摘要
根据免疫学原理提出了一种适合于故障诊断的算法———疫苗算法.当出现未知抗原时提取系统内记忆的此类抗原特征并对其进行分析,再将此人工抗原输入到系统中刺激系统并产生相应的抗体,当此抗原再次出现时作出免疫应答.论述了抗体和抗原结构的数学模型以及抗体数据的产生机理.运用此算法开发了具有边检测边学习功能的检测系统,能够充分捕捉反映被测车辆状态的信号特征.将此算法应用于汽车驱动桥的故障检测中,准确率达到95%.工程应用表明,此算法在小样本获取、连续学习、数据压缩等方面具有明显的优越性和有效性.
A new kind of immune algorithm named bacterin algorithm based on immune mechanism was presented. When an unknown antigen invades the system, the memorized features of the antigen is studied, then the artificial antigen is input to the system and activates the system to create the antibodies. When the antigen invades the body again, immune response will appear. The mathematical models of antibody and antigen were given, and the antibody generating mechanism was presented also. A detection system using the bacterin algorithm is a dynamic system which can detects and study at the same time, and can get the signal features of a detected vehicle＇s condition efficiently. Application of the algorithm in vehicle drive axle fault detection indicated that the exactness amounted to 95%. Engineering application proves that the bacterin algorithm is better a.nd efficient in small sample acquirement, continued learning and data compression.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2006年第2期298-303,共6页
Journal of Zhejiang University：Engineering Science
关键词
免疫学
免疫算法
疫苗算法
驱动桥
车辆故障检测
immunology
immune algorithm
bacterin algorithm
drive axle
vehicle fault detection
