Buck Converter Parameters Design Using Artificial Intelligence-Based Genetic Algorithm

The design of circuit parameters is critical to optimize the performance and efficiency of power converters. This research focuses on designing synchronous buck converters using an AI-based approach that combines machine learning models and genetic algorithms. The genetic algorithm run on the machine learning model operates within error bounds for voltage ripple (≤1%), current ripple (≤10%), and total component volume (≤7cm<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup>). It uses these criteria to evaluate fitness values. The results show that the AI-based design achieves higher efficiency (99.383%), lower voltage ripple (0.168%), and lower current ripple (9.757%) compared to the conventional design (99.354%, 1.1543%, and 11.661%). In addition, the AI-based method is more compact (5.732 cm<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> vs. 7.716 cm<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup>). Thus, AI is proven to be more effective in improving power converters’ efficiency and performance than conventional methods.