0
Research Papers

Enhanced Rotating Machine Fault Diagnosis Based on Time-Delayed Feedback Stochastic Resonance

[+] Author and Article Information
Siliang Lu, Haibin Zhang, Fanrang Kong

Department of Precision Machinery
and Precision Instrumentation,
University of Science and Technology of China,
Hefei, Anhui 230026, China

Qingbo He

Department of Precision Machinery
and Precision Instrumentation,
University of Science and Technology of China,
Hefei, Anhui 230026, China
e-mail: qbhe@ustc.edu.cn

1Corresponding author.

Contributed by the Technical Committee on Vibration and Sound of ASME for publication in the JOURNAL OF VIBRATION AND ACOUSTICS. Manuscript received June 3, 2014; final manuscript received April 3, 2015; published online May 20, 2015. Assoc. Editor: Patrick S. Keogh.

J. Vib. Acoust 137(5), 051008 (Oct 01, 2015) (12 pages) Paper No: VIB-14-1202; doi: 10.1115/1.4030346 History: Received June 03, 2014; Revised April 03, 2015; Online May 20, 2015

The fault-induced impulses with uneven amplitudes and durations are always accompanied with amplitude modulation and (or) frequency modulation, which leads to that the acquired vibration/acoustic signals for rotating machine fault diagnosis always present nonlinear and nonstationary properties. Such an effect affects precise fault detection, especially when the impulses are submerged in heavy background noise. To address this issue, a nonstationary weak signal detection strategy is proposed based on a time-delayed feedback stochastic resonance (TFSR) model. The TFSR is a long-memory system that can utilize historical information to enhance the signal periodicity in the feedback process, and such an effect is beneficial to periodic signal detection. By selecting the proper parameters including time delay, feedback intensity, and calculation step in the regime of TFSR, the weak signal, the noise, and the potential can be matched with each other to an extreme, and consequently a regular output waveform with low-noise interference can be obtained with the assistant of the distinct band-pass filtering effect. Simulation study and experimental verification are performed to evaluate the effectiveness and superiority of the proposed TFSR method in comparison with a traditional stochastic resonance (SR) method. The proposed method is suitable for detecting signals with strong nonlinear and nonstationary properties and (or) being subjected to heavy multiscale noise interference.

Copyright © 2015 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

The system model of the TFSR

Grahic Jump Location
Fig. 2

Potential shape with different time-delayed feedback parameters: (a) fixed τ and varying β and (b) fixed β and varying τ

Grahic Jump Location
Fig. 3

SNR versus noise intensity D

Grahic Jump Location
Fig. 4

SNR versus feedback intensity β at A = 0.001 and D = 0.1

Grahic Jump Location
Fig. 5

SNR versus time delay τ at A = 0.001 and D = 0.1

Grahic Jump Location
Fig. 6

Analyzed results of the unilateral attenuation impulses (equal amplitudes and durations) by using different methods: (a) pure signal; (b) noisy signal; (c) envelope signal; (d) optimal output of traditional SSR at h = 0.012 and b = 43500; and (e) optimal output of proposed TFSR method at h = 0.00001, β = −3000, and τ = 0.00043

Grahic Jump Location
Fig. 7

Analyzed results of the unilateral attenuation impulses (unequal amplitudes and durations) by using different methods: (a) pure signal; (b) noisy signal; (c) envelope signal (filter band (1000, 3000) Hz); (d) optimal output of traditional SSR at h = 0.032 and b = 300; and (e) optimal output of proposed TFSR method at h = 0.00002, β = −1500, and τ = 0.00085

Grahic Jump Location
Fig. 8

Pass band comparison between the envelope analysis method, the SSR method (h = 0.00005 and b = 1000), and the TFSR method (h = 0.00005, β = −2500, and τ = 0.0005)

Grahic Jump Location
Fig. 9

The bearing test stand from CWRU

Grahic Jump Location
Fig. 10

Analyzed results of the outer raceway defective signals by using different methods: (a) original signal; (b) envelope signal (filter band (2500–3250) Hz); (c) optimal output of traditional SSR at b = 2970 and h = 0.03; and (d) optimal output of proposed TFSR method at h = 0.00002, β = −2000, and τ = 0.0031

Grahic Jump Location
Fig. 11

Analyzed results of the inner raceway defective signals by using different methods: (a) original signal; (b) envelope signal (filter band (2500–4000) Hz); (c) optimal output of traditional SSR at b = 4940 and h = 0.06; and (d) optimal output of proposed TFSR method at h = 0.00002, β = −4000, and τ = 0.0015

Grahic Jump Location
Fig. 12

Analyzed results of the rolling element defective signals by using different methods: (a) original signal; (b) envelope signal (filter band (2500–3500) Hz); (c) optimal output of traditional SSR at b = 915 and h = 0.011; and (d) optimal output of proposed TFSR method at h = 0.000035, β = −2000, and τ = 0.0018

Grahic Jump Location
Fig. 13

The automobile transmission gearbox: (a) structure of the gearbox and (b) gearbox setup

Grahic Jump Location
Fig. 15

Performance comparison among the envelope spectral analysis method, the traditional SSR method, and the proposed TFSR method for addressing different practical defective signals

Grahic Jump Location
Fig. 14

Analyzed results of the defective gearbox signals by using different methods: (a) original signal; (b) envelope signal (filter band (150–350) Hz); (c) optimal output of traditional SSR at b = 88 and h = 0.021; and (d) optimal output of proposed method at h = 0.00002, β = −2000, and τ = 0.012

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In