切削颤振监测的小波参数估计法  

Cutting Chatter Detection via Wavelet-based Parameter Estimation

作　　者：胡健闻[1] 刘昶[1] 柴苍修[1] 汤翔[2] 

机构地区：[1]武汉理工大学机电工程学院,430070 [2]武汉大学电子与信息学院,430072

出　　处：《精密制造与自动化》2004年第1期15-17,共3页Precise Manufacturing & Automation

摘　　要：金属切削时的振动信号具有独立于尺度和随机的特征,与1/f过程相似。用1/f过程功率谱中的指数γ作为特征量可以用来检测颤振的发生。实验表明,切削颤振的产生伴随着参数γ的急剧的改变。本文提出了一种新的用于估计γ的基于小波的最小二乘算法。与极大似然估计法相比,它具有更广的适用性。通过实验结果对比可以看出,用γ的相对变化量作为特征量更为合适。Vibration signals in metal cutting are scale-invariant, random and similar to 1/f process. The exponentγin the power spectra of the 1/f process can serve as anindex used to detect the occurrence of chatter. Experimentsshow, the occurrence of cutting chatter accompanies sharpchange of exponentγ. A new wavelet-based least-squarealgorithm to estimateγis proposed in the paper. Comparedwith maximum likelihood algorithm, it is adaptable to awider range of applications. In addition, comparison amongexperiments points out that a relative changing quantity ofγis a more suitable index for chatter detection.

关 键 词：切削颤振 最小二乘法 小波变换 状态监测 机床 参数估计 

分 类 号：TG506[金属学及工艺—金属切削加工及机床]