ELECTROCHEMICAL LI-ION BATTERY MODELING FOR CONTROL DESIGN WITH OPTIMAL UNEVEN DISCRETIZATION

This paper proposes a reduced Li-ion battery model for design optimization and control design by implementing the electrode-averaged diffusion dynamics and uneven discretization of the particle radius for fast computation and accurate prediction of the Lithium intercalation dynamics. First, the unevenly discretized dynamics model is constructed from the ordinary differential equation (ODE) derived in the electrode-averaged battery model. Then, constrained optimization problems with multi-objectives are formulated to find the optimal uneven discretization. The cost function is evaluated under the wide battery operation data sets constructed by Latin hypercube sampling (LHS) to reduce the total number of cases. The optimally determined unevenly discretized model can predict the battery electrochemical dynamics with much smaller number of discretization steps compared to the evenly discretized electrode-averaged battery model without loss of physical interpretation of the diffusion dynamics in the electrode solid particles.