Global Transformer Optimization Method Using Evolutionary Design and Numerical Field Computation

This paper addresses the complex optimum transformer design problem, which is formulated as a mixed-integer nonlinear programming problem, by introducing an integrated design optimization method based on evolutionary algorithms and numerical electromagnetic and thermal field computations. The main contributions of this research are: i) introduction of a new overall transformer optimization method, minimizing either the overall transformer materials cost or the overall transformer materials and operating cost, ii) expansion of the solution space by innovative techniques that define the variation of crucial design variables such as the conductors' cross-section, ensuring global optimum transformer designs, and iii) incorporation of numerical field computation in order to validate the feasibility of the optimum designs. The proposed method is compared with a heuristic optimization method of the transformer manufacturing industry and the results demonstrate the robustness and the superiority of this new approach. Index Terms-Design method, finite element method (FEM), mixed integer nonlinear programming, optimization methods, transformer