Pattern matching for real time leak detection and location in pipelines

• US 6,668,619 B2
• Filed: 05/15/2002
• Issued: 12/30/2003
• Est. Priority Date: 05/27/1999
• Status: Expired due to Term

• Current Assignee: Acoustic Systems Incorporated
• Original Assignee: Acoustic Systems Incorporated
• Inventors: Recane, Marion, Yang, Bao-Wen
• Primary Examiner(s): Kwok, Helen
• Assistant Examiner(s): GARBER, CHARLES D
• Application Number:
  US10/146,745
• Publication Number:
  US 20020124633A1
• Time in Patent Office: 594 Days
• Field of Search: 73/405.R, 73/405.A, 735/70, 735/87, 735/92, 735/97, 702/51, 340/611
• US Class Current: 73/40.50R
• CPC Class Codes: G01M 3/243 for pipes

Abstract

A method and apparatus is provided for detecting and locating leaks in a pipeline. Pattern match filtering is used to reduce false alarm rate, increase sensitivity and improve leak location accuracy, while quickly detecting leaks by the acoustic signal generated from a leak event in pipelines containing gas or liquid under pressure. The pattern match filter technique detects a pressure wave generated by a leak, but discriminates against background noise and pressure disturbance generated by other non-leak sources that might otherwise be detected as a leak. The pattern match filter derives a sharp peaked output from the signal of the expansion wave which allows for a distinctive point of reference for a time stamp. This provides for improved accuracy in leak location calculations. The pattern match filter is incorporated into site processors located at multiple points along a pipeline, and at a central node processor which receives data from all site processors. The pattern match filter includes using previously recorded leak profiles. At site processes located at multiple points along a pipeline, a series of previously recorded signature leak profiles are continuously compared in real time against pipeline pressure signals. Data from each site processor are used collectively at a node processor and compared against multiple leak profiles to provide further false alarm rejection. The leak event data generated at each site processor is used by the node processor to declare a leak. By the application of this pattern match filter technique, the signal to noise ratio (S/N ratio) required to identify a leak event is reduced and the sensitivity of leak detection is increased.

Claims

1. A method for detecting leaks in pipelines comprising the steps of:

interconnecting sensors at a plurality of locations along a pipeline for sensing fluid acoustic pressure in the pipeline at the plurality of locations;

interconnecting local site processors with the sensors at the plurality locations along the pipeline;

obtaining past fluid acoustic pressure data associated with past leaks in pipelines;

storing the past fluid acoustic pressure data associated with past leaks in the local site processors;

sensing fluid acoustic pressure in the pipeline at the plurality of locations with the sensors;

comparing the sensed fluid acoustic pressure data with the past fluid acoustic pressure data in the local site processors; and

issuing an alarm if the sensed fluid acoustic pressure data matches the past fluid acoustic pressure data.

2. The method of claim 1 wherein the step of comparing the sensed fluid acoustic pressure data with the past fluid acoustic pressure data is conducted over a plurality of data points.

3. The method of claim 1 wherein the step of comparing the sensed fluid acoustic pressure data with the past fluid acoustic pressure data is performed by comparing amplitude, rate of decay and wave shape.

4. The method of claim 1 wherein the step of comparing the sensed fluid acoustic pressure data with the past fluid acoustic pressure data is performed by pattern match filtering.

5. The method of claim 1 further comprising interconnecting the local site processors with a node processor and storing past fluid acoustic pressure data associated with past leaks in the node processor, and comparing the sensed fluid acoustic pressure data and the past fluid acoustic pressure data in the node processor in response to an alarm issued by the local site processors.

6. The method of claim 5 further comprising transmitting the alarm signal from the node processor to a computer terminal when the comparisons from a plurality of site processors indicate that the sensed fluid acoustic pressure data matches the past fluid acoustic pressure data.

7. The method of claim 6 further comprising locating a leak in the pipeline by time stamping data for recording the time of an acoustic pressure profile at a plurality of sites, and calculating the leak location based on the time and velocity of sound in a fluid in the pipeline.

8. The method of claim 6 further comprising sensing fluid acoustic pressure data using low-frequency signals.

9. The method of claim 6 further comprising sensing fluid acoustic pressure data using low sample rates.

10. The method of claim 6 wherein two or more local site processors are interconnected with a node processor and the method further comprises confirming leaks by the node processor upon detection of a leak by at least two of the site processors.