Evaluation of transformer health condition using reduced number of tests

Abstract

This paper presents a method to find a reliable health index (HI) for oil-immersed transformers using reduced number of tests. The method depends on using the most important tests that are directly related to the transformer health. Field data for 90 working transformers were used to validate and test the proposed method. A linear relationship between the transformer HI and all available tests is proposed. The weights of the transformer tests are modeled exponentially according to the importance of each test. Subsequently, the HI is calculated using the three most important tests and the two most important tests. The results were compared with the results of existing research that uses a larger number of tests. The comparison shows that the proposed method gives acceptable results compared to other techniques that use a larger number of tests. As a result, the HI can be evaluated at lower cost, and a transformer health condition can be better monitored due to the ability to repeat HI evaluation at closer intervals.