摘要
在传统的油中溶解气体分析方法的基础上,利用模糊神经网络强有力的关系处理能力,研究提出牵引变压器全局故障诊断方法。依据模糊神经网络理论,通过对数值逻辑故障诊断模型和物理逻辑故障诊断模型2类模糊神经网络故障诊断模型的分析,考虑信息采集节点的向量特性、变化趋势特性以及模糊神经网络的反馈特性,给出牵引变压器全局故障诊断模型,以故障征兆特征变化趋势表征故障征兆与故障类别间的因果关系,确立增益参数、权系数判定矩阵与决策矩阵。试验结果表明:该方法能够更好地分析牵引变压器各类故障产生的原因,明确故障特征类型,避免用单一特征数据集诊断牵引变压器故障带来的局限性,可以提高故障诊断的准确率。
Based on traditional Dissolved Gases Analysis （DGA） method, global fault diagnosis method of traction transformer was introduced by using powerful relation processing capability of fuzzy neural network. On the basis of fuzzy neural network theory, global fault diagnosis model was established through analyzing two sorts of fault diagnosis models of fuzzy neural network, which were numerical value logic fault diagnosis model and physical logic fault diagnosis model, and considering the vector characteristics, variation trend characteristics of information acquisition nodes and the feedback characteristics of fuzzy neural network. Cause and effect relation between fault omen and fault types was expressed by variation trend of fault omen characteristics. And interrelated gain parameter, weight coefficient judging matrix and decision matrix were designed. The experiment result indicates that this method can better analyze the causes for different faults of traction transformer, ascertain the type of fault characteristics, avoid the localization to traction transformer fault diagnosis by single characteristics data set, and thus improve the precision of fault diagnosis.
出处
《中国铁道科学》
EI
CAS
CSCD
北大核心
2009年第1期103-107,共5页
China Railway Science
基金
国家“八六三”计划项目(2007AA11Z247)
关键词
故障诊断
牵引变压器
模糊神经网络
模糊综合决策
Fault diagnosis
Traction transformer
Fuzzy neural network
Fuzzy integrated decision-making
