Research on Condition Evaluation Algorithm of Oil-immersed Transformer Based on Naive Bayes

Abstract

The improved three ratio method is a general algorithm for condition evaluation of oil immersed transformer. It has high accuracy in pre-test and laboratory analysis. However, the accuracy of DGA online monitoring data is not high, resulting in the decline of the positive judgment rate of transformer condition evaluation using the improved three ratio method, which is difficult to support the development requirements of transformer intelligence. To solve this problem, a DGA state evaluation method based on Naive Bayesian algorithm is proposed. The algorithm first performs preprocessing such as median filtering on the DGA online monitoring data to remove invalid data, then uses a triple composed of three conditional attributes to describe the characteristics of the DGA data, and finally calculates the a priori probability of training samples and the a posteriori probability of test samples by naive Bayesian algorithm for state evaluation. The verification of the algorithm on the measured data set shows that the accuracy of the algorithm is better than the improved three ratio method, and the algorithm is feasible and effective