Early milling chatter identification by improved empirical mode decomposition and multi-indicator synthetic evaluation

Abstract

Chatter is a kind of self-excited vibration which reflects changes of frequency and energy distribution in machining process and it always leads to poor surface quality of the materials. An effective early chatter identification method is necessary to avoid the damage caused by chatter. The key technique of chatter identification is to capture the feature signatures. In this paper, the characteristics of milling chatter signal is analyzed in details in the aspect of time-frequency domain and morphological feature. An intelligent early chatter identification method which based on the improved empirical mode decomposition (EMD) and multi-indicator synthetic evaluation is proposed in this paper. The acceleration signal is decomposed into a series of intrinsic mode functions (IMFs) by the improved EMD, and then the IMFs which contain the chatter information are selected to reconstruct a new signal, then a three-dimensional characteristic vector which based on the multi-indicators (i.e., the standard deviation, power spectral entropy and fractal dimension of the new signal) is constructed for chatter identification. A support vector machine chatter identification model is obtained based on the multi-indicators. The results of milling experiment show that the proposed chatter identification method can recognize early milling chatter effectively.