【摘要】 为了预报车削颤振现象，构建了再生型车削颤振的预报函数。分析再生型颤振机理的基础上得到了其稳定性叶瓣图，为颤振预报提供可能。利用Elman神经网络对获取的数控车削过程中从稳定车削阶段到车削颤振阶段的时域信号进行训练和测试，提出了均方误差作为判别颤振的特征量。为了更准确的进行颤振预报，引入符号函数来构建再生型车削颤振的预报函数，确定了预报函数的阈值为5.625并试验了在该预报函数阈值时的预报准确率为92%。最后从频率域中分析，根据颤振的特点侧面验证了所构建预报函数的正确性。更多还原

【Abstract】 In order to forecast the turning chatter phenomenon, a turning chatter forecast function is constructed. After analyzing the mechanism of regenerative chatter, a stability lobes diagram is obtained and provides the possibility of chatter forecast. The Elman neural network is used to train and test the time-domain signal from the stable turning stage to the turning chatter during the process of turning. The mean square error is proposed as the feature quantity for judging chatter. In order to predict the chatter more accurately, the sign function is used to construct the forecast function of regenerative turning chatter and the threshold of the forecast function is determined to be 5.625 and the prediction accuracy at the threshold of the forecast function is 92%. Finally, in the frequency domain, according to the characteristics of chatter, the validity of the constructed forecast function is verified.更多还原