Chatter identification in milling process based on complementary ensemble empirical mode decomposition
Résumé
Chatter detection and identification are important in machining condition monitoring during machining to prevent product quality loss. This work presents a chatter identification method in the milling process of flexible work-piece. It is based on the method called complementary ensemble empirical mode decomposition (CEEMD), and it provides an indicator designed for detecting and identifying chatter from nonlinear vibration signals. Firstly, the signal on the flexible work-piece is collected during the milling process. In this paper, we used simulated data based on the milling process model to validate the proposed chatter identification method's effectiveness. Secondly, the collected vibration signal is preprocessed using a comb filter to remove rotation frequency interference, tooth passing frequency, and harmonics. Thirdly, the CEEMD method, based on a self-adaptive signal processing technique, is applied to decompose the filtered vibration signal into a set of intrinsic mode functions (IMFs). Sensitive IMFs are selected to describe chatter status. With the development of chatter, the amplitude of sensitive IMFs becomes large. Finally, an indicator called stand variance (SV) is proposed and extracted from sensitive IMFs. The effectiveness of the proposed method is demonstrated in experimental simulation. The results show that the proposed method based on CEEMD may effectively identify chatter in the milling process.