基于数据挖掘技术的台区合理线损预测模型研究

摘要

The forecast of reasonable line loss is the prerequi⁃site for lean management of transformer district. Therefore, the re⁃search status of transformer district line loss and correlative data mining technology is studied in this paper. On the basis of that, a prediction method and algorithm flow for reasonable line loss is proposed. Firstly, transformer district data is classified according to district characteristic by K⁃means. Secondly, linear regression models are established separately using the characteristic data in each class. Lastly, forecasting of line loss rate and error analysis are executed by these regression models. The applicability, rapidi⁃ty and convenience of this method have been verified by data calcu⁃lation using actual power consumption data.%台区合理线损预测是实现台区线损精益化管理的前提。通过对国内外台区线损管理的研究现状以及相关数据挖掘技术的分析研究，提出基于聚类分析和线性回归的合理线损预测方法与算法流程。利用K均值聚类（K⁃means）将台区线损数据按照台区特征进行分类，对每一个数据类分别进行线性回归，利用线性回归模型进行线损率预测与误差分析。通过实际的用电数据，具体分析了模型输出结果，论证了所提方法的适用性、快速性、便捷性。