基于QPSO-KELM的变压器绕组热点温度预测  

Prediction of transformer winding hot spot temperature based on QPSO-KELM

作　　者：胡丹 HU Dan(Shiyan Power Supply Company of State Grid Hubei Electric Power Co.,Ltd.,Shiyan 442000,China)

机构地区：[1]国网湖北省电力有限公司十堰供电公司,湖北十堰442000

出　　处：《黑龙江电力》2023年第2期172-177,共6页Heilongjiang Electric Power

摘　　要：为提高变压器绕组热点温度预测精度,采用灰色关联分析确定绕组热点温度的关键影响因素为负载电流、环境温度、顶层油温和油箱上死角温度。采用QPSO算法对KELM的受惩罚系数和核参数进行优化,以负载电流、环境温度、顶层油温和油箱上死角温度为输入量,绕组热点温度为输出量,建立基于QPSO-KELM的变压器绕组热点温度预测模型。以变压器温升试验进行仿真分析,结果表明,QPSO-KELM模型的平均相对百分误差、均方根误差和决定系数分别为3.12%、2.25和0.995,均优于其他模型,验证了该模型的正确性和优越性。In order to improve the prediction accuracy of transformer winding hot spot temperature,grey correlation analysis is used to determine that the key influencing factors of winding hot spot temperature are load current,ambient temperature,top oil temperature and top dead corner temperature of oil tank.QPSO algorithm is used to optimize the penalty coefficient and nuclear parameters of KELM.Taking load current,ambient temperature,top oil temperature and top dead corner temperature of oil tank as the input and winding hot spot temperature as the output,a transformer winding hot spot temperature prediction model based on QPSO-KELM is established.The simulation analysis is carried out by using the transformer temperature rise test.The results show that the average relative percentage error,root mean square error and determination coefficient of QPSO-KELM model are 3.12%,2.25 and 0.995 respectively,which are better than other models,which verifies the correctness and superiority of the model.

关 键 词：变压器 绕组热点温度 量子粒子群算法 核极限学习机 灰色关联分析 

分 类 号：TM411[电气工程—电器]