A Signal Processing Technique for Chatter Stability Prediction

The frequency response function (FRF) of a CNC machine tool plays a key role the onset of regenerative chatter. The majority of modal testing techniques to measure the tool’s FRF rely on skilled operators and expensive equipment, but can still suffer from poor repeatability and excessive machine down-time. Some efforts have been performed to overcome these issues using non-contact devices but in general these methods do not capture the FRF during the metal cutting process. Consequently the effects of the spindle rotation and preload due to the cutting forces cannot be simultaneously considered. The present paper presents a novel method of predicting the tool’s FRF by exciting the system while cutting, and measuring the forces acting on the tool. A signal processing procedure is used study the stability of the system. Numerical simulations are used to illustrate two possible uses of the approach: prediction of the full stability lobes with no prior knowledge of the material cutting parameters; and prediction of the stable depth of cut at the current spindle speed. The results are then discussed with particular emphasis on the issues concerning practical implementation of the approach.