Equipment Deployment of Direct Tool-Condition Monitoring Based on Improved Information Entropy
-
摘要:
在不拆刀情况下，基于机器视觉的在线刀具状态监测系统可完成刀具磨损测量和状态评估，但与在线捕获刀具图像质量息息相关的系统部署参数选择却鲜有研究. 为解决上述问题，本文构建基于改进信息熵的多项式回归模型以实现刀具状态监测系统的最优部署. 首先，使用自适应阈值方法去除捕获刀具图像中背景要素干扰，并通过信息熵指标评估图像中刀具磨损区域的成像质量；然后，构建相机工作距离、曝光时间与所提出评价指标之间的多项式回归模型以描述部署参数与提出评价指标的映射关系；最后，应用最小二乘法求取多项式模型系数获得最优部署参数. 在确保自变量的因子水平涵盖最优部署参数情况下设计正交实验，实验结果表明：提出的评价指标与工作距离、曝光时间等部署参数之间均存在主效应关系，符合光学成像系统的变化规律；与支持向量机、决策树和K近邻（K-nearest neighbor, KNN）算法等非线性回归预测模型相比，三次多项式回归模型预测误差最小，其平均绝对误差、均方误差、均方根误差分别为0.022631, 0.00068, 0.026069；在多项式回归模型求解的最优部署参数下，所捕获的刀具图像的测量精度达到96.76%，提高0.74%，满足刀具状态监测的精度要求.
Abstract:In-situ tool-condition monitoring system based on machine vision realizes tool wear measurement and condition assessment without removing the tool. However, the system deployment parameters that are closely related to the quality of the tool image are rarely studied. To this end, a polynomial regression model based on improved information entropy is constructed to realize the optimal deployment of the tool-condition monitoring system. First, the adaptive threshold method is used to remove the interference of background elements in the captured tool image, and the imaging quality of the tool wear area is evaluated by the information entropy metric. Then, a polynomial regression model with respect to the camera working distance, exposure time, and the proposed evaluation metric is constructed to describe the mapping relationship between the deployment parameters and the proposed evaluation metric. Finally, the least squares method is used to solve the coefficients of the polynomial model and obtain the optimal deployment parameters. Orthogonal experiments are designed to ensure that the factor levels of independent variables cover the optimal deployment parameters. The experimental results show that there is a main effect relationship between the proposed evaluation metric and deployment parameters, such as working distance and exposure time, which is in line with the changing rule of optical imaging systems. Compared with nonlinear regression prediction models such as support vector machine, decision tree and K-nearest neighbor (KNN), the cubic polynomial regression model has the smallest prediction error, with its mean absolute error, mean square error, and root mean square error being 0.022631, 0.00068, and 0.026069, respectively. The measurement accuracy of the tool image captured under the optimal deployment parameters reaches 96.76%, increased by 0.74%, demonstrating that it meets the accuracy requirements of tool condition monitoring.
-
图 2 在不同参数设置下刀具图像特点
Figure 2. Tool image characteristics under different parameter settings
图 5 不同图像评价指标下的主效应对比
Figure 5. Comparison of main effects under different indexes of image quality assessment
图 7 不同部署情况下捕获图像的分割效果
Figure 7. Segmentation effect of captured images in different deployment conditions
表 1 相机工作距离和曝光时间的水平
Table 1. Levels of camera working distance and exposure time
因子水平 工作距离/mm 曝光时间/μs 1 −231.175 500 2 −231.375 700 3 −231.575 900 4 −231.775 1100 5 −231.975 1300 6 −232.175 1500 7 −232.375 1700 8 −232.575 1900 9 −232.775 2100 
下载: 导出CSV
表 2 不同最高项次数下的模型预测误差
Table 2. Model prediction errors under different highest terms
最高项数/次 eMAE eMSE eRMSE 2 0.025847 0.000758 0.027525 3 0.022631 0.000680 0.026069 4 0.023176 0.000799 0.028260 5 0.023122 0.000756 0.027500
下载: 导出CSV
表 3 不同回归模型的预测误差
Table 3. Model prediction errors under different regression models
回归模型 eMAE eMSE eRMSE DT 0.047296 0.003525 0.059371 SVM 0.081945 0.008955 0.094628 KNN 0.076204 0.009435 0.097136 PR 0.022631 0.000680 0.026069
下载: 导出CSV
-
[1] 高宏力,许明恒,傅攀. 基于集成神经网络的刀具磨损量监测[J]. 西南交通大学学报,2005,40(5): 641-644, 653.GAO Hongli, XU Mingheng, FU Pan. Tool wear mon-itoring based on integrated neutral networks[J]. Journal of Southwest Jiaotong University, 2005, 40(5): 641-644, 653. [2] 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.LU Zhiyuan, MA Pengfei, XIAO Jianglin, 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. [3] ZHOU J J, YU J B. Chisel edge wear measurement of high-speed steel twist drills based on machine vision[J]. Computers in Industry, 2021, 128: 103436.1-103436.12. doi: 10.1016/j.compind.2021.103436 [4] 胡一星,许黎明,范帆,等. 曲线磨削砂轮廓形的原位视觉检测和误差补偿[J]. 上海交通大学学报,2019,53(6): 654-659.HU Yixing, XU Liming, FAN Fan, et al. In-situ vision detection and compensation of wheel profile error in profile grinding[J]. Journal of Shanghai Jiao Tong University, 2019, 53(6): 654-659. [5] WANG P, LIU Z, GAO R X, et al. Heterogeneous data-driven hybrid machine learning for tool condition prognosis[J]. CIRP Annals, 2019, 68(1): 455-458. doi: 10.1016/j.cirp.2019.03.007 [6] 朱锟鹏,李刚. 基于刀具磨损映射关系的微细铣削力理论建模与试验研究[J]. 机械工程学报,2021,57(19): 246-259. doi: 10.3901/JME.2021.19.023ZHU Kunpeng, LI Gang. Theoretical modeling and experimental study of micro milling force based on tool wear mapping[J]. Journal of Mechanical Engineering, 2021, 57(19): 246-259. doi: 10.3901/JME.2021.19.023 [7] GUO C Z, MA Z L, GUO X, et al. Fast auto-focusing search algorithm for a high-speed and high-resolution camera based on the image histogram feature function[J]. Applied Optics, 2018, 57(34): F44-F49. doi: 10.1364/AO.57.000F44 [8] KIM J, CHO Y, KIM A. Exposure control using Bayesian optimization based on entropy weighted image gradient[C]//2018 IEEE International Conference on Robotics and Automation (ICRA). Brisbane: IEEE, 2018: 857-864. [9] 陈健,李诗云,林丽,等. 模糊失真图像无参考质量评价综述[J]. 自动化学报,2022,48(3): 689-711.CHEN Jian, LI Shiyun, LIN Li, et al. A review on no-reference quality assessment for blurred image[J]. Acta Automatica Sinica, 2022, 48(3): 689-711. [10] 曹同坤,徐英涛,谈庆瑶. 在刀-屑界面持续润滑刀具切削45钢的性能及润滑机理[J]. 中国机械工程,2021,32(20): 2411-2417,2426.CAO Tongkun, XU Yingtao, TAN Qingyao. Cutting performance and lubrication mechanism of cutting 45 steel with tool continuously lubricated at the tool-chip interface[J]. China Mechanical Engineering, 2021, 32(20): 2411-2417,2426. [11] 李杰. 空域无参考图像质量评价方法研究[D]. 武汉: 武汉大学, 2017. [12] YOU Z C, GAO H L, GUO L, et al. Machine vision based adaptive online condition monitoring for milling cutter under spindle rotation[J]. Mechanical Systems and Signal Processing, 2022, 171: 108904.1-108904.18. doi: 10.1016/j.ymssp.2022.108904 [13] KUMAR A, CHINNAM R B, TSENG F. An HMM and polynomial regression based approach for remaining useful life and health state estimation of cutting tools[J]. Computers & Industrial Engineering, 2019, 128: 1008-1014. [14] 吕晓玲, 宋捷. 大数据挖掘与统计机器学习[M]. 北京: 中国人民大学出版社, 2016. -
下载:
点击查看大图
计量
- 文章访问数: 393
- HTML全文浏览量: 127
- PDF下载量: 58
- 被引次数: 0