可能性模糊均值聚类的变压器故障诊断研究

摘要

The control strategy of the Possibilistic Fuzzy C-Means is proposed for Transformer Fault Diagnosis in this paper. Firstly, the relationship between the fault diagnosis of transformer and the gas dissolved from insulating oil has been analyzed, and then, the basis of fault diagnosis is derived. Secondly, this paper has analyzed the principle of the traditional fuzzy clustering algorithm, and points out that the traditional algorithm is sensitive to the data noise interference. On the basis of this, an improved algorithm is applied to transformer fault diagnosis and the objective function of the fault model of the transformer system is established. Finally, the experimental results show that: the improved algorithm has achieved good results in transformer fault diagnosis.%为提高变压器故障诊断准确率，采用一种可能性模糊均值聚类算法。首先，对变压器的故障与绝缘油中溶解的气体之间关系进行分析，推导出采用特征气体法判断变压器故障依据。其次，分析传统模糊均值聚类算法的原理，指出传统模糊均值聚类在变压器故障诊断中存在对数据噪音干扰敏感，影响诊断结果等问题，并在此基础上提出采用一种可能性模糊聚类均值算法，建立变压器系统故障模型的目标函数，确定算法步骤流程图。实验对比表明，采用可能性模糊均值聚类算法对变压器故障取得了良好的诊断效果。