留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

刀具磨损感知数据驱动下的DBN预测模型研究

刘子安 刘建春 苏进发 秦昆

刘子安, 刘建春, 苏进发, 秦昆. 刀具磨损感知数据驱动下的DBN预测模型研究[J]. 机械科学与技术, 2021, 40(7): 1043-1050. doi: 10.13433/j.cnki.1003-8728.20200178
引用本文: 刘子安, 刘建春, 苏进发, 秦昆. 刀具磨损感知数据驱动下的DBN预测模型研究[J]. 机械科学与技术, 2021, 40(7): 1043-1050. doi: 10.13433/j.cnki.1003-8728.20200178
LIU Zian, LIU Jianchun, SU Jinfa, QIN Kun. Study on DBN Prediction Model Driven by Tool Wear Sensing Data[J]. Mechanical Science and Technology for Aerospace Engineering, 2021, 40(7): 1043-1050. doi: 10.13433/j.cnki.1003-8728.20200178
Citation: LIU Zian, LIU Jianchun, SU Jinfa, QIN Kun. Study on DBN Prediction Model Driven by Tool Wear Sensing Data[J]. Mechanical Science and Technology for Aerospace Engineering, 2021, 40(7): 1043-1050. doi: 10.13433/j.cnki.1003-8728.20200178

刀具磨损感知数据驱动下的DBN预测模型研究

doi: 10.13433/j.cnki.1003-8728.20200178
基金项目: 

2018年福建省科技计划项目高校产学合作项目 2018H6025

厦门市科技计划项目 3502Z20183051

详细信息
    作者简介:

    刘子安(1995-), 硕士研究生, 研究方向为数控机床在线监测技术、刀具磨损预测分析, 470684386@qq.com

    通讯作者:

    刘建春, 教授, 博士, xmjcliu@163.com

  • 中图分类号: TH16;TG659

Study on DBN Prediction Model Driven by Tool Wear Sensing Data

  • 摘要: 针对制造车间数控刀具在连续作业过程中易出现过度使用或提前置换的现状, 对刀具磨损感知数据获取方法和磨损预测模型构建进行研究。为避免传感器噪声影响, 采用OPC技术直接与机床协同完成数控通信, 并设计一套双镜头垂直分布的感知数据获取系统; 为增强预测模型泛化能力, 采用Dropout优化后深度信念网络(DBN)作为预测模型, 先在特征提取阶段重构出优化权值, 再引入标签量训练特征匹配阶段。结果显示, 改进的DBN算法平均预测准确度约96.0%, 在预测精度和稳定性方面较传统模型显著改善。
  • 图  1  RBM网络架构

    图  2  DBN预测网络整体架构

    图  3  刀具磨损视觉检测机构实物图

    图  4  刀具磨损检测图像获取

    图  5  刀具全生命周期内磨损率变化

    图  6  RBM迭代过程

    图  7  DBN改进效果

    图  8  预测精度对比

    图  9  车间加工流程图

    表  1  刀具全生命周期内检测数据对比

    行程/mm 实验值/mm 真实值/mm 绝对误差/mm
    1 000 0.035 0.039 0.004
    2 000 0.122 4 0.119 0.003 4
    3 000 0.146 4 0.145 0.001 4
    11 000 0.395 1 0.386 0.009 1
    12 000 0.518 1 0.514 0.004 1
    下载: 导出CSV

    表  2  刀具磨损状态识别

    刀面磨损VB/mm 刀具磨损状态 刀面磨损VB/mm 刀具磨损状态
    [0, 0.04) 初期磨损阶段① 0.21, 0.24) 急剧磨损阶段⑤
    [0.04, 0.11) 初期磨损阶段② [0.24, 0.3) 严重磨损阶段⑥
    [0.11, 0.16) 正常磨损阶段③ [0.3, 0.4] 严重磨损阶段⑦
    [0.16, 0.21) 急剧磨损阶段④ 大于0.4 废置阶段⑧
    下载: 导出CSV

    表  3  样本数据采集

    序号 因素变量 水平度
    1 主轴转速N/(r·min-1) 3
    2 切削速度Vc/(mm·min-1) 3
    3 进给量f/(mm·r-1) 3
    4 背吃刀量ap/mm 3
    5 进给速度Vf/(mm·min-1) 3
    6 冷却情况(开/关) 2
    7 刀具磨损状态 7
    下载: 导出CSV

    表  4  对比模型误差比较

    网络模型 预测准确度/% MSE MAPE
    DBN 92.6 3.42×10-4 0.074
    改进的DBN 96.0 1.05×10-4 0.040
    BP 72.8 3.8×10-3 0.272
    深度BP 88.5 1.05×10-3 0.115
    SVR 91.8 3.01×10-4 0.082
    下载: 导出CSV
  • [1] KONG D D, CHEN Y J, LI N, et al. Relevance vector machine for tool wear prediction[J]. Mechanical Systems and Signal Processing, 2019, 127: 573-594 doi: 10.1016/j.ymssp.2019.03.023
    [2] 李聪波, 万腾, 陈行政, 等. 基于切削功率的数控车削批量加工刀具磨损在线监测[J]. 计算机集成制造系统, 2018, 24(8): 1910-1919

    LI C B, WAN T, CHEN X Z, et al. On-line monitoring method of tool wear for NC turning in batch processing based on cutting power[J]. Computer Integrated Manufacturing Systems, 2018, 24(8): 1910-1919 (in Chinese)
    [3] NIAKI F A, MEARS L. A probabilistic-based study on fused direct and indirect methods for tracking tool flank wear of Rene-108, nickel-based alloy[J]. Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture, 2018, 232(11): 2030-2043 doi: 10.1177/0954405416683432
    [4] LU X H, HU X C, WANG H, et al. Research on the prediction model of micro-milling surface roughness of Inconel718 based on SVM[J]. Industrial Lubrication and Tribology, 2016, 68(2): 206-211 doi: 10.1108/ILT-06-2015-0079
    [5] 曹翔, 赵培轶, 王鹏程, 等. 基于高斯过程回归方法的钛合金铣削刀具磨损预测[J]. 制造技术与机床, 2019(6): 55-59

    CAO X, ZHAO P Y, WANG P C, et al. A novel method for tool wear prediction in titanium milling by Gaussian process regression method[J]. Manufacturing Technology & Machine Tool, 2019(6): 55-59 (in Chinese)
    [6] 库祥臣, 郭跃飞, 段明德, 等. 利用振动频谱预测刀具磨损量[J]. 机械设计与制造, 2017(10): 113-116 https://www.cnki.com.cn/Article/CJFDTOTAL-JSYZ201710029.htm

    KU X C, GUO Y F, DUAN M D, et al. Predicting tool wear by vibration frequency spectrum[J]. Machinery Design & Manufacture, 2017(10): 113-116 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JSYZ201710029.htm
    [7] 毕长波, 王宇浩, 马廉洁, 等. 基于GA-BP算法的刀具磨损预测模型[J]. 组合机床与自动化加工技术, 2018(10): 145-146, 150 https://www.cnki.com.cn/Article/CJFDTOTAL-ZHJC201810037.htm

    BI C B, WANG Y H, MA L J, et al. Tool wear prediction model based on GA-BP algorithm[J]. Modular Machine Tool & Automatic Manufacturing Technique, 2018(10): 145-146, 150 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZHJC201810037.htm
    [8] 王国锋, 董毅, 杨凯, 等. 基于深度学习与粒子滤波的刀具寿命预测[J]. 天津大学学报, 2019, 52(11): 1109-1116 https://www.cnki.com.cn/Article/CJFDTOTAL-TJDX201911001.htm

    WANG G F, DONG Y, YANG K, et al. Tool life prediction based on deep learning and particle filtering[J]. Journal of Tianjin University, 2019, 52(11): 1109-1116 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-TJDX201911001.htm
    [9] KEYVANRAD M A, HOMAYOUNPOUR M M. Deep belief network training improvement using elite samples minimizing free energy[J]. International Journal of Pattern Recognition and Artificial Intelligence, 2015, 29(5): 1551006 doi: 10.1142/S0218001415510064
    [10] MANNEPALLI K, SASTRY P N, SUMAN M. A novel adaptive fractional deep belief networks for speaker emotion recognition[J]. Alexandria Engineering Journal, 2017, 56(4): 485-497 doi: 10.1016/j.aej.2016.09.002
    [11] 彭锐涛, 降皓鉴, 徐莹, 等. 刀具磨损的机器视觉监测研究[J]. 机械科学与技术, 2019, 38(8): 1257-1263 doi: 10.13433/j.cnki.1003-8728.20180291

    PENG R T, JIANG H J, XU Y, et al. Study on tool wear monitoring using machine vision[J]. Mechanical Science and Technology for Aerospace Engineering, 2019, 38(8): 1257-1263 (in Chinese) doi: 10.13433/j.cnki.1003-8728.20180291
    [12] DAI Y Q, ZHU K P. A machine vision system for micro-milling tool condition monitoring[J]. Precision Engineering, 2018, 52: 183-191. doi: 10.1016/j.precisioneng.2017.12.006
    [13] WANG H Z, WANG G B, LI G Q, et al. Deep belief network based deterministic and probabilistic wind speed forecasting approach[J]. Applied Energy, 2016, 182: 80-93 doi: 10.1016/j.apenergy.2016.08.108
    [14] HINTON G E. Training products of experts by minimizing contrastive divergence[J]. Neural Computation, 2002, 14(8): 1771-1800 doi: 10.1162/089976602760128018
    [15] PIOTROWSKI A P, NAPIORKOWSKI J J, PIOTROWSKA A E. Impact of deep learning-based dropout on shallow neural networks applied to stream temperature modelling[J]. Earth-Science Reviews, 2020, 201: 103076 doi: 10.1016/j.earscirev.2019.103076
    [16] 卢志远, 马鹏飞, 肖江林, 等. 基于机床信息的加工过程刀具磨损状态在线监测[J]. 中国机械工程, 2019, 30(2): 220-225 doi: 10.3969/j.issn.1004-132X.2019.02.013

    LU Z Y, MA P F, XIAO J L, et al. On-line monitoring of tool wear conditions in machining processes based on machine tool data[J]. China Mechanical Engineering, 2019, 30(2): 220-225 (in Chinese) doi: 10.3969/j.issn.1004-132X.2019.02.013
  • 加载中
图(9) / 表(4)
计量
  • 文章访问数:  154
  • HTML全文浏览量:  39
  • PDF下载量:  18
  • 被引次数: 0
出版历程
  • 收稿日期:  2020-01-30
  • 刊出日期:  2021-07-01

目录

    /

    返回文章
    返回