Chatter Stability Prediction in Milling Using Speed-varying Cutting Force Coefficients

Milling process has improved in the last decades thanks to new tooling system, technologies and control, leading performance to high-level and introducing new challenges for the machine tool manufacturers and users. In this context the dynamics of the machine tool has a determining effect on the precision and removal rate of the machining process due to the possibility of chatter vibration occurrence. The prediction of this unstable phenomenon by means of stability lobe diagrams can lead to significant improvement in productivity; but the accuracy of the analytical prediction method is strongly affected by reliability of data entry, i.e. cutting force coefficients and frequency response of the system. Both these parameters are influenced by spindle speed: this is an issue for High-Speed Milling (HSM), where a wide range of cutting velocity is available. Generally stability lobe diagrams are calculated using as inputs the stationary tooltip frequency response function, and the cutting coefficients estimated at low spindle speed. Therefore stability prediction in a different speed range is not always reliable, because the extrapolation of the cutting force coefficients from the lower speed tests could introduce appreciable errors. Moreover the evaluation of such coefficients with high speed milling tests could lead to weak accuracy, due to the frequency bandwidth of commercial force sensors that is inadequate for high rotational speeds (dynamometer’s natural frequency limits measurements to low speed). In this paper the cutting coefficients have been identified by milling tests at different spindle speed: dynamometer signals have been compensated thanks to an improved technique based on Kalman filter estimator. The obtained cutting velocity dependent force coefficients have been used to improve the reliability of stability lobe diagrams for HSM, as proven by experimental tests.