A reduced vector model predictive controller for a three-level neutral point clamped inverter with common-mode...

Area of Science:

• Electrical Engineering
• Power Electronics
• Control Systems

Background:

• Conventional predictive control for three-level voltage source converters (NPC inverters) faces computational complexity and performance limitations.
• Existing methods struggle with efficiency and common mode voltage reduction, hindering renewable energy integration.

Purpose of the Study:

• To develop a novel predictive control architecture for NPC inverters that reduces computational complexity and common mode voltage.
• To enhance the overall efficiency and performance of predictive control techniques in three-level voltage source converters.

Main Methods:

• A statistical approach for proactive detection and selection of frequently used voltage vectors to limit the search space.
• A sector-splitting strategy based on statistical analysis to reduce the number of voltage vectors considered.
• Simulation and empirical evaluation across various operational scenarios.

Main Results:

• Demonstrated significant reduction in computational load compared to conventional methods.
• Achieved notable diminution in common mode voltage.
• Validated enhanced computational efficiency and reduced common mode voltage through simulations and experiments.

Conclusions:

• The proposed predictive control methodology effectively mitigates computational demands and common mode voltage in NPC inverters.
• The enhanced control strategy improves efficiency and facilitates the integration of renewable energy sources into the electrical grid.
• This approach offers a state-of-the-art solution for advanced power converter control.