基于前趋势相似度的细粒度用户用电负荷预测

FINE GRAINED USER POWER LOAD FORECASTING BASED ON PRIOR TREND SIMILARITY

在智能电网普及的大数据背景下,对电力数据进行准确地分析和预测具有重要意义。提出一种基于前趋势相似度的细粒度居民用电预测模型。根据用户的用电行为特征采用基于DTW距离的K-mediods方法对总体用户进行细粒度划分;在各个子类分别建立用电量预测模型;根据用户的用电行为具有周期性突变这一现象,采用基于前趋势相似度的BP神经网络模型对原BP网络进行改进。基于真实居民用电数据的实验表明,所提出的方法具有较好的预测效果。

In the context of big data popularizing in smart grids,it is of great significance to accurately analyze and predict power data. Aiming at this problem,a fine-grained residential electricity forecasting model based on the similarity of the previous trend was proposed. According to the characteristics of the user ＇s electricity consumption, the D-distance-based K-mediods method was used to fine-grain the overall user. In each subcategory, an electricity consumption prediction model was established. For the phenomenon that the user ＇s electricity consumption had a periodic mutation,the BP neural network model based on the similarity of the previous trend was used to improve the original BP network. Experiments based on real residential electricity data show that the propos method has a good predictive effect.