摘要
由于刀具磨损声发射信号的能量分布与刀具磨损状态密切相关,可以利用谐波小波包方法提取刀具磨损声发射信号的特征能量,对各频段能量做归一化处理,与切削三要素组成特征向量输入到Elman神经网络,通过神经网络判别刀具磨损状态。实验结果表明,刀具磨损产生的声发射信号频率主要集中在10Hz^130k Hz之间,将谐波小波包和Elman神经网络结合的方法可以有效地识别刀具磨损状态。
Because of tool wear acoustic emission signal energy distribution was closely related to tool wear state. Use the harmonic wavelet packet to extract the AE signal characteristic energy of the tool wear. The normalized data of energy spectrum and the three elements of cutting were seen as Elman neural network input vector. Through neural network discrimination tool wear state.The experiment result express that the acoustic emission signals of tool wear are mainly concentrated in the frequency of 10 Hz^130 k Hz. The method of tool wear based on harmonic wavelet packet and Elman neural network can effectively recognize the tool wear state.
出处
《现代制造工程》
CSCD
北大核心
2015年第12期78-81,85,共5页
Modern Manufacturing Engineering
基金
辽宁省重点实验室项目(LS2010117)
关键词
刀具磨损
声发射信号
谐波小波包
神经网络
tool wear
acoustic emission signal
harmonic wavelet
neural network
