摘要
针对铣刀切削时磨损特征声音信号较微弱、难以分解和提取的问题,提出改进变分模态分解VMD边际谱能量熵比与最小二乘支持向量机LS-SVM相结合的铣刀磨损状态监测方法,以多尺度排列熵为目标函数,利用自适应遗传算法优化VMD参数,获得信号的最佳分解。通过监测铣刀磨损状态,分析三种状态铣削声音信号,采用改进VMD提取信号的边际谱能量熵比特征,输入到LS-SVM中确定铣刀磨损状态。结果表明,该方法可以有效提取铣刀磨损状态信息,在小样本情况下,对铣刀磨损状态的监测识别准确率可达100%,具有较好的实用性。
This paper aims to address the weaker sound signal of the wear characteristics of the milling cutter and the resulting difficulties decomposing and extracting the signals and proposes a method of milling cutter wear state monitoring based on improved variational mode decomposition(VMD)marginal spectrum energy entropy ratio and least squares support vector machine(LS-SVM).This method works by taking multi-scale permutation entropy as the objective function and using adaptive genetic algorithm to optimize VMD parameters to obtain the best signal decomposition.The study involves analyzing the three-state milling sound signals by monitoring the wear status of the milling cutter,extracting the marginal spectrum energy-entropy ratio characteristics of the signal using the improved VMD,and inputting it into LS-SVM to determine the wear status of the milling cutter.The results show that the method enables the effective extraction of the wear status information of the milling cutter,and boasts a better practicability thanks to its 100%accuracy of monitoring and identifying the wear status of the milling cutter in the case of a small sample.
作者
周广林
舒贝贝
刘培江
Zhou Guanglin;Shu Beibei;Liu Peijiang(School of Mechanical Engineering, Heilongjiang University of Science & Technology, Harbin 150022, China)
出处
《黑龙江科技大学学报》
2021年第4期458-464,共7页
Journal of Heilongjiang University of Science And Technology
关键词
铣刀状态监测
变分模态分解
最小二乘支持向量机
边际谱能量熵比
声音信号
milling cutter status monitoring
variational modal decomposition
least squares support vector machine
marginal spectrum energy-entropy ratio
sound signal
