基于理性样条插值的经验模态分解及其在轴承故障诊断中的应用

Empirical Mode Decomposition and Its Application in Bearing Fault Diagnosis Based on Rational Spline Interpolation

针对现场采集的滚动轴承振动信号表现出的较强的非线性、非平稳特性,提出了基于理性样条插值法的经验模态分解算法(Improved Empirical Mode Decomposition,IEMD)以解决曲线拟合问题。首先取得信号的极值点,利用镜像延拓法延拓两端极值;其次引入具有保形控制参数的有理三次样条插值法对延拓后的极值点进行曲线拟合;最后将迭代信号减去取得的均值包络线,使迭代后的信号满足分解条件。实验表明,改进算法有效改善了曲线拟合中的过包络和欠包络问题。结合希尔伯特边界谱分析,将该算法应用于轴承数据处理中,结果表明,该算法准确提取了故障特征频率,提高了经验模态分解算法的有效性,具有较高的工程实用价值。

Considering the strong nonlinearity and non-stationary characteristics of the rolling bearing vibration signals collected on site, the improved empirical mode decomposition （IEMD） algorithm based on rational spline interpolation was proposed to solve the problem of curve fitting. Firstly, having signals＇ extreme point obtained and extended at the start and end data points by adopting the mirror extension method; secondly, hav- ing an rational cubic spline interpolation method with shape control parameters introduced to the curve fitting of the extreme points extended ; finally, having the iterative signal subtracted from the obtained mean envelopes so that the iterated signals can satisfy EMD＇s conditions. Experimental results show th can improve the over-enveloping and under-enveloping problems in curve fitting. Co at, the improved algorithm mbining with the boundary spectrum analysis and applying this algorithm to processing of the bearing data indicate that, this algorithm can accurately extract the fault frequency and improve EMD effectiveness and it has high engineering utility.