基于神经网络的通风机性能监测组合建模  

Combination Model of Performance Monitoring for Fan Based on Neural Network

作　　者：侯军虎[1] 王松岭[1] 安连锁[1] 王大勇[1] 丁千玲[1] 

机构地区：[1]华北电力大学

出　　处：《风机技术》2002年第3期29-32,共4页Chinese Journal of Turbomachinery

摘　　要：通过对电力工业中广泛应用的通风机性能曲线的曲面、超曲面特性的分析 ,提出应用神经网络进行通风机性能曲线的曲面拟合。并以实验室4- 73№ 8D通风机为研究对象 ,建立了神经网络与机理的组合模型 ,并提出与DCS结合的通风机性能监测及通风系统管网故障诊断模型。The curved surface and hypersurface characteristic of fans used widely in the power industry is analyzed.On the basis of it,a method using neural network to accomplish the curved surface fit is put forward.After that,the integrated model of neural net and mechanism has been constructed.In the end combining with DCS,the monitoring model of fan performance and the diagnosis model of fan system is put forward. Key word:Centrifugal fan Neural network simulation Diagnosis

关 键 词：神经网络 性能监测 组合建模 离心式通风机 火电厂 

分 类 号：TH432[机械工程—机械制造及自动化] TM621.73[电气工程—电力系统及自动化]