基于图论诊断法与小波包变换的数控机床进给系统故障诊断研究

Fault diagnosis of NC machine tool feed system based on graph theory diagnosis method and wavelet packet transform

在复杂机械系统未知故障诊断中,单一故障诊断方法诊断精度低、推理速度慢,不利于工程实际应用。针对这一问题,提出了基于小波包变换与图论诊断法的数控机床进给系统故障联合诊断方法。首先利用图论诊断法建立数控机床进给系统的故障模型,并根据故障原因定位算法快速确定引起故障发生的可能原因,缩小诊断范围;然后对振动信号进行小波包变换,进一步精确定位故障原因,最终实现数控机床进给系统中未知故障的准确识别与诊断。试验证明,基于图论诊断法与小波包变换的故障联合诊断方法,弥补了图论诊断法故障冗余解多和诊断精度低的缺陷,实现了未知故障的快速筛选与精确定位,可以高效、准确的诊断数控机床进给系统中的未知故障。

In unknown fault diagnosis of complex mechanical system,the single fault diagnosis method has low diagnosis accuracy and slow reasoning speed,it is not conducive to practical application of engineering.Here,to solve this problem,a joint fault diagnosis method for NC machine tool feed system based on the wavelet packet transform and the graph theory diagnosis method was proposed.Firstly,the fault model of NC machine tool feed system was established by using the graph theory diagnosis method,and possible causes of fault were quickly determined according to the fault cause locating algorithm to narrow diagnosis range.Then,vibration signal was transformed using wavelet packet transform to further accurately locate fault cause.Finally,the correct identification and diagnosis of unknown faults were realized in the feed system of NC machine tool.Tests showed that the joint fault diagnosis method based on graph theory diagnosis method and wavelet packet transform can make up for the graph theory diagnosis method’s defects of many redundant solutions and low diagnosis accuracy,realize rapid screening and accurate positioning of unknown faults,and efficiently and accurately diagnose unknown faults in NC machine tool feed system.