Model Predictive Power Control for Fault-tolerant Grid-connected Converter Using Reconstructed Currents

The reliability and stability of a grid-connected converter is vital to the power conversion system. The faults of switching device and current sensor lead to operation failure for a grid-connected converter. An improved model predictive power control (MPPC) based on reconstructed currents is proposed for fault-tolerant operation of a grid-connected converter. Firstly, the three-phase currents are reconstructed using the sampled dc current and normal sensor information according to the relationship between the voltage vector and output currents. Secondly, the predictive power model is built based on fault-tolerant three-phase four-switch topology. Then the reconstructed current is used for the MPPC method. To keep the balance of dc capacitor voltages, the cost function is optimised and the switching vector, which minimises the cost function, is applied to the fault-tolerant converter. Finally, the experiment platform and simulation models are built to verify the proposed method. The results show that the proposed MPPC using reconstructed currents can keep the fault-tolerant operation of the grid-connected converter with sinusoidal currents and stable output power. The proposed method achieves the fault-tolerant operation under switching device and ac current sensor fault, which improve the reliability of the system.