摘要
基于免疫系统的信息处理特点和传统免疫算法的不足,通过建立准自体抗体集SS和准非自体抗体集SNS来改进传统免疫算法,并结合故障信号的小波包分解特点,提出了一种基于小波包免疫算法的故障检测系统。首先利用小波包将检测信号进行分解,获取检测信号能量的特征向量。然后以信号能量的特征向量作为免疫系统的原始抗原,利用阴性选择算法对原始抗原进行自体-非自体分析。最后,将此故障检测系统成功应用于汽车驱动桥的故障检测。
Based on the characteristics of information process in a immune system and the shortage of traditional immune algorithm,stochastic self(SS) and stochastic non-self(SNS) are developed to improve the traditional immune algorithm.Combined with the feature of wavelet packet analysis of fault signal,a new fault diagnosis system based on wavelet packet and immune algorithm is presented.First,wavelet packet algorithm is used to decompose the non-stable testing signal to obtain the energy eigenvectors of the signal.Then,these energy eigenvectors are treated as the original antigens of immune system,and the negative selection algorithm is used to analyze the original antigens to distinguish self from non-self.Finally,the immune system is successfully applied to the fault testing of an automobile drive-axle.
出处
《机械科学与技术》
CSCD
北大核心
2010年第4期480-483,共4页
Mechanical Science and Technology for Aerospace Engineering
关键词
小波包
特征提取
免疫系统
免疫算法
故障检测
wavelet packet
feature extraction
immune system
immune algorithm
fault testing