基于模糊神经网络的火电厂热控系统故障诊断方法  

Fuzzy Neural Network-based Fault Diagnosis for Thermal Control System of Thermal Power Plant Based on Fuzzy Neural Network

作　　者：杨雄毅 张小亮 高广洲 于昌明 YANG Xiongyi;ZHANG Xiaoliang;GAO Guangzhou;YU Changming(Guoneng Shenhua Jiujiang Power Generation Co.,Ltd.,Jiujiang 332504,China;Guoneng Fengcheng Power Generation Co.,Ltd.,Fengcheng 331100,China;Qinhuangdao Huadian Measurement and Control Equipment Co.,Ltd.,Qinhuangdao 066000,China)

机构地区：[1]国能神华九江发电有限责任公司,江西九江332504 [2]国能丰城发电有限公司,江西丰城331100 [3]秦皇岛华电测控设备有限公司,河北秦皇岛066000

出　　处：《电工技术》2024年第6期82-84,共3页Electric Engineering

摘　　要：火电厂热控系统运行是由众多组件配合实现的,故障诊断难度较大,为此提出基于模糊神经网络的火电厂热控系统故障诊断方法。从火电厂热控系统自身构成和运行属性的角度出发,结合模糊控制与5层等价型BP神经网,构建了火电厂热控系统模糊神经网络模型。在模型内进行火电厂热控系统运行状态变量的前向传播和误差的反向传播,实现对火电厂热控系统的故障诊断。测试结果表明,该设计方法对设置的7种不同类型火电厂热控系统故障状态均实现了准确诊断。The operation of thermal control systems in thermal power plants is achieved through the cooperation of numerous components,making fault diagnosis difficult.Therefore,a fuzzy neural network-based fault diagnosis method was proposed.From the perspective of composition and operational attributes of the thermal control system in thermal power plants,a fuzzy neural network model for thermal control system in thermal power plants was constructed by combining fuzzy control with a 5-layer equivalent BP neural network.Forward propagation of operating state variables and back propagation of errors were carried out within the model to achieve fault diagnosis.The test results indicate that the designed method has achieved accurate diagnosis of the fault states of seven different types of thermal control systems in thermal power plants.

关 键 词：模糊神经网络 火电厂热控系统 故障诊断 模糊控制 5层等价型BP神经网 运行状态变量 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]