Genetic Algorithm for Optimized Design of Flyback Transformers

The paper presents an approach into designing transformers for flyback converters. The approach is based on the concept of computer aided design, where at its core is the utilization of a genetic algorithm. The genetic algorithm allows for optimal selection of critical transformer components and parameters, where the objective function - based on user preference - may include efficiency, cost, geometric size, thermal dissipation, weight and other parameters critical to the design. Each step of the design approach is presented and explained in details. The algorithm is realized and integrated into a computer application. The approach and the developed application are verified over a test design, where a selected example is presented in detail in order to highlight each step of the algorithm. The approach allows for the development of optimal designs over several algorithm generations.