两级自适应调频模式分解-同步提取变换的故障诊断方法
Fault Diagnosis Method Based on Two-level Adaptive Chirp Mode Decomposition and Synchro-extracting Transform
葛丽英 1李志农 1胡志峰 1毛清华 2张旭辉2
作者信息
- 1. 南昌航空大学 无损检测技术教育部重点实验室,南昌 330063
- 2. 西安科技大学 陕西省矿山机电装备智能监测重点实验室,西安 710054
- 折叠
摘要
同步提取变换(Synchroextracting Transform,SET)处理强干扰信号分量时缺乏自适应性而易发生频率模糊,导致难以精确提取快速波动的瞬时频率.针对此问题,结合自适应调频模式分解(Adaptive Chirp Mode Decomposition,ACMD)的自适应先验信息和贪婪算法的优势,将ACMD引入到SET中,构造一种两级ACMD-SET故障诊断方法.在提出的方法中,将基于基尼指数(Gini Index,GI)最大化准则的分量选择重组算法和第一级ACMD结合,提取出强干扰下的多模态故障脉冲信号的模式.然后,利用SET对第二级ACMD分离出的时变频率故障特征进行高精度的时频表示.将此方法应用到仿真调频-调幅信号中,得到高分辨率的故障特征,方法的有效性得到验证.最后,将所提方法应用于航空发动机高速滚动轴承的振动信号分析中,结果表明,所提方法能有效地提取高速滚动轴承振动信号的时变故障特征频率,其结果明显优于SET方法.
Abstract
Synchroextracting transform(SET)lacks adaptivity in handling strong interfering signal components,which leads to frequency ambiguity and makes it difficult to accurately extract the instantaneous frequencies with rapid fluctuations.In view of this problem,a fault diagnosis method is constructed by introducing adaptive chirp mode decomposition(ACMD)into the SET.The adaptive prior information of ACMD is combined with the advantages of the greedy algorithm.And a double grade fault diagnosis method is established based on ACMD-SET.In this method,the component selection recombination algorithm based on the Gini index(GI)maximization criterion and the Level 1 ACMD are combined to extract the patterns of multi-modal fault pulse signals with strong disturbances.The time-varying fault feature frequencies separated by level 2 ACMD are then represented in time-frequency with high accuracy using SET.To verify the effectiveness,high-resolution fault features of a simulated AM-FM signal are obtained by the proposed method.Finally,this method is applied to the vibration signal analysis of high-speed rolling bearings in aero-engines.The results show that the proposed method can effectively extract the time-varying fault characteristic frequencies of high-speed rolling bearing vibration signals,and the obtained results are significantly better than the SET method.
关键词
故障诊断/同步提取变换/自适应调频模态分解/滚动轴承Key words
fault diagnosis/synchroextracting transform/adaptive chirp mode decomposition/rolling bearing引用本文复制引用
基金项目
国家自然科学基金(52075236)
江西省自然科学基金重点项目(20212ACB202005)
陕西省矿山机电装备智能监测重点实验室开放基金(SKL-MEEIM201901)
出版年
2024
国家科技期刊平台
NSTL
万方数据