Error estimation of Bernardi heat evaluation of Li-ion cells under drive cycle operation

Accurate estimation of heat generation is critical for EV batteries. Conventional means of heat estimation (i.e. Bernardi equation) works only for constant current discharge. However, fluctuating current condition is the normal mode of operation for cells in EV applications. The current study attempts a comprehensive investigation of heat generation and temperature evolution for a cylindrical lithium iron phosphate cell under standard drive cycles (FUDS, US06). An experimentally parameterized electrochemical-heat transfer model is developed. Experimental voltage curve of FUDS drive cycle is used to obtain unknown electrochemical parameters while its fidelity is tested against US06. It is observed that Bernardi method overpredicts heat generation rate by approximately 25% under drive cycle operation. It is noted that Bernardi method is an indirect method of heat estimation based majorly on electrical parameters. It is also observed that the reversible part of Bernardi heat is negligible under drive cycle operation. Moreover, SOC and temperature evolution are accurately estimated by the electrochemical-heat transfer model - errors lie within 7% and 2 degrees C, respectively. Thus, the present study demonstrates the capability of the electrochemical-heat transfer model for estimating heat generation rate of Liion cells under drive cycle operation where Bernardi equation is shown to be an inadequate method.