基于LabVIEW的轴心轨迹故障自动识别系统

摘要

针对目前旋转机械故障诊断的计算量大、识别准确度不高、自动化程度低等问题,提出一种基于LabVIEW的轴心轨迹故障自动识别的新方法.对比小波与传统去噪算法,选用效果更优的小波提纯仿真轴心轨迹.通过改进的HU不变矩函数提取轴心轨迹的特征值,保证比例缩放不变性.两路相互垂直的位移传感器连接西门子LMS采集振动信号,结合关联度算法,在LabVIEW轴心轨迹故障自动识别系统上进行转子不对中故障测试,识别的结果与外8字轴心轨迹关联度高达97%,同时信号的Matlab时域轴心轨迹图为外8字,信号频谱图主要为一倍频和二倍频,均符合转子不对中故障特征.结果表明:该系统能够进行在线故障识别,为旋转机械的智能故障诊断提供参考依据.%For the problems of large computational quantity, low recognition precision and low automatization of rotating machinery fault diagnosis, a new method of automatic recognition system for shaft orbit faults based on LabVIEW is proposed. The wavelet and traditional denoising algorithms are compared and the shaft orbit simulation with wavelet purification of better effects is selected. The characteristic value of the shaft orbit is extracted by the improved HU invariant-moment function to ensure the invariance of scaling. Two mutually vertical displacement sensors are connected with Siemens LMS to acquire vibration signals. With the correlation degree algorithm, fault test is conducted for the automatic recognition system for shaft orbit faults based on LabVIEW. The results show that the correlation between the recognized results and the external 8-character shaft orbit reaches as high as 97%. Meanwhile, the signal's Matlab time domain shaft orbit is external 8-character and the signal frequency spectrum is mainly of one time frequency and doubled frequency, fully according with the fault characteristic of rotor misalignment. The results show that the system can recognize the faults on line and it provides a reference for intelligent fault diagnosis of rotating machinery.