摘要
钛合金具有弹性模量小、导热性能差等特点,加工过程中刀具易发生严重磨损,导致加工精度下降,表面粗糙度上升。针对上述问题,提出了基于深度神经网络的钛合金铣削刀具状态监测方法。首先设计并搭建刀具状态监测软硬件系统;采集加工过程中的振动与功率数据用于模型训练与状态监测;最后基于深度置信网络建立刀具状态监测模型,实验结果表明模型的平均准确率达到97.85%,相对于传统机器学习方法具有明显性能优势。提出的方法可以降低实际加工过程中对工人经验的依赖,具有较大的应用价值。
Because titanium alloys have the characteristics of small elastic modulus and poor thermal conductivity,the tools are prone severe wear during the machining of titanium alloys,which leads to a decrease in processing accuracy and an increase in surface roughness.In view of the problems above,a method for Titanium alloy milling tools condition monitoring based on deep neural network is proposed.First,the software and hardware systems for tool condition monitoring are built,and the vibration and power data during the machining process are collected for model training and condition monitoring.Then a tool condition monitoring model based on the deep learning method is built.The experimental results show that the average accuracy of the mon⁃itoring model reach 97.85%,which has obvious performance advantages over traditional machine learning methods.The method proposed can reduce the dependence on workers'experience in the actual processing process,and has great application value.
作者
周丹
ZHOU Dan(School of Intelligent Manufacturing,Panzhihua College,Sichuan Panzhihua 617000,China)
出处
《机械设计与制造》
北大核心
2025年第4期367-369,374,共4页
Machinery Design & Manufacture
基金
四川省教育厅科技项目(18CZ0046)—钒钛微合金钢特种轧制成形控制技术研究及产业化。
