[1]林 杨,高思煜,刘同舜,等.基于深度学习的高速铣削刀具磨损状态预测方法[J].机械与电子,2017,(07):12-17.
　LIN Yang,GAO Siyu,LIU Tongshun,et al.A Deep Learning-Based Method for Tool Wear State Prediction in High Speed Milling[J].Machinery & Electronics,2017,(07):12-17.
点击复制

基于深度学习的高速铣削刀具磨损状态预测方法
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
期数:
2017年07期
页码:
12-17
栏目:
设计与研究
出版日期:
2017-07-25

文章信息/Info

Title:
A Deep Learning-Based Method for Tool Wear State Prediction in High Speed Milling
文章编号:
1001-2257(2017)07-0012-06
作者:
林 杨1; 高思煜2; 刘同舜1; 朱锟鹏2
(1.中国科学技术大学 信息科学与技术学院自动化系,安徽 合肥 230026; 2.中国科学院 合肥物质科学研究院先进制造技术研究所,江苏 常州 213164)
Author(s):
LIN Yang1; GAO Siyu 2; LIU Tongshun 1; ZHU Kunpeng2
(1.Department of Automation,School of Information Science and Technology,University of Science and Technology of China, Hefei 230026,China; 2.Institute of Advanced Manufacturing Technology, Hefei Institutes of Physical Science, Chinese Academy of Sciences
关键词:
高速铣削;  刀具磨损;  状态预测;  深度学习;  稀疏自编码
Keywords:
high-speed milling;  tool wear;  state prediction;  deep learning;  sparse auto-encoder
分类号:
TP181; TG506
文献标志码:
A
摘要:
由于铣刀在高转速下进行不连续切削,刀具磨损迅速且难于监测,并且刀具磨损严重影响加工精度与产品质量。针对高速铣削刀具磨损难以在线预测问题,提出了一种基于深度学习的高速铣削刀具磨损预测的新方法。通过小波包变换提取铣削力信号在不同频段上的能量分布作为初始特征向量; 采用无监督学习对稀疏自编码网络进行特征学习,并将单层网络堆栈构成深度神经网络; 最后利用有监督学习对整个深度网络进行微调训练,建立铣削刀具磨损预测模型。实验结果表明,所提出的方法对刀具磨损状态预测准确率达到93.038%。
Abstract:
Due to discontinuous cutting of the milling cutters operating at high rotational speed, the milling tools wear quickly with difficult monitoring, which seriously affect the machining precision and product quality. In order to solve the on-line prediction problems of high speed milling tool wear, a deep learning-based method for predicting tool wear state in high speed milling is proposed in this paper. Firstly, a wavelet based method was used to extract the energy distribution of cutting force at different frequency bands as the initial feature vectors; Secondly, an unsupervised learning method was employed to learn the features of the sparse auto-encoder network, and the single layer networks were stacked to construct the deep neural network; Finally, the whole deep learning network was fine-tuned by a supervised learning method, and the prediction model of tool wear state was established. The experimental results show that the prediction accuracy of the proposed method is 93.038%.

参考文献/References:

[1] 陈明,安庆龙,刘志强.高速切削技术基础与应用[M].上海:上海科学技术出版社,2012.
[2] Waydande P, Ambhore N, Chinchanikar S. A review on tool wear monitoring system[J].Journal of Mechanical Engineering and Automation, 2016, 6(5A): 49-54.
[3] Sick B. On-line and indirect tool wear monitoring in turning with artificial neural networks: a review of more than a decade of research[J].Mechanical Systems and Signal Processing,2002,16(4):487-546.
[4] Tansel I N, Arkan T T, Bao W Y, et al. Tool wear estimation in micro-machining:Part II: neural-network-based periodic inspector for non-metals[J].International Journal of Machine Tools and Manufacture, 2000, 40(4): 609-620.
[5] Bao W Y, Tansel I N.Modeling micro-end-milling operations:Part III: influence of tool wear[J]. International Journal of Machine Tools and Manufacture, 2000, 40(15): 2193-2211.
[6] Zhu K, Wong Y S, Hong G S.Multi-category micro-milling tool wear monitoring with continuous hidden Markov models[J].Mechanical Systems and Signal Processing, 2009, 23(2): 547-560.
[7] Zhu K, Vogel-Heuser B. Sparse representation and its applications in micro-milling condition monitoring: noise separation and tool condition monitoring[J].The International Journal of Advanced Manufacturing Technology, 2014, 70(1): 185-199.
[8] Roth J T, Djurdjanovic D, Yang X, et al. Quality and inspection of machining operations: tool condition monitoring[J]. Journal of Manufacturing Science and Engineering, 2010, 132(4): 575-590.
[9] Zhang J Z, Chen J C.Tool condition monitoring in an end-milling operation based on the vibration signal collected through a microcontroller-based data acquisition system[J].The International Journal of Advanced Manufacturing Technology, 2008, 39(1): 118-128.
[10] Hinton G E, Salakhutdinov R R.Reducing the dimensionality of data with neural networks[J].Science, 2006, 313(5786): 504-507.
[11] LeCun Y, Bengio Y, Hinton G E. Deep learning[J]. Nature, 2015, 521(7553): 436-444.
[12] Hinton G E, Deng L, Yu D, et al. Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups[J]. IEEE Signal Processing Magazine, 2012, 29(6): 82-97.
[13] Chan T H, Jia K, Gao S, et al. PCANet: A simple deep learning baseline for image classification[J].IEEE Transactions on Image Processing, 2015, 24(12): 5017-5032.
[14] He X, Gao J, Deng L. Deep learning for natural language processing and related applications(Tutorial at ICASSP)[C]//IEEE International Conference on Acoustics, Speech, and Signal Processing(ICASSP),2014.
[15] Al Rahhal M M, Bazi Y, AlHichri H, et al. Deep learning approach for active classification of electrocardiogram signals[J]. Information Sciences, 2016, 345: 340-354.
[16] Cire

备注/Memo

备注/Memo:
收稿日期:2017-03-21
作者简介:林 杨(1990-),男,福建莆田人,硕士研究生,研究方向为深度学习、数据融合等; 高思煜(1986-),男,河南永城人,博士,助理研究员,研究方向为精密/超精密加工装备; 刘同舜(1989-),男,安徽合肥人,硕士,博士研究生,研究方向为机械设备故障诊断; 朱锟鹏(1977-),男,湖北黄冈人,博士研究生导师,研究员,研究方向为精密微细切削加工理论与智能控制技术、制造信息学、3D金属打印成型过程建模与智能控制,通信作者。
更新日期/Last Update: 2017-07-25