The design of power transformers is a demanding process that requires extensive knowledge and experience. The purpose of the master's thesis is the development of a genetic algorithm that facilitates the work of designers in the design of power transformers.
This master's thesis presents the fabrication and optimisation of an elitist multi-objective genetic algorithm with constraints for transformer design. The basics of transformer design and calculation are presented. Genetic algorithms are described, implemented extensions and improvements are presented. The setting of metaparameters to the optimal value is shown. The dependencies of the results on metaparameters are presented. The results are compared with transformers that are already manufactured by the company Kolektor Etra and with the results of the adapted hill-climbing algorithm.
The implemented algorithm returns reasonable solutions which can help the designer in the process of the design of the transformer. In some instances, the algorithm solutions are better than already manufactured transformers. The solutions of the genetic algorithm are much better than the solutions of the customised hill-climbing algorithm.
|
