汇报题目：A HYBRID SPARSE REPRESENTATION METHOD FOR DIAGNOSIS OF BEARINGS
汇报时间：2019年7月1日（星期一） 20:00
汇报地点：科技园西五楼北楼228会议室
汇报人：于金涛
会议名称：2019 International Conference on Structural Engineering Dynamics
会议时间：24-26 June 2019
会议地点：Viana do Castelo, Portugal
会议简介：
ICEDyn 2019 (International Conference on Structural Engineering Dynamics)是两年一届的结构健康监测的国际盛会。会议汇聚了世界各国的研究学者，在几天的时间中一起研讨交流学术思想。
会议交流工作
Oral Presentation: A HYBRID SPARSE REPRESENTATION METHOD FOR DIAGNOSIS OF BEARINGS
报告人：于金涛
参加论文信息
Title: A HYBRID SPARSE REPRESENTATION METHOD FOR DIAGNOSIS OF BEARINGS
Author: Jintao Yu, Zhibo Yang, Ruobin Sun, Junpeng Zhang, Baijie Qiao
Abstract: Bearings are one of the most important components for rotating machinery, the faults of which usually result in the failure on structural level. Condition monitoring is now a popular tool to reduce the possible losses caused by faults. The bearings, however, always operate in non-stationary condition, which induces the failure of classical demodulation methodologies. Sparse representation method is a new and alternative tool for above problem, whose superior performance has been verified in non-stationary condition. Nevertheless, the implementation of sparse representation algorithm requires a non-real-time calculation as it is trying to match the inspected signal with the atoms in a redundant dictionary. The key problem of sparse representation is to design an appropriate dictionary for inspected signal. Large dictionary reduces efficiency and small dictionary decreases accuracy. Hence, in order to address this issue, this paper proposed a hybrid approach to improve the sparse representation. We construct the over-complete dictionary by the unit impulse response function of underdamped second-order mass-spring-damper system. The modal parameters, including natural frequency and relative damping ratio, are directly identified through Ensemble Empirical Mode Decomposition (EEMD) and Random Decrement (RD) technique. The effectiveness and superiority of the proposed method are verified by simulation and experiment.
