Leak detection of gas pipelines based on Wigner distribution

Leak detection of gas pipelines based on acoustic signals has drawn intensive attention in recent years because of its advantages over the widely adopted method based on negative pressure wave. However, detection of a small leakage in a long gas pipeline with a low pressure and complex disturbances is still a challenging problem, because in this case, the leak acoustic signal that arrives at a transducer may be weak and thus thoroughly covered by disturbances. This paper reports on the application of the WignerVille distribution (WVD) method in leak detection of gas pipelines based on acoustic signals. The WVD of the acoustic transducer signal is used to find the sensitive frequency band corresponding to a leak and to detect the leak acoustic signal from strong disturbances. To suppress the influence of cross terms of the WVD, we adopted a time-averaging WVD. Offline tests based on historical leak experimental data of a real pipeline show the effectiveness of the method in detecting leak faults of a long gas pipeline with a low pressure and strong disturbances. (c) 2012 Curtin University of Technology and John Wiley & Sons, Ltd.