A new method of leak location for the natural gas pipeline based on wavelet analysis

This paper analyzes the properties of wavelet transform and its potential application in detecting the leakage location of gas pipeline. An entity-part method, which is proposed for the accurate leaking point, has been proved valid in the stimulation experiment. In addition, method of Romberg and Dichotomy Searching are also adopted for the computational analysis of leaking point. In detail, the effect of gas velocity is taken into consideration in the location formula and we propose to consider the average velocities of two parts caused by the leaking point, respectively. Finally, we test these three methods in a real gas pipeline experiment and the result shows the improvement to some extent in contrast to the traditional approach.