基于密度估计的异常电力负荷数据辨识与修正

Identification and Correction of Abnormal Electric Load Data Based on Density Evaluation

SCADA系统数据库中一般会有一些异常的电力负荷数据,直接用其来进行短期负荷预测将影响预测结果的准确性,因此有必要对这些异常数据进行辨识和修正.文中同时考虑负荷的横向连续性和纵向连续性,先把负荷数据按照日期排列成二维数据集,然后采用基于密度的方法,在两个维度中对异常数据进行辨识与修正,最后通过实例分析验证了该方法的有效性.

In general, there exist some abnormal data in the electric load database derived from the SCADA sys- tem. As these abnormal data may reduce the accuracy of the short-term load forecasting, they should be identified and corrected before their employment. In this paper, both the horizontal and the vertical continuities of electric loads are taken into consideration, and a two-dimension load data set is arranged according to the date. Then, a density evaluation-based method is presented to identify and correct the abnormal data in two dimensions, and a case study is finally performed to verify the effectiveness of the proposed method.