On-line Chatter Detection Method Based On Drive Motor Current Signal In Turning

Chatter is a self-excited vibration that occurs during machining operations andbecomes a common limitation to productivity and part quality. Chatter occurrence also hasseveral negative effects, such as excessive noise, machine tool damage, disproportionatetool wear, damage of the tooling structure and the spindle bearing. For these reasons,chatter avoidance is a topic of enormous interest for many researchers who have focusedon the solutions to the problem of chatter. But most of the chatter detection methods canonly be used in laboratory, and be difficult to use in practical production. Hence, in orderthat the on-line chatter detection technique can be applied to manufacturing production,further research need to be done.In order to realize the engineering application of chatter detection technique, amethod of chatter detection based on motor current signal is proposed. By determining theexperimental program, experiment design of chatter detection in NC vertical lathe wasfinished. The current signal monitoring experimental system was established, and a seriresof experiment was completed.The response characteristics of different motor current signal to the cutting state wasanalyzed. Mode of spindle and feed drive system was established. Experiment wasdesigned to identificate parameters of spindle and feed drive system. Rapid assessmentmethod was given to judge the applicability of the proposed chatter detection based oncurrent signal and how to choose the motor current which is more sensitive.According to the low signal to noise ratio of current signal, feature extraction wascarried out in the time domain and time-fraquency domain. The feature vector of currentsignal used for chatter detection is constructed on the basis of empirical modedecomposition. In addition, acceleration signal of tool is taken as reference and used tocompare with the results from current signal.In order to detect chatter accurately, a support vector machine is designed for patternclassification based on the feature vector constituted by energy index and kurtosis index.The intelligent chatter detection system composed of the feature extraction and the SVMhas an accuracy rate of above95%for the detection of cutting state after being trained byexperimental data.