Machine Learning Based Robust Control Solutions for Buck Converters: ANN and Fuzzy Logic Methods

Machine learning (ML) techniques are increasingly explored in power electronics converters for enhancing control solutions. This study investigates robust control for buck converters using Artificial Neural Networks (ANN), Fuzzy Logic Control (FLC), and Model Predictive Control (MPC). Implemented in MATLAB Simulink, four scenarios were considered: Open Loop Buck Converter (OLBC), Fuzzy Logic Controller-Based Buck Converter (FLBC), ANN-Based Buck Converter (ANNBC), and MPC-Based Buck Converter (MPCBC). Performance metrics such as rise time, settling time, overshoot, and peak values were evaluated under various input voltages and load conditions. Results demonstrate that ANN-based control significantly outperforms other methods, with ANNBC achieving an extremely low overshoot of 0.0781 %, compared to 57.7626% for OLBC, 47.8383% for FLBC, and 105.3098% for MPCBC. ANNBC also maintained robust performance with minimal variations in rise time, settling time, and peak values across different conditions. Despite a longer rise time and settling time compared to some techniques, ANNBC's precision, stability, and robustness justify this tradeoff. The study highlights that while MPCBC offers effective control, it exhibits higher overshoot and longer settling time compared to ANNBC. This study underscores the potential of ML, particularly ANN, in providing robust and high-precision control for buck converters, making it preferable for demanding applications.