运用时序贝叶斯知识库的电网故障诊断方法

发布时间：2018-01-10 17:15

本文关键词：运用时序贝叶斯知识库的电网故障诊断方法　出处：《电网技术》2014年03期 　论文类型：期刊论文

【摘要】：电网故障时有大量报警产生,充分利用报警信号及其时序信息,处理好保护与断路器误动、拒动、信息缺失等不确定性情况,对于电网故障诊断显得非常重要。时序贝叶斯知识库(temporal Bayesian knowledge bases,TBKB)能够清晰表达多个事件之间的时序约束关系,并具备贝叶斯网络的推理能力。建立了基于TBKB的电网故障诊断模型,提出了元件故障与保护动作、保护动作与相应断路器跳闸等之间的时序因果关系(TCR)表达、时序约束一致性检查方法。根据电网结构,可先在线搜索出疑似元件,再对它们自动构造TBKB模型。针对信息缺失节点的状态进行假设,形成假设状态组合。针对这些状态组合,通过贝叶斯反向、正向推理,可判断故障元件,误动与拒动的保护与断路器。多个算例验证了该方法的有效性。
[Abstract]:When a large number of power grid fault alarm, make full use of the alarm signal and timing information, protect and circuit breaker misoperation, maloperation, lack of information uncertainty, it is very important for power system fault diagnosis. The timing of knowledge base (temporal Bayesian knowledge Bias bases, TBKB) to express temporal constraints between multiple events clearly, and have the ability to reason the Bias network. A fault diagnosis model based on TBKB, presents a fault and protection, protection and corresponding circuit breaker of the timing between causality (TCR) expression, timing constraints consistency check method. According to the structure of power grid, the first online search the suspected component, they automatically construct TBKB model. The assumptions for the missing information node state, form hypotheses state combinations. For these States, by Tony Juliu reverse, forward reasoning, can judge the fault element, the malfunction and the rejection of the protection and circuit breaker. Many examples verify the effectiveness of the method.

【作者单位】：西南交通大学电气工程学院;四川省电力科学研究院;
【基金】：国家自然科学基金项目(51377137,51377136)~~
【分类号】：TM711
【正文快照】： 0引言电网故障诊断在故障元件识别、故障后快速恢复、防止连锁跳闸等方面起着重要作用。已有神经网络[1]、模式匹配[2-3]、Petri网[4-6]等人工智能方法用于电网故障诊断。贝叶斯网络通过因果推理与概率计算,擅长在信息不完备条件下处理复杂问题的不确定性,较符合电网故障诊断

【参考文献】