| Summary: | This paper presented a dissolved gas analysis (DGA) based prediction of transformer health index (HI) for transformer fault determination and diagnosis over the life of power transformers. The transformer health index was invented and has been used in the industry for a long time. However, it lacks an accurate model and calculation because the transformer parameters are complex and the calculation needs to be validated by experts. Therefore, the DGA prediction algorithm to determine the health of the transformer is needed to solve this problem. In this project, gap analysis, DGA based health index calculation and performance analysis were used to determine the transformer health condition and the best ANN algorithm to predict the transformer HI. The study shows that Bayesian Regularization is the most effective compared to other prediction algorithms. In conclusion, the study shows that the proposed DGA prediction is feasible for further applications in predicting the transformer HI instead of relying on complex computation and expert validation. © 2023 IEEE.
|
|---|