Fast Zone Model Predictive Control for High Capacity Battery Subpack of Electric Vehicle

Abstract

As the demand for electric vehicles (EVs) increases, battery thermal management is required to guarantee safety and improve driving performance. The batteries need to be operated within an appropriate temperature range while minimizing energy consumption. We propose a fast zone model predictive control (MPC), which determines the optimal flow rate and inlet temperature of the coolant to control the temperatures of 48 cells in subpack. When the battery temperatures are within the proper temperature range, the proposed zone MPC focuses on minimizing power consumption while maintaining the temperature within the zone. When the temperature is outside the zone, the set-point MPC with terminal cost is used to determine the optimal input sequence. To achieve computational efficiency, a control-oriented battery thermal model is first established considering the temperature distribution of all cells. Second, the zone MPC formulation is converted into quadratic programming (QP). The nondifferentiable objective function of zone nonlinear MPC (NMPC) is approximated with a soft-plus function, and then, new variables are introduced to convert the nonlinear objective function into the quadratic form. Finally, the nonlinear dynamics are handled with the successive linearization method, which leads to QP formulation. By applying the proposed MPC, total energy consumption for cooling and heating cases under 1C-rate discharge was reduced by 27.11% and 78.73%, respectively, compared with the set-point MPC.