基于RVM和ANF的电压暂降检测识别  

Detection and identification of voltage sags based on RVM and ANF

作　　者：吕干云[1] 吴传满[1] 孙维蒙[1] 汪晓东[1] 蔡秀珊[1] 

机构地区：[1]浙江师范大学数理与信息工程学院,浙江金华321004

出　　处：《电力系统保护与控制》2011年第12期65-68,共4页Power System Protection and Control

摘　　要：提出了一种基于关联向量机和自适应陷波器的电压暂降检测识别方法。通过一自适应陷波器,快速完成电压暂降特征检测(包括持续时间、幅值、相位跳变等),并从其输出提取出9个暂降关键特征信息。采用关联向量机分类模型对提取的特征信息进行训练与识别。仿真算例证明,该法快速有效完成电压暂降分类,识别出故障、电机启动和变压器激磁引起的三种电压暂降,精度高、实时性好,可应用于电能质量实时监测系统。An approach for detection and identification of voltage sags based on relevance vector machines（RVM）and adaptive notch filter（ANF）is presented.Firstly,three characteristics of sags,such as magnitude,duration and phase jump are detected through an ANF,and the information of nine key features are extracted from the output of the ANF.Finally,the extracted feature information is trained and indentified by RVM classifier model.The simulation results confirm that the approach is effective to identify the voltage sags generated by short-circuit faults,motor starting or transformer saturation with high accuracy and short test time,thus it can be used in power quality real-time monitoring system.

关 键 词：电压暂降 自适应陷波器 关联向量机 检测 识别 

分 类 号：TM714.2[电气工程—电力系统及自动化]