Abstract:

Against the problem of traditional grey relation analysis method, a novel method was proposed by introducing dynamic identification and factor weight coefficient. Compared with traditional method,this method has two advantages: ①
reducing the dependence of man-made identification and weight coefficient; ②strengthening  the reliability and veracity. Application of this method to fault recognition of NC machine tool spindle shows a better effect on the reliability and
veracity over traditional method and neural network,this method can also be applied to many other fields.

Key words: fault recognition, grey relation analysis, dynamic identification coefficient, factor weight coefficient, NC machine tool

摘要：

针对传统灰关联分析方法存在的问题,引入动态分辨系数和因子权重系数,提出新型灰关联分析方法。与传统灰关联分析方法相比，该方法具有两个优点:一是降低对人为确定分辨系数和权重系数的依赖性;二是提高识别结果的可靠性和准确性。最后将该方法应用于数控机床主轴故障识别中，并与传统灰关联分析方法和神经网络识别结果进行对比分析。结果显示，新型灰关联分析方法识别结果更准确可靠。

关键词: 故障识别, 灰关联分析, 动态分辨系数, 因子权重系数, 数控机床