A mechanical failure early warning methodology based on dynamic self-learning threshold and trend filtering techniques

ZHANG Ming;FENG Kun;JIANG Zhi-nong

Journal of Vibration and Shock ›› 2014, Vol. 33 ›› Issue (24) : 8-14.

PDF(1513 KB)
PDF(1513 KB)
Journal of Vibration and Shock ›› 2014, Vol. 33 ›› Issue (24) : 8-14.
论文

A mechanical failure early warning methodology based on dynamic self-learning threshold and trend filtering techniques

  • ZHANG Ming, FENG Kun, JIANG Zhi-nong
Author information +
History +

Abstract

Because that the alarm mode of current online monitoring system is hard to warn early mechanical failure, an early warning methodology is proposed in this paper. Based on statistical analysis of mass data in online monitoring system, taking advantage of dynamic self learning threshold algorithm calculates warning threshold, and using ℓ1 trend filtering technology obtains filtered trend which has eliminated random errors. With dynamic self-learning threshold instead of general alarm threshold on monitoring system, early warning can be acquired by comparing the self-learning warning threshold and filtered trend. It can be shown form the project cases that early warning of mechanical failure can be achieved by the proposed method, which plays an important role in preventing occurrence of mechanical failure.

Key words

self-learning threshold / fault early warning / nonparametric test / beta distribution / / 1 trend filtering

Cite this article

Download Citations
ZHANG Ming;FENG Kun;JIANG Zhi-nong. A mechanical failure early warning methodology based on dynamic self-learning threshold and trend filtering techniques[J]. Journal of Vibration and Shock, 2014, 33(24): 8-14
PDF(1513 KB)

Accesses

Citation

Detail

Sections
Recommended

/