APPARATUS AND METHODS OF ANALYSIS OF PIPE AND ANNULUS IN A WELLBORE

• US 20140114892A1
• Filed: 06/24/2011
• Published: 04/24/2014
• Est. Priority Date: 06/24/2011
• Status: Active Grant

• Current Assignee: Halliburton Energy Services Incorporated (Halliburton Company)
• Original Assignee: Halliburton Energy Services Incorporated (Halliburton Company)
• Inventors: Quirein, John, Fox, Philip Edmund, Truax, Jerome Allen, Chen, Dingding
• Granted Patent: US 9,519,865 B2
• Time in Patent Office: Days
• US Class Current: 706/20
• CPC Class Codes: E21B 47/005 Monitoring or checking of cementation quality or level
• CPC Class Codes: G01V 11/002 Details, e.g. power supply systems for logging instruments, transmitting or recording data, specially adapted for well logging, also if the prospecting method is irrelevant means for transmitting well survey signals E21B47/12; signal transmission systems in general G08C; transmission in general H04B
• CPC Class Codes: G06N 5/047 Pattern matching networks; Rete networks

Abstract

Various embodiments include apparatus and methods to provide pipe analysis, annulus analysis, or one or more combinations of pipe analysis and annulus analysis with respect to one or more pipes in a wellbore. The analysis can include application of clustering and classification methods with respect to the status and the environment of the one or more pipes in the wellbore. In various embodiments, the clustering and classification can be used in characterizing borehole annular material including cement bond quality evaluation. Additional apparatus, systems, and methods are disclosed.

Claims

1. A method comprising:

acquiring waveforms and/or other sensor responses under the control of a processor unit, the acquired waveforms and/or other sensor responses being associated with a plurality of tools operating with respect to a pipe or multiple pipes in a borehole, each waveform and/or other sensor responses being data collected from a respective tool, each waveform or sensor response having an attribute correlated to the respective tool from which the attribute is generated;

extracting values of each attribute in selected intervals of each respective waveform or sensor response by the processor unit operating on the waveform or sensor response and storing the extracted values in a memory device;

performing clustering and classification analysis by the processor unit operating on the attributes and the extracted values of the associated waveforms or sensor responses; and

outputting, to a memory, data correlated to status of the pipe or multiple pipes in the borehole or character of annular material surrounding the pipe or multiple pipes in the borehole.

2. The method of claim 1, wherein outputting, to the memory, data correlated to status of the pipe or multiple pipes includes outputting data correlated to status of cement bonding of the pipe or multiple pipes in the borehole.

3. The method of claim 1, wherein outputting, to the memory, data correlated to character of annular material surrounding the pipe or multiple pipes in the borehole includes outputting data correlated to characterization of annular materials surrounding the pipe or multiple pipes in the borehole.

4. The method of claim 1, wherein outputting data includes outputting cement bond and/or annular material characterization classes to the memory.

5. The method of claim 1, wherein extracting values of each attribute in selected intervals of each respective waveform or other sensor responses includes extracting values in a time domain or a frequency domain.

6. The method of claim 1, the method includes normalizing the extracted values of each attribute prior to performing the clustering and classification analysis.

7. The method of claim 1, wherein acquiring waveforms and other sensor responses includes acquiring the waveforms from one or more source data files.

8. The method of claim 7, wherein outputting data includes outputting cement bonding and/or annular material characterization classes to the one or more source data files.

9. The method of claim 1, wherein performing clustering and classification analysis includes processing the attributes and the extracted values of the associated waveforms and other sensor responses by the processing unit arranged to generate one or more of a self organized feature map, hierarchical clustering, or weighted response function classification.

10. The method of claim 9, wherein processing the attributes and the extracted values includes using stored default settings or settings generated user preferences.

11. The method of claim 1, wherein performing clustering and classification analysis includes quantitatively evaluating cluster quality of a cluster based on a centroid vector generated by the processor unit correlated to the cluster.

12. The method of claim 11, wherein the method includes generating cluster centroid vectors from weighting vectors of each neuron in a self organized feature map analysis or generating an average of the values of attributes in each cluster to form a cluster centroid vector of the respective cluster in a hierarchical clustering analysis.

13. The method of claim 1, wherein the method includes identifying patterns of cement bond and/or annular material characterization using a material signature at a specified depth.

14. The method of claim 1, wherein the method includes the performing clustering analysis using a first analysis technique, performing clustering analysis using a second analysis technique, and comparing results of the first analysis technique with results of the second analysis technique.

15. The method of claim 1, wherein performing the clustering classification analysis includes performing the clustering classification analysis using a weighted response function analysis such that weights of the weighted response function analysis are adjusted to determine a best match with respect to one or more of an expected bound pipe response, an expected free pipe response, or a transitional condition.

16. The method of claim 1, wherein performing the clustering and classification analysis includes merging clusters to form classes based upon a service requirement.

17. The method of claim 16, wherein merging clusters includes merging clusters with a number of classes specified by a user or merging clusters with the classes derived from natural grouping.

18. The method of claim 16, wherein merging clusters includes merging the cluster based on a probability index of each class, the probability index calculated from a weighted response function analysis.

19. The method of claim 1, wherein the method includes refining results from performing clustering and classification analysis such that the refining includes using information in addition the extracting values and associated waveforms or other sensor responses and/or using knowledge based criteria.

20. The method of claim 1, wherein the method includes:

validating a classification resulting from performing clustering and classification analysis, the validation using one or more cross plots on a selected training well and a well being tested, each cross plot being a plot of one of the attributes with respect to another one of the attributes; and

if validity on the well being tested meets a threshold level, storing cluster weights derived from the training well in a memory such that the stored cluster weights can be used directly to classify data of a new well without re-clustering.

21. The method of claim 20, wherein the method includes, if validity on the well being tested meets the threshold level, storing rules for cluster merger derived from the training well in a memory such that the stored rules for cluster merger can be used directly to classify data of the new well without re-clustering.

22. The method of claim 1, wherein the method includes displaying, on a display unit, cement bond/annular material classification classes and/or one or more log views correlated to conducting one or more of acquiring the waveforms or other sensor responses, extracting the values of each attribute, performing the clustering and classification analysis, or outputting the cement bond/annular material classes.

23. The method of claim 1, wherein the attributes include a refracted acoustic amplitude, a filtered cement bond index, a sum of magnitude at a selected frequency band of an acoustic waveform of a micro seismogram, a far detector count rate from a dual-spaced pulsed neutron logging tool, a ultrasonic average impedance, or combinations thereof.

24. A machine-readable storage device having instructions stored thereon, which, when performed by a machine, cause the machine to perform operations, the operations comprising:

acquiring waveforms and/or other sensor responses under the control of a processor unit, the acquired waveforms and/or other sensor responses being associated with a plurality of tools operating with respect to a pipe or multiple pipes in a borehole, each waveform and/or other sensor responses being data collected from a respective tool, each waveform or sensor response having an attribute correlated to the respective tool from which the attribute is generated;

extracting values of each attribute in selected intervals of each respective waveform or sensor response by the processor unit operating on the waveform or sensor response and storing the extracted values in a memory device;

performing clustering and classification analysis by the processor unit operating on the attributes and the extracted values of the associated waveforms or sensor responses; and

outputting, to a memory, data correlated to status of the pipe or multiple pipes in the borehole or character of annular material surrounding the pipe or multiple pipes in the borehole.

30. The machine-readable storage device of claim 24, wherein the operations include:

validating a classification resulting from performing clustering and classification analysis, the validation using one or more cross plots on a selected training well and a well being tested, each cross plot being a plot of one of the attributes with respect to another one of the attributes; and

if validity on the well being tested meets a threshold level, storing cluster weights derived from the training well in a memory such that the stored cluster weights can be used directly to classify data of a new well without re-clustering.

31. The machine-readable storage device of claim 30, wherein the operations include, if validity on the well being tested meets the threshold level, storing rules for cluster merger derived from the training well in a memory such that the stored rules for cluster merger can be used directly to classify data of the new well without re-clustering.

32. The machine-readable storage device of claim 24, wherein performing clustering and classification analysis includes quantitatively evaluating cluster quality of a cluster based on a centroid vector generated by the processor unit correlated to the cluster.

33. The machine-readable storage device of claim 32, wherein the operations include generating cluster centroid vectors from weighting vectors of each neuron in a self organized feature map analysis or generating an average of the values of attributes in each cluster to form a cluster centroid vector of the respective cluster in a hierarchical clustering analysis.

25. A system comprising:

a processor unit;

a memory unit operably coupled to the processor unit; and

an analysis unit, wherein the processor unit, the memory unit, and the analysis unit are configured to operate to;

acquire waveforms and/or other sensor responses under the control of the processor unit, the acquired waveforms and/or other sensor responses being associated with a plurality of tools operating with respect to a pipe or multiple pipes in a borehole, each waveform and/or other sensor responses being data collected from a respective tool of the plurality of tools, each waveform or sensor response having an attribute correlated to the respective tool from which the attribute is generated;

extract values of each attribute in selected intervals of each respective waveform or sensor response by the processor unit operating on the waveform or sensor response and storing the extracted values in a memory device;

perform clustering and classification analysis by the processor unit operating on the attributes and the extracted values of the associated waveforms or sensor responses; and

output, to the memory unit, data correlated to status of the pipe or multiple pipes in the borehole or character of annular material surrounding the pipe or multiple pipes in the borehole.

26. The system of claim 25, wherein the analysis unit includes memory in addition to the memory unit, the memory unit located on a network.

27. The system of claim 25, wherein the memory unit includes electronic source files arranged to store waveform and other sensor response data from measurement tools operating with respect to the pipe or multiple pipes within the measurement volume.

28. The system of claim 25, wherein the processor unit includes a plurality of processors.

29. The system of claim 25, wherein the system includes the plurality of tools.