摘要
针对传统增量配电网线损率预测模型预测效果波动大、精准度差的问题,本文提出了一种以极限学习机为切入点的预测模型优化方案。基于ELM神经网络结构,结合配电网的环境条件分析、选择适应度函数,对传统的电量参数进行优化处理。以改进极限学习算法为依据,参考增量配电网的结构构建相应的配电样本,在明确线损分布特征以及流程细则的基础上,落实增量配电网线损率预测改进模型的设计,并开展相应的仿真应用实验。结果表明,以极限学习机为切入点的增量配电网线损率预测改进模型能够全面适应增量配电网环境中的各种复杂情景,精准收集电量数据序列并分析其特征,高效预测增量配电网线损率。
In view of the problems of the line loss rate prediction model of the traditional incremental distribution network and poor accuracy,this paper proposed an optimization scheme of the prediction model with the extreme learning machine as the breakthrough point.Based on the ELM neural network structure,combined with the environmental condition analysis of the distribution network and the fitness function,the traditional power parameters are improved and optimized.Based on the improvement of the limit learning algorithm,the corresponding distribution samples are constructed according to the structure of incremental distribution network.On the basis of special diagnosis of line loss distribution and process details,the design of the improvement model of line loss rate prediction of incremental distribution network is implemented,and the corresponding simulation application experiments are carried out.Results show that the limit learning machine as the breakthrough point of incremental distribution network line loss rate prediction improvement model can fully adapt to the incremental distribution network environment,accurately collect power data sequence and analyze its characteristics,and efficiently predict of incremental distribution network line loss rate.
作者
王肖亚
孙奉杰
谢志平
WANG Xiaoya;SUN Fengjie;XIE Zhiping(Zibo Power Supply Company,State Grid Shandong Electric Power Company,Zibo,Shandong 255000,China)
出处
《自动化应用》
2023年第22期44-46,49,共4页
Automation Application
关键词
增量配电网
线损率
预测模型
ELM
incremental distribution network
line loss rate
prediction model
ELM