Study of the tradeoff in coupling multiple electrochemical with a 3D thermal battery model

MUÑOZ, PEDRO MATÍAS; LEVITAN, DAVID; FALAGUERRA, TOMÁS; HUMANA, RITA MARIANGELES; CORREA PERELMUTER, GABRIEL

Workshop; 7 th International Workshop on Lithium, Industrial Minerals and Energy; 2020

The lithium ion batteries are one of the most relevant devices during the transition towards a more sustainable generation and consumption of energy as it is a key component in the electro- mobility field and used as energy reservoirs in solar and wind power plants. Li-ion batteries are electrochemical cells whose performance and cycle life depend on the conditions of operation, one of which is the temperature. To ensure a safe operation, the temperature management of the cells is required. In this context, the development of thermal coupled models becomes important as a nexus between investigation and product development. Lithium-ion battery modelling has been subject of interest in electro-chemistry and engineering for a long time. Thermal models of li-ion batteries are used in product development to design suitable thermal management systems (TMS) for battery packs that ensure proper conditions during battery operation. This model consists in the coupling of an electrochemical (EC) model and a thermal model that work at different scales.While the EC model describes the phenomena at the electrode level, the thermal model deals with the temperature distribution and heat evacuation at the full cell scale.A full cell 3D model for a cylindrical battery that considers the electrochemical and thermalphenomena while modeling the wounded current collectors, electrodes and separator would be extremely expensive from a computational point of view. Different methods are proposed to reduce the model computational time and, at the same time, retain an accurate depiction of the real battery phenomena. The main simplifications applied are (i) modelling the jelly roll as one composite material with averaged properties and (ii) using an homogeneous heat generation profile for the 3D thermal model and an average temperature of the cell for the EC model.In works where an electrochemical (EC) model is coupled with a thermal model, in mostcases, both simplifications are applied [1]. That is, the heat generation from the EC model isaveraged along the electrode length and this heat per unit of volume is considered as ahomogeneous source in the thermal model. Likewise, the temperature from the thermal model is averaged and passed on to the EC model.On the other hand, there are publications that do not use both simplifications. For the firstcategory, Capron et al. [2] consider a homogeneous heat generation from an EC model for two different thermal models, (i) one where the jellyroll is considered as a homogeneous material with averaged properties and (ii) another model where the jelly roll is modelled as discrete layers. While the discrete model results in higher temperatures in the cell core and a greater temperature difference inside the battery, the data from an analytical formula to predict the core temperature from experimental measurements on the battery surface favors the homogenous model as it shows a better agreement with the experimental data.On the second category, while using an homogeneous jelly roll model, Kupper and Bessler[3], propose the coupling of 5 P2D EC models to simulate the battery temperature and heat generation at different depths of a cylindrical 26650 battery, resulting in a 1D+1D+1D, pseudo-3D or P3D model. Tahir [4] follows a similar approach using 6 P2D EC models to describe a heat generation surface for a 2D axisymmetric model of a 26650 battery. Trantel et al. [5] proposes the use of 1300 single particle models (SPMs), which are less accurate than P2D models, distributed along the cross section perpendicular to the cylinder axis in what is called a ?N+1D? model. Those models are used later to study the current distribution inside the spiral wound and the temperature distribution along the cross section. The question remains, what is the tradeoff between the use of these more detailed models and the use of the simplification described previously (homogeneous heat generation profile).In this work, a similar approach to that of Kupper and Bessler [3], and Tahir [4] is used toevaluate the temperature of a 26650 cylindrical battery using different numbers of P2D EC models to generate heat generation profiles for a 2D axisymmetric thermal model. These electrochemical and thermal models are then simulated under different discharge conditions to evaluate thedifference in heat generation, temperature distribution and computational time against a model previously published [5].