Mesoscopic modeling and parameter estimation of a lithium-ion battery based on LiFePO4/graphite

• Published in: Journal of power sources 2018-03, Vol.379, p.84-90
• Main Authors: Jokar, Ali, Désilets, Martin, Lacroix, Marcel, Zaghib, Karim
• Format: Article
• Language: English
• Subjects: Cylindrical Li-ion battery; Electrochemical parameters; Inverse method; LiFePO4 (LFP) positive electrode material; Mesoscopic model; Parameter estimation
• ISSN: 0378-7753; 1873-2755
• DOI: 10.1016/j.jpowsour.2018.01.035

A novel numerical model for simulating the behavior of lithium-ion batteries based on LiFePO4(LFP)/graphite is presented. The model is based on the modified Single Particle Model (SPM) coupled to a mesoscopic approach for the LFP electrode. The model comprises one representative spherical particle as the graphite electrode, and N LFP units as the positive electrode. All the SPM equations are retained to model the negative electrode performance. The mesoscopic model rests on non-equilibrium thermodynamic conditions and uses a non-monotonic open circuit potential for each unit. A parameter estimation study is also carried out to identify all the parameters needed for the model. The unknown parameters are the solid diffusion coefficient of the negative electrode (Ds,n), reaction-rate constant of the negative electrode (Kn), negative and positive electrode porosity (εn&εn), initial State-Of-Charge of the negative electrode (SOCn,0), initial partial composition of the LFP units (yk,0), minimum and maximum resistance of the LFP units (Rmin&Rmax), and solution resistance (Rcell). The results show that the mesoscopic model can simulate successfully the electrochemical behavior of lithium-ion batteries at low and high charge/discharge rates. The model also describes adequately the lithiation/delithiation of the LFP particles, however, it is computationally expensive compared to macro-based models. •A mesoscopic approach for LiFePO4(LFP) is considered in the Single Particle Model.•Nine electrochemical parameters are estimated by the inverse method.•The results show a good agreement with the experimental data from 1C to 15C.•A comparison is made between the mesoscopic model and a macro-based model.