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
杨雄毅 1张小亮 2高广洲 3于昌明3
作者信息
- 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
- 折叠
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.
Abstract
The operation of thermal control systems in thermal power plants is achieved through the cooperation of numer-ous 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 prop-agation 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Key words
fuzzy neural network/thermal control system of thermal power plant/fault diagnosis/fuzzy control/5-layer equivalent BP neural network/operating state variable引用本文复制引用
出版年
2024
NETL
NSTL
万方数据