A Unified Model Predictive Control for the Grid Integration of Vanadium Redox Flow Batteries

This paper proposes a unified model predictive control algorithm for a double-stage isolated bidirectional converter interfacing a vanadium redox flow battery with the power grid. The proposed control algorithm relies on a single cost-function optimization routine to control both the grid-side and the battery-side converters. A dynamic model of the battery based on a variable-parameter electrical circuit equivalent is included in the predictive controller to improve the predictions, and thus the control performance, and to explicitly integrate advanced battery management features to fully exploit the potential of the vanadium redox flow technology. The proposed model predictive controller is compared with a conventional PI-based control system to assess its performance. Furthermore, an example of flow rate optimization based on the integrated battery model is presented, which aims at maximizing the system efficiency over a battery charge and discharge cycle.