基于ADASYN-RF的用电安全隐患自适应分类识别方法  

Adaptive Classification and Identification Method for Electrical Safety Hazards Based on ADASYN-RF

作　　者：康洁滢 舒一飞 樊博 史强 杨琦 KANG Jieying;SHU Yifei;FAN Bo;SHI Qiang;YANG Qi(State Grid Ningxia Electric Power Co.,Ltd.Marketing Service Center(State Grid Ningxia Electric Power Co.,Ltd.Measurement Center),Yinchuan 750001,China)

机构地区：[1]国网宁夏电力有限公司营销服务中心(国网宁夏电力有限公司计量中心),宁夏银川750001

出　　处：《微型电脑应用》2024年第12期250-254,共5页Microcomputer Applications

摘　　要：为了减少用电安全隐患可能带来的损失,为设备维修提供有效的参考数据,提出自适应合成抽样(ADASYN)—随机森林(RF)的用电安全隐患自适应分类识别方法。根据不同类型用电安全隐患的产生原理,设置对应的电流、电压特征作为分类识别标准。利用ADASYN算法自适应采集用电设备运行数据,提取电流谐波畸变率、电压不平衡度等用电设备运行特征。构建RF分类器,确定当前用电安全隐患类型,实现用电安全隐患的自适应分类识别。通过与传统识别方法的比较,优化设计方法的精准率、召回率和平均F值分别提高了0.016、0.01和0.013,具有更优的识别性能。In order to reduce the possible loss caused by electrical safety hazards and provide effective reference data for equipment maintenance,an adaptive classification and identification method for electrical safety hazards based on adaptive synthetic sampling(ADASYN)-random forest(RF)is proposed.According to the generation principle of different types of electrical safety hazards,this paper sets the corresponding current and voltage characteristics as the classification and identification criteria.ADASYN algorithm is used to adaptively collect the operating data of electrical equipment,and extract the operating characteristics of electrical equipment,such as current harmonic distortion rate and voltage imbalance.The RF classifier is constructed to determine the type of current electrical safety hazards,so as to realize the adaptive classification and identification of electrical safety hazards.Through comparison with traditional identification methods,it is concluded that the accuracy,recall and mean F value of the optimal design method are improved by 0.016,0.01 and 0.013,which has better identification performance.

关 键 词：ADASYN-RF 用电安全隐患 自适应分类 安全隐患识别 

分 类 号：TM75[电气工程—电力系统及自动化]