Desain Parameter Buck Converter Berbasis Artificial Intelligence

Published in Bachelor's Thesis, Universitas Gadjah Mada, 2024

This research investigates an artificial intelligence (AI)-based approach for designing the circuit parameters of a synchronous buck converter. The study develops a machine learning model using simulation data generated in PLECS to predict key converter characteristics, including voltage ripple, current ripple, inductor current, and power losses. A genetic algorithm is then integrated with the machine learning model to optimize the converter parameters under specified design constraints.

Key highlights:

  • AI-Based Parameter Design: Uses machine learning to support the design and optimization of buck converter circuit parameters.
  • Machine Learning Model: Develops a neural-network model from simulation data to predict key converter performance characteristics.
  • Genetic Algorithm Optimization: Optimizes converter parameters under constraints on voltage ripple, current ripple, and total component volume.
  • Performance Evaluation: Compares the AI-based design with a conventional design in terms of efficiency, ripple, and component volume.