摘要
以超高斯函数为基础,构造出一类用递推公式进行小波变换的小波基,提出一个新的递归小波基,并对其时频特性进行了分析。基于框架小波神经网络理论,利用连续函数介值定理,构造出一种紧致型小波网络,并对其初始化与学习算法进行了研究。最后,对刀具AE信号进行递归小波分解,提取特征并应用小波网络识别刀具状态,识别率达到100%。
Based on the Super-Gaussian function, a general method of recursive mother wavelet was introduced, and an optimal method of wavelet construction was proposed. The time-frequency characteristics of recursive wavelet were analyzed. A close-typo wavelet network was constructed using the theory of frame wavelet network and the intermediate value theorem of continuous functions. The AE signals of tool conditions were decomposed using a recumive wavelet from which the features were extracted and delivered to the wavelet network for fault recognition, and the recognition rate is up to 100%.
出处
《机床与液压》
北大核心
2007年第2期172-175,共4页
Machine Tool & Hydraulics
关键词
递归小波
小波网络
刀具状态
在线监测
Recursive wavelet
Wavelet network
Tool condition
On-line monitoring
