摘要
针对水电机组现有在线监测方式的缺陷,设计了一种基于多智能体的故障诊断系统模型。与经典MAS故障诊断系统相比,该模型将任务控制与任务分解融合于一个Agent子系统中,提高了控制信号的传输率并简化系统;状态监测子系统直接从监控现场提取异常特征信号以提高故障监测的灵敏性;在故障诊断子系统中设计了多Agent并行工作结构以满足系统诊断实时性与高可靠性的要求;采用基于java语言的JAFMAS软件构建开发平台建立多A-gent协调机制。实验结果表明了该方法的有效性。
A MAS-based fault diagnosis system model was designed for improving existing defects of hydropower unit on-line monitor.Comparing to classical MAS-based fault diagnosis system,it integrates the task control and decomposition in an agent subsystem.This system can increase the transmission rate of control signal and simplify the system,increase sensitivity of fault monitoring as getting abnormal signals directly from site,and meet the requirement on real-time and sensitivity of fault monitoring after using a multi-Agent parallel structure.The JAFMAS software based on java was used to build a platform for multi-agent coordination.
出处
《水力发电》
北大核心
2010年第3期72-74,共3页
Water Power
关键词
故障诊断
智能体
模糊矩阵
水电机组
fault diagnosis
intelligent agent
fuzzy matrix
hydropower generating unit
