Data Modeling and Optimal Control of DC/DC Converters Based on AI

This paper focuses on the modeling and control of DC/DC converters in microgrids and proposes a data modeling and optimization control based on Artificial Intelligence (AI-DMOC) for DC/DC converter. As key components in power conversion, accurate modeling and control of these converters are essential for enhancing microgrid performance. Traditional mechanism-based models are labor-intensive and less accurate, while existing data-driven methods often require long training times. AI-DMOC addresses these issues by employing the LightGBM algorithm, which offers high accuracy with reduced data, faster training, and lower memory consumption. It also integrates Particle Swarm Optimization (PSO) to determine optimal circuit parameters for maximum efficiency. Simulation and hardware tests show AI-DMOC achieves 99.86% modeling accuracy using just 20% of the data and reaches 98.51% efficiency—significantly outperforming traditional PID control. In summary, AI-DMOC provides a highly efficient and intelligent solution for controlling DC/DC converters in microgrids, demonstrating strong potential for real-world applications.