Application of lifting wavelet transform in oil theft signal detection

Abstract

In oil pipeline, when a theft alarm signal is generated, the strong vibration signal will be brought in stress wave. Then singularity will be introduced, which contains rich information about oil theft signal. When the detecting distance increases to a certain extent, the singularity is drowned in noise. At the same time, oil theft signal distributes mainly in low frequency band. Firstly, the system to collect stress wave signal of oil theft was briefly introduced, and on-the-spot data collection steps were given. Secondly, oil theft signal is analyzed in wavelet domain and time domain. In wavelet domain, features of energy distribution in different bands are extracted. In time domain, the stress wave signal is denoised by hard threshold method, and then the characteristics of the singularity are abstracted. The research provides a new method to monitor oil stolen events in real-time. And it is easily realized on hardware and has very good practical value.