An Experimentally-Verified Thermal-Electrochemical Simulation Model of a 21700 Cell Using a Lumped Semi-Empirical Battery Model

TY - CHAP

T1 - An Experimentally-Verified Thermal-Electrochemical Simulation Model of a 21700 Cell Using a Lumped Semi-Empirical Battery Model

AU - Sarmadian, Alireza

AU - Yu, Yifei

AU - Marco, James

AU - Shollock, Barbara

AU - Restuccia, Francesco

PY - 2022/8/8

Y1 - 2022/8/8

N2 - This paper addresses, for the first time, lumped electrochemical-thermal coupled model requirements for the core temperature prediction of a cylindrical 21700 cell. This approach reduces the total number of necessary input parameters in comparison with fully physics-based models. A Newman P2D or a Newman simplified electrochemical model of Single Particle (SP) require approximately 40 and 20 parameters. By contrast, the lumped semi-empirical electrochemical-thermal model approach uses 3 fitting and 10 input parameters, simplifying the detailed knowledge required and also enabling the reduction of the load on the battery management system. COMSOL Multiphysics has been used to achieve a coupled electrochemical-thermal model which is both accurate and with fast response (computation time: ~18 seconds) in the prediction of 21700 NMC/Si-Graphite cell potentials and core temperatures. The fitting parameters of dimensionless charge exchange, the ohmic overpotential at 1C, and the diffusion time for the calibrated model were obtained using the experimental battery data at 1C. The model is validated against a different set of data for discharge tests at 0.3 and 0.7 C-rates. These include experimental cell potential curves and core temperature measurements by using a fibre optic sensing system. Results showed that the standard deviations obtained for the cell voltage and the core temperature are respectively 0.12 V and 0.8 K.

AB - This paper addresses, for the first time, lumped electrochemical-thermal coupled model requirements for the core temperature prediction of a cylindrical 21700 cell. This approach reduces the total number of necessary input parameters in comparison with fully physics-based models. A Newman P2D or a Newman simplified electrochemical model of Single Particle (SP) require approximately 40 and 20 parameters. By contrast, the lumped semi-empirical electrochemical-thermal model approach uses 3 fitting and 10 input parameters, simplifying the detailed knowledge required and also enabling the reduction of the load on the battery management system. COMSOL Multiphysics has been used to achieve a coupled electrochemical-thermal model which is both accurate and with fast response (computation time: ~18 seconds) in the prediction of 21700 NMC/Si-Graphite cell potentials and core temperatures. The fitting parameters of dimensionless charge exchange, the ohmic overpotential at 1C, and the diffusion time for the calibrated model were obtained using the experimental battery data at 1C. The model is validated against a different set of data for discharge tests at 0.3 and 0.7 C-rates. These include experimental cell potential curves and core temperature measurements by using a fibre optic sensing system. Results showed that the standard deviations obtained for the cell voltage and the core temperature are respectively 0.12 V and 0.8 K.

UR - https://hefat2022.org

M3 - Conference paper

SN - 978-0-7972-1886-4

SP - 128

EP - 133

BT - 16th International Conference on Heat Transfer, Fluid Mechanics and Thermodynamics (HEFAT2022)

A2 - Meyer, JP

PB - HEFAT

ER -