特征筛选与SVM结合的风机轴承故障诊断研究  被引量:11

Research on Bearing Fault Diagnosis of Fan Based on Feature Selection and SVM

在线阅读下载全文

作  者:常新宇 李琦[2] CHANG Xin-yu;LI Qi(School of Chemical Process Automation Shenyang University of Technology,Liaoning Shenyang 110023,China;School of Mechanical Engineering and Automation Northeastern University,Liaoning Shenyang 110819,China)

机构地区:[1]沈阳工业大学化工过程自动化学院,辽宁沈阳110023 [2]东北大学机械工程与自动化学院,辽宁沈阳110819

出  处:《机械设计与制造》2022年第10期71-74,80,共5页Machinery Design & Manufacture

摘  要:由于城市轨道交通频繁运行,地铁风机滚动轴承的故障特征极易被复杂多变的背景噪声所掩盖。针对这一问题,提出了一种基于特征筛选与支持向量机(SVM)的故障诊断方法。首先,从轴承原始监测信号中提取多尺度特征,构建轴承常见故障状态下的健康指标特征集;其次,利用拉普拉斯评分对处于故障状态下的健康指标特征集与正常状态下的健康指标特征集进行对比,获得健康指标敏感程度权重分数,筛选出敏感故障特征;最后,运用SVM算法对筛选出的特征进行故障识别,从而准确地实现地铁风机滚动轴承的故障诊断。通过在实际轴承故障数据集上的故障诊断实验,证明了提出方法的有效性和优越性。Due to frequent operation of urban orbit traffic,the fault feature of rolling bearing for the metro fan is extremely easy to be masked by complex and variable background noise.To solve this problem,a fault diagnosis method based on feature selection and support vector machine(SVM)is proposed.Firstly,the multi-scale features from the original bearing monitoring signals are extracted,which is used to construct a set of health indicators under health fault conditions of bearings;Secondly,the health indicators feature sets under fault conditions and those under normal conditions are compared by Laplace scores.The feature set is compared to obtain the weight score of the health index sensitivity degree,and the sensitive features are selected out;finally,the SVM algorithm is used to identify the faults of the selected features,so as to accurately realize the fault diagnosis of the rolling bearing of the metro fan.Through the fault diagnosis experiment on the actual bearing fault dataset,the effectiveness and superiority of the proposed method are proved.

关 键 词:地铁风机 轴承 敏感特征 故障诊断 

分 类 号:TH16[机械工程—机械制造及自动化] TP311.52[自动化与计算机技术—计算机软件与理论]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象