Assessment of Asset Health Index on Power Transformers Using AI Technique

Asset health index (AHI) is a methodological technique that combines field inspection, monitoring, field testing and laboratory testing to assess and rank the power transformer. The prediction of the power transformer’s AHI was performed using 639 data sets from utilities. The proposed model uses the Artificial Intelligent (AI) method to predict and diagnose transformer AHI based on DGA test results. The results show that the Random Forest method is more accurate with 96,1% accuracy, 100% precision, 96% recall and 98% F1 Score. The high performance of Random Forest was effective in handling complex patterns and large data sets where the relationships between variables are not easily modeled by traditional algorithms. The optimization of hyper parameters and cross validation is very effective for identifying the best set of hyper parameters for the model, ensuring better performance and robustness in dataset. The condition of the transformer health index can be used as a recommendation for preparing power transformer maintenance plans and setting priority maintenance scales to optimize operational costs and asset maintenance costs.