基于先进诊断策略的网络化机组轴系监测系统研究

A STUDY ON A NETWORK TURBO-UNIT SHAFT MONITOR BASED ON THE ADAVANCED DIAGNOSTIC STRATAGEM

先进的系统智能诊断策略与兼具安全性与开放性的系统网络体系是当前电厂机组轴系安全、经济运行监测系统发展的趋势。该文介绍的汽轮发电机组轴系监测系统是基于改进的RBFNN神经网络故障诊断模型,该模型采用多输入、单输出改进组合子网络结构改进了原有RBFNN神经网络结构输出节点数不易确定、脱机时间较长的缺陷。系统经过验证可以准确判别汽轮发电机组多种轴系典型故障,并具有判断新故障及不断修正样本的功能 。监测系统整体框架采用 Browser/Server网络体系,在保证信息安全的前提下实现设备运行参数在设备管理层(电厂主控室)、生产管理层及远端专家的信息共享。

It is the inexomble trend for the current power station to adopt the advanced intelligent systematic theory as well as the safety and opening network system to realize the safety and economic operation of the turbo-generator shaft system. This paper introduces a set of shaft-monitoring system based on the improved RBFNN neural network fault diagnosis model, which adopt advanced mixed sub-network structure with multi-input and single-output. This kind of model can easily solve some problems such as the difficulty in determining the number of node and the over-long time of off-line operation. It has been certified that the model has the ability to detect exactly all kinds of typical fault of the turbo-shafting system, and it is a function for detecting the new fault so as to modify the sampling value on time .The framework of this monitor system is suppied with Brower/Server which can on one hand guarantee a safety operation of the equipment, on the other hand, realize the information sharing of the operation parameters among the following levels: equipment management station (MCR of the power station),operation management station and the remote expert station, and so on..