Transformer Hot Spot Temperature Prediction Method Based on LM-UGO Algorithm

The transformer hot spot temperature determines the operating state and useful life of the transformer, and accurately predicting the hot spot temperature can effectively improve the transformer maintenance efficiency. Therefore, a 10 kV oil-immersed stereo roll core amorphous metal transformer was used as the research object in this paper, and its multiphysics transient coupling analysis simulation model was established. Based on the results of multiphysics simulation, the Levenberg-Marquardt (LM) algorithm combined with the universal global optimization (UGO) was used to correct the correlation coefficient in the empirical formula of hot spot temperature calculation, thus to improve the accuracy of temperature prediction. The results show that the proposed method can improve the efficiency of the actual operation and maintenance of transformers.