Application Of Information Entropy In Transformer Fault Diagnosis And Condition Based Planned Maintenance

With the continuous expansion of the grid scale, increasing the number of electrical equipment, traditional regular maintenance overhaul based model has been unable to meet the requirements of grid development. In recent years, state maintenance has been more and more attention, This method through on-line monitoring data to detect equipment abnormal, and repair fault location targeted. But the structure of transformer is complex and the field data is numerous, theoretical research on transformer condition assessment and fault diagnosis there are still many shortcomings by analysis data.This article is based on information entropy theory combines with fault tree model and oil chromatographic diagnostic techniques for transformer maintenance optimization studies. According to information entropy values for each detection method, optimize transformer maintenance program and establish a state based on information entropy theory transformer maintenance strategy optimization model.In considering the test results, test costs and conditional probability of failure transformer test items to be sorted. Improve maintenance efficiency, minimize maintenance times and reduce maintenance costs.According to statistical data transformer fault history and transformer oil dissolved gas parameters is the feature vector. Combined with information entropy get a new transformer fault diagnosis method based on the traditional gray correlation theory. By calculating the entropy of the gas value for each feature, the traditional characteristics of the gas reference sequence weighting factor values are revised. This new reference on accident probability factor, to avoid the influence of human factors previously gray theory diagnosis, improve the objectivity and accuracy of fault diagnosis.