变压器顶层油温预测及其数学模型的参数辨识.pdf

　2004 年 6 月 　　　　　　　　　　　　高 　电　压 　技 　术 　　　　　　第 30 卷第 6 期 ·9 · 变压器顶层油温预测及其数学模型的参数辨识 1 2 赵 　斌 , 张霄元 ( 1 许继电气主设备状态监测部 ,许昌 46 1000 ;2 北京许继电气有限公司 ,北京 100085) Para meter Identif ication of Transf ormer Top Oil Temperature Model and Prediction of Top Oil Tempeature 1 2 Zhao Bin , Zhang Xiaoyan ( 1. Dep art ment of Monitoring of Elect rical Equip ment of XJ Elect ric CO L t d , Beij ing China ; 2 . XJ Elect ric Co L t d , Beij ing , China) Abstract 　The paper introduces the purpose and development of 因此模型参数的辨识方法是现场应用它监测和诊断 measuring and predicting transformer top oil temperature. From 变压器的状态的关键 。本文从变压器顶层油温模型 MIT transformer top oil temperature model , using the least square method , the paper gives a new algorithm for the parameters identi 参数的辨识入手 ,推导出一种参数辨识的算法 ,并通 fication of the model . The algorithm and the model are applied on 过变电站现场采集的数据对它进行分析研究 ,得出了 the data on site in Shuanghu substation , Xinzheng city , Henan 有指导性的结论 。 province. The good result for predicting the transformer top oil temperature proves the validity of the algorithm. Furthermore , the 1 　顶层油温模型参数辨识与温度预测 paper analyzes some factors that may produce errors , such as the quantity of the data , the noise of the data. Actual measured data 11 　M IT 改进的顶层油温模型与参数辨识 accompanied by white noise may inspire fully the mode , and is good M IT 改进的顶层油温模型考虑了环境温度变 for parameter identification . Key words 　transformer 　top oil temperature model 　parameters 量 ,实际应用时需采用离散模型 ,对于强迫冷却 ,取 n identification 　predicting