一种结合模糊TOPSIS法和BP神经网络的变压器故障诊断方法

对变压器绝缘油采取色谱分析并进行故障诊断是目前变压器故障诊断的重要环节。针对传统故障诊断技术中降低状态检测与诊断系统的复杂度和增加系统采集的特征信息量的问题，首次提出将模糊TOPSIS 法用于状态检测信息处理。首先将TOPSIS 法在Vague集下进行扩展，把样本数据的多属性群问题转化为模糊多属性决策问题。然后计算Hausdorff矢量距离，得到可能解与理想解的相对贴近度，进行样本的优劣排序和聚类。在此基础上建立了简单的BP神经网络，实现了不同类型的模式识别。该方法可以快速、准确地判断变压器的运行状态，进而提高了变压器的安全、稳定及经济运行水平。

A new method based on fuzzy TOPSIS and BP neural network for power transformer fault diagnosis

Presently,dissolved gas analysis and fault diagnosis are the important segments of power transformer.A new method based on fuzzy TOSIS and BP neural network for power transformer fault diagnosis is put forward to deal with the problem,which is treated as the collision between lowering complex process of the diagnosis and increasing quantity of monitoring information.Firstly,the TOPSIS method is extended with vague set,and the multiattribute group problem of sample data is turned into fuzzy multiple attribut...