随着源网荷储四侧资源的建设,电力部门积累了海量用户的用电数据,如何有效挖掘这些数据的潜在信息,促进电力用户的精细化管理,是当前电力系统分析的一个重要问题.基于此,提出了基于对称KL散度(Kullback-Leibler Divergence)的电力用户负荷聚类方法.首先,利用高斯混合模型(Gaussian Mixture Model,GMM)表示用户的日常用电规律,并通过最大均值差异(Maximum Mean Discrepancy,MMD)检验模型的合理性;然后,将对称KL散度作为相似性判据,从划分聚类的角度出发,对GMM表示的电力负荷用户进行聚类;最后,以所提算法对MNIST以及某小区用户数据进行分析,实验结果表明所提方法具有可行性和有效性.
Electric Customer Load Clustering Research Based on Symmetric KL Divergence
With the construction of resources on the four sides of the source,grid,load and storage,the power depart-ment has accumulated a large number of users'electricity consumption data;how to effectively mine the potential infor-mation of these data and promote the fine management of customer is an important problem in current power system anal-ysis.Based on this,an electric customer load clustering method based on kullback-Leibler (KL)divergence was put for-ward:First,the gaussian mixture model was used to represent the customers'daily electricity usage and the rationality of the model was tested by maximum mean difference.Then,symmetrical KL divergence is used as the similarity criterion to cluster the electric customer represented by gaussian mixture model from the perspective of clustering.Finally,the pro-posed algorithm is used to analyze MNIST and customer data of a cell,and the experimental results show that the pro-posed method is feasible and effective.
electric loadGaussian mixture modelmaximum mean differenceKL divergenceclustering
万方数据
维普
