Comparative Performance Analysis of the DC-AC Converter Control System Based on Linear Robust or Nonlinear PCH...

Area of Science:

• Electrical Engineering
• Control Systems
• Renewable Energy Integration

Background:

• Microgrids require robust DC-AC converters for stable grid connection.
• Existing controllers like PI and PCH face challenges with dynamic load variations.
• Advanced control strategies are needed to optimize converter performance and grid stability.

Purpose of the Study:

• To evaluate and compare the performance of different DC-AC converter control systems.
• To investigate the enhancement of Port Controlled Hamiltonian (PCH) controllers using Reinforcement Learning-Twin Delayed Deep Deterministic Policy Gradient (RL-TD3) agents.
• To ensure stable voltage regulation under balanced, unbalanced, and nonlinear three-phase load conditions.

Main Methods:

• Implementation of PI, linear/nonlinear PCH controllers for DC-AC converter active elements (IGBT/MOSFET).
• Integration of a Reinforcement Learning-Twin Delayed Deep Deterministic Policy Gradient (RL-TD3) agent for controller performance enhancement.
• Performance evaluation using metrics like steady-state error, error ripple, and Total Harmonic Distortion (THD) via Matlab/Simulink simulations.

Main Results:

• Nonlinear PCH controllers demonstrated superior performance compared to traditional PI controllers.
• The RL-TD3 agent significantly improved the performance of all tested controllers.
• Reduced steady-state error, error ripple, and Total Harmonic Distortion (THD) were observed with the enhanced control strategies.

Conclusions:

• Nonlinear PCH controllers offer robust and effective control for DC-AC converters in microgrids.
• RL-TD3 agents provide a valuable mechanism for enhancing the performance of existing control systems.
• The proposed control strategies ensure reliable microgrid integration and power quality.