CFD-based 3-D optimization of the mutual coil configuration for the effective cooling of an electrical transformer

An optimal mutual configuration of coils and cooling ducts for the effective cooling of a dry-type transformer is presented in this paper based on the method developed by the author. In the optimization procedure, a computational fluid dynamics (CFD) and a genetic algorithm are combined to optimize the diameters of both the ducts and the coils. The method was applied to cool a special dry-type unit to minimize the hot-spot temperature of the windings. These simulations were performed using various sets of optimized shape parameters and copper mass constraints in a real 3-D transformer geometry. The objective function value is computed using a CFD model that accounts for all three heat transfer modes. In the proposed model, the thermal properties of the coils and core are treated as anisotropic and temperature-dependent quantities, and the power losses are treated as heat sources and are computed based on the coupled CFD-electromagnetic (EMAG) model. Due to a shape change, both coil properties and power losses vary with each generated coil configuration. The results show that the nonuniform positioning of the wires and air ducts and an optimal splitting of high- and low-voltage coils can significantly lower the hot-spot temperature and improve the heat dissipation in comparison with current transformer designs.