基于改进证据推理规则的变压器故障诊断

Transformer's Fault Diagnosis Based on Improved Evidence Reasoning Rule

为对变压器进行准确、高效的故障诊断,克服已有证据推理规则处理冲突证据时收敛慢甚至不收敛的缺陷,提出了一种按相关性对证据进行修正及对冲突信息合理分配的改进证据推理规则,并将其应用于变压器故障诊断;数值算例中,与传统证据推理规则相比,无论对于一致性证据还是冲突性证据,改进证据推理规则提高了变压器故障诊断的收敛速度、降低了故障误判率、取得了良好的诊断效果;改进证据推理规则,可以充分利用变压器故障中的一致性或冲突性多源故障征兆信息,准确、高效对其进行故障诊断。

To diagnose the transformer＇s fault accurately and efficiently,overcome the defects of slow convergence,even non-convergence of existing evidence reasoning rules dealing with conflict evidences,the paper proposes an improved evidence reasoning rule and utilizes it in fault diagnosis of transformer,the improved rule modifies the evidence information according to their correlation and reasonably allocates the conflict information.In the numerical examples,compare with the traditional evidence reasoning rule,the improved one can increases convergence speed and reduces misjudgment ratio of the transformer＇ s fault diagnosis and have a good applied effect,whether for consistent or conflicting evidence.As a result,the improved evidence reasoning rule can utilize adequately the consistent or conflicting fault symptom information,and diagnose the transformer＇ s fault accurately and efficiently.