One-dimensional physics-based reduced-order model of lithium-ion dynamics

We present a method to produce a physics-based one-dimensional discrete-time state-space reduced-order model (ROM) of a lithium-ion cell. The resulting ROM can predict the five variables of a standard porous-electrode model-reaction flux, solid and electrolyte lithium concentration, and solid and electrolyte potentials-at any location across the cell cross section, as well as cell terminal voltage. The method to generate the model involves first linearizing the porous-electrode-model equations, and then deriving closed-form Laplace-domain transfer functions from the linearized equations. Next, the discrete-time realization algorithm (DRA) is used to convert the transfer functions into an optimal discrete-time state-space realization. Advantages of this approach include that the DRA avoids nonlinear optimization and gives a straightforward method for selecting the system order for the ROM. Simulation results demonstrate that the ROM cell voltage predictions and the ROM internal electrochemical variable predictions match very closely with results obtained by simulating the full nonlinear porous-electrode partial differential equations. (C) 2012 Elsevier B.V. All rights reserved.