论文部分内容阅读
将刀具磨损状态的在线监测作为模式识别中的两类模式分类问题,从切削振动信号中抽取特征向量;根据投影原理构造了最佳特征平面.在此基础上提出了一种具有自学习功能的G(D)判别函数,对车削试验的磨损状态进行分类,确诊率达95%,漏诊率小于0.6%,判别时间少于15s,适用于在线监测。
The on-line monitoring of tool wear state is taken as two types of pattern classification problems in pattern recognition, and the feature vector is extracted from the vibration signal of cutting. The best feature plane is constructed according to the principle of projection. Based on this, a self-learning function G (D) discriminant function to classify the worn state of turning test. The diagnosis rate is 95%, the rate of missed diagnosis is less than 0.6% and the discriminant time is less than 15s, which is suitable for on-line monitoring.