Chatter detection in milling processes using frequency-domain Rényi entropy

Abstract

Chatter is a kind of self-excited vibration and causes negative effects in machining processes. This paper presents a practical method to identify the chatter with cutting force signals in milling processes. Since the spectrum of the chatter signal exhibits discrete spectral lines around the chatter frequencies and the Renyi entropy is an effective index to characterize the randomness of data series, the frequency-domain Renyi entropy is proposed as a chatter indicator. As the chatter severity level grows, the signal components at the chatter frequencies become more and more significant, which means a reduction of the randomness of the spectral series. As a result, the value of the Renyi entropy-based indicator decreases rapidly at the onset of the chatter. In order to eliminate the interference of the normal signal components, i.e., the spindle speed-related frequency components, the spectrum is preprocessed to filter out those components first. Various milling experiments are conducted. The results show that the value of the proposed indicator changes sharply at the onset of chatter in various milling conditions with different spindle speeds and cutting depths. Also, the proposed indicator is compared with the commonly used Shannon entropy-based indicator and verified to have a larger difference between the stable and chatter statuses and is higher sensitivity to the chatter.