基于信息融合技术的变压器故障诊断

Transformer Fault Diagnosis Based on Information Fusion Technology

利用变压器绝缘油中溶解气体分析,检测油浸式变压器内部早期故障,已经成为变压器监测的一种重要手段,以确保变压器的安全可靠运行,从而对变压器的定期维修方式转变为内部状态预知维修方式。以变压器油中溶解特征气体为基础,运用BP神经网络、Elman神经网络和概率神经网络对故障变压器进行初步诊断,再利用D-S证据理论进行数据信息融合来确定变压器的运行故障。

Using dissolved gas analysis in transformer insulation oil to detect early internal faults of oil-immersed transformers has become an important means of transformer monitoring to ensure the safety and reliable operation of transformers.As a result,the regular maintenance mode of transformers is transformed into the internal state predictive maintenance mode.Based on the dissolved characteristic gas in transformer oil,BP neural network,the study makes use of Elman neural network and probabilistic neural network to make a preliminary diagnosis of the fault transformer,and D-S evidence theory is used to fuse data and information to determine the operation fault of transformer.