Investigation On The Analysis Of Cutting Tool Life Data And Tool Wear Condition Monitoring In Milling Of Titanium Alloy

With the continuous development and innovation of manufacturing technology in aviation industry,titanium alloy has become the main material of aeronautical monolithic components because of its high strength-to-weight ratio,good corrosion resistance and high temperature strength.However,titanium alloy is a typical difficult-to-machine material,which has low thermal conductivity and small deformation coefficient.The cutt:ing temperature is high and the cutting force is large in milling of titanium alloy.Consequently,severe tool wear can be found in the process of machining titanium alloys,results in difficulty in predicting cutting tool life.With the application of various data measuring and storage technologies,a large amount of process data will be gathered during the running process of titanium alloy machining system.It is of great importance to study on the analysis methods of the process data so as to accurately predict the tool life and evaluate the tool wear status.In this paper,the data processing method was studied based on the large amount of tool usage data gathered from the process of machining titanium alloy aeronautical structural parts.Firstly,facing on the process data of tool wear,an equivalent evaluation model of tool life was established for cutting tools used at changing parameters.Milling tests on titanium alloy were carried out in order to predict the tool wear at changing parameters.Based on the theory of zero-failure data reliability assessment,under the premise of 2-parameter Weibull distribution,the approaches of failure probability estimation based on matching distribution curve method and least square parameter estimation were studied.Reliability analysis for cutting tools under typical working conditions with zero-failure data obtained from the production site was performed.The reliable life of cutting tools was predicted.Based on the monitoring signals in milling experiments,the changing regularity of force signals,vibration signals and sound signals with tool wear was studied.The feature extraction was carried out in time domain,frequency domain and time-frequency domain.The information measurement based method was used to reduce the dimension of extracted features.On the basis of the selected features after dimension reduction,the intelligent recognition of tool wear state was carried out by using established support vector machine model and BP neLural network model.The precision of the models was verified,and the recognition results of the two models were compared.