由于受工作时长和环境因素的影响,电容式电压互感器(capacitor voltage transformer,CVT)在运行过程中误差稳定性不高,易出现电能计量失准现象.为此,该文提出了一种基于融合K-近邻算法(fusion K-nearest neighbor algorithm,FKNN)的电压互感器在线评估方法.该方法利用互感器的历史运行数据构建虚拟标准器,通过改进K-近邻算法对互感器实时状态进行监测,实现对异常情况的报警.同时,提出了一种加权移动时间窗的方法,自适应更新异常阈值,有效削弱电网不平衡波动的影响.实验结果表明,该文方法能够准确监测互感器的0.2级误差漂移.
Abstract
Due to the influence of working hours and environmental factors,capacitor voltage transformer(CVT)error stability during operation is low,and it is prone to inaccurate energy measurement.Therefore,this paper proposes an online evaluation method for voltage transformers based on the fusion K-nearest neighbor algorithm(FKNN).The method uses the historical operation data of the transformer to construct a virtual standard instrument.It monitors the real-time operation status of the transformer through an improved K-nearest neighbor algorithm to achieve abnormal condition alarms.At the same time,a weighted moving time window method is proposed to adaptive update the abnormal threshold value,effectively weakening the influence of unbalanced fluctuations in the power grid.Experimental results show that the proposed method can accurately monitor the 0.2-level error drift of the transformer.
关键词
电压互感器/虚拟标准器/K-近邻算法/自适应更新
Key words
voltage transformer/virtual standard instrument/K-nearest neighbor algorithm/adaptive updates
维普
万方数据