基于改进BP神经网络法电力负荷短期预测  

Short-term Power Load Forecasting Based on Improved BP Neural Network Method

作　　者：张晋轩 柳影 黄冬梅 ZHANG Jin-xuan;LIU Ying;HUANG Dong-mei(Faculty of Science and Engineering,Guangxi University Xingjian College of Science and Liberal Arts,Nanning 530000,China)

机构地区：[1]广西大学行健文理学院理工学部,南宁530000

出　　处：《武汉理工大学学报》2021年第1期103-106,共4页Journal of Wuhan University of Technology

基　　金：广西高校中青年教师科研基础能力提升项目(2020KY54016)。

摘　　要：电力网络结构日益庞大,电力安全可靠运行难度不断增大,须提高电力负荷预测精度,以保证电力设备正常稳定运行。电力负荷变化具有非线性的特点,通过传统建立的模型难以精确预测电力负荷的变化,针对此难点,文中采用具有非线性特征的改进BP神经网络法进行短期负荷预测,在进行负荷预测算法时,为消除训练样本顺序的影响,将整个样本集替代单系列样本进行学习。选取某县城电力负荷历史数据作为样本,Matlab编程仿真,得出预测与期望结果比较接近,部分数据较精确,说明该算法具有一定的参考意义。The net of electricity become gigantic,which is difficult to run safety and reliable for the electric energy,so,it has to predict power of electric A more accurate.In view of the difficulty in establishing a suitable mathematical model to accurately express the relationship to calculate power load,and power load changes is nonlinearity.the paper explain the topology structural and mechanism process about BP neural netork,To remove the affect of the training samples’ s order.It is proposed to apply the learning of the entire sample set instead of the series of sample learning to the load forecasting algorithm.,it use software of matlab to simulate,the predict almost equally expect,part digital is accurate,which has reference value.

关 键 词：负荷 神经网络 电力 

分 类 号：TM933[电气工程—电力电子与电力传动]