摘要
针对复杂系统的符号有向图(signed directed graph,SDG)模型结构复杂,搜索路径"组合爆炸"等缺点,将SDG符号识别能力和知识库易于维护的特点相结合,提出了模糊概率SDG单元模型,建立了模糊概率SDG单元库,并描述了基于SDG单元库的故障推理方法。复杂系统的测点不完备性和参数耦合性是故障诊断中亟需解决的问题,在建立SDG单元库时,充分考虑了这两个问题。以某600MW机组高加系统为例,采取仿真与现场经验相结合建立SDG单元库,并通过现场故障实例验证了该故障推理方法的有效性。SDG单元模型结构简单,推理快速,有效避免了"组合爆炸"的问题,故障诊断准确率高,能有效克服现场测点缺失和系统参数耦合问题。
The structure of signed directed graph(SDG) of complex system is complexity,and fault diagnosis reasoning of SDG will lead to combination explosion problem of search path.Fuzzy probabilistic SDG unit model was proposed and fuzzy probabilistic SDG unit library was built with the signed recognition capability of SDG combined with knowledge base.And the fault diagnosis reasoning method was described.The incomplete information of measuring points and the coupling of system were two problems of fault diagnosis and these two problems were considered fully on the process of SDG modeling.A comprehensive and reliable SDG unit library was built for fault diagnosis reasoning of high pressure heater system of a 600MW unit,which was obtained by combination of simulation and field experience.And this method was valid by field fault of high pressure heater system.The structure of SDG unit model is simple and the speed of diagnosis reasoning is fast.The result shows that this method can avoid the combination explosion problem and solve the lack of measuring points and the coupling of system.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2011年第2期104-110,共7页
Proceedings of the CSEE
关键词
符号有向图
模糊概率
电站
高加系统
故障推理
signed directed graph
fuzzy probability
power plant
high pressure heater system
fault detection