基于惩罚变权的RDA同期线损数据质量评估模型

同期线损数据质量评估对线损数据治理、线损分析至关重要。针对目前线损数据质量评估中评估准确率和评估效率不可兼得的矛盾，提出基于惩罚变权的“秩和”差异评估（rank difference assessment，RDA）同期线损数据质量评估模型。首先，基于传统“秩和”检验不依赖样本容量大小和指标选取的特点，建立“秩和”差异评估（RDA），以正“秩和”统计量直接对同期线损与理论线损在相同时间尺度下的电量差值数据序列进行整体数据质量评估；其次，以“秩和”差异数据序列的差异概率密度函数为指数，对“秩和”差异评估模型的权值进行修正，建立基于惩罚变权的RDA同期线损数据质量评估模型，综合考虑异常数据数量及异常幅度对数据质量评估的影响程度，对RDA评估值进行修正。该方法既不依赖指标选取，可以准确评估，又不依赖样本容量，评估效率较高。最后，以甘肃实际运行数据进行的仿真结果表明，所提惩罚变权评估模型对同期线损数据质量评估的有效性和优越性。

Evaluation Model of Synchronous Line Loss Data Quality Based on Penalty Variable Weight RDA

The evaluation of synchronous line loss data quality is very important for line loss data management and line loss analysis. In view of the contradiction between the accuracy and efficiency of line loss data quality evaluation, a novel evaluation model of line loss data quality based on penalty variable weight “rank sum” difference assessment (RDA) is proposed. Firstly, based on the traditional “rank sum” test, which does not depend on sample size and index selection, a RDA model is established to directly evaluate the overall quality of synchronous line loss data. Secondly, using the difference probability density function of the “rank sum” difference data series as the index, the weight of RDA is modified, with consideration of the influence of the number and magnitude of abnormal data on the data quality evaluation. And the method does not depend on the selection of indicators and the sample size. Finally, the simulation of Gansu grid actual operation data proves the validity and superiority of the evaluation model of synchronous line loss data quality based on penalty variable weight RDA proposed in this paper.