Leakage detection for oil pipelines based on Independent Component Analysis
Zhigang Chen, Xiangjiao Lian, Zhihong Yu · Chinese Control Conference · 2010
Independent Component Analysis (ICA) technique develops from blind source separation (BSS). Mixed signal under certain conditions can be separated effectively by means of ICA technique. In this paper, the principal theory and algorithm of ICA is introduced, application in signal separation and filtering with ICA is studied in leakage detection for oil pipelines and imitation and field examples are given. The experiments show that signal-to-noise ratio of processed signal was raised obviously. Pressure wave in different running modes were analyzed and wave structure recognition was proposed to judge pipeline condition. Characteristic vectors depicting wave were selected and then mode base was established in limit samples. The test results show reliability and validity of this method.