Comparative Study on Defects and Faults Detection of Main Transformer Based on Logistic Regression and Naive Bayes Algorithm

Abstract

Abstract The proper work of main transformers plays an important role in the stability of the power grid. The development of power data allows us to evaluate the instant status of the main transformer. If we use appropriate data classification algorithm, the laws and values hidden behind the mass power data can be discovered. This paper took basic power data, main transformers’ status data and oil chromatography data into consideration, and used Logistic regression and Naive Bayes algorithm to evaluate 110kV main transformers’ defects and faults from regression analysis and probability statistics seperately. The results of these two algorithms were obtained and compared. Finally, a more accurate algorithm of evaluating main transformer fault was selected, which provided a new reference for the application of power big data.