【作者】 曹飞；

【导师】 吕红兵；

【作者基本信息】 浙江大学 ， 计算机应用技术， 2008， 硕士

【摘要】 断路器是变电站机械和电气动作最频繁的设备,在运行过程中若出现故障,极易引起电网事故,造成较大的经济损失,因此对其可靠性提出了非常高的要求。断路器在线状态监测系统能及时了解断路器的工作状态和缺陷部位,减少过早或不必要的停电试验和检修,减少维护工作量,降低维修费用,提高检修的针对性,可显著提高电力系统可靠性。状态监测是实施状态检修的基础,以科学的方法对设备运行状态进行监测和诊断是实施状态检修的关键。我国对于电力设备状态检修的研究多集中于变压器,发电机一类的大型设备,对于断路器的状态监测和诊断的研究还处于起步阶段,现有的在线监测系统中普遍存在以下问题:监测量少,功能单一,缺乏系统性和综合性;采用的分析方法少,对断路器的运行状态的识别缺乏有效的数学方法和技术手段。断路器在线监测研究的内容十分广泛,本课题着眼于断路器在线监测系统中的一些信号处理和分析方法,主要是对振动信号和线圈电流进行数据分析和处理,同时深入研究小波消噪方法和数据预处理方法。利用绍兴电力局提供的断路器在线监测数据,应用小波分析理论和各种信号处理方法,重点分析振动信号与线圈电流波形,给出断路器的故障诊断结果,并建立起断路器的正常指纹库和故障指纹库。因此,本人主要从以下几个方面开展研究工作:1)查阅国内外的大量文献资料,深入研究振动信号与线圈电流的各种波形分析方法,寻找有效的数据分析方法。2)利用MATLAB强大的计算功能,实现数据分析算法,对振动信号波形与线圈电流波形进行MATLAB仿真分析,确定能用于工程实践的数据分析方法。3)开发断路器在线监测数据分析系统,将研究成果应用于生产实践。本课题对断路器在线监测系统的数据进行了深入的数据挖掘,研究并实现了一些比较有效的断路器在线监测数据分析诊断方法,更对变电所其他设备的在线监测数据开展分析研究作出了有益的探索。数据分析系统为绍兴局提供正常指纹库,并可以维护故障指纹库,取得了令人满意的运行结果。数据分析系统给出的断路器故障诊断结果可供技术人员参考,进一步提高了断路器故障诊断水平。更多还原

【Abstract】 Circuit breaker is the device of many mechanical and electric actions in the substation. If circuit breaker has troubles, electric power system fails easily, and it will cause greater ecnomic losses. Therefore the reliability of circuit breaker is very important. On-line working condition monitoring system can find out working condition and fault position of circuit breaker just in time. It also can decrease earlier or unnecessary power cut experiment and repair, reduce maintenance and the charge of repair, increase the pertinence of inspection and repair, enhance the reliability of electric powder system distinctly.Condition monitoring is the base of Condition Based Maintenance. Scientific method is the linchpin of monitoring and diagnose for running conditon of device. In the area of power device condition maintenance, research in our country focuses on large devices, such as transformer and dynamotor, but weak in circuit breaker’s condition maintenance and diagnose. There are some problems in the existing on-line monitoring system: a few of monitoring quantity and functions, lacking systemic orgnization and integration, a few of analytic method, and lacking methematic methods and techniques to identify circuit breaker’s running condition.Rich in the content of research of circuit breaker on-line working conditon monitoring, this project focuses on some signal disposal and analytic method of circuit breaker on-line working conditon monitoring system, disposing vibration signal and winding electric current mostly, researching on wavelet denoise method and data pretreatment deeply. Use the data of circuit breaker on-line working conditon monitoring of ShaoXing Power Bureau, applicate wavelet analysis theory and all kinds of signal dispoing methods, focus on vibration signal and winding electric current wave, at last, provide the fault diagnose result of circuit breaker, and build circuit breaker’s standard finger library and fault finger library. Therefore the paper’s main content includes:1) study much literature in the world, do research deeply on all kinds of wave analysis methods of vibration signal and winding electric current deeply, and find out effective data ayalysis methods.2) use MATLAB’s strong computing function, implement data analysis algorithm, do tated analysis to vibration signal and winding electric current using MATLAB, confirm the data analysis methods on project practice.3) develop circuit breaker on-line working conditon monitoring data analysis system, and apply the research results to production.This project have done an in-depth data mining to the data of circuit breaker on-line working conditon monitoring system, and implement some effective circuit breaker on-line working conditon monitoring data analysis diagnose methods, promote the development of the analysis and research of other devices’s on-line working conditon monitoring in the substation too. Data analysis system provides standard finger library and satisfied fault diagnose results for ShaoXing Power Bureau, and the fault finger library is maintainable. Finally the level of circuit breaker’s fault diagnose is much improved.更多还原