An intelligent dc current minimization method for transformerless grid-connected photovoltaic inverters

Area of Science:

• Electrical Engineering
• Power Electronics
• Control Systems

Background:

• Transformerless grid-connected inverters can inject DC components into the AC grid due to sensor errors, voltage imbalances, and switching device tolerances.
• This DC injection degrades overall grid power quality.
• Existing methods like blocking capacitors and feedback control have limitations in minimizing DC injection.

Purpose of the Study:

• To propose and evaluate an intelligent control strategy for suppressing DC current injection into the AC grid.
• To enhance the power quality of grid-connected systems.
• To demonstrate the effectiveness of a novel control approach.

Main Methods:

• Development of an adaptive-back-propagation (ABP) neural network PID controller.
• Implementation of the proposed control strategy for DC current injection suppression.
• Comparative performance evaluation against traditional and existing methods.

Main Results:

• The proposed ABP neural network PID controller effectively suppresses DC current injection.
• The intelligent control strategy significantly improves grid power quality compared to conventional methods.
• Experimental validation on a 2-kW three-phase grid-connected inverter confirmed the scheme's performance.

Conclusions:

• The adaptive-back-propagation neural network PID controller offers a superior solution for mitigating DC injection from transformerless inverters.
• This intelligent control strategy enhances the reliability and efficiency of grid-connected power systems.
• The validated method provides a practical approach to maintaining high grid power quality.