【作者】 王洋；

【导师】 顾煜炯； 肖官和；

【作者基本信息】 华北电力大学(北京) ， 能源动力硕士（专业学位）， 2023， 硕士

【摘要】 新型电力系统的建设使得电源结构和电网结构发生了深刻变化,这对电力设备的可靠性提出了新的考验,其中包括了电网中的核心设备变压器。变压器的运行状态直接关系到电网运行的安全性和稳定性。因此,在新型电力系统下,变压器需要具备更高的可靠性和安全性,以满足电网运行的要求。所以变压器状态检修与故障诊断是一项极其重要的工程。为了更加系统直观的分析变压器的故障模式,提高变压器的可靠性成为了电网越来越重视的问题,本文在对电力变压器典型故障知识研究的基础上,结合RCM理论和贝叶斯网络等相关理论,对变压器故障知识进行了知识化表达,继而开展以可靠性为中心的变压器状态检修与故障诊断关键技术研究。首先,针对变压器设备结构与故障间关联关系,方便更加系统性的分析变压器设备,提出了基于RCM理论的变压器故障诊断分析。运用FTA和FMEA分析方法全面地获取变压器故障知识;在此基础上,采用RCM理论对变压器的主要的故障进行了系统的分析,在RCM过程中,FTA和FMEA通常用于确定哪些故障模式可能导致设备失效,以及应该采取哪些预防措施来减少或消除潜在的故障;结合本体理论对变压器设备和故障知识进行结构化表达,提高知识管理效率。通过案例形成了以500kV变压器为对象,建立了涵盖六种组部件的故障知识库。其次,基于贝叶斯网络高准确性和实时性等特点,构建基于贝叶斯网络的变压器故障诊断。通过RCM分析收集的变压器历史运行数据中的状态参数与故障模式,对其进行离散化处理,确定贝叶斯网络节点。基于贝叶斯定理,构建了基于贝叶斯网络的故障诊断模型;结合基于RCM提取的故障特征和已建立的故障知识本体,运用贝叶斯网络模型分析计算了变压器故障发生的概率;通过某变压器案例验证了模型的有效性,并确定了故障原因。最后,基于变压器故障诊断的理论研究工作,设计并开发了变压器故障诊断原型系统,为后续变压器状态检测与故障诊断系统的工程应用提供了支撑。更多还原

【Abstract】 The construction of new power systems has brought profound changes to the power supply structure and power grid structure,posing new challenges to the reliability of power equipment,including the core equipment transformers in the power grid.The operating status of transformers directly affects the safety and stability of power grid operation.Therefore,in the new power system,transformers need to have higher reliability and safety to meet the requirements of grid operation.So transformer condition maintenance and fault diagnosis is an extremely important project.In order to analyze the fault mode of transformers more systematically and intuitively,and improve the reliability of transformers has become an increasingly important issue for the power grid.On the basis of the research on the typical fault knowledge of power transformers,this paper combines RCM theory and Bayesian network and other related theories to express the transformer fault knowledge in a knowledge-based way,and then carries out the research on the key technologies of transformer condition based maintenance and fault diagnosis centered on reliability.Firstly,aiming at the correlation between the structure and faults of transformer equipment,and facilitating a more systematic analysis of transformer equipment,a transformer fault diagnosis analysis based on RCM theory is proposed.Obtain transformer fault knowledge comprehensively using FTA and FMEA analysis methods;On this basis,the main faults of transformers are systematically analyzed using RCM theory.During the RCM process,FTA and FMEA are commonly used to determine which failure modes may cause equipment failures,and what preventive measures should be taken to reduce or eliminate potential faults;Structured representation of transformer equipment and fault knowledge based on ontology theory improves knowledge management efficiency.Through cases,a fault knowledge base covering six groups of components was established with a 500kV transformer as the object.Secondly,based on the high accuracy and real-time characteristics of Bayesian networks,a transformer fault diagnosis method based on Bayesian networks is constructed.RCM is used to analyze the status parameters and fault modes in the collected historical operation data of transformers,discretize them,and determine Bayesian network nodes.Based on Bayesian theorem,a fault diagnosis model based on Bayesian network is constructed;Combining the fault features extracted based on RCM and the established fault knowledge ontology,the probability of transformer fault occurrence is analyzed and calculated using Bayesian network model;The effectiveness of the model is verified through a transformer case,and the cause of the fault is determined.Finally,based on the theoretical research work of transformer fault diagnosis,a prototype system for transformer fault diagnosis was designed and developed,providing support for the subsequent engineering applications of transformer condition detection and fault diagnosis systems.更多还原