基于支持向量机与粒子滤波的刀具磨损状态识别  

Tools' wear state recognition based on support vector machine and particle filtering

作　　者：程灿 李建勇[1,2] 徐文胜 聂蒙[1,2] CHENG Can;LI Jianyong;XU Wensheng;NIE Meng(School of Mechanical,Electronic and Control Engineering,Beijing Jiaotong University,Beijing 100044,China;MOE Key Lab of Vehicles’Advanced Manufacturing,Measuring and Control Technology,Beijing 100044,China)

机构地区：[1]北京交通大学机械与电子控制工程学院,北京100044 [2]载运工具先进制造与测控技术教育部重点实验室,北京100044

出　　处：《振动与冲击》2018年第17期48-55,71,共9页Journal of Vibration and Shock

基　　金：中央高校基本科研业务费(M16JB00210)

摘　　要：为了提高小样本情况下刀具磨损量识别的精度,提出一种基于支持向量机和粒子滤波的刀具磨损量识别方法。针对支持向量机的输入特征选择和参数选择难题,建立支持向量机输入特征与参数优化双层规划模型,并组合遗传算法和人工蜂群算法进行求解。之后,利用粒子滤波方法对支持向量机回归得到的结果进行修正。实验结果表明,在小样本情况下,基于支持向量机和粒子滤波的刀具磨损量识别方法具备良好的学习能力,能够精确地识别刀具的磨损量。To improve the accuracy of tools’wear recognition under the case of small samples,a tools’wear recognition method based on support vector machine(SVM)and particle filtering was proposed.Aiming at difficult problems of SVM’s feature selecting and parametric selecting,a bi-level programming model for SVM’s input features and parametric optimization was developed and solved using the genetic algorithm and the artificial bee colony algorithm.Then,the particle filtering method was used to modify tools’wear recognition results obtained using the regression algorithm of SVM.Test results showed that under the case of small samples,the tools’wear recognition method based on SVM and particle filtering has a good learning capability and can recognize tools’wear precisely.

关 键 词：刀具磨损 支持向量机 双层规划 遗传算法 人工蜂群 粒子滤波 

分 类 号：TG71[金属学及工艺—刀具与模具] TH117[机械工程—机械设计及理论]