Title :
The fault diagnosis method of pipeline leakage based on neural network
Author :
Zhao, Jiang ; Li, Dan ; Qi, Huan ; Feng Sun ; An, Feng Suni Ran
Author_Institution :
Inst. of Electr. & Inf., Hebei Univ. of Sci. & Technol., Shijiazhuang, China
Abstract :
In this paper, BP neural network is applied to fault pattern recognition of pipeline leakage. When the pipeline pressure falls suddenly, the pressure sensors on both sides of the pipeline get pressure signals. The fundamental principal of using wavelet transform to decompose the pressure signal is introduced, using wavelet transform in pressure de-noising and pipeline feature vector extraction, and the feature vector of pipeline operation state is established based on the energy of frequency bandwidth, BP neural network with the input matrix composed by these eigenvectors is used to establish fault models of the classification of pipeline operation conditions in order to identify the leakage fault. The capability of this method has been proved by experiments, which can highlight fault information and improve the accuracy of the leak fault diagnosis.
Keywords :
backpropagation; fault diagnosis; feature extraction; mechanical engineering computing; neural nets; pipelines; signal denoising; wavelet transforms; BP neural network; fault diagnosis method; feature vector extraction; frequency bandwidth; input matrix; pattern recognition; pipeline leakage; pipeline pressure; pressure sensors; pressure signals; wavelet transform; Pipelines; Training data; Neural network; feature vector; leakage fault; wavelet transform;
Conference_Titel :
Computer, Mechatronics, Control and Electronic Engineering (CMCE), 2010 International Conference on
Conference_Location :
Changchun
Print_ISBN :
978-1-4244-7957-3
DOI :
10.1109/CMCE.2010.5610502