Development and parameterization of a control-oriented electrochemical model of lithium-ion batteries for battery-management-systems applications

A precise electrochemical battery model is critical for advanced battery management systems to improve the safety and efficiency of electric vehicles. This paper presents a novel methodology to develop and parameterize the electrochemical model through cell teardown and current/voltage data estimation. The partial differential equations of ionic electrolyte and potential dynamics in the solid and liquid phases are solved and reduced to a low-order system with Pade & PRIME; approximation. The systematic identification procedure is proposed by first dividing the parameters into fixed geometric properties, thermodynamics, and kinetics. Then the cells are dismantled. Subsequent chemical and thermodynamic analyses, including half-cell tests, are conducted for parameter extraction. Next, the parameterized model is validated with extensive experimental data, illustrating the superior capability of predicting cell voltage with root-mean-square errors of 8.90 mV at 2C and 13.98 mV for Urban Dynamometer Driving Schedule profile at 0 ?. The accuracy of the cell internal electrochemical states of the reduced model is verified as well. Comparative studies concerning model accuracy and computation efficiency on hardware reveal that the model is 31% more accurate than equivalent circuit models but occupies similar computation resources. Finally, the need and advantages of combining cell teardown and parameter estimation in achieving a precise electrochemical model are addressed.

Automated summary

This paper presents a novel methodology to develop and parameterize an electrochemical battery model through cell teardown and current/voltage data estimation. The accuracy and computation efficiency of the model are validated using extensive experimental data. The results show that the model can accurately predict cell voltage and demonstrate improved accuracy compared to equivalent circuit models.