改进的LMD及在齿轮故障诊断的应用

IMPROVED LMD AND ITS APPLICATIONS ON GEAR FAULT DIAGNOSIS

作为机械传动系统的关键部件之一,齿轮的工作状态直接影响设备的性能,其故障在传动系统故障中占有很大的比例。针对齿轮故障信号大多为非平稳调制信号的特点,提出一种基于样条和边界特征延拓的改进LMD算法,通过采用三次样条插值,对信号的上下极值点进行插值以获得上下包络线,并采用边界局部特征尺度延拓的方法以减少端点效应问题,提高了分解精度和计算效率。仿真信号和实验平台故障数据的仿真研究表明,改进LMD算法分解后的PF分量与理想值相比误差更小,分解时间更短,能准确地分析出齿轮的故障特征。

As one of the key components of mechanical drive system,gear fault holds a large proportion in the transmission system fault,its working status directly affect the performance of the equipment.Aiming at the characteristics of the gear fault signal mostly for the non-stationary modulation signal,proposed based on spline and an improved algorithm of LMD continuation boundary characteristics,through the adoption of cubic spline interpolation,the signal of extreme value point interpolation to get the upper and lower envelope,and local characteristic scale boundary continuation method is adopted to reduce the endpoint effect,improve the precision and computational efficiency of decomposition.The simulation study of simulated signals and experimental platform of fault data show that improved LMD algorithm is proposed in this paper that decomposition PF components is smaller error compared with the ideal value,shorter time decomposition,accurately analyzing the fault characteristics of the gear.