基于组合赋值法的稳态电能质量预警阈值研究

Combinatorial Weighting Method for Early Warning Thresholding of Power Quality

对电网中电能质量监测数据进行超标和异常数据挖掘,并给出相应预警,可实现问题早发现,早解决。但目前对异常数据预警阈值研究欠缺,且现有的阈值判断算法大多采用单一赋值现状,本文提出了基于组合赋值法的稳态电能质量预警阈值判定方案。该方案利用基于距离的聚类分析法和改进的最大类间方差法(Otsu)确定客观阈值,同时使用专家打分法计算出主观阈值,并按比例集权方式确定综合阈值。最后,基于实测数据对相关算法进行了算例验证,结果表明所提方法可对异常数据预警阈值进行较好设定,并能将其有效应用于稳态电能质量预警系统。

Conducting over-limit and abnormal data mining for power quality （PQ） monitored data, and givingcorresponding early warning can help find PQ problems earlier. However, few researches on setting the threshold ofabnormal data can be found, and existing threshold judgment algorithms mainly concentrates on single assignment. Inthis paper, the combinatorial weighting method for early warning thresholding of PQ is put forward. The method usesdistance based clustering analysis method and improved Otsu method to determine the objective weight. At the sametime, expert scoring system is applied to calculate the subjective weight, and the final weight is determined bycombinatorial weighting method. Calculation example based on measured data shows that the above method can setPQ early warning threshold accurately, and can be effectively used in the early warning system for PQ abnormaldata.