Indirect Prediction Study Of Milling Force Based On Spindle And Feed Axis Currents

Milling force is an important parameter to characterize the working state and machining accuracy of CNC machine tools.Based on the current signal,this paper establishes a prospective prediction model of milling force applicable to five-axis CNC machine tools.Through the experiment of milling thin S-shaped samples in the 5-axis CNC milling machining center of Tupu VMC-C50,the prediction accuracy of the model under different rotational speeds is analyzed,and the instantaneous milling force prediction system for five-axis CNC machine tools is established.Firstly,the prediction model of the X direction instantaneous milling force is established based on the spindle current signal.Based on the Devavit Hartenberg(D-H)method,the kinematics modeling of the five-axis machine tool is carried out.The mapping relationship between the driving motor torque of the feed axis and the instantaneous milling force in X,Y and Z directions is obtained through the force Jacobian matrix.Secondly,the milling current signal and milling force signal are obtained through the five axis milling S-type sample experiment.Based on cross correlation analysis,milling current signal is a delayed signal of milling force signal.After verifying the prediction accuracy of the two models,a combined prediction method of instantaneous milling force based on the optimal variable weight method of spindle and feed axis current is proposed,which improves the prediction accuracy of instantaneous milling force by 12% compared with the feed model.Then,the complementary integrated empirical mode decomposition(CEEMD)method is used to decompose the Y and Z direction milling force signals predicted based on the feed shaft torque and the X direction milling force predicted based on the optimal variable weight method,and the eigenmode functions and trend sequences of each layer are obtained.The prediction model of Extreme Gradient Boosting(XGBoost)optimized by Sparrow Search Algorithm(SSA)is constructed to predict the instantaneous milling force signal prospectively.The result shows that SSA-XGBoost model can predict more points in a certain range of error than the unoptimized XGBoost model,with higher prediction accuracy and better fitting effect.By inputting multiple groups of unequal length data for comparative verification,it is proved that the prediction accuracy of the model will improve with the increase of the number of samples.Finally,combined with Matlab software,a five axis linkage CNC machine tool instantaneous milling force prediction system is established,which realizes the visualization of the predictive model of instantaneous milling force,and provides software support for realizing the online prediction of instantaneous milling force.