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

摘要:

为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。数值算例表明,改进证据推理规则可以充分利用变压器故障诊断中的一致性或冲突性多源故障征兆信息,提高变压器故障诊断的收敛速度,降低故障误判率。

Abstract:

To 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. The numerical examples show that the improved evidence reasoning rule can utilize adequately the consistent or conflicting fault symptom information, increases convergence speed and reduces misjudgment ratio of the transformer’s fault diagnosis.