Dissolved gas analysis based feedback cloud entropy model for power transformer fault diagnosis
XU Hui-ju · Power System Protection and Control · 2013
In order to solve the problem of randomness and fuzziness in power transformer fault diagnosis, a new fault diagnosis method for power transformer based on feedback cloud entropy model is proposed. After the statistic analysis of the collected fault examples of chromatographic data for transformer oil, which is put into Bayesian feedback backward cloud generator as cloud drop,the transformer fault diagnosis standard normal cloud model is built based on parameter values of fault characteristic gases cloud model. The model integrates cloud correlation coefficient and information entropy theory, reduces the dependence on the single standard normal cloud model and digs more information of the dissolved gases in transformer oil, and improves the accuracy of transformer fault diagnosis. By increasing training samples and correction cloud model parameter, the effectiveness of the model can be further enhanced. The results of example show that the model has higher accuracy, and has well theoretical value and application prospects.