Research On Sensorless Monitoring Technology Of The Drill Wear State In The BTA Deep-hole Drilling Process

With the application of advanced manufacturing technology,there will be only a few or no more people take part in the machining process while manufacturing tasks will replaced by automation equipment.While the tool is the weakest link in the entire automated manufacturing system.Thus,to make the process automation and stability,research and development on tool state monitoring technology in the cutting process is particularly important.In this paper,the BTA deep hole drilling process,whose cutting area is a closed space,was interned as the research object.In view of the problem that current research on sensor acquisition data and the need for an external decision module,We proposed the scheme that extract spindle motor current signal in machine process from the numerical control system by CNC system its abundant condition information,based on the intrinsic relationship between current signal characteristic parameters and drill wear state,automatic identify the drill wear state in machie process of sensorless monitoring.What's more,the extraction of the feature parameters and the intelligent recognition technology based on the feature parameters are studied.In order to facilitate offline signal research,current signal acquisition system based on CNC system communication module of the spindle motor is established,and the spindle motor current signal and bit wear law information can be obtained by drilling experiment.Processing and analyzing of the spindle motor current signal in the time domain and frequency domain respectively.Study on the signal and its characteristic parameters change rules in the process of drilling,discusses the relationship between the characteristic parameters and BTA drill wear.The results show that the time domain statistical features such as the mean,RMS,kurtosis and skewness is relevant to the drill wear rules,the kurtosis and skewness as dimensionless parameters can better reflect the essential characteristics of the signal.In the frequency domain,the power spectral density of spindle motor current is relevant to the drill wear rules.The frequency band energy and frequency gravity can also show the relevance to drill wear.Aiming at the complex component of the current signal,and the certain similarity between the local and the whole,the fractal characteristics of different frequency components in the signal are studied by the method of wavelet fractal.In order to describe the self similarity between the internal components and each component under the different scales,the wavelet fractal box dimension and wavelet fractal dimension were extracted and study the relationship between them and drill wear was reserched.The results show that the wavelet fractal box dimension of the spindle motor current signal increases gradually from low frequency to high frequency band,which means the different components information in diferent frequency band.What's more,the box dimension of the first layer of the details shows the relevance to drill wear in a certain degree.However,the wavelet fractal dimension of the spindle motor current signal showing a good inverse correlation to the drill wear rules,while three stage of drill wear was well characterized on the fitting curve.In view of the small sample size of characteristics,the relevance vector machine(RVM)classification model was used to identify the wear state of the BTA drill,and the influence of kernel function and kernel parameters on the recognition rate of RVM classification model was studied.Kurtosis,skewness,and wavelet fractal dimension is selected to form a feature vector as training samples to trained the relevance vector machine(RVM)classification model,then the identification of the wear state was tested by the RVM classification model though test samples.The results show that the RVM classification model for drill wear state recognition in deep hole machining process with high accuracy,which can meet the requirements of real-time identification.This work provides a new way of drill wear condition monitoring in drilling process for BTA deep hole drilling.