Study on On-line Recognition Method for Thermal Deformation State of End Mill
-
摘要: 机床在铣削加工时总伴随着大量的切削热产生, 为了减少热变形对刀具加工精度的影响, 本文以硬质合金立铣刀作为研究对象, 提出一种基于BP神经网络的立铣刀在铣削过程中热变形实时状态的识别方法。搭建测试平台进行立铣刀热误差实验, 并设计了一种立铣刀热变形变形量的直接测量法。采集机床连续加工期间主轴的温度信号和立铣刀热误差变形量, 对温度信号进行特征提取。将刀具不同热变形状态及相关的特征值, 输入到8-4-2的三层BP神经网络模型进行训练。实验结果表明: 刀具热变形识别系统识别率在为87.2%左右。Abstract: In order to reduce the influence of the thermal deformation on the machining accuracy of the cutter, taking the carbide end milling cutter as the research object, a method based on BP neural network to identify the real-time thermal error state of the end milling cutter in the milling is proposed. A test platform was built to carry out the thermal error experiment of end milling cutter, and a direct measurement method of thermal error deformation of end milling cutter was designed to collect the temperature signal of spindle during continuous addition of machine tool and thermal error deformation of end milling cutter. The different thermal deformation states of the cutter and their associated eigenvalues were input into the 8-4-2 three-layer BP neural network model. The experimental results show that the recognition rate of the tool thermal deformation recognition system is about 87.2%.
-
Key words:
- end milling cutter /
- sensor /
- thermal deformation /
- BP neural network
-
表 1 切削参数
主轴转速/ (r·min-1) 进给速度/ (mm·min-1) 切削深度/mm 切削宽度/mm 3 000 800 0.05 6 表 2 传感器规格
传感器名称 工作温度/℃ 工作湿度/RH 测量范围 测量精度 HL-G1激光位移传感器 -10~45 35~85 50 mm 0.5 μm PT100热电阻传感器 - - -30~200 ℃ 0.03 ℃ 表 3 重复定位实验数据
实验组号 测量距离/mm 1 66.897 2 66.895 3 66.895 4 66.892 5 66.894 表 4 刀具热变形值
测量次数 测量值/mm 基准值/mm 热变形值/mm 1 66.658 0 66.658 0 2 66.670 0 66.658 0.012 0 3 66.673 0 66.658 0.015 0 4 66.675 5 66.658 0.017 5 5 66.674 0 66.658 0.016 0 6 66.676 0 66.658 0.018 0 7 66.677 0 66.658 0.019 0 8 66.676 5 66.658 0.018 5 9 66.680 0 66.658 0.022 0 10 66.679 0 66.658 0.021 0 11 66.680 0 66.658 0.022 0 12 66.678 0 66.658 0.020 0 ⋮ ⋮ ⋮ ⋮ 表 5 归一化后的特征
传感器编号 平均值 有效值 方差 1 0.619 74 0.192 43 0.131 32 2 0.721 16 0.182 21 0.141 35 3 0.879 60 0.188 00 0.153 58 4 0.713 02 0.201 47 0.170 32 表 6 不同隐含层的模型对应训练误差
隐含层的层数 第1层节点数 第2层节点数 第3层节点数 训练误差 1 4 0 0 0.32 1 8 0 0 0.28 1 12 0 0 0.24 2 4 2 0 0.26 2 8 2 0 0.24 2 12 2 0 0.25 3 4 4 2 0.24 3 8 4 2 0.19 3 12 4 2 0.22 表 7 BP神经网络的鲁棒性测试结果
实验序号 刀具需要补偿的识别率 刀具不需要补偿的识别率 1 0.873 0.911 2 0.863 0.893 3 0.888 0.902 4 0.867 0.911 平均值 0.872 0.904 -
[1] 王时龙, 祁鹏, 周杰, 等. 数控滚齿机热变形误差分析与补偿新方法[J]. 重庆大学学报, 2011, 34(3): 13-17 https://www.cnki.com.cn/Article/CJFDTOTAL-FIVE201103002.htmWANG S L, QI P, ZHOU J, et al. Thermal deformation error analysis and a novel compensation method for NC gear hobbing machine tools[J]. Journal of Chongqing University, 2011, 34(3): 13-17 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-FIVE201103002.htm [2] 刘涛, 韩晓静. 基于热源法的涡旋盘高速精密铣削切削热研究[J]. 机械设计与制造工程, 2018, 47(2): 25-29 https://www.cnki.com.cn/Article/CJFDTOTAL-JXZZ201802006.htmLIU T, HAN X J. Research on the cutting heat of high speed milling scroll plates based on the heat source method[J]. Machine Design and Manufacturing Engineering, 2008, 47(2): 25-29 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JXZZ201802006.htm [3] 关立文, 杨亮亮, 王立平, 等. "S"形试件间歇性切削温度场建模与分析[J]. 清华大学学报(自然科学版), 2016, 56(2): 192-199 https://www.cnki.com.cn/Article/CJFDTOTAL-QHXB201602012.htmGUAN L W, YANG L L, WANG L P, et al. Modeling and Analysis of intermittent cutting temperature field for the "S" test specimens[J]. Journal of Tsinghua University (Science and Technology), 2016, 56(2): 192-199 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-QHXB201602012.htm [4] 刘胜. 钛合金正交切削的温度场和切削力仿真与试验研究[D]. 南京: 南京航空航天大学, 2007LIU S. FEM Simulation and experiment research of cutting temperature and force in orthogonal cutting of titanium alloys[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2007 (in Chinese) [5] 迟晓明, 张小栋, 张凯. 高速数控车床刀具热变形的计算分析[J]. 机械设计, 2011, 28(11): 72-76 https://www.cnki.com.cn/Article/CJFDTOTAL-JXSJ201111019.htmCHI X M, ZHANG X D, ZHANG K. Calculation and analysis on turning tool thermal deformation of high-speed NC lathe[J]. Journal of Machine Design, 2011, 28(11): 72-76 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JXSJ201111019.htm [6] 李静, 王占礼, 王伟冰. 车削加工仿真中刀具热变形对工件加工精度的影响[J]. 工具技术, 2009, 43(6): 67-69 https://www.cnki.com.cn/Article/CJFDTOTAL-GJJS200906027.htmLI J, WANG Z L, WANG W B. Influence of tool thermal deformation on workpiece machining accuracy in turning machining simulation[J]. Tool Engineering, 2009, 43(6): 67-69 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-GJJS200906027.htm [7] 曹岩, 刘新户, 付雷杰, 等. 整体式立铣刀铣削加工中温度场有限元分析[J]. 西安工业大学学报, 2017, 37(5): 378-385 https://www.cnki.com.cn/Article/CJFDTOTAL-XAGY201705007.htmCAO Y, LIU X H, FU L J, et al. Finite element analysis of milling temperature field of an integral end milling cutter[J]. Journal of Xi'an Technological University, 2017, 37(5): 378-385 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-XAGY201705007.htm [8] 王占礼, 杜爽, 王尧. 仿真加工中刀具热变形引起的加工误差分析[J]. 工具技术, 2006, 40(6): 47-49 https://www.cnki.com.cn/Article/CJFDTOTAL-GJJS200606013.htmWANG Z L, DU S, WANG Y. Analysis of machining error caused by thermal distortion of cutting tool in simulation machining[J]. Tool Engineering, 2006, 40(6): 47-49 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-GJJS200606013.htm [9] 王晓强, 张云, 周华民, 等. 基于隐马尔科夫模型的刀具磨损连续监测[J]. 组合机床与自动化加工技术, 2016(10): 87-90 https://www.cnki.com.cn/Article/CJFDTOTAL-ZHJC201610023.htmWANG X Q, ZHANG Y, ZHOU H M, et al. Continuous tool wear monitoring based on hidden Markov model[J]. Modular Machine Tool & Automatic Manufacturing Technique, 2016(10): 87-90 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZHJC201610023.htm [10] 何一千, 黄民, 孙巍伟. 基于BP神经网络的数控机床刀具磨损状态识别方法[J]. 设备管理与维修, 2019(17): 37-40 https://www.cnki.com.cn/Article/CJFDTOTAL-SBGX201917020.htmHE Y Q, HUANG M, SUN W W. A method of tool wear identification for NC machine tool based on BP neural network[J]. Plant Maintenance Engineering, 2019(17): 37-40 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-SBGX201917020.htm [11] 关山, 石志标, 刘炎. 基于多特征融合的刀具磨损识别方法[J]. 振动、测试与诊断, 2014, 34(3): 576-584 https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCS201403032.htmGUAN S, SHI Z B, LIU Y. Identification method of tool wear based on multi-feature fusion[J]. Journal of Vibration, Measurement & Diagnosis, 2014, 34(3): 576-584 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCS201403032.htm [12] ZHU K P, HONG G S, WONG Y S. A comparative study of feature selection for hidden Markov model-based micro-milling tool wear monitoring[J]. Machining Science and Technology, 2008, 12(3): 348-369 [13] 陈世平, 刘状, 闫文, 等. 干切滚齿机热变形误差模型及其误差补偿技术研究[J]. 制造技术与机床, 2019(11): 119-124 https://www.cnki.com.cn/Article/CJFDTOTAL-ZJYC201911024.htmCHEN S P, LIU Z, YAN W, et al. Study on thermal deformation error model and compensation technology of dry-cutting hobbing machine[J]. Manufacturing Technology & Machine Tool, 2019(11): 119-124 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZJYC201911024.htm [14] 闫文. 干切滚齿机床热误差实时检测与补偿的方法研究[D]. 重庆: 重庆理工大学, 2020YAN W. Research on real-time detection and compensation method for thermal error of dry-cut hobbing machine[D]. Chongqing: Chongqing University of Technology, 2020 (in Chinese) [15] 郭艳. 基于时频分析与智能算法的滚动轴承智能诊断方法[D]. 南昌: 华东交通大学, 2015GUO Y. Intelligent rolling bearing diagnosis method based on time-frequency analysis and intelligent algorithm[D]. Nanchang: East China Jiaotong University, 2015 (in Chinese)