基于奇异谱分析的用户用电行为变化检测方法

A Change Detection Method of Customer’s Electricity Consumption Behavior Based on Singular Spectrum Analysis

用户的用电行为代表着用户的用电特征,是进行用电预测分析的基础条件。现实生活中由于外部因素导致的用电行为变化,会使得用电特征随之发生变化,在进行用电预测分析时就需要区分历史用电特征和当前用电特征。为此本文提出了基于奇异谱分析的用户用电行为检测方法,对检测点前后的用户用电数据进行奇异谱分析,通过计算奇异值向量与历史特征超平面夹角余弦值,得到检测点的变化系数并作为用户用电行为变化的判断依据。针对本年度发生增减容业务变更的用户算例进行用电行为变化的判断,验证了判断方法的灵敏性,结果表明所述方法可实现用电行为变化和变化时间的准确判断。

The electricity consumption behavior of customers represents their electricity consumption characteristics,which are the basic condition for the electricity consumption forecast analysis.In real life,the changes of electricity consumption behaviors caused by external factors will cause the changes of electricity consumption characteristics.It is necessary to distinguish the historical characteristics from current characteristics of electricity consumption during the electricity consumption forecast analysis.Therefore,a detection method for consumer’ s electricity consumption behavior based on singular spectrum analysis is proposed in this paper to perform analysis on the electricity data of customer before and after the detection point.By calculation of the cosine value of the angle between the singular value vector and the historical feature hyperplane,the change coefficient of the detection point is obtained and used as the basis for judging the change of customer’ s electricity consumption behavior.The sensitivity of the method is verified by judgment of the electricity consumption behavior changes based on the cases of the users whose capacity change business occurred this year.The results show that the proposed method can make accurate judgment of the change of electricity consumption behavior and the change time.