基于MFCC与PCA的滚动轴承故障诊断  

Fault Diagnosis of Rolling Bearing Based on MFCC and PCA

作　　者：王前 于嘉成 宁永杰 

机构地区：[1]中国矿业大学信息与控制工程学院,江苏徐州221116

出　　处：《组合机床与自动化加工技术》2017年第12期103-105,共3页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家重点研发计划资助(2017YFC0804404)

摘　　要：针对振动传感器不易安装、传统分类算法训练时间较长等问题,提出了基于美尔倒谱系数(MFCC)与主成分分析(PCA)的滚动轴承故障诊断方法。首先利用声音传感器采集滚动轴承声音信号,而后提取声音信号的MFCC特征,最后将MFCC特征作为PCA分类器的输入进行故障分类,并与反向传播神经网络(BPNN)、支持向量机(SVM)进行比较研究。实验结果表明:MFCC系数可以有效反应轴承不同工作状态下的信号特征;基于MFCC与PCA的轴承故障诊断方法能够准确、有效地识别轴承故障类型。Aimed at the problem that the vibration sensor is difficult to be installed and the conventional classification algorithms take a long time to be trained,this paper proposes a method of fault diagnosis for rolling bearing based on MFCC and PCA. Firstly,acoustic sensor is used to acquire acoustic signals of rolling bearing. Then,MFCC features are extracted from the signals. Finally,MFCC features are used as the inputs of PCA classifier for fault classification and its performance is compared with BPNN and SVM. The experimental result shows that MFCC can represent the features of rolling bearing in different conditions and the method based on MFCC and PCA can detect the fault category accurately and effectively.

关 键 词：轴承 故障诊断 美尔倒谱系数 

分 类 号：TH165.3[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]