基于EMD的微风振动在线监测系统误差来源分析  

Error Source Analysis of On-line Monitoring System of Aeolian Vibration Based on Empirical Mode Decomposition

作　　者：张军[1,2] 贺瑞娟 龙嘉川 周玮[2] 卢冰[2] 王先培[3] 

机构地区：[1]华中科技大学强电磁工程与新技术国家重点实验室,湖北武汉430074 [2]中国电力科学研究院,湖北武汉430074 [3]武汉大学电子信息学院,湖北武汉430072

出　　处：《仪表技术与传感器》2016年第8期117-122,共6页Instrument Technique and Sensor

基　　金：国家自然科学基金项目(50677047);国家电网公司科技项目(5442JL140008)

摘　　要：微风振动在线监测系统因其能够获得导线振动幅值、频率等信息而被广泛应用于高压电力架空线路。但是研究表明,由于工作环境所限,其测量数据往往受到温度、湿度、电磁场等因素的影响,因此系统数据的可靠性难以保证。为分析微风振动在线监测系统所测数据中的误差来源,文中提出了利用功能模块分解的方法对系统误差进行溯源。首先,建立系统各单元模块相应的误差指纹库,然后利用经验模态分解的方法对现场采集数据进行多层分解,最后对分解的特征分量利用自相关和互相关算法与误差指纹库进行特征匹配辨识,以此实现在线监测的误差溯源。实验结果表明:所提方法能够准确有效定位系统误差来源,对装置的优化设计具有借鉴意义。Owning to its access to wire vibration amplitude,frequency and other information,on-line monitoring system of aeolian vibration was widely applied to high voltage overhead lines.However,it’s shown that the measuring data was heavily influenced by factors such as temperature,humidity,electromagnetic field,so it’s hard to ensure the reliability of the system’s data.In order to analyze the measurement errors of on-line monitoring system of aeolian vibration,the method of function module decomposition for tracing the system error was presented in this paper.Firstly,a corresponding error fingerprint database of each unit module was established.Secondly,empirical mode decomposition（EMD） method was used to decompose the field data acquisition into many layers.Finally,the decomposed characteristic component is matched with error fingerprint database by using autocorrelation and cross-correlation algorithm to realize the error source analysis of the system.Experiment results show that the proposed method can effectively find the source of system error and has reference to the optimization design of the device.

关 键 词：微风振动 经验模态分解 误差来源分析 自相关函数 互相关函数 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]