摘要
用电采集负荷数据反映了用户的用电特性及用电习惯,通过用电负荷数据分析识别用电离群用户在工业生产中具有重要意义。本文根据高维用电负荷数据的特点,提出了一种基于改进One-Class SVM算法的电力离群用户检测方法,同时采用蚁群算法对支持向量机的训练参数进行优化,可以在样本分布不均匀、样本分布未知的环境下有效识别电力离群用户。通过对某市纺织业用户的数据进行实践证明,改进的算法能够有效提高收敛速度,并有效地识别离群的用电用户。
The electricity collection load data reflects the user’s power consumption characteristics and power usage habits.It is of great significance in industrial production to identify outlier users by using electric load data analysis.According to the high-dimensional electric load data,a method of outlier user detection is proposed based on improved One-Class SVM algorithm.At the same time,Ant Colony Algorithm is used to optimize the training parameters of Support Vector Machine,which can detect outlier users of electricity effectively with unevenly or unknow distributed sample.Through the practice of data from users in textile industry of a certain city,the improved algorithm can obviously accelerate the convergence speed and effectively identify outliers.
作者
黄宇腾
裴旭斌
孔历波
李波
殷杰
Huang Yuteng;Pei Xubin;KONG Libo;LI Bo;YIN Jie(State Grid Zhejiang Electric Power Co.,Ltd.Information and Communication Branch,Hangzhou 310000,China;State Grid Hangzhou Electric Power Supply Co.,Ltd.,Hangzhou 310000,China;Zhejiang Huayun Information Technology Co.,Ltd.,Hangzhou 310000,China)
出处
《自动化与仪器仪表》
2019年第5期111-114,共4页
Automation & Instrumentation
