摘要
针对转子系统偏心、不对中及碰磨引起的耦合故障,采用卷积混合的频域盲源分离转子故障诊断方法。在利用转子实验台模拟转子系统偏心、偏心—碰磨耦合故障的基础上,通过频域盲解卷方法对耦合故障的观测信号进行分离,并对分离后的源信号进行FFT分析,从而实现转子系统复杂耦合故障的精确定位与诊断。仿真实验初步验证了算法的有效性,分离波形和二次残差(VQM)均证实了该算法能较好实现混合信号分离。同时,转子系统故障诊断实例结果表明:该方法在成功诊断出偏心故障与碰磨故障的同时还能诊断出不对中、轴承松动的故障,提高了故障诊断的精确性。
Aiming at the eccentricity, misalignment and rubbing faults in rotor system, a vibration fault diagnosis methodbased on multi-signal hybrid convolution and blind de-convolution（BD）was proposed. The rotor test bench was used tosimulate the faults such as eccentricity and eccentricity-rubbing coupling. Then, the fault signals were separated through BDmethod in frequency-domain, and the separated fault signals were diagnosed by using FFT. Therefore, the precise localizationand diagnosis of the complex coupling faults in the rotor system was realized. The simulation experiments verified theeffectiveness of the algorithm. The separated waveforms and secondary residuals show that the algorithm can separate themixed signals effectively. The results of the rotor system fault diagnosis show that this method can diagnose the faults of eccentricity,rubbing, misalignment and bearing looseness, and improve the accuracy of fault diagnosis.
出处
《噪声与振动控制》
CSCD
2015年第1期219-224,共6页
Noise and Vibration Control
关键词
振动与波
故障诊断
频域盲解卷积法
转子偏心
不对中
vibration and wave
fault diagnosis
blind de-convolution in frequency-domain
rotor eccentricity
misalinment
