Model Predictive Control Strategies Applied to Battery Power Limit Estimation in Electric Vehicles

As the automotive industry transitions towards cleaner and more sustainable technologies, the shift from internal combustion engine (ICE) vehicles to electric vehicles (EVs) has gained momentum. This shift, driven by increasing environmental concerns and stringent regulations, highlights the need for advanced energy management systems, particularly in battery technology. A crucial component in ensuring optimal battery performance and longevity is the Battery Management System (BMS), which plays a pivotal role in monitoring and controlling critical parameters like temperature, State of Charge (SoC), and State of Power (SoP). Among these, SoP estimation is vital for high-performance applications such as electric vehicles, as it dictates how much power the battery can safely provide under current operating conditions. Unlike instantaneous estimates, predictive estimates of SoP offer a significant advantage by allowing for load scheduling over a future time horizon. This proactive approach requires continuous updating of power limits based on changing conditions, which is where Model Predictive Control (MPC) comes into play. MPC is an advanced control strategy that leverages a "look-ahead" technique to predict future dynamic behaviors and compute optimal inputs. This allows the controller to achieve the best possible performance while ensuring that the battery operates within its safe operating limits. The primary advantage of MPC is its ability to take into account both current and future constraints, optimizing the balance between power delivery and battery safety. In this thesis, MPC is applied to estimate and control the power limits of a lithium-ion battery, used in EVs. A model-based approach is employed, where a single battery cell is graphically modeled using Simulink, a widely used simulation tool in the MATLAB environment. By integrating the MPC toolbox of MATLAB with the battery model, a control algorithm is designed that adjusts the power limits in real-time, aiming to follow a reference current while maintaining the battery within its design constraints. Various simulations and tests are conducted to validate the effectiveness of this control strategy, including tests with different prediction horizons. Two types of algorithms are implemented: a nonlinear MPC, since the battery is inherently a nonlinear system, and an adaptive MPC, most suitable for the battery’s time varying parameters. In conclusion, this work demonstrates that MPC can be a powerful tool for managing the power limits of batteries in electric vehicles. By incorporating future predictions and constraints into the control strategy, it offers a promising solution for managing EV battery power limits efficiently, ensuring safe operation while enhancing performance.