Recognition of Power Transformer Defect Identification Based on Dissolved Gas Analysis using Support-Vector Machine Approach

The primary purpose of the proposed research is to identify the different types of defects in high voltage transformers using a support vector machine (SVM) technique with tuned hyperparameters employing artificial bee colony (ABC) optimization, particle swarm optimization (PS 0), and genetic algorithms (GA) optimization. Based on the dissolved gas measurement, several fault types in the transformer can be indented to improve the accuracy of defect prediction using the Duval triangle and conventional methods. Performance of the transformer improvement and future defect identification is employed; in the support-vector machines method, the hyperparameters are optimized using three distinct optimization techniques. This proposed study also uses four unique SVM kernel functions for more research. An intelligent maintenance approach is advised for efficient planning to prevent significant damage to the power transformer.