基于模糊神经网络的凝汽器故障诊断及其性能监测  

Fault diagnosis and performance supervision of condenser based on fuzzy neural network

作　　者：滕丕忠[1] 张聘[2] 陈荣生 陈群[2] 石永恒[1] 禹宝宁[2] 杨亚平[1] 

机构地区：[1]东南大学动力工程系,江苏南京210096 [2]福建永安火电厂,福建永安364013 [3]泉州电力技术学校动力系,福建泉州364000

出　　处：《电力自动化设备》2007年第1期90-92,共3页Electric Power Automation Equipment

摘　　要：凝汽器是凝汽式汽轮机的主要辅助设备,其凝汽器系统运行中出现故障的原因与故障征兆之间是非线性关系,具有复杂性、模糊性和随机性,难以用数学公式表示。针对此情况,结合模糊理论与神经网络2种故障诊断方法的优势,提出采用串联型模糊神经网络为凝汽器故障诊断模型,用Matlab 6.5矩阵式运算语言开发故障诊断及其性能监测软件。故障诊断软件包括征兆参数的获得、故障诊断及结果柱状图显示;性能分析软件主要是相关参数的计算及正常与变化工况曲线的比较。给出了某电厂100 MW汽轮发电机组诊断实例。Condenser is the primarily auxiliary equipment of condensing turbine,and relations between condenser system operation faults and their symptoms are nonlinear,complicated,fuzzy and random,which are difficult to be represented by math equations. Combining advantages of fuzzy theory and neural network ,a serial fuzzy neural network is presented as condenser fault diagnosis model and the relevant software of fauh diagnosis and performance supervision is developed with Matlab 6.5. The fault diagnosis software includes symptom parameter derivation,fauh diagnosis and result display. The performance supervision software is used to calculate relevant parameters and compare changed operation state with normal one. An actual example of a 100 MW steam generation unit is given.

关 键 词：凝汽器 模糊神经网络 故障诊断 动态数据交换 

分 类 号：TM311[电气工程—电机]