Author(s): Muhammad Hanafi Yusop, Mohd Fairusham Ghazali, Mohd Fadhlan Mohd Yusof, Muhammad Aminuddin Piremli
Linked Author(s): Muhammad Hanafi Yusop
Keywords: Leak detection, pressure transient, I-kaz, Hilbert Huang Transform (HHT)
Abstract: Signal processing is an important tool for diagnostic of non-stationary and non-linear data. Many techniques of analysis are available to process these types of data, such as STFT, Wavelets Transform, and demodulation analysis. In the recent study, the analysis of pressure transient signal can be seen as an accurate and low-cost method for the leak and pipe feature detection in water distribution system. Transient phenomena occurred due to sudden changes in fluid propagation filled in pipelines system, which is caused by rapid pressure and flow fluctuation in a system, such as closing and opening valve rapidly or pump failure. Various method of pressure transient analysis has been applied by several groups of researcher, such as cepstrum analysis, cross-correlation, and wavelets transform. In this research, the application of Hilbert Huang Transform (HHT) was used as the method to analyse the pressure transient signal. However, this method has the difficulty in selecting the suitable IMF for the next data post-processing method, which is Hilbert Transform (HT). Previous researchers normally perform the selection of the IMF on visually basis according to user's experience and used the statistical approach that allows the automatic selection of intrinsic mode function (IMF). This paper proposed the implementation of Integrated Kurtosis-based Algorithm for z-filter Technique (I-kaz) to kurtosis ratio (I-kaz-Kurtosis) for that allows automatic selection of intrinsic mode function (IMF) should be used. This work demonstrates on 67. 9 meters Medium High Density Poly Ethylene (MDPE) pipe installed with single artificial leak simulator. The analysis results using I-kaz-kurtosis ratio revealed that the method can be used as an automatic selection of intrinsic mode function (IMF), although the noise level ratio of the signal is lower. I-kaz-kurtosis ratio is recommended and advised to be implemented as automatic selection of intrinsic mode function (IMF) through HHT analysis
Year: 2017