粒子群优化极限学习机的短路电流预测技术  

Short-circuit current prediction technology based on particle swarm optimization extreme learning machine

作　　者：王梦娇[1] 魏新劳[1] WANG Meng-jiao;WEI Xin-lao(Key Laboratory of Engineering Dielectrics and Its Application,Ministry of Education,Harbin University of Science and Technology,Harbin 150080,China)

机构地区：[1]哈尔滨理工大学工程电介质及其应用教育部重点实验室,黑龙江哈尔滨150080

出　　处：《电机与控制学报》2022年第1期68-76,共9页Electric Machines and Control

基　　金：国家重点基础研究发展“973”计划项目(2012CB723308);高等学校博士学科点专项科研基金(20122303110007)。

摘　　要：针对传统极限学习机预测模型精度低和稳定性能差的问题,提出了一种利用粒子群算法优化极限学习机的短路电流峰值预测模型。建立超高压输电线路仿真模型,分析短路故障波形特点,获取全相角短路故障电流历史数据,利用平均相对误差、均方根误差、灰色绝对关联度三种精度检验法作为粒子群算法的适应度函数,构建粒子群算法与极限学习机算法相结合的电流预测模型。实验结果表明,以灰色绝对关联度作为适应度函数的粒子群优化极限学习机算法,对短路电流峰值预测精度较高、速度较快,当故障点位置未知时,采用粒子群优化极限学习机算法依然可以准确预测短路电流峰值,提升了算法应用的实际意义,为超、特高压线路快速限制、消除短路故障奠定理论基础。Aiming at the problems of low precision and poor stability of traditional extreme learning machine,a new short-circuit current peak prediction model based on particle swarm optimization was proposed.The simulation model of EHV transmission line was established,the characteristics of short-circuit fault waveform were analyzed,and the historical data of full phase angle short-circuit fault current were obtained.Three precision testing methods,namely average relative error,root mean square error and grey absolute correlation degree,were used as fitness functions of particle swarm optimization algorithm,and a current prediction model combining particle swarm optimization algorithm with extreme learning machine algorithm was constructed.The experimental results show that the PSO extreme learning machine algorithm with grey absolute correlation degree as fitness function can predict the peak value of short-circuit current with high accuracy and fast speed.When the location of the fault point is unknown,the PSO extreme learning machine algorithm can still accurately predict the peak value of short-circuit current,which improves the practical significance of application of the algorithm and lays a theoretical foundation for quickly limiting and eliminating short-circuit fault for EHV and UHV transmission lines.

关 键 词：粒子群算法 极限学习机 短路电流 峰值预测 灰色绝对关联度 适应度函数 

分 类 号：TM411[电气工程—电器]