基于主轴振动信号的铣刀磨损监测研究

Research of milling cutter wear monitoring based on spindle vibration Signal

为实现复杂零件制造过程中的刀具磨损状态在线监测与识别,提出了一种基于机床主轴振动信号,利用近似解析小波包变换和径向基(RBF)神经网络的监测方法。采用近似解析小波包变换对监测信号进行多尺度分解,以峭度为指标寻找周期性冲击特征明显的频带,进而构造了包含8个频带重构信号的峰峰值和有效值的特征向量。建立刀具不同磨损状态下多特征向量样本数据库,训练RBF神经网络,实现刀具磨损状态的精确识别。

In order to realize on-line monitoring and identification of tool wear state in the manufacturing process of complex parts, a monitoring method based on machine tool spindle vibration signal and approximate analytical wavelet packet transform and RBF neural network is proposed. The approximated wavelet packet transform is used to perform multiscale decomposition of the monitoring signal. The kurtosis is used as the index to find the frequency band with obvious periodic impact characteristics. Then the eigenvectors of the peak-to-peak and RMS values of the reconstructed signals of the eight bands are constructed. A multi-feature vector sample database with different wear states of the tool is established ,and the RBF neural network is trained to accurately identify the degree of tool wear.