Global classification of sonic logs

• US 7,337,068 B2
• Filed: 12/10/2004
• Issued: 02/26/2008
• Est. Priority Date: 11/08/2001
• Status: Expired due to Term

• Current Assignee: Schlumberger Technology Corporation (Schlumberger NV)
• Original Assignee: Schlumberger Technology Corporation (Schlumberger NV)
• Inventors: Valero, Henri-Pierre, Brie, Alain, Endo, Takeshi
• Primary Examiner(s): Nghiem; Michael P.
• Assistant Examiner(s): Le; Toan M.
• Application Number:
  US11/010,204
• Publication Number:
  US 20050143922A1
• Time in Patent Office: 1,173 Days
• Field of Search: 702/14, 702/6, 702/11, 702/18, 703/5, 367/29, 367/25, 367/27, 367/28, 367/30, 367/33, 181/106, 181/113
• US Class Current: 702/14
• CPC Class Codes: G01V 1/48 Processing data

Abstract

A method of determining the sonic slowness of a formation traversed by a borehole comprising generating tracks from sonic waveform peaks received at a plurality of depths wherein the peaks that are not classified prior to tracking is set forth. A method for generating a slowness versus depth log is generated for waveform arrivals by classifying long tracks, classifying small tracks; classifying tracks that overlap; filling in gaps; and creating a final log is disclosed. In further improvements, non-classified tracks and interpolation are used to fill in gaps.

Claims

1. A method of determining the sonic slowness of a formation traversed by a borehole comprising:

receiving sonic waveform data for a plurality of depth intervals;

generating tracks as individual objects at least for a selected depth interval from the sonic waveform data received;

classifying the generated tracks within the selected depth interval, wherein the classification is based on the tracks comprising a predetermined number of peaks and all the tracks are generated for the selected depth interval before classification; and

determining the sonic slowness of a formation and outputting the determined results.

2. The method of claim 1 wherein the classifying the generated tracks comprises classifying long tracks;

classifying small tracks;

classifying tracks that overlap;

filling in gaps; and

creating a final log.

3. The method of claim 2, wherein said filling in gaps further comprises using non-classified tracks to fill gaps.

4. The method of claim 3, wherein filling in gaps further comprising performing interpolation.

5. The method of claim 2, wherein said filling in gaps further comprises performing interpolation.

6. The method of claim 5, wherein said interpolation is linear.

7. The method of claim 6 wherein linear interpolation is done if the gaps are less than 6 frames.

8. The method of claim 2, wherein classifying the long tracks further comprises:

fitting a distribution function on peaks of the track;

calculating a mean and variance of the distribution;

comparing distribution of the data with a distribution of a model data; and

classifying the long track according to the model data if said comparison determines that the track data and model data are consistent.

9. The method of claim 8 wherein slowness and time are treated having Gaussian probability distribution.

10. The method of claim 9 wherein 2D Gaussian probability distribution of slowness and time peaks is measured at one depth based on measurements at a previous depth.

11. The method of claim 9 wherein said measurement is done by a 2D Kalman filter process.

12. The method of claim 2 wherein the classifying the small tracks further comprises:

computing a 2-D median of the track, said median being a point defined by corresponding coordinates in a slowness and time domain;

determining an intersection of the slowness and time domain with a model data distribution;

defining the model in the slowness and time domain as an ellipse; and

classifying the small track based on a position of the peak in relation to the model data.

13. The method of claim 12, wherein said model is one of a compressional model and shear model.

14. The method of claim 2, wherein step of filling in the gaps further comprises:

determining if there is a gap in a selected track at a depth range covered by the selected non-classified track;

deleting the track if no gap is found; and

filling the gap in the selected track after determining that the selected non-classified track can be used to fill the gap.

15. The method of claim 14, wherein said determining if the selected track can be used to fill the gap is done by evaluating if the selected track is between upper part and lower part of a skeleton, wherein said skeleton comprises tracks that have been classified so far.

16. The method of claim 2, wherein said long track comprises more than 20 frames.

17. The method of claim 2, wherein said small track comprises less than or equal to 20 frames.

18. The method of claim 1 wherein said peaks are defined using semblance, time and slowness.

19. The method of claim 18 wherein only time and slowness are used for classification.

20. The method of claim 19, wherein a probability of a track being one of a compressional and shear is determined using all points forming the track.

21. The method of claim 20, wherein classification of one track is independent of classification of a track different from said one track.

22. The method of claim 1, wherein the sonic waveform peaks are received at more than two depths.

23. A computer system for performing a method of determining the sonic slowness of a formation traversed by a borehole comprising:

receiving sonic waveform data for a plurality of depth intervals;

generating tracks as individual objects at least for a selected depth intervals from the sonic waveform data received;

classifying the generated tracks within the selected depth interval, wherein the classification is based on the tracks comprising a predetermined number of peaks and all the tracks are generated for the selected depth interval before classification; and

determining the sonic slowness of a formation and outputting the determined results, wherein the method is implemented in a program stored on a storage media and the output of the determined sonic slowness is applied to at least one output device.

24. A method of determining the sonic slowness of a formation traversed by a borehole comprising:

generating tracks from sonic waveform peaks received at a plurality of depths, wherein the step of generating tracks comprises;

a) classifying long tracks;

b) classifying small tracks;

c) classifying tracks that overlap;

d) filling in gaps; and

e) creating a final log.

25. The method of claim 24, wherein said filling in gaps further comprises using non-classified tracks to fill gaps.

26. The method of claim 25, wherein filling in gaps further comprising performing interpolation.

27. The method of claim 24, wherein said filling in gaps further comprises performing interpolation.

28. The method of claim 27, wherein said interpolation is liner.

29. The method of claim 28, wherein linear interpolation is done if the gaps are less than 6 frames.

30. The method of claim 24, wherein tracks are considered as individual objects comprising peaks.

31. The method of claim 30, wherein said peaks are defined using semblance, time and slowness.

32. The method of claim 31, wherein only time and slowness are used for classification.

33. The method of claim 32, wherein a probability of a track being one of a compressional and shear is determined using all points forming the track.

34. The method of claim 24, wherein classification of one track is independent of classification of a track different from said one track.

35. The method of claim 24, wherein step of classifying the long tracks further comprises:

fitting a distribution function on peaks of the track;

calculating a mean and variance of the distribution;

comparing distribution of the data with a distribution of a model data; and

classifying the long track according to the model data if said comparison determines that the track data and model data are consistent.

36. The method of claim 35, wherein slowness and time are treated having Gaussian probability distribution.

37. The method of claim 36, wherein 2D Gaussian probability distribution of slowness and time peaks is measured at one depth based on measurements at a previous depth.

38. The method of claim 36, wherein said measurement is done by a 2D Kalman filter process.

39. The method of claim 24, wherein step of classifying the small tracks further comprises:

computing a 2-D median of the track, said median being a point defined by corresponding coordinates in a slowness and time domain;

determining an intersection of the slowness and time domain with a model data distribution;

defining the model in the slowness and time domain as an ellipse; and

classifying the small track based on a position of the peak in relation to the model data.

40. The method of claim 39, wherein said model is one of a compressional model and shear model.

41. The method of claim 24, wherein step of filling in the gaps further comprises:

determining if there is a gap in a selected track at a depth range covered by the selected non-classified track;

deleting the track if no gap is found; and

filling the gap in the selected track after determining that the selected non-classified track can be used to fill the gap.

42. The method of claim 41, wherein said determining if the selected track can be used to fill the gap is done by evaluating if the selected track is between upper part and lower part of a skeleton, wherein said skeleton comprises tracks that have been classified so far.

43. The method of claim 24, wherein said long track comprises more than 20 frames.

44. The method of claim 24, wherein said small track comprises less than or equal to 20 frames.