ARTIFICIAL INTELLIGENCE-BASED ADAPTIVE CONTROL IN AC/DC MICROGRID ENERGY MANAGEMENT

Abstract

The integration of artificial intelligence into adaptive control systems is a novel strategy for enhancing energy management in AC/DC microgrids. The research constructs an extensive simulation model in MATLAB Simulink that replicates a hybrid microgrid system linked to the utility grid, focusing on renewable energy sources such as solar panels and wind turbines. Instead of relying on traditional fixed-parameter Proportional-Integral (PI) controllers, reinforcement learning (RL) agent have been used which constantly adjusts control settings based on changes in load and generation conditions. The Twin-Delayed Deep Deterministic Policy Gradient (TD3) method proved to be a much better fit for the job. The RL-based controller helped stabilize the DC link voltage and reduce those annoying overshoots during transients, making the system more efficient and adaptable. Experimental results show that, unlike traditional methods, this approach does a better job of handling system nonlinearity and the unpredictable nature of renewable energy sources. Overall, this adaptive control system highlights how AI-driven solutions can improve the stability, reliability, and overall performance of complex microgrid systems.