模糊专家系统在电力变压器在线故障诊断中的应用初探

在经故障诊断专家系统的实现遇到的瓶颈问题是：在线可获得的专家知识相对匮乏，以及专家知识的有效性。本文分析了在线可获得的原始特征数据，首次提出了利用在线获得的油气量和局放信号相结合来进行故障诊断，提高了诊断的精确性。利用油色谱进行故障分析需要很多的非定量的经验知识，利用模糊推理，很好的定量化了这些知识，在此之上，构造了一个全新的变压器在线故障诊断专家系统。

A PRIMITIVE STUDY ON ON-LINE FAULT DIAGNOSIS OF POWER TRANSFORMERS BASED ON FUZZY EXPERT SYSTEM

The critical problems existing in the process of realizing an on line fault diagnosis expert system are the lack of the expert knowledge obtained on line and the validity of the expert knowledge.The original character data which can be obtained on line are analyzed in this paper,and the new method put forward firstly to perform fault diagnosis,which is combined oil gas and partial discharge,raise the accuracy of diagnosis.Much qualitative experiential knowledge is required when oil chromatogram is used to analyzed diagnosis,and the knowledge can be quantificational by fuzzy consequence.Finally,a bran new system of on line fault diagnosis of power transformers is constructed.