基于LE-SVM的切削颤振识别

Cutting chatter recognition based on LE-SVM

为了减少切削颤振对机械加工的影响,提出了结合流形学习和支持向量机的颤振识别方法。首先采集切削加工中的振动信号,对信号进行小波滤波和归一化处理;然后对预处理后的信号进行数据流形降维;最后把降维后数据输入到支持向量机,建立切削加工振动信号的识别模型。试验结果表明,该方法具有较好的识别效果和较高的运算速度。

In order to reduce the vibration of cutting chatter in machine processing, a new method based on mani- fold learning and support-vector-machine is proposed. First, the signal in the cutting is acquired, and then is filtered and normalized. Finaly the preprocessed signal is processed by LE, and is applied as in- put of SVM to build the model of cutting vibration signals. The experimental results indicated that the new method has better recognition effect and higher computation speed.