This paper compares various sensors and shows that a microphone is an excellent sensor to be used for chatter detection and control. Comparisons are made between the microphone and some other common sensors (dynamometers, displacement probes, and accelerometers) regarding sensing of unstable milling. It is shown that the signal from the microphone provides a competitive, and in many instances a superior, signal tht can be utilized to identify chatter. Using time domain milling simulations of low-radial-immersion, low-feed, finishing operations it is shown that for these cuts (especially at relatively high speeds) chatter is not adequately reflected in the force signal because of the short contact time, but that it is clearly seen in the displacement signal. Using the dynamics of existing production milling machines it is shown how the microphone is more suitable to chatter detection than other remotely placed displacement sensors, especially in cases that involve flexible tooling and workpieces. Aspects important for practical implementation of a microphone in an industrial setting are discussed. Limitations of the microphone are addressed, such as directional considerations, frequency response, and environmental sensitivity (i.e., workspace enclosure, room size, etc). To compensate for expected unwanted noises, commonly known directionalization techniques such as isolation, collection, and intensity methods are suggested to improve the ability of the microphone to identify chatter by reducing or eliminating background and extraneous noises. Using frequency domain processing and the deterministic frequency domain chatter theory, a microphone is shown to provide a proper and consistent signal for reliable chatter detection and control. Cutting test records for an operating, chatter recognition and control system, using a microphone, are presented; and numerous examples of chatter control are listed which include full and partial immersion, face-and end-milling cuts.
A Hybrid Chatter Detection Method Based on WPD, SSA, and SVM-PSO
