基于粒子滤波信号降噪的齿轮箱故障诊断研究

齿轮和轴承是机械设备中重要的部件,它的健康状况直接影响着机械设备的运行状态。研究粒子滤波对齿轮箱的振动信号进行降噪处理,需要知道信号的特性和噪声的统计特性。针对此,先对齿轮箱的齿轮进行振动分析,建立振动模型,而后再研究粒子数算法,对振动信号进行EMD分解,把分解的第一个IMF分量和残余分量进行重构,组建AR模型,经AR模型系数作为粒子滤波状态方程的系数。最后采用小波变换阈值降噪思想,把分解提出的噪声信号放到粒子滤波观测方程中,对信号进行降噪处理。

Research on the Gear Box Failure Diagnosis Based on Particle Filtering Technology

Gears and bearings are the important components of machinery and equipment, its health status directly affects the oper- ation of mechanical equipment. This paper studies the noise reduction processing by particle filter for gearbox vibration signal, and knows the characteristics of signal and the statistical properties of noise. Firstly, this article makes the vibration analysis for the gear in gearbox and establish vibration model, then studies the particle number algorithm to make the EMD decomposition for vibration signal. The decomposition of the first IMF component and residual component is reconstructed to form the AR model, and the AR factor model coefficients is taken as the coefficient of particle filtering equation of state. Finally, this paper uses wavelet transform threshold noise reduction ideas to put the decomposed noise signal into the proposed particle filter observation equation, and makes the noise reduction to the signal.