基于用电大数据的低压电力客户电费异常识别模型

Identification Model of Abnormal Electricity Charges for Low Voltage Power Customers Based on Big Data of Electricity Consumption

针对传统电费识别模型中电费异常识别耗时较长、误差较大等问题,提出基于用电大数据的低压电力客户电费异常识别模型。将低压电力客户用电数据转换为数据集形式,清洗该数据集中脏数据,通过多阶拉格朗日内插法修补日负荷曲线的缺失值;利用流聚类技术设置阈值,获取用电数据集中初始簇的中心,采用平方和误差最小函数确定具有最大关联特征的用电数据;通过对n维空间中所有电力特征参数的处理,获取时间序列矩阵,构建低压电力客户电费异常识别模型。实验结果表明:采用所提模型识别低压电力客户电费异常的耗时最短约为0.1 s,且误差最高约为2.1%。

Aiming at the problems of long time consuming and large error in the traditional electricity charge recognition model, a low-voltage electricity customer electricity charge anomaly recognition model based on electricity big data is proposed. The power consumption data of low-voltage power customers are transformed into data set form, the dirty data in the data set are cleaned, and the missing values of daily load curve are repaired by multi-stage Lagrange interpolation method. The flow clustering technology is used to set the threshold to obtain the center of the initial cluster in the power consumption data set, and the least square sum error function is used to determine the power consumption data with the largest correlation feature. Through the processing of all the power characteristic parameters in the n-dimensional space, the time series matrix is obtained, and the low-voltage power customer electricity charge anomaly recognition model is constructed. The experimental results show that the shortest time for the proposed model to identify the abnormal tariff of low-voltage power customers is about 0.1 s, and the highest error is about 2.1%.