基于时间序列的电力业务承载能力自动感知方法研究  

Research on automatic perception method of power service carrying capacity based on time series

作　　者：许放 周林林 周宇洋 李伯宇 段金磊 XU Fang;ZHOU Linlin;ZHOU Yuyang;LI Boyu;DUAN Jinlei(Beijing Fibrlink Communications Co.,Ltd.,Beijing 100071,China)

机构地区：[1]北京中电飞华通信有限公司,北京100071

出　　处：《粘接》2024年第10期181-184,共4页Adhesion

摘　　要：随着电力业务数据逐步增加,电力业务承载能力分析的重要性日益凸显。为解决电力业务承载能力感知准确性差的问题,提出了一种基于时间序列的电力业务承载能力感知方法。采用最小二乘法对历史电力业务数据进行预处理,消除错误数据对模型带来的影响;通过近邻传播聚类,获取电力业务承载能力的典型特征,并以此建立评估指标。在此基础上,基于时间序列检测机制,实现电力业务承载能力感知和风险识别。在某计量中心进行了验证,其电力业务承载能力感知准确率为99.56%。所提方法能有效提高电力业务承载能力感知能力。With the gradual increase of power business data,the importance of analyzing the carrying capacity of power business is becoming increasingly prominent.To address the issue of poor accuracy in perception of power business carrying capacity,a time series based perception of power business carrying capacity was proposed.The least squares method was used to preprocess historical power business data and eliminate the impact of erroneous data on the model.Through nearest neighbor propagation clustering,typical characteristics of power business carrying capacity were obtained,and evaluation indicators were established based on this.On this basis,based on the time series detection mechanism,the perception of power business carrying capacity and risk identification were achieved.It was verified in a certain metering center,and the perceived accuracy of its power business carrying capacity was 99.56%.The proposed method can effectively improve the perception ability of power business carrying capacity.

关 键 词：时间序列 电力业务 承载能力 感知 多线位置距离 

分 类 号：TM76[电气工程—电力系统及自动化] TP39[自动化与计算机技术—计算机应用技术]