Sequential Scenario MPC for Battery Thermal Management Optimization of Electric Vehicles on Real-World Driving Cycles

Battery performance is susceptible to temperature and efficient battery thermal management (BTM) system needs to be designed to keep battery temperature, which also consumes considerable energy. However, the optimization for BTM require a long-term reliable prediction of the future velocity to improve performance. For real-world driving cycles, it is a much more challenging task. To cope with these uncertainties and to reduce computational time, this paper presents a sequential scenario-based model predictive control (Se-SMPC) strategy for BTM system to maintain the battery of the battery and reduce the energy consumption of the cooling system. The upper-level optimizer is used to generate a rough long-horizon control plan based on the value of the nominal velocity prediction uncertainties, while the low-level optimizer tracks long-term control sequence and provides the current control sequence by considering all the possible realisation scenarios of these uncertainties. In low-level, we extract random samples from the uncertainty intervals around the nominal velocity profiles and optimize over multiple scenarios. The effectiveness of the proposed approach is verified by simulation. The results show that Se-SMPC method reduces the BTM energy consumption by 1.46%, 3.79% and 17.36% compared to nominal MPC, PID and Rule on real-world cycles.