Assessment Method of Power Transformer Winding Damage under Short-Circuit Conditions Based on Machine Learning

摘要

Power transformers take the impact from grid failures. How to effectively assess damage to the windings during design is a key issue. In this paper, a method for assessing winding damage under short circuit conditions is proposed. Historical experimental data and finite element simulation results are used to form the raw data. A machine learning approach is introduced to obtain the evaluation model. This assessment modeling approach reduces the number of costly and destructive experiments.