摘要
针对目前水电机组故障诊断中存在的建模复杂、样本需求量大及诊断学习缺乏自主连续性等问题,提出了基于免疫原理的故障诊断方法。以状态征兆为抗原,各种故障模式下的故障检测器作为抗体,通过反向选择机制判别正常/异常状态,利用克隆选择原理进化学习获得能识别抗原结构的记忆抗体,根据最大故障隶属度诊断故障类型。以机组振动为诊断对象的仿真结果表明,该方法识别故障的准确率高,非常适合故障样本难以获得的小样本故障诊断。
In order to solve the problems such as complex modeling, large sample demands and discontinuous learning in the fault diagnosis techniques, a new method based on immunology principles is proposed. Taking state symptoms as antigens, and detectors under different fault modes as antibodies, the normal conditions and anomaly are distinguished via negative selections. Memory antibodies that can identify the structures of antigens are generated via the mechanism of clonal selection. Fault types are determined by the maximum fault membership. The proposed method was applied to the vibration diagnosis of hydroelectric generating unit. Results indicated that the high diagnosis accuracy with low demands on fault samples, and the merits of this method in application as well.
出处
《水力发电》
北大核心
2007年第3期54-56,共3页
Water Power
基金
国家自然科学基金重点项目(50539140)
国家自然科学基金项目(50579022)
高等学校博士学科点专项科研基金项目(20050487062)
关键词
免疫原理
水电机组
故障诊断
反向选择
克隆选择
immunology principles
hydro-generating unit
fault diagnosis
negative selection
clonal selection
