基于循环神经网络的深孔钻削镗床刀具振动监测  

Tool Vibration Monitoring of Deep Hole Drilling and Boring Machine Based on Recurrent Neural Network

作　　者：李仕存 Li Shicun(Department of Automotive Engineering,Zhengzhou City Vocational College,Zhengzhou Henan 452370,China)

机构地区：[1]郑州城市职业学院,河南郑州452370

出　　处：《机械管理开发》2024年第8期15-17,共3页Mechanical Management and Development

摘　　要：为克服刀具状态监测缺陷,采集得到加工期间产生的振动以及声音等多种音频信号,通过平衡样本训练深循环神经网络(RNN)。利用深孔镗床开展模型验证,分析刀具状态的变化情况。研究结果表明:为实现模型监测效果的验证分析,利用深孔镗床完成模型的测试。测试结果达到了98.4%的准确率,获得了理想监测效果。相比较模糊PID与BP神经网络,采用循环神经网络进行处理时除了可以实现降噪功能以外还能够有效保留信号本身的有用信息。In order to overcome the defect of tool condition monitoring,a variety of audio signals such as vibration and sound generated during machining were collected,and deep cycle neural network(RNN)was trained by balancing samples.Deep hole boring machine was used to carry out model verification,and the change of tool state was analyzed.The results show that in order to verify and analyze the monitoring effect of the model,a deep hole boring machine is used to test the model.The accuracy of the test results reached 98.4%,and the ideal monitoring effect was obtained.Compared with fuzzy PID and BP neural network,the recurrent neural network can not only reduce the noise but also retain the useful information of the signal itself.

关 键 词：深孔钻削 状态监测 循环神经网络 诊断准确率 

分 类 号：TG156[金属学及工艺—热处理]