Method and system for inspecting flaws using ultrasound scan data

• US 7,234,355 B2
• Filed: 12/27/2004
• Issued: 06/26/2007
• Est. Priority Date: 12/27/2004
• Status: Active Grant

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Sengupta, Anandraj, Dewangan, Sandeep Kumar, Dutta, Amitabha
• Primary Examiner(s): Williams; Hezron
• Assistant Examiner(s): Miller; Rose M.
• Application Number:
  US11/022,726
• Publication Number:
  US 20060137451A1
• Time in Patent Office: 911 Days
• Field of Search: 736/22, 736/23
• US Class Current: 73/622
• CPC Class Codes: G01N 2291/044 Internal reflections (echoes), e.g. on walls or defects
• CPC Class Codes: G01N 2291/105 two or more emitters, two or more receivers
• CPC Class Codes: G01N 2291/2675 Seam, butt welding
• CPC Class Codes: G01N 29/11 by measuring attenuation of acoustic waves

Abstract

A method and system for detecting weld signatures from an ultrasound scan data obtained from scanning a pipeline is provided. The method includes a step for mapping multiple amplitude responses from the ultrasound scan data, each amplitude response being representative of a respective sensor signal. Continuous amplitude responses are located from the amplitude responses and corresponding signatures are identified for the continuous amplitude responses. The method also includes a step for tagging the corresponding signatures as weld signatures.

Claims

1. A method for detecting weld signatures from ultrasound scan data obtained from scanning a pipeline, the method comprising:

mapping a plurality of amplitude responses from the ultrasound scan data, each mapped amplitude response being representative of a respective sensor signal;

locating a plurality of continuous amplitude responses from the plurality of mapped amplitude responses;

locating a long weld angle co-ordinate from the continuous amplitude responses;

identifying at least one corresponding signature for the continuous amplitude responses; and

tagging the corresponding signature as a weld signature.

2. The method of claim 1 wherein locating the long weld angle co-ordinate comprises:

selecting a first angularly oriented sensor corresponding to a first amplitude reflection signal close to a first surface of the pipeline; and

selecting a second differently angularly oriented sensor adjacent to the first, that senses a second amplitude reflection signal close to the first surface of the pipeline.

3. The method of claim 2 further comprising computing the long weld angle co-ordinate from respective angular values of the first angularly oriented sensor and the second differently angularly oriented sensor.

4. The method of claim 1 further comprising mapping a plurality of signatures from the ultrasound scan data onto a three-dimensional histogram for locating the plurality of continuous amplitude responses.

5. The method of claim 4 further comprising using at least one pre-determined threshold for locating the continuous amplitude responses.

6. The method of claim 5 wherein the pre-determined threshold is at least one of length, standard deviation or spatial deviation of the mapped amplitude responses.

7. The method of claim 4 further comprising identifying at least one reference continuous amplitude response and a corresponding reference weld signature.

8. The method of claim 4 further comprising:

selecting an adaptive band of a pre-determined width around the reference weld signature; and

tagging the corresponding signatures in the adaptive band as weld signatures.

9. The method of claim 1 further comprising suppressing the weld signature and obtaining a weld-suppressed image.

10. The method of claim 9 further comprising using a noise filter for filtering noise in the weld-suppressed image to obtain a weld and noise suppressed image.

11. The method of claim 10 wherein the noise filter is a multi-directional probabilistic filter with an adjustable width to capture noise in multi-directions.

12. The method of claim 11 wherein the noise is attenuated by using a transfer function.

13. The method of claim 12 further comprising comparing standard deviation with a predefined threshold to suppress noise.

14. The method of claim 10 further comprising processing the weld and noise suppressed image to retrieve a flaw signature.

15. The method of claim 14 further comprising:

comparing the weld and noise suppressed image with the ultrasound scan data; and

estimating an approximate quantification of signal loss for flaw signatures.

16. The method of claim 15 further comprising selecting a threshold such that a minimum flaw length is recognized as a valid signature and restoring a corresponding signature.

17. An apparatus for identifying weld signatures from an ultrasound scan data, the apparatus comprising:

an amplitude processing component for mapping a plurality of amplitude responses from the ultrasound scan data, and for locating a plurality of continuous amplitude responses, and for locating a long weld angle co-ordinate from the continuous amplitude responses, each mapped amplitude response being representative of a respective sensor signal; and

a weld signature tagging component for identifying corresponding signatures for the continuous amplitude responses and for tagging the corresponding signatures as weld signatures.

18. An ultrasound imaging system used for detecting flaw signatures, the system comprising:

a plurality of sensors disposed around a region of interest and configured to transmit and receive signals from the region of interest;

a data acquisition system for acquiring ultrasound scan data from the plurality of sensors, the data being representative of signals received by the plurality of sensors from the region of interest;

an amplitude processing component for mapping a plurality of amplitude responses from the ultrasound scan data, and for locating a plurality of continuous amplitude responses, each mapped amplitude response being representative of a respective sensor signal; and

a weld signature tagging and suppressing component for identifying corresponding signatures for the continuous amplitude responses and for tagging the corresponding signatures as weld signatures, and for suppressing the weld signatures and obtaining a weld-suppressed image.

19. The imaging system of claim 18 further comprising a noise filter for filtering noise in the weld-suppressed image to obtain a weld and noise suppressed image.