The Algorithm Study On Event Detection And Deinterleave Based On Optical Fiber Vibration Sensor

People rely heavily on the energy, and pipeline transportation is the main way for the energy transport. Pipeline leakage is usually caused by human activities, such as oil stolen by drilling hole or construction work caused pipeline broken. Pipeline leakage can cause economic losses of the energy companies, and cause environmental pollution. It is more terrible that the oil and gas will often cause fire or explosion. Therefore, for preventing the leakage accidents, it is an urgent need to design the pre-warning system for the safety of oil and gas pipeline.The pattern recognition which based on the data feature was utilized to detect vibration event in most of the existing methods. There are some problems for these methods. Firstly, these methods were only applied at high signal to noise power ratio region (SNR).Secondly, there is difference in the geology of different places, so the same vibration event can engender different signals in different places. This is not conducive to identify. Thirdly, the noise jamming is various, the existing methods cannot guarantee constant false alarm rate detection. Finally, the existing methods is based on the assumption that the event is single, which is not according with the true condition.Based on the above reasons, an algorithm of event detection and deinterleave is proposed. Firstly, the adaptive constant false alarm rate detection and the sequential probability ratio test(SPRT) were applied to detect vibration event. Secondly, delay factor is introduced to overcome the problem that the SPRT cannot detect the pulse event. In the module of event deinterleave, the generalized statistical distance was applied to solve data association problem.The test experiments show that the proposed algorithm can effectively detect single events or multiple-events, and be effective in false alarm rejection. Moreover, sequential process can satisfy the requirement of the real-time system.