基于迭代自适应多尺度形态滤波的滚动轴承故障诊断  

Iterative Adaptive Multiscale Morphological Filtering-based Fault Diagnosis for Rolling Bearings

作　　者：姜万录[1,2] 李扬[1,2] 郑直[1,2] 朱勇[1,2] 

机构地区：[1]燕山大学河北省重型机械流体动力传输与控制重点实验室,秦皇岛066004 [2]先进锻压成形技术与科学教育部重点实验室(燕山大学),秦皇岛066004

出　　处：《中国机械工程》2015年第23期3192-3199,共8页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(51475405);国家重点基础研究发展计划(973计划)资助项目(2014CB046405);河北省自然科学基金资助项目(E2013203161)

摘　　要：针对工业现场强噪声背景下振动信号特征信息提取困难和单尺度形态滤波时尺度选择的盲目性和随意性的问题,基于自适应多尺度形态分析(AMMA)的思想提出了一种迭代自适应多尺度形态分析(IAMMA)的滤波方法。该方法对振动信号进行多尺度形态差值迭代运算,每次采用的结构元素尺度逐渐增大,然后求多次滤波结果的平均值,达到滤除噪声成分的目的。对仿真信号和滚动轴承故障信号进行分析,结果表明,IAMMA较AMMA能够选取更为合适的结构元素尺度,提取更多的故障特征信息,滤波效果更佳,与Hilbert包络解调方法相比处理过程更加简捷,从而为轴承的故障诊断提供了一种有效的方法。Aiming at the problems of extracting characteristics from vibration signals in the intense industrial background noise and the blindness and randomness of single scale morphology filter in scale selection,a novel filtering method of iterative adaptive multiscale morphology analysis（IAMMA） was proposed.The method was based on adaptive multiscale morphology analysis （AMMA）.In order to filter out noise components,the multiscale morphological difference iteration was used to process the signals with gradually increase structure elements.Then,the average value of multiple filtering reaults was calculated.Furthermore,the simulation signals and fault signals of the rolling bearings were analyzed.The results demonstrate that,comparing with AMMA,the IAMMA can select better scale and extract more feature information and get better filtering results.Moreover,comparing with Hilbert,the processing procedure of IAMMA is more simple.It is a valid method for fault diagnosis of rolling bearings.

关 键 词：迭代自适应 多尺度 形态差值滤波 滚动轴承 故障诊断 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]