基于粗糙集理论与模糊Petri网络的油浸电力变压器综合故障诊断

电力变压器故障诊断因果关系的复杂性与模糊性，采用单一智能方法难以准确描述。文中基于智能互补融合的思想，将粗糙集理论与模糊Petri网络有机结合在一起进行油浸电力变压器故障诊断。利用粗糙集信息表简化技术来实现对专家知识的简化与故障特征的压缩，获得最小诊断规则，基于最小诊断规则的Petri网络模型可以有效降低网络结构的复杂性与故障特征获取的难度。同时利用模糊Petri网络实现并行模糊推理，便于描述故障特征的变化及对变压器运行特性的快速分析。故障实例分析表明，文中所提出的智能方法可以有效地进行模糊推理，减小诊断信息的冗余性，诊断效率高，计算快速、准确，结果易于被人理解。

SYNTHETIC FAULT DIAGNOSIS OF OIL-IMMERSED POWER TRANSFORMER BASED ON ROUGH SET THEORY AND FUZZY PETRI NETS

Transformer fault reasons are complex and fuzzy, which can not be described by single intelligent approach accurately. Based on complementary strategy, rough sets theory (RST) and fuzzy Petri nets (FPN) are integrated for synthetic fault diagnosis of oil-immersed power transformer. Through reduction approach of RST information table to simplify expert knowledge and reduce fault symptoms, the minimal diagnostic rules can be mined. According to the minimal rules, complexity of FPN structure and difficulties of fault symptom acquisition are largely lessened. Meanwhile, parallel and fuzzy reasoning can be realized by FPN, which can be used to describe changes of fault symptoms and analyze operating status of the transformer. The correctness and effectiveness are validated by the analysis result of practical fault examples.