基于改进谐波小波和分形的碰摩故障诊断研究  

Rub-Impact Fault Diagnosis Based on Improved Harmonic Wavelet and Fractal

作　　者：宋圣霖 雷文平[1] 郝旺身[1] 陈宏[1] SONG Sheng-lin;LEI Wen-ping;HAO Wang-shen;CHEN Hong(institute of Vibration Engineering,Zhengzhou University,He'nan Zhengzhou 450001,China)

机构地区：[1]郑州大学振动工程研究所,河南郑州450001

出　　处：《机械设计与制造》2022年第6期180-183,共4页Machinery Design & Manufacture

基　　金：河南省高等学校精密制造技术与工程重点学科开发实验室开发基金资助项目(PMTE201301A)。

摘　　要：为了提高分形维数提取碰摩故障特征的区分度和稳定性,提出了一种基于改进谐波小波和分形的算法。采用高斯包络改进的谐波小波对旋转机械信号进行处理,运用G-P关联维数计算分形维数。在Bently转子实验台上分别采集正常工况、单点碰摩、不平衡激励的碰摩三种数据,计算得出改进谐波小波处理后的关联维数能够很好的识别出故障,保真性较高,稳定性相比原始信号以及谐波小波处理信号高29.17%、14.81%以上。结果表明,改进谐波小波和分形能够更准确稳定得诊断碰摩故障信息,优于传统的关联维数算法。In order to improve the discrimination and stability of the feature extraction of rub-impact fault by fractal dimension,an algorithm based on improved harmonic wavelet and fractal is proposed.The improved harmonic wavelet with Gauss envelope is used to process the signal of rotating machinery,and G-P correlation dimension is used to calculate the fractal dimension. In Bently rotor test-bed,three kinds of data of normal working condition,single point rub-impact and imbalanced rub-impact are collected,and the results show that the correlation dimension of the improved harmonic wavelet processing can well identify the fault and has a high fidelity.Compared with the original signal and harmonic wavelet processing signal,the stability is 29.17% and 14.81% higher.The results show that the improved harmonic wavelet and fractal are more accurately and stably than the traditional correlation dimension algorithm.

关 键 词：谐波小波 分形 关联维数 碰摩 故障诊断 

分 类 号：TH16[机械工程—机械制造及自动化] TH133