融合残差块与Swin-Transformer机制的刀具磨损监测方法*

doi:10.16731/j.cnki.1671-3133.2024.08.016

现代制造工程 ›› 2024, Vol. 527 ›› Issue (8): 126-135.doi: 10.16731/j.cnki.1671-3133.2024.08.016

• 仪器仪表/检测/监控 • 上一篇下一篇

融合残差块与Swin-Transformer机制的刀具磨损监测方法^*

李泽稷¹, 周学良¹, 孙培禄²

1 湖北汽车工业学院机械工程学院,十堰 442002;
2 运城学院机械工程学院,运城 044000

收稿日期:2023-11-13 出版日期:2024-08-18 发布日期:2024-08-30
通讯作者: 周学良,教授,硕士研究生导师,主要研究方向为数字化设计与制造、制造系统调度与优化。E-mail:zhouxl@huat.edu.cn
作者简介:李泽稷,硕士研究生,主要研究方向为深度学习、刀具磨损。E-mail:2404563430@qq.com
基金资助:
^*国家自然科学基金资助项目(52075107);湖北省高等学校优秀中青年科技创新团队计划项目(T2020018)

Tool wear monitoring methods incorporating residual block and Swin-Transformer mechanisms

LI Zeji¹, ZHOU Xueliang¹, SUN Peilu²

1 School of Mechanical Engineering,Hubei Institute of Automotive Technology,Shiyan 442002,China;
2 School of Mechanical Engineering,Yuncheng University,Yuncheng 044000,China

Received:2023-11-13 Online:2024-08-18 Published:2024-08-30

摘要/Abstract

摘要： 为进一步提高切削加工过程刀具磨损值监测的精度,提出一种融合残差块与Swin-Transformer模型的刀具磨损监测模型。首先,采用分组卷积残差块提取信号的特征;然后,利用Swin-Transformer模型中的分块滑动窗口自注意力机制对提取的特征进行平移融合;最后,通过回归层实现刀具磨损值监测。试验结果表明,融合一层残差块与一层stage机制的Swin-Transformer模型可以有效融合刀具磨损状态监测信号的全局信息,相比其他Swin-Transformer模型,不仅模型结构简单,而且具有更高的监测精度,利用PHM2010数据集验证的MSE、MAE和R²分别达到4.471 9、1.467 5和0.995 8。

关键词: 刀具, 磨损监测, 残差卷积神经网络, Swin-Transformer模型

Abstract: To further improve the accuracy of tool wear value monitoring in the cutting machining process,a tool wear monitoring model that integrated the residual block and Swin-Transformer model was proposed.Firstly,the grouped convolutional residual block was used to extract the features of the signal.Then,the chunked sliding window self-attention mechanism in the Swin-Transformer model was used to translate the extracted features.Finally,the tool wear value prediction was realized through the regression layer.The experimental results show that the Swin-Transformer model fusing a layer of residual blocks with a layer of stage mechanism can effectively fuse the global information of tool wear state monitoring signals,which not only has a simple model structure but also has a higher monitoring accuracy compared with other Swin-Transformer models,and the MSE,MAE,and R² verified by utilizing the PHM2010 dataset reached 4.471 9,1.467 5,and 0.995 8,respectively.

Key words: cutting tool, wear monitoring, residual convolutional neural networks, Swin-Transformer model

中图分类号:

TG713
TH162

李泽稷, 周学良, 孙培禄. 融合残差块与Swin-Transformer机制的刀具磨损监测方法^*[J]. 现代制造工程, 2024, 527(8): 126-135.

LI Zeji, ZHOU Xueliang, SUN Peilu. Tool wear monitoring methods incorporating residual block and Swin-Transformer mechanisms[J]. Modern Manufacturing Engineering, 2024, 527(8): 126-135.

参考文献

[1] REHORN A G,JIANG J,ORBAN P E.State-of-the-art methods and results in tool condition monitoring:a review[J].The International Journal of Advanced Manufacturing Technology,2005,26(7):693-710.
[2] ELSHEIKH A,YACOUT S,OUALI M S.Bidirectional handshaking LSTM for remaining useful life prediction[J].Neurocomputing,2019,323:148-156.
[3] ABELLAN-NEBOT J V,ROMERO SUBIRÓN F.A review of machining monitoring systems based on artificial intelligence process models[J].The International Journal of Advanced Manufacturing Technology,2010,47:237-257.
[4] LI X,LIM B S,ZHOU J H,et al.Fuzzy neural network modelling for tool wear estimation in dry milling operation[C]//Annual Conference of the PHM Society.San Diego,Cal.,USA:Prognostics and Health Management,2009:1.
[5] 张栋梁,莫蓉,孙惠斌,等.基于混沌时序分析方法与支持向量机的刀具磨损状态识别[J].计算机集成制造系统,2015,21(8):2138.
[6] SATEESH B G,ZHAO P,LI X L.Deep convolutional neural network based regression approach for estimation of remaining useful life[C]//Proceedings of 21st International Conference(Part I 21)Database Systems For Advanced Applications(DASFAA).Dallas:Springer International Publishing,2016:214-228.
[7] LU L,SHIN H,ROTH H R,et al.Deep convolutional neural networks for computer-aided detection:CNN architectures,dataset characteristics and transfer learning[J].IEEE Transactions on Medical Imaging,2016,35(5):1285-1298.
[8] 曹大理,孙惠斌,张纪铎,等.基于卷积神经网络的刀具磨损在线监测[J].计算机集成制造统,2020,26(1):74-80.DOI:10.13196/j.cims.2020.01.008.
[9] 戴稳,张超勇,孟磊磊,等.采用深度学习的铣刀磨损状态预测模型[J].中国机械工程,2020,31(17):2071.
[10] 滕瑞,黄海松,杨凯,等.基于图像编码技术和卷积神经网络的刀具磨损值在线监测方法[J].计算机集成制造系统,2022,28(4):1042.
[11] 邓祥力,吴高珍,魏聪聪,等.基于多源数据融合的AlexNet神经网络大电网故障诊断[J].现代电力,2023,40(2):161-169.DOI:10.19725/j.cnki.1007-2322.2021.0260.
[12] 方宇伦,陈雪纯,杜世昌,等.基于轻量化深度学习VGG16网络模型的表面缺陷检测方法[J].机械设计与研究,2023,39(2):143-147.DOI:10.13952/j.cnki.jofmdr.2023.0068.
[13] 曾飞,陶玉衡,苏俊彬,等.融合ResNet18和Deconvolution的输送带横向跑偏检测方法[J].现代制造工程,2023(8):121-126.DOI:10.16731/j.cnki.1671-3133.2023.08.017.
[14] 赵昀杰,张太红,姚芷馨.基于EfficientNet的实时目标检测模型[J].计算机应用与软件,2023,40(8):255-264,297.
[15] IOANNOU Y,ROBERTSON D,CIPOLLA R,et al.Deep Roots:Improving CNN Efficiency with Hierarchical Filter Groups[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Honolulu:[s.n.],2017:5977-5986.DOI:10.1109/CVPR.2017.633.
[16] CARION N,MASSA F,SYNNAEVE G,et al.End-to-end object detection with transformers[C]//European Conference on Computer Vision.Cham:Springer International Publish-ing,2020:213-229.
[17] LIU Z,LIN Y,CAO Y,et al.Swin Transformer:Hierarchical Vision Transformer using Shifted Windows[C]//2021 IEEE/CVF International Conference on Computer Vision(ICCV). Montreal:[s.n.],2021:10012-10022.
[18] 扶兰兰,黄昊,王恒,等.基于 Swin Transformer 模型的玉米生长期分类[J].Transactions of the Chinese Society of Agricultural Engineering,2022,38(14):191-200.
[19] 叶铭亮,周慧英,李建军.基于改进Swin Transformer的森林火灾检测算法[J].中南林业科技大学学报,2022,42(8):101-110.DOI:10.14067/j.cnki.1673-923x.2022.08.010.
[20] 李俊杰,易诗,何润华,等.基于窗口注意力聚合Swin Transformer的无人机影像语义分割方法[J/OL].计算机工程与应用:1-14[2023-09-20].http://kns.cnki.net/kcms/detail/11.2127.TP.20230812.1057.002.html.
[21] 李紫桐,赵健康,徐静冉,等.基于改进Swin Transformer的遥感图像融合方法[J/OL].光子学报:1-15[2023-09-20].http://kns.cnki.net/kcms/detail/61.1235.O4.20230918.1631.039.html.
[22] 王庆明,程耀东,朱国辉.刀具磨损型式与磨钝标准VB的选取[J].机械,1996(6):37-38.
[23] 程胜明,王雅君,张昕晨,等.采用残差结构和卷积神经网络的铣刀磨损研究[J/OL].机械科学与技术:1-8[2023-09-06].https://doi.org/10.13433/j.cnki.1003-8728.20230108.
[24] IOANNOU Y,ROBERTSON D,CIPOLLA R,et al.Deep Roots:Improving CNN Efficiency with Hierarchical Filter Groups[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Honolulu:[s.n.],2017: 5977-5986.DOI:10.1109/CVPR.2017.633.
[25] PHM Society.PHM data challenge 2010[DB/OL].[2023-11-13].https://www.phmsociety.org/competition/phm/10.
[26] 何彦,凌俊杰,王禹林,等,基于长短时记忆卷积神经网络的刀具磨损在线监测模型[J].中国机械工程,2020,31(16):1959-1967.
[27] 刘会永,张松,李剑峰,等.采用改进CNN-BiLSTM模型的刀具磨损状态监测[J].中国机械工程,2022,33(16):1940-1947,1956.

融合残差块与Swin-Transformer机制的刀具磨损监测方法^*

Tool wear monitoring methods incorporating residual block and Swin-Transformer mechanisms

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	李浩, 赵威, 李小睿, 陈啸宇. 基于刀具结构的T800-CFRP铣削刀具磨损抑制和加工质量优化^*[J]. 现代制造工程, 2024, 525(6): 101-110.
[2]	陈汉阳, 唐清春, 杨宇航, 李晓宇, 乔鹏, 周乐安. 一种涡轮盘榫槽的组合加工刀具设计研究^*[J]. 现代制造工程, 2024, 523(4): 109-122.
[3]	杨宇, 杨发展, 刘朝伟, 赵烁, 隋潇斌. 标题非对称织构对YT15硬质合金刀具切削性能及其衍生切削的影响^*[J]. 现代制造工程, 2024, 521(2): 76-85.
[4]	赵东旭, 袁志响, 易思广, 潘加港, 张云鹏, 卢文壮. 基于双路并行卷积信息融合的刀具磨损识别^*[J]. 现代制造工程, 2024, 520(1): 124-129.
[5]	王栋梁;张继林;易湘斌. 切削速度对微织构涂层刀具切削性能的影响[J]. 现代制造工程, 2023, 514(7): 106-111.
[6]	荆成利;郑光明;王成;刘丰江;苏红星. 清洁高速硬切削模具钢的仿真与试验研究[J]. 现代制造工程, 2023, 511(4): 110-115.
[7]	赵文硕;郑光明;张鸿滨;纪任可;李勋. 刀具磨损对FGH96试件加工表面完整性及疲劳性能的影响[J]. 现代制造工程, 2023, 510(3): 8-14.
[8]	黄智坤，刘丽冰，张晶，袁军，杨泽青. 视听信息融合的CNC铣刀磨损状态监测方法[J]. 现代制造工程, 2022, 503(8): 93-100.
[9]	周慧琳，于汇泳，赵向阳. 铝合金材料切削用织构刀具性能研究[J]. 现代制造工程, 2022, 501(6): 85-89.
[10]	王珉，唐小聪，朱奔驰，信连甲，杨吟飞，赵国龙. 基于干冰低温冷却的大进给铣削TC4钛合金刀具磨损研究[J]. 现代制造工程, 2022, 499(4): 94-101.
[11]	孙作民，郑光明，李玉龙，颜培，刘丰江，苏红星，李祥. 液氮低温铣削GH4169合金时涂层刀具表面完整性及磨损研究[J]. 现代制造工程, 2022, 506(11): 89-92.
[12]	韩新港;张海云;林伟盛;赵玉刚;蒲业壮;孟建兵;曹辰;苗旭. 基于田口方法的Al2O3陶瓷激光辅助车削实验研究[J]. 现代制造工程, 2022, 505(10): 8-13.
[13]	谢峰，张晨晨，李凡. 自适应随机共振算法的立铣刀磨损信号检测[J]. 现代制造工程, 2021, 492(9): 1-8.
[14]	于洲，郑光明，程祥，刘焕宝，杨先海. 微喷砂处理对AlTiN涂层刀具表面性能的影响研究[J]. 现代制造工程, 2021, 489(6): 7-11.
[15]	钟江，宁诗铎，郑晓军，鲍洪阳. 面向非连续多去除量的城轨底架滑槽加工优化[J]. 现代制造工程, 2021, 489(6): 63-68.