Advanced predictive direct power control for grid-connected WECS-PMSG: comprehensive simulation and real-time...

Area of Science:

• Electrical Engineering
• Renewable Energy Systems
• Control Systems

Background:

• Wind energy is crucial for the global energy mix, necessitating efficient and stable Wind Energy Conversion Systems (WECS).
• Conventional Direct Power Control (DPC) offers fast response but suffers from power ripples, high Total Harmonic Distortion (THD), and variable switching frequency.
• Permanent Magnet Synchronous Generators (PMSG) are preferred for their efficiency and operational flexibility in WECS.

Purpose of the Study:

• To propose and validate an Advanced Predictive Direct Power Control (APDPC) strategy for grid-connected WECS using PMSG.
• To address the limitations of conventional DPC, specifically power ripples, THD, and switching frequency variations.
• To enhance power regulation, dynamic performance, and grid integration of WECS under variable wind conditions.

Main Methods:

• Implemented a unified control framework for both Generator-Side Rectifier and Grid-Side Inverter using APDPC.
• Selected three voltage vectors based on predicted active and reactive power evolution, optimizing their application times with a multi-objective cost function.
• Validated the strategy via MATLAB/Simulink simulations and real-time implementation on a dSPACE DS1104 platform.

Main Results:

• Achieved a 37% faster dynamic response, eliminated overshoot/undershoot, and reduced steady-state error by 61% compared to conventional DPC.
• Reduced torque and power ripples by 40-45% and THD in grid current/voltage by up to 50%.
• Demonstrated superior tracking accuracy and overall performance against recent control strategies.

Conclusions:

• The proposed APDPC strategy significantly enhances the dynamic performance and power quality of grid-connected PMSG-based WECS.
• APDPC offers a practical and effective solution for robust WECS operation under variable wind speeds.
• The unified control framework and predictive vector selection provide a robust and efficient alternative to conventional DPC.