Thermal Analysis of High-Power Prismatic LiFePO4 batteries: Modeling and Experiment

Monday, 6 October 2014: 16:50
Sunrise, 2nd Floor, Galactic Ballroom 4 (Moon Palace Resort)
M. Yazdanpour, P. Taheri, and M. Bahrami (Laboratory for Alternative Energy Conversion (LAEC), School of Mechatronic Systems Engineering, Simon Fraser University)
A 90Ah prismatic LiFePO4/graphite battery (Sinopoly Battery Ltd., China) is used for the experiments and thermal modeling [cf. Fig. 1]. 

In the experimental study, a power processing system is employed to cycle the experimental battery. Constant-current discharge tests at 1C-rate (90A), 2C-rate (180A), and 3C (270A) are performed to measure the battery voltage and surface temperatures.

In the modelling study, a three-dimensional transient model has been developed to predict the thermal behaviour of the battery during galvanostatic processes. A transient heat conduction equation is sufficient to describe the thermal phenomenon of domain (1); while the heat transfer mechanism in domain (2) is modelled by the natural convection. A convective-radiation boundary condition is considered at the boundaries of the domain.

The above-mentioned models are performed in a finite element solver (COMSOL Multiphysics, Version 4.4), and the accuracy of the model is validated through comparison with the experimental data.

Some Results

The rate of heat generation which is the main parameter calculation in the thermal model, is approximated from [1],

g=I/ν[(V-Voc)+T dVoc/dT]                          (1)

where I and V denote the battery operational voltage and current, respectively, ν represents the battery volume, Voc is the open circuit voltage of the battery and T is temperature. The heat generation terms can be determined by using experimental tests or electrochemical models, which the first approach is applied in this study.  A constant value of -0.4337 mV/K adapted from Ref. [2], is considered for term dVoc/dT. The rate of heat generation caused by the polarization losses, V-Voc, is defined from the measured electrical performance of the battery.

The heat generation rates inside the battery (domain (1)) during three different galvanostatic discharge processes (90 A, 180 A, and 270 A) are calculated based on Eq. (1) and the battery electrical performance, and then the obtained expressions are implemented into the thermal model.

In Fig. 2, the modeling results for average surface temperature of the battery (shown by solid lines) are compared to the corresponding experimental data (symbols), during constant-current discharge processes with 90 A, 180 A, and 270 A. The comparison shows an excellent agreement between the predicted results and the data from the experiments.

In future, an electrical model will be coupled to the presented model to describe the ohmic heating in the electrodes as a result of the electrical constriction resistance. 

[1]         D. Bernardi, J. Electrochem. Soc., Vol. 132, pp. 5-12, (1985).

[2]         A. Samba et al. , J. Electrochimica Acta, Vol. 117, pp. 246–254, (2014).

[3]         P. Taheri et al., J. Electrochem. Soc., Vol. 160, pp. A1731–A1740, (2013).