A Robust and Sleek Electrochemical Battery Model Implementation: A MATLAB® Framework

Abstract

A Doyle-Fuller-Newman electrochemical battery model implementation in a robust and sleek MATLAB® framework for lithium-ion batteries as an open-access MATLAB code is presented. The Doyle-Fuller-Newman (DFN) model, in the form of partial differential equations, is first numerically discretized then converted to a differential algebraic equation (DAE). The most efficient way to implement a DAE system is through the adoption of standard DAE solvers provided by available commercial software. MATLAB® is a widely used software in the control community, and to the best of our knowledge, its standard solvers have failed to successfully simulate the DFN model when the battery undergoes high C-rates of operations. One critical issue with DFN model simulation is related to the inconsistency of initial conditions. In 2015, a robust single-step iteration-free initialization approach, enabling solving DAE systems using a standard solver using Maple® symbolic environment, was proposed by Lawder et al. A symbolic environment enables direct and efficient derivation of implicit ordinary differential equations from algebraic equations during initialization. We perform the single-step iteration-free initialization approach in MATLAB® environment by adopting the MATLAB® symbolic toolbox and simulate the DFN model with the ode15s solver. This framework allows users to robustly simulate the DFN model and identify model parameters directly after numerical discretization, utilizing a standard solver.