Detecting Leak in Gas Pipeline using Continuous Wavelet Transform and Kurtosis

• Main Authors: N. F., Adnan, M. F., Ghazali
• Format: Book Section
• Language: English
• Published: Universiti Teknikal Malaysia Melaka (UTeM) 2015
• Subjects: TJ Mechanical engineering and machinery
• Online Access: http://umpir.ump.edu.my/id/eprint/7562/1/Detecting_Leak_in_Gas_Pipeline_using_Continuous_Wavelet_Transform_and_Kurtosis.pdf; http://umpir.ump.edu.my/id/eprint/7562/; http://ocs.utem.edu.my/index.php/mucet2014/MUCET2014/paper/viewFile/132/39

The detection of the leak detection is the main investigation issue in order to get the fast and reliable leak detection method. Even thought the reasons for these leaks are very well known, some of the current method is quite complicated and not precise. In addition, it is all about time consuming and cost of instalment. In this paper, we proposed a leak detection method using acoustic. The chirp signal injected into the pipeline system and the estimation of the leak detection from the delay time passing by the reflection of pressure wave in the pipeline if there have a leak. Using wavelet as the noise filtering, there can give a useful signal to verify the leak. Wavelet is the tool to de-noise the noise from the original signal and then tuned using maximum values of kurtosis. The main idea is the echoes detection of the pressure wave from the signal given by the original signal. Kurtosis plays the main role as the component to choose the filter parameter because of their nature to measure spikiness. The result shows that the highest value of kurtosis for the pipeline with leak is 6.465 while for the pipe without leak, the highest value for the kurtosis 5.3214.