Research On Tool Condition Monitoring Technology In Drilling

Tool condition monitoring technology is deemed as one of the essential technology in intelligent process,which has great significances for ensuring the machined surface quality,reducing production costs,improving production efficiency and achieving continuous automated machining.Drilling is a widely used processing method in the machinery manufacturing industry.However,the semi-closed or closed cutting conditions and special geometry of tool make the wear rate of drilling tool higher than other processing methods under the same cutting condition,and the tool condition cannot be directly observed.Based on the analysis of tool condition monitoring technology at home and abroad,this paper researches on the tool condition monitoring technology in drilling.The main contents and conclusion are shown as follows:(1)Through analyzing the rules of tool wear and the features of the drilling process,the experimental scheme and system for tool wear condition monitoring in drilling based on the cutting force signal and vibration signal are established.The relevant experiments are performed to collect the tool condition monitoring data.(2)Research on preprocessing technology of tool condition monitoring signal in drilling.This paper proposed to use the wavelet threshold denoising method to reduce the noise of the cutting force signal and the vibration signal.By analyzing the denoising results of monitoring signals under different wavelet basis functions and decomposition layers,the denoising method with the number of decomposition layer of 1 and wavelet basis function of coif1 and coif5 was determined to preprocess the monitoring signals.(3)Research on feature extraction technology of tool condition signal in drilling.In this paper,time domain,frequency domain and harmonic wavelet packet analysis technology are adopted to extract features of the processed signals.Through the correlation analysis between signal features and the tool state,a total of 31 typical features are selected as the input vector of tool state recognition.(4)Research on tool condition monitoring technology in drilling.An adaptive particle swarm optimization algorithm was introduced to automatically optimize the penalty factors?and kernel parameters2? of the kernel function of least squares support vector machine,and the optimal parameter combination ? =189.83 and2? =13.64 were selected.On this basis of this,a tool wear condition recognition model based on the least square support vector machine is established to recognize the tool wear condition.The obtained characteristic sample data is taken as the input vector and the corresponding tool wear state is taken as the output vector.The results show that the average absolute error of the drilling tool wear condition recognition model is 0.91% and the model has higher recognition accuracy through the reasonable feature selection methods and identification model parameter optimization.