基于自回归模型的螺纹连接松动诊断  

Autoregressive Model-based Bolted Joints Fault Diagnosis

作　　者：张言[1] 曾国英[1] 赵登峰[1] 栗明燕[1] 

机构地区：[1]西南科技大学制造科学与工程学院,绵阳621010

出　　处：《机械科学与技术》2012年第3期483-486,共4页Mechanical Science and Technology for Aerospace Engineering

基　　金：国家自然科学基金-中国工程物理研究院联合基金项目(10876034)资助

摘　　要：应用自回归模型分析法兰螺纹连接上下壳体的差值信号,进行螺纹连接松动诊断。在螺纹连接模型上,改变某一个螺钉的预紧力,将正常工况下的差值信号作为参考总体,并以此建立AR模型,其他预紧力下的差值信号为待检总体,分别计算参考总体与待检总体之间的Euclide距离和Mahalanobis距离。分析结果表明:基于自回归模型的几何距离是有效的螺纹连接松动指标,可以用于螺纹连接故障诊断。An autoregressive model-based technique is proposed to detect flange screw-connected joints fault.The AR model can be validated by reference population which is defined as difference signal of normal conditions.The difference signal obtained in other preload conditions is defined as inspection population.The analyses of Euclide distance and Mahalanobis distance between reference population and inspection population indicate that autoregressive model-based geometrical distance is considered as effective index of screw-connected joints fault and it can be used to detect screw-connected joints fault.

关 键 词：自回归模型 螺纹连接 故障诊断 预紧力 几何距离 

分 类 号：TH131[机械工程—机械制造及自动化]