Power Transformers Modelling, Finite Element Method, Training Dataset, Inter-turn Short Circuit
Abstract
Monitoring and diagnostic systems for dry-type transformers are particularly important in the oil and gas industry, where power supply is more susceptible to disruptions due to failures in these equipment. During their operation, transformers face electrical, thermal, mechanical, and environmental stresses, which accelerate the degradation of insulation materials, especially in the windings. Identifying these defects early potentially reduces the severity of failures and minimizes operation and maintenance costs. This study proposes a modeling approach for dry-type transformers to train algorithms for detecting interturn short-circuits based on machine learning. This article presents the 2D modeling of a transformer implemented using the finite element method to obtain data on interturn short-circuit conditions and its experimental validation.
