基于PSO优化LS-SVM的刀具磨损状态识别

发布时间：2018-04-05 10:42

本文选题：刀具状态识别　切入点：时域特征值　出处：《清华大学学报(自然科学版)》2017年09期

【摘要】：为监测刀具的磨损状态,该文建立了一个基于声发射的刀具磨损状态监测系统。在刀具磨损状态监测实验中,采集加工过程中的声发射信号,提取方根幅值、绝对值均值、均方根、最大值作为反映刀具磨损的时域特征值。针对人工神经网络容易陷入局部极小值、结构难以确定、学习收敛速度慢等缺点,提出最小二乘支持向量机(least square support vector machine,LS-SVM)的刀具磨损状态识别方法。针对LS-SVM性能依赖于惩罚因子和核参数,利用粒子群优化(particle swarm optimization,PSO)算法对LSSVM参数进行自动寻优,建立PSO优化LS-SVM模型进行刀具磨损状态识别。结果表明:与LS-SVM识别模型相比,优化后的LS-SVM模型具有更高的识别率。
[Abstract]:In order to monitor tool wear, a tool wear monitoring system based on acoustic emission is established in this paper.In the tool wear monitoring experiment, the acoustic emission signals are collected and the square root amplitude, the absolute mean value, the mean square root and the maximum value are taken as the time domain eigenvalues to reflect the tool wear.In view of the disadvantages of artificial neural network (Ann), such as easy to fall into local minima, difficult to determine the structure and slow learning convergence, a tool wear recognition method based on least square support vector machine (LS-SVM) is proposed.For LS-SVM performance depends on penalty factor and kernel parameter, particle swarm optimization (PSO) algorithm is used to optimize LSSVM parameters automatically, and PSO optimization LS-SVM model is established to identify tool wear state.The results show that compared with the LS-SVM recognition model, the optimized LS-SVM model has a higher recognition rate.
【作者单位】：清华大学机械工程系;清华大学精密超精密制造装备及控制北京市重点实验室;电子科技大学机械电子工程学院;海军航空工程学院飞行器工程系;
【基金】：国家科技重大专项(2013ZX04001-021)
【分类号】：TG71;TH871.6;TP18

【相似文献】