Abstract

For relatively large grid-connected power converters, good low voltage ride through (LVRT) capability is mandatory. LVRT operation is typically achieved using control systems based on several resonant or proportional integral (PI) regulators, which are designed using simplified models that do not consider all the cross-couplings and interactions between the currents and voltages of different electrical sequences produced during a grid fault. As an alternative to conventional regulators, this article presents a novel continuous control set model predictive control (CCS-MPC) strategy for grid-connected power converters operating under LVRT conditions. The proposed controller achieves zero steady-state error in tracking sinusoidal current references through an augmented state-space model that incorporates resonant poles, thereby addressing a fundamental limitation of conventional MPC implementations. Some of the advantages and innovations of the proposed control methodology include: 1) a constrained optimization formulation with polytope-based current limits for fault scenarios; 2) a discussion of the discretization methodology required to ensure numerical stability of the predictive control system with embedded resonant poles; and 3) an analytical tuning methodology linking cost function weights to closed-loop dynamics. The controller's performance is validated under both symmetrical and asymmetrical voltage dip scenarios through comprehensive simulations and experimental tests on a 2 kW prototype. © IEEE. 1982-2012 IEEE.