基于神经网络信息融合的铣刀磨损状态监测

为了获得铣削加工过程中铣刀后刀面磨损的全面评价，用铣刀后刀面磨损带面积作为衡量刀具磨损量的一个评价指标。提取和精选了8个对铣刀后刀面磨损状态敏感的无量纲特征参数并经归一化处理后，作为基于神经网络信息融合的铣刀磨损状态监测系统的输入信号。采用3层BP神经网络模型，利用其多传感器信息融合功能在线监测了铣刀后刀面磨损带宽度和磨损带面积。监测系统的输出结果与实际测量结果基本吻合。

Wear Condition Monitoring of Helical Cutters Based on Neural Network Information Infusion Method

To obtain comprehensive evaluations of major flank wear of a helical cutter in the milling process, the wear land area was proposed as an index for estimating the wear out of milling cutters. In the research, 8 dimensionless characteristic parameters, which are sensitive to major flank wear condition of the cutter, were extracted, selected and normalized as input signals of the wear condition monitoring system based on neural network information infusion method. By three-layer back propagation neural network model, with its capability of multi-sensor information infusion, major flank wear land width and wear land area of the helical cutter were monitored online. The output results of the monitoring system were consistent with the tested data.