A procedure for weighted summation of the derivatives of reflection coefficients in adaptive Schur filter with application to fault detection in rolling element bearings

作者：Makowski Ryszard; Zimroz Radoslaw^*

来源：Mechanical Systems and Signal Processing, 2013, 38(1): 65-77.

DOI：10.1016/j.ymssp.2012.05.005

• Damage detection
• Vibration analysis
• Adaptive Schur filter
• Reflection coefficients
• Weighted sum

A procedure for feature extraction using adaptive Schur filter for damage detection in rolling element bearings is proposed in the paper. Damaged bearings produce impact signals (shocks) related with local change (loss) of stiffness in pairs: inner/outer race-rolling element. If significant disturbances do not occur (i.e. signal to noise ratio is sufficient), diagnostics is not very complicated and usually envelope analysis is used. Unfortunately, in most industrial examples, these impulsive contributions in vibration are completely masked by noise or other high energy sources. Moreover, impulses may have time varying amplitudes caused by transmission path, load and properties of noise changing in time. Thus, in order to extract time varying signal of interest, the solution would be an adaptive one. The proposed approach is based on the normalized exact least-square time-variant lattice filter (adaptive Schur filter). It is characterized by an extremely fast start-up performance, excellent convergence behavior, and fast parameter tracking capability, making this approach interesting. Schur adaptive filter consists of P sections, estimating, among others, time-varying reflection coefficients (RCs). In this paper it is proposed to use RCs and their derivatives as diagnostic features.
However, it is not convenient to analyze simultaneously P signals for P sections, so instead of these, weighted sum of derivatives of RCs can be used. The key question is how to find these weight values for summation procedure. An original contributions are: application of Schur filter to bearings vibration processing, proposal of several features that can be used for detection and mentioned procedure of weighted summation of signal from sections of Schur filter.
The method of signal processing is well-adapted for analysis of the non-stationary time-series, so it sounds very promising for diagnostics of machines working in time varying load/speed conditions.

• 出版日期2013-7-5