About my thesis - Model predictive control for flow batteries

The project aims to develop a model predictive control (MPC) framework for fuel cell stacks (FBs) through a structured approach. First, key FB parameters will be compiled, and a basic control framework established. Various modeling strategies will be explored, including a detailed white-box model based on computational fluid dynamics (CFD), a reduced grey-box electrochemical model (ECM), and a data-driven black-box model using machine learning. These models will support the design of a nonlinear MPC framework, which will then be integrated with a system model. Additionally, stack behavior (ENX) will be incorporated into the MPC framework to enhance accuracy. Validation will be performed either theoretically using a pseudo FB system or experimentally on an ICT laboratory cell.

The expected outcomes include the creation of a foundational MPC unit for automated FB operation, identification of the most effective modeling approach, and the establishment of an advanced MPC framework with improved control performance. The framework is expected to enable automated operation, trajectory tracking, management of complex interdependencies, and adaptability to varying operating conditions, ultimately demonstrating the practical applicability of the proposed control strategy.