Monitoring cutting tool wear based on spindle current signal multi-feature fusion
-
摘要: 为实现在正常生产条件下进行刀具磨损的长期在线监测, 提出了基于主轴电流信号和粒子群优化支持向量机模型(PSO-SVM)的刀具磨损状态间接监测方法。首先对数控机床主轴电机电流信号进行分析, 将与刀具磨损相关的主轴电流信号多个特征参数和EMD能量熵进行特征融合作为输入特征向量; 其次, 通过粒子群寻优算法(PSO)对支持向量机模型(SVM)参数进行优化, 建立基于主轴电流信号融合特征和PSO-SVM理论的刀具磨损状态识别模型; 最后, 通过实验采集某立式加工中心主轴在刀具不同磨损状态下电流信号进行验证, 并与传统SVM模型、BP神经网络模型进行了对比分析。结果表明, 所提出的方法具有较高的准确率和较好的泛化能力。能够实现正常生产条件下对刀具磨损的长期在线监测。Abstract: In order to gain long-term online monitoring cutting tool wear data under normal cutting conditions, a new method for monitoring cutting tool wear was proposed based on the spindle current signal and particle swarm optimization support vector machine (PSO-SVM) model. Firstly, the spindle motor current signals of CNC machine tool were analyzed. Multiple feature parameters related to tool wear and the EMD energy entropy were fused as an input feature vector. Secondly, the SVM model parameters were optimized by the PSO algorithm. The tool wear condition recognition model was established based on the spindle current signal and PSO-SVM theory. Finally, the spindle current signals of the vertical machining center under different tool wear conditions were collected by the experiment. The proposed method was compared with traditional SVM model and BP neural network model. The results show that the proposed method has higher recognition accuracy and better generalization ability. The proposed method can realize the long-term online monitoring of the tool wear condition.
-
表 1 各统计特征计算公式
名称 公式 名称 公式 平均值 $ \overline x=\frac{1}{N} \sum\limits_{1}^{N} x_{i}$ 均方根 $\sqrt{\frac{1}{N} \sum\limits_{i=1}^{N} x(i)^{2}}$ 波形因子 $\frac{\sqrt{\frac{1}{N} \sum\limits_{i=1}^{N} x(i)^{2}}}{\frac{1}{N} \sum\limits_{i=1}^{N}|x(i)|}$ 波峰因子 $\frac{x_{\max }}{\sqrt{\frac{1}{N} \sum\limits_{i=1}^{N} x(i)^{2}}}$ 
下载: 导出CSV
表 2 电流信号特征值与刀具状态数据
序号 1 2 ⋯ 61 62 ⋯ 平均值 7.729 7.876 ⋯ 8.056 8.137 ⋯ 均方根 8.381 8.5412 ⋯ 8.841 8.817 ⋯ 波峰因子 1.597 1.615 ⋯ 1.698 1.707 ⋯ 波形因子(×10-7) 6.085 6.102 ⋯ 6.174 6.213 ⋯ EMD能量熵 0.712 0.719 ⋯ 0.479 0.471 ⋯ 刀具状态 1 1 ⋯ 2 2 ⋯
下载: 导出CSV
-
[1] 李斌, 张琛, 刘红奇. 基于主轴电流的铣削力间接测量方法研究[J]. 华中科技大学学报, 2008, 36(3): 5-11. doi: 10.3321/j.issn:1671-4512.2008.03.002 [2] 马旭, 陈捷. 数控铣床刀具磨损量与主轴电流关系研究[J]. 组合机床与自动化加工技术, 2009(10): 88-90. doi: 10.3969/j.issn.1001-2265.2009.10.024 [3] 王军平, 敬忠良, 王安. 基于随机模糊神经网络的刀具磨损量软测量技术[J]. 信息与控制, 2002, 31(6): 534 -537. doi: 10.3969/j.issn.1002-0411.2002.06.012 [4] 刘智鹏. 基于多传感器信息融合的刀具磨损状态模式识别和寿命预测研究[D]. 成都: 西南交通大学, 2015. [5] 李锡文, 杜润生, 杨叔子. 立铣刀渐进磨损过程中主电机功率信号的时域特性研究[J]. 工具技术, 2000, 34(12): 7-9. doi: 10.3969/j.issn.1000-7008.2000.12.002 [6] 董慧. 基于HHT的刀具磨损状态监测技术研究[D]. 沈阳: 沈阳航空航天大学, 2012. [7] 贺灵敏. EMD和支持向量机在刀具故障诊断中的应用[D]. 成都: 西华大学, 2010. [8] 武滢, 舒启林. 基于时频特征和PSO-SVR模型的零部件剩余寿命分析[J]. 组合机床与自动化加工技术, 2021(5): 1-5. https://www.cnki.com.cn/Article/CJFDTOTAL-ZHJC202105002.htm [9] García Nieto P J, García-Gonzalo E, Sánchez Lasheras F, et al. Hybrid PSO–SVM-based method for forecasting of the remaining useful life for aircraft engines and evaluation of its reliability [J]. Reliability Engineering and System Safety, 2015, 138: 219-231. doi: 10.1016/j.ress.2015.02.001 [10] Tran H D, Li Hai-zhou. Sound event recognition with probabilistic distance SVMs [J]. IEEE Transactions on Audio, Speech, and Language Processing, 2011, 19(6): 1556-1568. doi: 10.1109/TASL.2010.2093519 [11] 于雪晶, 麻肖妃, 夏斌. 动态粒子群优化算法[J]. 计算机工程, 2010, 36(4): 193-197. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC201004070.htm [12] 王春和. 基于主轴电流的车刀寿命预测技术研究[D]. 武汉: 华中科技大学, 2017. -
下载:
点击查看大图
图(4) / 表(3)
计量
- 文章访问数: 49
- HTML全文浏览量: 83
- PDF下载量: 30
- 被引次数: 0
