The Research On Monitoring Technology Of Micro Milling Tool Wear And Tool Broken Condition

Micro milling is a new processing technology which can directly produce high precision 3D free surface and complex micro parts,it has been widely used in the field of aerospace,medical,automotive,electronics and other fields.With the increase in the hardness of materials used in these areas as well as the miniaturization of the size of the micro milling cutter,the tool failure occurs more frequently than that of the ordinary materials.This paper focuses on the tool wear and tool breakage,and in order to establish the micro milling tool wear condition monitoring system,the micro milling tool wear and tool breakage monitoring algorithm is studied.The main research contents of this paper are as follows:(1)The research background at home and abroad of micro milling tool condition monitoring is presented.From four aspects of the selection of sensors,signal processing algorithms,feature selection algorithms and tool state recognition,the research background of micro milling tool condition monitoring is analyzed.An effective tool condition monitoring system is very necessary in micro milling process.(2)The micro milling tool life experiment is analysed,and the micro milling tool wear and broken experiment are designed.The boundary value of tool wear is determined by micro milling cutting tool life test.A Small micro milling test platform is established to designe micro milling tool wear and tool broken monitoring experimental system with three cutting parameters of spindle speed,cutting depth and feed rate.The vibration signal of the workpiece under different cutting tool wear and tool broken condition are collected.(3)The signal processing algorithm of micro milling tool condition monitoring is studied.The feature extraction method and feature selection method of tool condition monitoring signal are studied.First,the high frequency noise are eliminated using wavelet threshold denoising method for monitoring signal and improve the SNR;Then the feature extraction of the monitoring signal in time domain is carried out,and the feature that reflects the state of the tool are obtained.Finally,the statistical mean and variance of Lipschitz of signal are obtained using wavelet singularity analysis.The original state feature space composed of the effective time domain feature and the feature of singularity is obtained.(4)A feature selection algorithm based on principal component analysis is studied.The computational efficiency of the algorithm is reduced due to the high dimensionality of the original feature space,it is necessary to reduce the dimension of the feature space.The feature selection algorithm based on principal component analysis algorithm is studied to get the optimal feature subset reflect the tool state.(5)The pattern recognition algorithm of micro milling tool condition monitoring signal is studied and the algorithm is evaluated.Firstly,the principle of multi classification support vector machine is described,and the kernel function type of support vector machine is determined.The support vector machine identification model is established by the parameter optimization of genetic algorithm.Then,the state identification algorithm of the tool wear and tool broken is evaluated by using the test data.Finally,the recognition accuracy is improved by the adaptive optimization of the recognition model,and the method is applied to the micro milling tool condition monitoring.