基于双谱的刀具振动信号特征值提取

Using Bispectrum Analysis to Extract Characteristic Values of the Tool Vibration Signals

针对数控机床刀具磨损故障诊断中振动信号特征值提取问题,提出了使用双谱间接估计法对振动信号进行分析处理来提取特征值的方法。对于双谱分析结果数据量庞大的问题,使用奇异值分解的方法对双谱分析后的结果进一步压缩提取优化,从而得到不同磨损阶段的刀具振动信号特征值,为后期的模式识别系统提供了数据基础。

In view of the problem that characteristic value was extracted from tool vibration signals in CNC machine tool wear fault diagnosis,the method was put forward that vibration signal was processed by indirect estimation method of bispectrum analysis.Then,the bispectmm result was decomposed by singular value decomposition for solving its vast data,resulting in different stages of the wear of cutting tool vibration signal characteristic value.So,there is data foundation for late pattern recognition system.