiples on the transformed seismic data set to generate a multiples template;

wherein said multiples template is usable in analyzing the formation.

32. The method of claim 1, wherein said generating 3-dimensional (3-D) prestack traces from the plurality of stack traces comprises:

for each stack seismic trace location of a plurality of specified stack seismic trace locations;

for a stack seismic trace at the stack seismic trace location, using prestack recording geometry to create an initial all-zero predicted multiple gather with traces at the locations of traces contributing to the stack seismic trace;

using a stacking velocity, performing inverse moveout of the stack seismic trace to each stack trace location in a specified neighborhood to generate corresponding inverse moveout corrected traces at said stack trace locations in the specified neighborhood; and

using the stacking velocity, performing inverse moveout of a stack trace at each stack trace location in said specified neighborhood to each location in the predicted multiple gather, thereby generating corresponding inverse moveout corrected traces at each location in the predicted multiple gather; and

wherein said convolving the generated 3-D prestack traces to compute predicted multiples comprises;

for each stack seismic trace location of said plurality of specified stack seismic trace locations;

for each location in the predicted multiple gather;

convolving each inverse moveout corrected trace at said stack trace locations in the specified neighborhood with the inverse moveout corrected trace at the location in the predicted multiple gather; and

summing the convolved traces into the predicted multiple trace at the location in the predicted multiple gather.

33. The method of claim 32, further comprising:

stacking the predicted multiple gathers to generate a predicted multiple stack; and

adaptively subtracting the predicted multiple stack from the stacked seismic data.

34. The method of claim 32, further comprising:

receiving a stacked seismic data set acquired over said formation;

transforming the seismic data set and a stack of the predicted multiple gather of to generate a transformed seismic data set and transformed predicted multiples,

overlaying the transformed predicted multiples on the transformed seismic data set to generate a multiples template;

wherein said multiples template is usable in analyzing the formation.

35. The method of claim 32, further comprising:

using prestack seismic data comprising prestack gathers, adaptively subtracting the predicted multiple gathers from the prestack gathers.

36. The method of claim 1, wherein said generating 3-dimensional (3-D) prestack traces from the plurality of stack traces comprises:

for each stack seismic trace location of a plurality of specified stack seismic trace locations;

for a stack seismic trace at the stack seismic trace location, using prestack recording geometry to create an initial all-zero predicted multiple gather with traces at the locations of traces contributing to the stack seismic trace;

synthesizing a spike trace at each stack trace location in a specified neighborhood;

using a stacking velocity and reflector times, performing inverse moveout of the stack seismic trace to said stack trace locations, thereby generating corresponding inverse moveout corrected traces at said stack trace locations in the specified neighborhood; and

using the stacking velocity and reflector times, performing inverse moveout of the spike trace at each stack trace location in said specified neighborhood to each location in the predicted multiple gather, thereby generating corresponding inverse moveout corrected spike traces at each location in the predicted multiple gather; and

wherein said convolving the generated 3-D prestack traces to compute predicted multiples comprises;

for each stack seismic trace location of said plurality of specified stack seismic trace locations;

for each location in the predicted multiple gather;

convolving the inverse moveout corrected trace at each stack trace location in the specified neighborhood with the inverse moveout corrected spike trace at the location in the predicted multiple gather; and

summing the convolved traces into the predicted multiple trace at the location in the predicted multiple gather.

37. The method of claim 36, further comprising:

stacking the predicted multiple gathers to generate a predicted multiple stack; and

adaptively subtracting the predicted multiple stack from the stacked seismic data.

38. The method of claim 36, further comprising:

receiving a seismic data set acquired over said formation;

transforming the seismic data set and the predicted multiple gathers to generate a transformed seismic data set and transformed predicted multiples.

overlaying the transformed predicted multiples on the transformed seismic data set to generate a multiples template;

wherein said multiples template is usable in analyzing the formation.

39. The method of claim 36, further comprising:

using prestack seismic data comprising prestack gathers, adaptively subtracting the predicted multiple gathers from the prestack gathers.

40. The method of claim 1, wherein said generating 3-dimensional (3-D) prestack traces from the plurality of stack traces comprises:

for each stack seismic trace location of a plurality of specified stack seismic trace locations;

for a stack seismic trace at the stack seismic trace location, using prestack recording geometry to create an initial all-zero predicted multiple gather with traces at the locations of traces contributing to the stack seismic trace;

using a reflector and velocity model, ray tracing arrivals of reflectors in said model for said stack seismic trace to each stack trace location in a specified neighborhood, thereby generating inverse moveout corrected traces at said stack trace locations in the specified neighborhood;

using said reflector and velocity model, ray tracing arrivals of reflectors in said model for a stack trace at each stack trace location in said specified neighborhood to each location in the predicted multiple gather, thereby generating corresponding inverse moveout corrected traces at each location in the predicted multiple gather; and

wherein said convolving the generated 3-D prestack traces to compute predicted multiples comprises;

for each stack seismic trace location of said plurality of specified stack seismic trace locations;

for each location in the predicted multiple gather;

convolving the inverse moveout corrected trace at each stack trace location in the specified neighborhood with the inverse moveout corrected trace at the location in the predicted multiple gather; and

summing the convolved traces into the predicted multiple trace at the location in the predicted multiple gather.

41. The method of claim 40, further comprising:

stacking the predicted multiple gathers to generate a predicted multiple stack; and

adaptively subtracting the predicted multiple stack from the stacked seismic data.

42. The method of claim 40 further comprising:

receiving a seismic data set acquired over said formation;

transforming the seismic data set and the predicted multiple gather to generate a transformed seismic data set and transformed predicted multiples.

overlaying the transformed predicted multiples on the transformed seismic data set to generate a multiples template;

wherein said multiples template is usable in analyzing the formation.

43. The method of claim 40, further comprising:

using prestack seismic data comprising prestack gathers to adaptively subtract the predicted multiple gathers from the prestack gathers.

44. The method of claim 1, further comprising:

analyzing the formation to determine petroleum production potential.

45. The method of claim 1, wherein said convolving the generated 3-D traces to compute predicted multiples comprises:

applying time-variant scaling and/or time-variant filtering to the generated 3-D traces to generate scaled and/or filtered 3-D traces;

convolving the scaled and/or filtered 3-D traces to generate convolved traces; and

summing the convolved traces to generate the predicted multiples.

46. The method of claim 1, wherein said convolving the generated 3-D traces to compute predicted multiples comprises:

convolving the 3-D traces to generate convolved traces;

performing time-variant scaling and/or time-variant filtering on the convolved traces to generate modified traces; and

summing the modified traces to generate the predicted multiples.

47. The method of claim 1, wherein said convolving the generated 3-D traces to compute predicted multiples comprises:

applying time-variant scaling and/or time-variant filtering to the generated 3-D traces to generate scaled and/or filtered 3-D traces;

convolving the scaled and/or filtered 3-D traces to generate convolved traces;

performing time-variant scaling and/or time-variant filtering on the convolved traces to generate modified traces; and

summing the modified traces to generate the predicted multiples.

48. A system for processing seismic data from a formation, the system comprising:

a processor;

a memory medium coupled to the processor;

an input coupled to the processor and the memory medium; and

an output coupled to the processor and the memory medium;

wherein the input is operable to;

receive stacked seismic data, wherein the stacked seismic data comprise a plurality of stack traces;

wherein the memory medium stores program instructions which are executable by the processor to;

generate 3-dimensional (3-D) prestack traces from said plurality of stack traces; and

convolve and combine the generated 3-D prestack traces to compute predicted multiples, wherein the predicted multiples are usable in analyzing the formation.

49. The system of claim 48, wherein the program instructions are further executable by the processor to:

perform inverse moveout of each stack trace to each of a plurality of locations to compute a plurality of corresponding inverse moveout corrected traces comprising the 3-D prestack traces.

50. A system for processing seismic data from a formation, the system comprising:

means for receiving stacked seismic data, wherein the stacked seismic data comprise a plurality of stack traces;

means for generating 3-dimensional (3-D) prestack traces from said plurality of stack traces by performing inverse moveout of each stack trace to each of a plurality of locations to compute a plurality of corresponding inverse moveout corrected traces comprising the 3-D prestack traces; and

means for convolving and combining the generated 3-D prestack traces to compute predicted multiples, wherein the predicted multiples are usable in analyzing the formation.

51. A carrier medium which stores program instructions for processing seismic data from a formation, wherein the program instructions are executable to perform:

receiving stacked seismic data, wherein the stacked seismic data comprise a plurality of stack traces;

generating 3-dimensional (3-D) prestack traces from said plurality of stack traces by performing inverse moveout of each stack trace to each of a plurality of locations to compute a plurality of corresponding inverse moveout corrected traces comprising the 3-D prestack traces; and

convolving and combining the generated 3-D prestack traces to compute predicted multiples, wherein the predicted multiples are usable in analyzing the formation.