Transformers are critical components, and early detection of potential failures plays an important role in the reliable operation of the power system. This article describes a novel approach for power transformer failure prediction based on dissolved gas analysis (DGA) by applying the TDCG method with Random Forest algorithm. DGA data from operational transformers are used to train and test the predictive model. The Random Forest method based on TDCG enables comprehensive analysis of dissolved gas changes in transformer oil, thus enabling early detection of failure conditions. Experimental results show that the predictive model using the model created by applying hyperparameter tuning for optimal parameter tuning to have high accuracy, the accuracy obtained reaches 96% in detecting potential failures, the standard used for accuracy presentation uses confusion matrix as the accuracy of the predictive model. In addition, it can optimize time efficiency in analyzing failures and prevent human error when calculating gas fault identification or potential failures.
Copyrights © 2024