摘要
针对在外部扰动作用下传统孤岛检测方法易受到影响,导致错误判断,降低孤岛检测准确率问题,提出基于能量特性与随机森林的孤岛辨识模型.设计了多种干扰运行状态,并提取孤岛和干扰运行状态下的公共耦合点电压信号的特征能量,做归一化处理后组成辨识模型的输入特征向量;通过训练随机森林决策树构造孤岛辨识分类器,从而构建孤岛辨识模型;通过对比不同组数训练样本准确率和不同检测方法的辨识结果,验证该模型的快速性和准确性.研究结果表明:该模型能够快速辨识多种扰动与孤岛运行状态,更全面地保留检测信号的状态信息,为准确辨识孤岛效应的研究提供参考.
In view of the problem that the traditional islanding detection method is susceptible to be influenced under external disturbance and makes a wrong judgment to reduce the accuracy of islanding detection,an islanding identification model based on energy characteristics and random forest is proposed.Various interference and islanding operation states are designed,and the characteristic energy of PCC voltage signal are extracted and normalized to form the input characteristic vector of the identification model.Islanding identification classifiers are constructed by training a random forest decision tree to construct an islanding identification model.By comparing the accuracy of different groups of training samples and the identification results of different detection methods,the rapidity and accuracy of the model are verified.The simulation results show that the model can quickly identify a variety of disturbances and islanding operation states,retain the status information of detection signals more comprehensively,and provide a reference for the research on accurately identifying the islanding effect.
作者
付华
韩冰
崔鹏
孟祥云
FU Hua;HAN Bing;CUI Peng;MENG Xiangyun(School of Electrical and Control Engineering,Liaoning Technical University,Huludao 125105,China;Panjin Electric Power Supply Company,State Grid Liaoning Electric Power Supply Company Limited,Panjin 124000,China;Jinzhou Electric Power Supply Company,State Grid Liaoning Electric Power Supply Company Limited,Jinzhou 121000,China)
出处
《辽宁工程技术大学学报（自然科学版）》
CAS
北大核心
2021年第1期41-47,共7页
Journal of Liaoning Technical University (Natural Science)
基金
国家自然科学基金(51974151)
关键词
外部扰动
孤岛检测
能量特性
随机森林
辨识模型
external disturbance
islanding detection
energy characteristics
random forest
identification model
作者简介
付华(1962-),女,辽宁辽阳人,博士,教授,主要从事电力系统故障诊断等方面的研究.
