Grid-Voltage Sensorless Model Predictive Control of Three-Phase PWM Rectifier Under Unbalanced and Distorted Grid Voltages

Abstract

The sensors in a three-phase pulse width modulation (PWM) rectifier control system are the basis of high-performance methods. Line voltage sensors are utilized to obtain information about the grid voltage amplitude and phase. To eliminate the grid voltage sensors and to decrease the cost of PWM rectifier control systems, various grid-voltage sensorless methods have been developed. In this article, a virtual-flux-based grid voltage estimation method is utilized in the model predictive control (MPC) of a PWM rectifier. The extended reactive power is introduced in the proposed MPC method; the MPC is then able to realize the control targets of eliminating input-side power oscillations and ensuring sinusoidal currents under unbalanced network conditions without any power compensation terms. The cascaded delayed signal cancellation (CDSC) method is applied to obtain the fundamental rectifier voltages in the virtual flux-based grid-voltage sensorless method. The proposed method can operate under unbalanced and distorted network conditions. Moreover, a startup current suppression method is applied in this method to avoid over-currents during the startup process. Experimental tests are performed to verify the effectiveness of the proposed method.