Predictive Direct Power Control of Grid-Connected PV Systems

Abstract

This paper presents an effective grid-injected power control method for a photovoltaic (PV) system using the theory of model predictive
control (MPC). The system consists of two power converters including a DC/DC boost converter for maximum point point tracking (MPPT)
and a DC/AC converter formed by a two-level voltage-source three-phase inverter. Future values of the grid-injected active and reactive
power can be predicted by using the inverter model and the discrete-time model of the load. A cost function is defined based on the differences
of the reference and predictive power. The optimal switching state of the inverter at each sampling period is selected corresponding to the
minimum value of the cost function. In addition, the desired grid-injected active and reactive power can be obtained by simply setting the
reference active and reactive power in the cost function. The main advantage of this control method is that it does not require the use of any
external current controller and pulse width modulation (PWM) techniques such as sinusoidal pulse width modulation (SPWM) or space vector
pulse width modulation (SVPWM). Moreover, the predictive direct power control does not use a phase locked loop (PLL) usually required
for the conventional current control methods of the grid-connected inverter. The simulation results obtained using MATLAB/Simulink verify
the effectiveness of the proposed method.