摘要
由于红外温度传感器不能直接测量带屏蔽罩结构的GIS触头温度,因此提出了通过直接测量屏蔽罩温度来间接获得触头部位温度的方法。建立了GIS开关传热模型,获得了触头处温度场分布。研究了GIS开关触头温度与屏蔽罩温度、环境温度关系,并且运用支持向量机算法和神经网络算法预测了触头温度和屏蔽罩温度之间的映射关系,结果表明神经网络算法的拟合预测准确度更高。
As the infrared temperature sensor can not directly measure the temperature of shielded GIS (Gas Insulated Switchgear) contacts,it is proposed to measure the shield temperature for indirect getting the contacts temperature. The heat transfer model of GIS switch is established and the temperature-field distribution around the contacts is obtained. The relationship among contacts,shield and environmental temperatures is studied and the mapping relationship between contacts and shield temperatures is fitted respectively by SVM(Support Vector Machine) method and neural network algorithm for forecasting the contacts temperature. Results show the neural network algorithm has higher accuracy.
出处
《电力自动化设备》
EI
CSCD
北大核心
2014年第9期148-153,共6页
Electric Power Automation Equipment
基金
国家高技术研究发展计划(863计划)资助项目(2011AA05A121)~~
关键词
GIS
红外传感
红外辐射
触头
温度测量
监测
神经网络
屏蔽罩
模型
GIS
infrared sensing
infrared radiation
contacts
temperature measurement
monitoring
neural networks
shielding
models