GB2421311A

GB2421311A - Assessing the size of a leak in a pipeline by detecting leak noise and pressure

Info

Publication number: GB2421311A
Application number: GB0523328A
Authority: GB
Inventors: Charles Gerard Harris
Original assignee: Metrika Ltd
Current assignee: Metrika Ltd
Priority date: 2004-11-16
Filing date: 2005-11-16
Publication date: 2006-06-21
Anticipated expiration: 2025-11-16
Also published as: GB0425211D0; GB0523328D0; GB2421311B

Abstract

Two accelerometers and/or hydrophones are deployed on a pipeline so that the straddle a suspected leak position, and acoustic signals emanating from the leak are recorded. A pressure transducer is also deployed on the pipeline in the vicinity of one or both of the two accelerometers and/or hydrophones. The pressure pattern, or a mathematically adjusted version of it, is correlated with each of the two accelerometer/hydrophone vibration patterns and the time delays maximising the cross correlation function in each case are used to estimate the position of the leak. Measuring the pressure signals allows the information relating to leak noise to be extracted from the vibration signals even if this information is obscured by ambient noise.

Description

Field of the Invention A method of assessing the size of a leak in a pipeline This invention relates to the location of leaks in water pipelines. Correlators are used to locate hidden leaks on buried water pipelines. Typically two accelerometers are placed on accessible fittings on a leaking pipeline so that they straddle the suspected leak position. Leaks usually emit acoustic noise and cause the pipeline to vibrate randomly. This vibration is detected by the accelerometers and the degree of synchronisation of the two vibration patterns measured over a range of time delays is revealed by cross-correlation. Knowing the dimensions of the pipeline and the likely speed of noise propagation, the time delay maximising the correlation function can be used to deduce the distance from the sensors to the leak. This procedure works remarkably well on predominantly metal pipes of less than typically 300 mm. diameter. However, on very large diameter pipes (commonly called trunk mains) or on plastic pipes of any size, it rarely works. The main problem is that acoustic signals emanating from leaks in these situations do not propagate well. By the time that the signals reach the sensors, they are obscured by ambient noise, such as that from passing traffic or other even quite minor surface disturbances. It is an object of the present invention to use the pressure noise to extract information relating to leak noise from acoustic signals in these and other circumstances. It is the existence of pressure in a pipeline or other system that causes leaks to leak! It may be expected that variations in pressure in the vicinity of a leak will cause the volume of water escaping from the leak to vary. In Patent Specification No. 2 406 654 (to which reference should be made) there is described a method of assessing the size of a leak in a buried water distribution pipe which includes measuring the surface vibration level on the ground surface above the suspected or known position of the leak in the buried water distribution pipe. The method described in Patent Specification No. 2 406 654 also includes measuring the local water pressure and leak power. The present invention is concerned with the measurement of pressure data in the vicinity of a leak to reveal the vibration patterns associated with the leak in preference to ambient unrelated vibration patterns.

Summary of the Invention According to the present invention there is provided a method of assessing the size of a leak in a pipeline in which two accelerometers and/or hydrophones are deployed on the pipeline so that they straddle the suspected leak position and a pressure transducer is also deployed on the pipeline in the vicinity of one or both of the two accelerometers and/or hydrophones. The pressure transducer is preferably located substantially midway between the two accelerometers and/or hydrophones. Suitable locations may include customer boundary boxes, fire hydrants or other fittings. Simultaneous measurements (or as simultaneous as is practically possible) are taken of the two vibration signals and the pressure signal. All three signals are sampled at a typical rate of 4Khz and for periods of typically 30 seconds. Longer periods or different frequency ranges may be used. The pattern of pressure will affect the shape and/or other characteristics of the vibration signals being emitted by the leak and there will be a detectable correlation between the pressure variation as measured by the pressure transducer and characteristics of the vibration caused by the leak measured at the accelerometer and/or hydrophone locations. In one method, the pressure pattern (or a mathematically adjusted version of it) is correlated with each of the two accelerometer and/or hydrophone vibration patterns. The time delays maximising the cross-correlation function in each case are recorded. The difference in these time delays is related to the difference in time of flight of the originating leak noise to each of the two sensors and, knowing the propagation speed of the noise (or an estimate of it) and the pipe dimensions, the spatial position of the source of the leak noise may be calculated. To understand the likely advantage of this procedure on plastic pipes or large diameter metal pipes, it is useful to first consider why the current methodology fails. In the conventional approach, which relies only on vibration signals, it is very difficult to identify those parts of the signals relating to the leak. This is because they are highly degraded and mixed with ambient noise such as that caused by passing traffic. On plastic pipes and large diameter mains, as the vibration signals travel from the leak towards the sensors, the energies in the signals are quickly absorbed and their sizes diminish. They arrive in differently corrupted states at their respective sensors. When they are cross-correlated, each corrupted pattern is checked against the other for likeness. However, because both are corrupted differently, and the common leak signal strength is low, it is unlikely that the patterns will match each other and the correlation will not peak. If it does, this is more than likely due to chance.In either event, the result is useless to the technician searching for the leak. By contrast, with the method of the present invention, by measuring the high frequency pressure pattern, the correlator is provided with a view of the shape of the forcing function driving the leak noise for which it is searching. The pressure pattern is useful because it relates directly to the hydraulic conditions in the pipe (rather than external noise), it is relatively easy to measure and is unlikely to be degraded significantly by ambient noise. Armed with such a pattern, a correlator has a better chance of locking onto the relevant leak noise in the vibration signatures that it is analysing.

Further applications of pressure data measurement A) Correlation and Pressure Spectrum Information On conventional correlators, facilities are sometimes provided to pre-filter raw vibration signatures before correlating so that irrelevant noise signals can be discarded. One approach is to display the spectra of the vibration signals being monitored. Users may then set filters to focus on common regions of spectral activity in the frequency domain (on the assumption that these relate to the leak noise). A fundamental problem with this approach is that areas of common spectral acoustic activity do not necessarily relate to leak noise. With the present invention, a facility is provided for showing the spectrum of the pressure signal monitored in the region of the leak. The benefit of doing so relates to the fact that it is the pressure variation in the pipe that has the most influence on the leak. It is expected that areas of significant spectral activity in the pressure spectrum will be the most useful ones to concentrate on in the correlation process. The pressure spectrum may be shown by itself or together with any of the vibration/acoustic spectra described above. To calculate the pressure spectrum the pressure signal may be sampled at a rate of up to 4Khz. Other frequency ranges may be used. Data from the pressure spectrum may be used to set filters optimally in any of the correlation processes described above. Means may be provided for analysing the pressure spectrum and setting filters automatically or manually. B) Pressure wave reflections It is known that enlargements of pipe section cause increases in static pressure (Bernouilli's equation). Pressure waves travelling down pipes are reflected by any section change that causes an abrupt increase in static pressure. With the present invention, analysis of the pressure patterns measured by one or more pressure transducers placed on the pipeline in the vicinity of leaks is used to detect section changes. The analysis may include auto-correlation of one pressure signal and/or cross correlation of two or more signals. For example, the auto-correlation of a pressure signal may reveal the degree of self-similarity of the signal as it is compared with delayed versions of itself. Plotting the auto-correlation function of the pressure signal for a range of time delays may reveal peaks in the auto-correlation function corresponding to pressure signal reflections. Knowing the geometry of the pipeline and the typical speed of propagation of pressure waves, peaks on the auto-correlation plot may be used to identify important section changes in the pipe such as customer draw-offs, diameter changes, sharp bends and leaks. This information is useful when combined with conventional leakage detection methods or with any of the methods described above. A plot showing significant pressure wave reflection points may be superimposed on a correlation plot aiding the interpretation of correlation peaks. C) Applications in Area Noise Logging (Acoustic Loggers) All the techniques discussed above may be applied to Correlating Acoustic Loggers. These are synchronised noise logging instruments typically deployed by scattering them over a network of pipes. Correlations are performed between each pair of loggers to reveal network sections with common leakage noise. With the present invention, one or more high frequency pressure loggers can be distributed on the network in and amongst the acoustic loggers or combined with them. Analysis of the pressure data may be used to automatically set frequency windows and/or provide time domain reference patterns in the manner described above. Figure 1 of the accompanying drawings shows the carrying out of a measurement using a pair of accelerometers and a pressure transducer located substantially midway between the two accelerometers, and Figure 2 illustrates the leak noise correlation theory. Potential benefits include improved operational noise discrimination and improved leak detection accuracy.

Claims

Claims:-

1. A method of assessing the size of a leak in a pipeline in which two accelerometers and/or hydrophones are deployed on the pipeline so that they straddle the suspected leak position and a pressure transducer is also deployed on the pipeline in the vicinity of one or both of the two accelerometers and/or hydrophones.

2. A method as claimed in Claim 1, in which the pressure transducer is located between the two accelerometers and/or hydrophones.

3. A method as claimed in Claim 1, in which the pressure transducer is located substantially midway between the two accelerometers and/or hydrophones.

4. A method as claimed in any one of the preceding claims, in which simultaneous measurements (or as simultaneous as is practically possible) are taken of the two vibration signals and the pressure signal.

5. A method as claimed in Claim 4, in which all three signals are sampled at a typical rate of 4Khz and for periods of typically 30 seconds.

6. A method as claimed in any one of the preceding claims, in which the pressure pattern (or a mathematically adjusted version of it) is correlated with each of the two accelerometer and/or hydrophone vibration patterns.

7. A method as claimed in Claim 6, in which the time delays maximising the cross-correlation function in each case are recorded.

8. A method of assessing the size of a leak in a pipeline substantially as hereinbefore described.