Structural optimisation of oil immersed transformer winding block washers based on WOA-DRBF surrogate model optimisation strategy

Abstract

In order to reduce the transformer winding hot spot temperature (HST), a dynamic radial basis function surrogate model optimisation strategy is proposed to optimise the structure of the forced oil circulation transformer winding's block washers in this paper. The whale optimisation algorithm (WOA) is introduced to obtain the optimal hyperparameter, and the WOA-DRBF surrogate model optimisation strategy is used to improve the fitting and optimisation ability. After optimisation, the HST corresponding to the size of the block washers decreased by 3.67°C, and the maximal temperature rise decreased by 12.93%, effectively ameliorating the phenomenon of excessive local temperature rise within each partition. Comparing the above optimisation results with those of the WOA-RBF static surrogate model and the genetic algorithm (GA), the proposed method shows superior performance in both optimisation search capability and efficiency. Comparing the optimisation efficiencies of the three methods, the number of model calls for the WOA-DRBF method is 23.89% of the WOA-RBF method and 1.77% of the GA method, respectively. The proposed method significantly enhances optimisation efficiency while improving global optimisation search capability. It offers a feasible solution for efficiently and accurately solving the optimisation problem of the transformer winding's block washer.