基于非线性Volterra核辨识的转子裂纹故障诊断方法研究  

Fault Diagnosis Method Based on Volterra Kernel Identification for Rotor Crack

作　　者：蒋静[1,2] 李志农[1,2] 唐高松[2] 

机构地区：[1]南昌航空大学无损检测技术教育部重点实验室,江西南昌330063 [2]郑州大学机械工程学院,河南郑州450001

出　　处：《机床与液压》2010年第23期138-140,147,共4页Machine Tool & Hydraulics

基　　金：国家自然科学基金项目(50775208);河南省教育厅自然科学基金项目(2006460005;2008C460003)

摘　　要：Volterra级数模型是一类能够完全描述系统的非线性传递特性的模型。将Volterra核的辨识应用于转子裂纹的故障诊断,提出一种基于Volterra核的转子系统故障诊断方法。该方法利用系统的输入输出同步采样信号,采用RLS算法进行非线性系统Volterra时域核的辨识,利用得到的一阶核、二阶核和三阶核来判断系统当前所处的状态。以正常转子和有裂纹转子为例进行实验验证,实验结果表明:该方法不仅可以从一阶核的特性来判断系统的变化,而且还可以从二阶核、三阶核等来判断系统的变化,提供更丰富的故障信息。The transmission characteristic of nonlinear system can be completely described by Volterra series.Volterra kernels identification was used to the fault diagnosis of rotor crack,and a fault diagnosis method of rotor crack based on Volterra kernels was proposed.In the method,the input and output signals were synchronously sampled and used to identify Volterra series model of nonlinear system by recursive least square（RLS） algorithm,then the current state of system were judged by the obtained Volterra kernels.The proposed method was verified by the example for normal rotor and cracked rotor.The experiment results show that in the proposed method,the first order kernel,the second order kernel and third order kernel can be used to determine the change of system,and it provides abundant fault information.

关 键 词：转子裂纹 VOLTERRA级数 递推最小二乘算法 故障诊断 系统辨识 

分 类 号：TH17[机械工程—机械制造及自动化] TN911[电子电信—通信与信息系统]