Diego Gerardo Bernal Cobaleda

Diego Gerardo Bernal Cobaleda is a researcher at KU Leuven, affiliated with the Electrical Engineering Department (ELEKTRA) and the EnergyVille research institute. His work focuses on power electronics design automation, integrating artificial intelligence (AI) and machine learning (ML) techniques to optimize converter efficiency and component selection.

• Research Highlights: AI-driven semiconductor datasheet analysis, variable transformer design for wide voltage ranges, reinforcement learning for converter parameter optimization.
• Collaborators: Fanghao Tian, Wilmar Martinez, Camilo Suarez Buitrago, Miguel Vivert.

His recent publications address:

• Automated extraction of dynamic characteristics from MOSFET datasheets using computer vision.
• Development of resonant multi-input multi-output (MIMO) converters for aerospace applications.
• Integration of variable transformers in LLC and dual active bridge (DAB) topologies for improved efficiency.
• Application of deep reinforcement learning (DDPG) to optimize converter parameters.

He has contributed to journals including IEEE Transactions on Power Electronics, Energies, and IEEE Access, as well as conferences such as IEEE APEC, ECCE Asia, and ISIE.