Analysis of lithium-ion battery thermal models inaccuracy caused by physical properties uncertainty

Abstract

• Uncertainty of thermal conductivity, heat capacity and density is evaluated. • Inaccuracy of thermal model due to uncertainty of physical parameters is studied. • Uncertainty of thermal conductivity shows the largest influence on temperature. • The reliability and accuracy of a 1D model are analyzed. Thermal management is of upmost importance for the safe and efficient operation of lithium-ion batteries in electric vehicles. To this purpose it is required to develop reliable thermal models to assess the behavior of the battery under different operating and ambient conditions. In this work, it is proposed a three-dimensional thermal model of the 40Ah LiFePO 4 /graphite prismatic battery, which is a particular type of the lithium-ion battery (LIB), and it is analyzed the uncertainty related to the base data needed to run the model. The base data comprises, among others, the battery material physical properties and their dependence on the temperature, and the special “double-coated electrodes” structure. The charge and discharge processes of the 40Ah LiFePO 4 /graphite prismatic battery are tested experimentally to verify the reliability of the thermal model. The deviation of the thermal model predictions caused by the uncertainty of the physical parameters of the battery is fully investigated numerically. The influence of physical parameters on the predictions is verified for the battery surface temperature and temperature difference between the battery interior and the surface. For the range of the properties tested the highest deviations of the predicted surface temperature and the inside and surface temperature difference are 0.14 °C and 0.93 °C during discharge at room temperature, respectively. Based on this investigation, it is proposed a simplified one-dimensional thermal model, which has the potential of being an expedite way of calculating the temperature distribution inside most commercial prismatic batteries. The comparison of the temperature distribution predictions using the three-dimensional thermal model and the simplified model indicates that the temperature gradient predictions obtained with the two models are in close agreement; the maximum relative difference of the two models is only 0.5%. The simplified thermal model, in what concerns the uncertainty of the physical properties, may provide an easy-to-use preliminary tool to evaluate the temperature distribution related to prismatic batteries.