Combination of Support Vector Regression with Particle Swarm Optimization for Hot-spot temperature prediction of oil-immersed power transformer

The life expectancy and load capacity of oil-immersed power transformers are intimately associated with the winding hot spot temperature (HST). Thus, accurately predicting HSTs is essential in evaluating the life expectancy of power transformers and in preventing thermal failure. Previously, support vector machine (SVM) has been successfully employed to solve the regression problem of nonlinearity and small sample size. In the present study, given that the HSTs of transformers have a complex non-linear relationship with load information and environmental information, support vector regression (SVR) has been adopted to establish a model for the prediction of HSTs in power transformers. Among which, an improved particle swarm optimization (PSO) having passive congregation algorithm is utilized to determine the parameters of SVR. The PSO-SVR model has been applied to predict HSTs of a power transformer. Several experimental tests have been carried out involving a large power transformer in Sichuan Province, China, to verify the practicality and effectiveness of the proposed PSO-SVR model. In addition, PSO-SVR modeling results are compared with that of standard SVR and artificial neural network (ANN) by applying identical training and test samples. In conclusion, the PSO-SVR model has better prediction accuracy and generalization ability than both the standard SVR model and the ANN in the HST prediction of power transformers. Streszczenie. Artykul zajmuje się metodami przewidywania temperatury uzwojeń transformatora olejowego. W poprzedniej pracy do tego celu wykorzystano metodę SVM. W obecnej pracy uwzględniając nieliniowe zalezności od obciążenia i warunków zewnętrznych do przewidywania temperatury zastosowano metodę SVR (suport vector regression) oraz metodę optymalizacji wykorzystującą algorytmy mrówkowe.(Wykorzystanie metody SVR oraz algorytmówmrówkowych do przewidywania temperatury uzwojeń transformatora olejowego)