Chatter And Deformation Prediction In Milling Process Of Thin Walled Titanium Alloy Parts

Aerospace and other important fields are widely used titanium alloy thinwalled parts.When the chatter happened,tool skipping may even occur.The chatter may influences the machining productivity and quality.Therefore,chatter and deformation in milling process has become a hot issue in academic and industrial research.Aiming at the chatter in the machining of thin-walled parts,based on the mechanism of flutter,this paper do a research on the dynamic characteristics,the solution of dynamic equation and chatter prediction.Aiming at the problem of cutting deformation in thin-walled parts machining,this paper mainly do a research on the deformation principle and prediction of cutting deformation.The main contents are as follows:Frist,for the milling system,the dynamic cutting thickness models of rigid parts and thin-walled parts are discussed respectively.The dynamic cutting thickness models of rigid parts and thin-walled parts are established.Different instantaneous milling force and dynamic milling force models of two kinds of workpieces are proposed.Finally,different dynamic processes of the two kinds of workpieces are established.To get the modal parameters of workpiece and cutting tools,this paper used the modus of multi-point excitation and single point vibration pick-up.The acceleration signals of thin-walled parts and tools in different directions and the percussion force signals of impact hammer are obtained.Using peak value picking method get the modal parameters of workpiece and tool.The milling force parameters are got by changing the feed rate.By using the parameters and coefficients measured above,the stability lobe diagram is established.Next,to solve the noise problem on collecting signals,this article apply the EMD modus to screen the signal information.The eigen mode signals that meet the conditions are recombined to get a new signal.The proper value matrix is obtained by picking up the characteristic value of the signals.The characteristic value matrix is input into the support vector machine for learning to obtain the flutter prediction model.The measured signal is input into the flutter prediction model for testing,and the test accuracy reaches 97.50%.The influence of different kernel functions on flutter prediction model is studied.By discussing the advantages and disadvantages,training accuracy and test accuracy of different kernel functions,the RGB function is choosed.Then,the deformation problem in machining is studied.The deformation principle of rectangular thin-walled parts with three sides supported and one side free is introduced.Through the finite element simulation,the maximum deformation position of thin-walled parts is obtained by comparing the deformation at different points in milling.The milling force signal and machining process deformation are collected through milling experiment,and the signal is used as input to establish the support direction.The machining process deformation prediction model based on measuring machine is used to judge the input signal and obtain the processing deformation according to the input signal,and make compensation suggestions for the next machining.Finally,aiming at the problem of chatter and deformation in milling process,the chatter and deformation prediction system of frame parts are built by MATLAB/GUI.The signal read by GUI is processed by EMD.The read signal is input into the support vector machine to get whether chatter occurs or not and the deformation in the machining process,which provides a certain reference value for the actual processing.