Machining Chatter Monitoring Based on Wavelet Packet Energy Kurtosis Index of Vibration Signals

Chatter vibrations during machining operations always restrict the process quality and efficiency, leading to severe damage to the workpiece and shortened tool life. Therefore, accurate and timely diagnosis of chatter will help maintain stable operations. This paper presents an effective method for chatter monitoring in turning operations using wavelet packet energy kurtosis index of the online measured signals. The characteristics of vibration from stable to unstable state are compared in the time, frequency and time-frequency domains, respectively. Considering that the chatter information at its early stage is generally very weak, the wavelet transform is proposed to amplify this feature. Then, the feature information is further explored using wavelet packet decomposition. The wavelet packet energy ratio of each node is calculated to represent the evolution of chatter. Accordingly, the chatter monitoring and prediction indicators using the wavelet energy kurtosis index are constructed. Finally, experimental tests were carried out and the results showed that the proposed method enables to effectively improve the response to chatter vibrations and overcome uncertainties by noises. This study could be beneficial to increasing the time available for the control response to chatter occurrence.