基于贝叶斯知识库的牵引变电所故障诊断方法

A Diagnosis Method for Traction Substation Faults Based on Bayesian Knowledge Base

介绍了一种基于贝叶斯知识库TBKB的牵引变电系统故障诊断方法,运用贝叶斯网络在处理不确定性和不完备性问题时的推理能力,处理实际报警信息中继电保护信息的不确定情况;利用TBKB的表达能力表示保护、断路器间的约束关系;利用保护与断路器的逻辑关系与约束关系确定故障元件,并通过算例验证了方法和模型的可行性。

The paper introduces a diagnosis method for traction power supply system on the basis of TBKB. The reasoning ability for dealing with uncertainty and incompleteness by application of Bayesian network is adopted to deal with the uncertainty of relay protection information in the actual alarm information; TBKB＇s ability of expression is adopted for indicating the restricted relationship between protection device and circuit breaker; and the logical relationship and restricted relationship of the protection device and circuit breaker are adopted for determination of faulty components. The feasibilities of the method and the model are verified by the account case.