基于改进二叉树支持向量机的低压台区用户拓扑关系识别

Low Voltage Topological Connection Relation Identification Based on the Improved Binary Tree Support Vector Machine

针对当前电网公司用户拓扑连接关系缺失和不准确,提出了一种改进二叉树支持向量机的低压用户拓扑连接关系识别方法。从用电信息采集系统提取待识别台区所有用户最近一段时间的电压序列数据,计算每个用户与变压器A、B、C三相电压序列数据之间相关系数;基于电压曲线相关系数值在不同区间出现的频率,采用改进的二叉树支持向量机构建低压用户拓扑连接关系识别模型,可快速识别用户连接变压器相序,以及与变压器连接关系是否正确。经验证,该校验方法所需人力成本低,准确性高,可操作性强。

In view of the lack and inaccuracy of the low-voltage users’topological connection relation,an method for identifying the relation based on improved binary tree support vector machine(SVM)is proposed.First,the voltage sequence data of the latest period of all users in a low-voltage transformer area is extracted from the electricity information acquisition system.Secondly,the correlation coefficient value between each user and the A,B and C phase voltage sequence data is calculated.Thirdly,based on the frequency of the correlation coefficient value in different intervals,an improved binary tree support vector mechanism is used to build the identification model of the low-voltage user’s topological connection relation.It can quickly identify the phase sequence of the transformer connected by the user and whether the connection relation with the transformer is correct.It has been proved that this method requires low labor cost,high accuracy and strong maneuverability.