中文摘要：

为实现截齿截割过程中磨损程度的实时精准在线监测,提出了一种基于BP神经网络的截齿磨损程度多特征信号融合的检测方法。通过提取截割过程中不同磨损程度截齿的三向振动信号、红外温度信号和电流信号,建立了不同磨损程度截齿的多特征信号样本数据库,采用多特征信号样本对BP神经网络进行学习和训练,建立截齿磨损程度的识别模型,实现截齿磨损程度在线监测与精确识别。实验结果表明：基于BP神经网络的截齿磨损程度监测系统,网络判别结果和测试样本的实际磨损程度类别相符,该BP神经网络系统能够对截齿磨损程度类型进行准确的监测和识别。

英文摘要：

In order to realize the real-time monitoring degrees of pick wear in cutting processes, based on BP neural network a method was proposed for detection of multi-feature signal fusion. By extracting different wear degrees pick signals, such as three direction vibration signals, infrared temperature signals and current signals during the cutting processes, A database of multi-feature signal sampies of different pick wear degree was established, the BP neural network was learned and trained by using the multi characteristic pick wear degress signal samples, and the pick wear a recognition model for pick wear degress was established to achieve online monitoring and recognition. The results show that： Based on the BP neural network, the network discrimination results of monitoring system for the degree of pick wear is consistent with the actual wear degree category of the test samples, the establishment of BP neural network system may accurately monitor and identify the type of wear degrees.