摘要
提出了一种基于人工免疫的故障诊断进化学习模型及其相应的算法.通过对检测对象正常工作状态下获得的自己模式串的阴性选择,随机产生初始检测器;用基于人工免疫的进化学习机制实现对检测对象异常工作状态下获得的非己模式串的学习和记忆;利用进化学习结果和系统故障信息库知识区分和标记不同故障在状态空间上对应的区域.应用于机床齿轮箱故障检测和诊断问题,实验结果表明了所提出方法的有效性.
A evolutional learning model for detection and diagnosis of fault based on artificial immune theory is proposed. Initial detectors is produced at random combining reversed selection of self patterns which responses normal working situation of detecting object. Learning and memory of non-self patterns is realized with using mechanism of evolution leaning based on artificial immune theory. The Corresponding Zones of different faults on states space are distinguished and marked with the results of evolution learning and information warehouse of faults . Appling the methods to detection and diagnosis for faults of gear case of machine tools, the experiment results indicate that the method is effective.
出处
《南方冶金学院学报》
2005年第3期31-36,50,共7页
Journal of Southern Institute of Metallurgy
关键词
人工免疫
进化学习
异常检测
故障诊断
artificial immune
evolution and learning
anomaly detection
fault diagnosis
