综合振动信号能量比和幅值标准差的铣削颤振实时监测方法

The chatter phenomenon during the milling processing process of many thin-walled structural parts in aerospace equipment is one of the main factors affecting the surface quality of machining, a timely chatter detecting method in milling with integrated energy ratio and amplitude standard deviation of vibration signal is proposed to avoid the continuous occurrence of chatter. Firstly, the timely vibration signal of the spindle is measured by triaxial accelerometers during the milling processing process. Secondly, the energy ratio and auxiliary detecting indicator amplitude standard deviation are extracted from it. Then, feature layer fusion is performed and 6 detecting indicators are obtained. Finally, the milling state during the sampling period is output by the detection model. Test results show that the identification accuracy of the proposed method for chatter state can reach 99.07% at the change of all process parameters. Numerical experiment results show that the method can meet the timely requirements of online chatter detection. The method effectively solves the two major difficulties in the application of online detecting of milling chatter — poor generalization performance (process generalization ability) and low real-time performance.