Analytic Model Predictive Current Control of Grid-Connected Power Conversion System in Battery Energy Storage System

In this article, an analytic model predictive current controller is proposed for the grid-connected power conversion system (PCS) in the battery energy storage system (BESS). This controller is designed to provide the fast and robust dynamic response while maintaining the moderate computational cost, making it suitable for real-time industrial microcontrollers. An accurate discrete-time model of the grid-connected PCS is derived to enable a closed-form solution, and the 3-D current reference tracking problem, incorporating the varying dc-link voltage of the BESS, is formulated as a multiparametric quadratic program. For this optimization problem, the constrained optimal control law is shown to be the accurate orthogonal projection of the unconstrained solution onto the corresponding active voltage constraint. Subsequently, an efficient geometric method is developed to obtain the analytic optimal control in the general case, enabling the online application of MPC with a long prediction horizon for industrial PCS. Moreover, a disturbance observer is combined with the predictive current controller to ensure offset-free tracking and robustness against system uncertainties. Simulation and experimental results on a 20 kVA PCS demonstrate the fast dynamics, low total harmonic distortion, strong robustness, and high computational efficiency of the proposed current controller.