光纤传感器振动信号特征提取研究

Research on Vibration Signal Feature Extraction on Optical Fiber Sensor

由于环境干扰等问题造成振动光纤的漏报和误报是限制光纤应用推广的重要原因。如何提取能准确反映振动信号类型和强度的特征是降低系统误报率的关键。研究了基于马赫-泽德干涉原理的光纤传感器振动信号时域和频域的特征提取方法,分析了常见振动源时频特征参数变化的规律,并通过实验确定了常见振动源特征参数的阈值。实验表明,通过对振动信号时频特征参数设定多重阈值可以准确识别振动信号的类型,验证了文中特征参数提取方法的有效性和可靠性。

The false alert caused by environmental disturbances is the main reason for application limitation ofvibration optical fiber. One of the key issues to reduce the rate of false alert is how to extract features sensitive to thetype and amplitude of vibrations. The feature extraction method of vibration signals of an optical fiber sensor ontime-domain and frequency-domain based on the Mach-Zehnder interferometric principle is researched. The varia-tion rules of common vibration source time frequency feature parameters are analyzed, and the thresholds of featuresof common vibration source are determined through experiments. The experiments show that the types of vibrationsignals are identified correctly through setting multiple thresholds for time domain and frequency domain of the vi-bration signals, and the effectiveness and reliability of the proposed method for extracting the feature parameters arevalidated.