摘要
针对复杂设备系统中故障诊断知识获取困难的问题,借鉴生物体液免疫机理,提出了用于故障诊断的免疫学习模型。将检测器定义为B细胞及其所包含的若干抗体结构,采用B细胞和抗体双重学习机制概括在抗原数据中发现的模式,不但解决了因故障征兆的混叠导致故障难以辨别的问题,而且能够不断补充和完善诊断知识。实现已知故障和未知故障类型的检测与学习,使系统的诊断能力达到最优。通过异步电动机故障实验证明了该算法可以提高故障检测的效率与准确率。
In complex equipment system, the knowledge of fault diagnosis could hardly be precise and complete. Inspired by the biological humoral immunity, an immune learning model for fault diagnosis is proposed. The detectors are defined as Blymphocyte and antibody structures in the Blymphocyte. The patterns of antigen are generalized by using double learning mechanisms of Blymphocyte and antibody. The mechanism not only solves the problems that how to recognize the faults caused by the overlap of the omens, but also continuously supplements and improves the diagnostic knowledge. The system can detect and learn known and unknown fault types, and achieve optimal diagnostic results. Experiments were undertaken with induce motor to demonstrate the efficiency and accuracy of the fault detection.
出处
《南京航空航天大学学报》
EI
CAS
CSCD
北大核心
2013年第4期544-549,共6页
Journal of Nanjing University of Aeronautics & Astronautics
基金
国家自然科学基金重点(50335030)资助项目
山西省自然科学基金(2013011018-1)资助项目
关键词
人工免疫系统
故障诊断
学习机制
体液免疫
artificial immune system
fault diagnosis
learning mechanism
humoral immune