Cutter state monitoring method

The invention discloses a cutter state monitoring method, which comprises the following steps of: firstly, processing a material for multiple times by using a group of constant cutting parameters, dividing a cutter wear stage, determining a blunt grinding standard of a cutter, performing mechanism analysis on cutter wear, and determining main physical quantity factors influencing the cutter wear; secondly, cutter wear experiments under different cutting parameters are designed, signals of main physical quantity factors are collected, and the cutter wear loss after feeding each time is measured; thirdly, the collected signals are processed; and finally, a BP neural network is used for establishing a tool wear loss prediction model, a multi-granularity scanning cascade forest method is used for establishing a tool wear stage identification model, and the tool state is monitored. 本发明公开了一种刀具状态监测方法，首先用一组恒定的切削参数对材料进行多次加工，划分刀具磨损阶段并确定刀具的磨钝标准，对刀具磨损进行机理分析，确定影响刀具磨损的主要物理量因素；其次设计不同切削参数下的刀具磨损实验，采集主要物理量因素的信号并测量每次走刀后的刀具磨损量；再