基于小波包敏感频带选择的复材铣边颤振监测研究

Investigation on Chatter Monitoring of Composite Milling Edge Based on the Selection of Sensitive Frequency Band of Wavelet Packet

针对航空蒙皮类复合材料零件铣边颤振初期频率淹没问题,提出了一种基于小波包敏感频带选择的能量熵颤振监测方法。首先使用频率消除算法(Frequency elimination algorithm,FEA)对铣边采集信号进行去主轴转频、刀齿齿频及其倍频处理。其次对去频后的信号进行小波包分解(Wavelet packet decomposition,WPD),分析WPD的频带分布特性。然后基于频带相关系数ρ和能量比波动方差σ^(2)选取颤振敏感频带作为颤振研究对象,提取分解后信号能量熵作为颤振监测量并基于拉依达准则确定系统颤振监测阈值。最后通过仿真信号和试验验证了所提算法的有效性。结果表明,经FEA和敏感频带筛选后信号的能量熵值分辨率提高了205%,且相较于未去频信号系统颤振监测响应时间能提前0.5 s。

A novel energy entropy chatter monitoring method based on the selection of sensitive frequency band of wavelet packet is proposed to solve the problem of the initial chatter frequency submerged in the milling of large aviation skin composite materials.Firstly,the frequency elimination algorithm(FEA)is performed to remove the spindle rotation frequency,tooth passing frequency and their harmonics on the signals collected by the edge milling.Secondly,wavelet packet transform(WPD)is applied to decompose the filtered signal into a set of frequency bands and the frequency band distribution characteristics of WPD are analyzed.Then,sensitive frequency bands containing rich chatter information are selected as the research object based on the frequency band correlation coefficientρand energy ratio fluctuation varianceσ^(2) and energy entropy is extracted as the chatter monitoring feature.Meanwhile,PauTa criterion is used for determining the chatter monitoring threshold of the system.Finally,the simulation signal and experimental signal are employed to verify the effectiveness of the algorithm.The results showed that the energy entropy resolution ratio of the signal is increased by 202%after the FEA and sensitive frequency band selection,compared with the original signal,and the system chatter condition can be detected 0.5s earlier.