Offset free generalized model predictive control for 3-phase LCL-filter based grid-tied inverters

This paper presents a Generalized Model Predictive Control (GPC) in the natural reference frame for a 3-phase LCL-filter based grid-tied inverter. The proposed controller can track the sinusoidal reference without steadystate phase and amplitude offsets. This is achieved by parameterizing the control law matrices with considering the future reference trajectory. To cope with the LCL filter resonance, traditional controllers need a damping method. The proposed controller has inherent damping against the LCL filter resonance and extra sensors are avoided. The controller parameters are selected by a sensitivity analysis that guarantees the closed-loop stability. The proposed controller provides high accuracy in controlling active and reactive currents, robustness against grid inductance variations, low real-time computation, and a fixed switching frequency. The simulation and experimental results prove the performance of the controller in tracking the grid-side reference currents.