Feature Extraction of Vibration Signal Detected by Optical Fiber along Crude Oil Pipeline and Forewarning System Based on ICA
Lin Zhu, Zhoumo Zeng, Jingchuan Zhang, Shijiu Jin · 2009
The optical fiber distributed along the pipeline can be used to obtain vibration signal and furthermore to locate the leakage and Third-Party intervention, hence offer early warning about them. The vibration signal and the vibrating disturbance from natural environment and ground exist in the optical fiber at the same time. Therefore, the vibration can not easily be located accurately. The independent component analysis (ICA) method is widely used in signal processing. In this paper ICA is applied for separating the optical fiber vibration signals from each other, and the interested signal is extracted. First, basal principle of ICA is briefly introduced. Then, three types of vibration signal are chosen. After the noise is removed by wavelet transform and filter, the signal of the optical fiber is separated by ICA. Different types of vibration have the different features, through feature extraction, we could differentiate the types of vibration, so the separated signal will offer more details along the pipeline. On the other hand, the disposing signal by ICA is beneficial to locate the vibration source latter. By the experiments of real signal, we could get the conclusion that ICA could effectively separate the signal.