Buck Converter Parameters Design Using Artificial Intelligence-Based Genetic Algorithm

Date:

This research introduces an AI-based methodology for designing synchronous buck converters by combining machine learning models with genetic algorithms (GA). By setting strict fitness criteria for voltage ripple (≤ 1%), current ripple (≤ 10%), and component volume (≤ 7 cm3), the AI-driven approach successfully optimized the converter parameters beyond the capabilities of conventional manual design.

Key Results:

  • Efficiency: Achieved 99.383%, surpassing the conventional design.
  • Compactness: Reduced total component volume to 5.732 cm³ (a ~25% reduction).
  • Precision: Achieved a voltage ripple of only 0.168%, significantly lower than the 1.154% seen in standard methods.

Full Paper

Presentation Slides