AdaBoost Algorithm Enabled Integrated Algorithm of Staged Recognition of Cutting Tool Wear
-
摘要: 提出了一种基于AdaBoost(Adaptive boosting)集成算法的刀具磨损全阶段回归模型建模方法。首先,利用获取到的加工过程信号和刀具磨损值,建立刀具磨损拟合曲线,以实现对初期磨损、平稳磨损和急剧磨损3个阶段的准确划分;其次,对加工过程信号进行特征提取,并与相应的刀具磨损值形成3个阶段的数据样本,利用支持向量机分别建立3个磨损阶段的回归模型;再次,利用AdaBoost在全阶段上确定3个磨损阶段回归模型的权重,最终建立刀具磨损状态识别的回归模型;最后,以某铣刀切削过程采集的刀具磨损数据集验证所提出的模型和方法的有效性。Abstract: A regression method of full-stage wear of cutting tool based on the AdaBoost (Adaptive Boosting) integrated algorithm is proposed. Firstly, using the acquired machining process signals and tool wear values, a fitting curve of cutting tool wear is established to achieve an accurate division in the initial wear stage, smooth wear stage, and sharp wear stage. Secondly, the three-stage data samples with the corresponding wear values of cutting tool by extracting the data feature from the machining process signals are obtained. The regression models for the three stages are established with the support vector machine. Thirdly, the AdaBoost algorithm is used to determine the weights of the three regression models in the three stages, and a regression model is established of full-stage wear regression. Finally, the effectiveness of the present model and method is verified with the wear data of a cutting tool collected in the milling cutter.
-
Key words:
- cutting tool wear /
- regression model /
- integrated algorithm
-
表 1 时域特征参数
特征参数 计算公式 平均值 均方根 标准差 偏度 峭度因子 波形因子 峰值因子 脉冲因子 裕度因子 方差 表 2 频域特征参数
特征参数 计算公式 重心频率 频率方差 均方频率 表 3 磨损阶段划分结果
磨损阶段 加工时间/s 走刀次数 初期磨损 t < 324.338 1~75 平稳磨损 324.338≤t < 552.682 76~127 急剧磨损 t>552.682 128~315 表 4 全阶段刀刃1磨损识别结果
算法 平均绝对误差/μm AdaBoost 0.768 SVR 12.120 CART 10.146 表 5 分阶段刀刃1磨损识别结果
算法 初期阶段
MAE/μm平稳阶段
MAE/μm急剧阶段
MAE/μm加权综合
MAE/μmAdaBoost 0.985 0.000 0.414 0.480 SVR 1.957 0.004 0.812 0.949 CART 2.886 0.031 0.898 1.219 -
[1] 刘献礼, 刘强, 岳彩旭, 等. 切削过程中的智能技术[J]. 机械工程学报, 2018, 54(16): 45-61 https://www.cnki.com.cn/Article/CJFDTOTAL-JXXB201816005.htmLIU X L, LIU Q, YUE C X, et al. Intelligent machining technology in cutting process[J]. Journal of Mechanical Engineering, 2018, 54(16): 45-61 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JXXB201816005.htm [2] RIZAL M, GHANI J A, NUAWI M Z, et al. A review of sensor system and application in milling process for tool condition monitoring[J]. Research Journal of Applied Sciences, Engineering and Technology, 2014, 7(10): 2083-2097 doi: 10.19026/rjaset.7.502 [3] 陶耀东, 曾广圣, 李宁. 基于回归树和AdaBoost方法的刀具磨损评估[J]. 计算机系统应用, 2017, 26(12): 212-219 https://www.cnki.com.cn/Article/CJFDTOTAL-XTYY201712035.htmTAO Y D, ZENG G S, LI N. Tool wear evaluation based on decision tree regression and AdaBoost algorithm[J]. Computer Systems & Applications, 2017, 26(12): 212-219 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-XTYY201712035.htm [4] 孙玉晶. 钛合金铣削加工过程参量建模及刀具磨损状态预测[D]. 济南: 山东大学, 2014SUN Y J. Parametric modeling of milling Titannium alloy and prediction of tool wear state[D]. Jinan: Shandong University, 2014 (in Chinese) [5] 吴雪峰, 刘亚辉, 毕淞泽. 基于卷积神经网络刀具磨损类型的智能识别[J]. 计算机集成制造系统, 2020, 26(10): 2762-2771 https://www.cnki.com.cn/Article/CJFDTOTAL-JSJJ202010016.htmWU X F, LIU Y H, BI S Z. Intelligent recognition of tool wear type based on convolutional neural networks[J]. Computer Integrated Manufacturing Systems, 2020, 26(10): 2762-2771 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JSJJ202010016.htm [6] 丁宝洋, 白倩, 刘具龙, 等. 基于Simulink反馈方法的钛合金铣削刀具磨损预测[J]. 机械工程学报, 2019, 55(11): 224-232 https://www.cnki.com.cn/Article/CJFDTOTAL-JXXB201911027.htmDING B Y, BAI Q, LIU J L, et al. A novel method for tool wear prediction in titanium milling by Simulink feedback method[J]. Journal of Mechanical Engineering, 2019, 55(11): 224-232 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JXXB201911027.htm [7] 孙惠斌, 潘军林, 张纪铎, 等. 面向切削过程的刀具数字孪生模型[J]. 计算机集成制造系统, 2019, 25(6): 1474-1480 https://www.cnki.com.cn/Article/CJFDTOTAL-JSJJ201906015.htmSUN H B, PAN J L, ZHANG J L, et al. Digital twin model for cutting tools in machining process[J]. Computer Integrated Manufacturing Systems, 2019, 25(6): 1474-1480 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JSJJ201906015.htm [8] 秦国华, 谢文斌, 王华敏. 基于神经网络与遗传算法的刀具磨损检测与控制[J]. 光学精密工程, 2015, 23(5): 1314-1321 https://www.cnki.com.cn/Article/CJFDTOTAL-GXJM201505015.htmQIN G H, XIE W B, WANG H M. Detection and control for tool wear based on neural network and genetic algorithm[J]. Optics and Precision Engineering, 2015, 23(5): 1314-1321 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-GXJM201505015.htm [9] 郭一君, 周杰, 王时龙, 等. 基于改进人工蜂群算法和极限学习机的刀具磨损监测[J]. 重庆大学学报, 2018, 41(6): 1-8 https://www.cnki.com.cn/Article/CJFDTOTAL-FIVE201806001.htmGUO Y J, ZHOU J, WANG S L, et al. Tool wearmonitoring based on improved artificial bee colony algorithm and extreme learning machine[J]. Journal of Chongqing University, 2018, 41(6): 1-8 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-FIVE201806001.htm [10] 赵帅, 黄亦翔, 王浩任, 等. 基于随机森林与主成分分析的刀具磨损评估[J]. 机械工程学报, 2017, 53(21): 181-189 https://www.cnki.com.cn/Article/CJFDTOTAL-JXXB201721022.htmZHAO S, HUANG Y X, WANG H R, et al. Random forest and principle components analysis based on health assessment methodology for tool wear[J]. Journal of Mechanical Engineering, 2017, 53(21): 181-189 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JXXB201721022.htm [11] 卢治功, 贺鹏, 职连杰, 等. 基于最小二乘法多项式拟合三角测量模型研究[J]. 应用光学, 2019, 40(5): 853-858 https://www.cnki.com.cn/Article/CJFDTOTAL-YYGX201905023.htmLU Z G, HE P, ZHI L J, et al. Laser triangulation measurement model based on least square polynomial fitting method[J]. Journal of Applied Optics, 2019, 40(5): 853-858 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-YYGX201905023.htm [12] 宋伟杰, 关山, 庞弘阳. 基于希尔伯特-黄变换和等距特征映射的刀具磨损状态监测[J]. 组合机床与自动化加工技术, 2018(6): 114-118 https://www.cnki.com.cn/Article/CJFDTOTAL-ZHJC201806029.htmSONG W J, GUAN S, PANG H Y. Tool wear state monitoring based on Hilbert-Huang transform and isometric feature mapping[J]. Modular Machine Tool & Automatic Manufacturing Technique, 2018(6): 114-118 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZHJC201806029.htm [13] 靳同红, 莫正波, 郑德亮, 等. 时频分析技术及应用研究[J]. 机械科学与技术, 2009, 28(1): 75-78 https://www.cnki.com.cn/Article/CJFDTOTAL-JXKX200901018.htmJIN T H, MO Z B, ZHENG D L, et al. A review of time-frequency analysis and its application[J]. Mechanical Science and Technology for Aerospace Engineering, 2009, 28(1): 75-78 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JXKX200901018.htm [14] 魏大千, 王波, 刘涤尘, 等. 基于时序数据相关性挖掘的WAMS/SCADA数据融合方法[J]. 高电压技术, 2016, 42(1): 315-320 https://www.cnki.com.cn/Article/CJFDTOTAL-GDYJ201601042.htmWEI D Q, WANG B, LIU D C, et al. WAMS/SCADA data fusion method based on time-series data correlation mining[J]. High Voltage Engineering, 2016, 42(1): 315-320 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-GDYJ201601042.htm [15] 杨叶梅, 陈新. 利用惯性传感器和AdaBoost算法的步态识别方法[J]. 计算机应用研究, 2019, 36(4): 1212-1216, 1221 https://www.cnki.com.cn/Article/CJFDTOTAL-JSYJ201904058.htmYANG Y M, CHEN X. Gait recognition based on inertial sensor and AdaBoost algorithm[J]. Application Research of Computers, 2019, 36(4): 1212-1216, 1221 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JSYJ201904058.htm [16] SUN X F, LI Y. Denoising of desert seismic signal based on synchrosqueezing transform and Adaboost algorithm[J]. Acta Geophysica, 2020, 68(2): 403-412 doi: 10.1007/s11600-020-00408-1 [17] LI X, LIN B S, ZHOU J H, et al. Fuzzy neural network modelling for tool wear estimation in dry milling operation[C]//Proceedings of the Annual Conference of the Prognostics and Health Management Society. United States: PHM, 2009