航空学报 > 2020, Vol. 41 Issue (10): 423658-423658   doi: 10.7527/S1000-6893.2020.23658

一种MED最优滤波长度选择新方法及其应用

贺志远1, 陈果1, 何超1, 滕春禹2   

  1. 1. 南京航空航天大学 民航学院, 南京 211106;
    2. 中航工业中国航空综合技术研究所, 北京 100028
  • 收稿日期:2019-11-15 修回日期:2020-03-03 发布日期:2020-02-27
  • 通讯作者: 陈果 E-mail:cgzyx@263.net
  • 基金资助:
    国家自然科学基金(51675263);国家科技重大专项(2017-IV-0008-0045)

MED optimal filter length selection: New method and applications

HE Zhiyuan1, CHEN Guo1, HE Chao1, TENG Chunyu2   

  1. 1. College of Civil Aviation, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China;
    2. AVIC China Aero-polytechnology Establishment, Beijing 100028, China
  • Received:2019-11-15 Revised:2020-03-03 Published:2020-02-27
  • Supported by:
    National Natural Science Foundation of China (51675263);National Science and Technology Major Project (2017-IV-0008-0045)

摘要: 最小熵解卷积(MED)是旋转机械故障诊断领域广泛应用的有效方法,它可以从噪声中提取微弱的故障冲击成分。然而它的有效性依赖于滤波长度的选取,目前,针对MED滤波长度的自动选取并没有明确有效的方法,往往需要人为经验选择。因此,在MED的算法基础上,通过结合自相关函数,提出了一种MED最优滤波长度选择的新方法,该方法构建了一个能量判定标准来衡量输出信号的周期性,从而自适应地确定MED的最优的滤波长度以提升微弱故障信号中的周期脉冲成分,避免MED方法容易出现最大化单一随机脉冲现象的发生。该方法应用于滚动轴承故障微弱冲击特征提取,并利用两个实例进行了有效性验证:基于辛辛那提试验中心的滚动轴承全寿命疲劳加速试验;带机匣的航空发动机转子试验器模拟远离轴承振动源的故障试验。结果表明,所提方法可以消除传递路径影响,提升微弱冲击周期性特征,并且与最大相关峭度解卷积(MCKD)方法相比,诊断结果更具优势。

关键词: 最小熵解卷积(MED), 滤波长度, 滚动轴承, 微弱故障, 故障检测

Abstract: Minimal entropy deconvolution (MED), an effective method widely used in the field of rotating machinery fault diagnosis, can extract weak fault impact components from noise. However, its effectiveness depends on the selection of filter length. Currently, no clear and effective method exists for the automatic selection of MED filter length, thus often requiring experience based human selection. Based on the MED algorithm and combining the autocorrelation function, a new method for MED optimal filter length selection is proposed. It constructs an energy target to measure the periodicity of the output signal and adaptively determine the filter length to improve the periodic pulse components in the weak fault signal, hence avoiding maximization of a single random pulse. The effectiveness of the method was verified in two cases:the rolling bearing test in the Cincinnati based test center; the fault test of an aero-engine rotor tester with casing simulating the case of the signal being far away from the rolling bearing. The results show that the proposed method can eliminate the effect of the transmission path and enhance the weak periodic impulses. Compared with those of the maximum correlated kurtosis deconvolution method, the diagnostic results are more advantageous.

Key words: minimal entropy deconvolution (MED), filter length, rolling bearing, weak fault, fault detection

中图分类号: