Coupled Mechanical-Electrochemical-Thermal Modeling of Li-Ion Batteries

Lithium-ion batteries are currently the state-of-the-art power sources for a variety of applications, from consumer electronic devices to electric-drive vehicles.  Being an energized component, failure of the battery which can result in rupture, smoke, fire or venting, is an essential concern.  Existing mathematical models used to understand the crush response of batteries usually treat the cell as a homogenized entity with an average set of properties - which limits the utility of these models in capturing localized failure and has limited value in understanding the onset and propagation of failure [1-2].  Moreover, many of these models consider the mechanical response of the battery exclusively - and do not include the follow-on reactions and/or the energetics.  A third limitation in the homogenized battery models is that an electrical short-circuit is inherently assumed to follow mechanical failure of the cell. For example, if a critical strain value is used as the metric to determine mechanical failure, onset of an electrical short is presumed to follow as soon as any part of the cell crosses this value for the strain.  In these cases, the effects of the local electrical resistance is not considered.

To better understand the abuse response of Li-ion batteries, we recently built a coupled modeling methodology [3] that encompasses the mechanical, electrochemical and thermal (MECT) responses when a cell is subjected to a mechanical load. In our model, each individual component of the cell (i.e., active material coating, separator, current collector) is resolved.  A representative finite element model that enables simulation of localized damage and short-circuit initiation will be presented.  We present individual failure criteria for mechanical, electrical and thermal failure.  A maximum strain criterion is applied to the separator layer to represent the mechanical damage to the cell structure.  Subsequently, an electro-thermal simulation is conducted after the mechanical crush, where a criterion for electrical contact is defined based on the current flowing between the anode and the cathode active material layers, to allow for a direct prediction of the onset of electrical short-circuit.  Different abuse reactions contribute to the evolution of the thermal response - depending on the local current density values and instantaneous temperatures.  A third, thermal failure criterion that takes into account phase change of the components, is used to assess the extent of failure propagation.  Whereas the conventional short-circuit models use arbitrary values for the contact area and short-circuit resistance [4], the short circuit area is computed in our present work as part of the explicit mechanical simulations and our electrical simulations compute the evolution of the locali resistances during a short-circuit.

The results from these simulations allow us to distinguish between load conditions that result in benign thermal behavior and those that result in thermal runaway.  A parameteric sweep that helps us understand the threshold loads the cell can withstand at a given value for the initial state-of-charge and ambient temperature before raising causes for concern will be presented.  Simulation results will be compared against experimental data for different cell types and test configurations.  These comparisons will highlight the advantages of our MECT approach to simulate the abuse response of lithium ion cells, as well as identify the technical gaps in making quantitative interpretations of the experimental data.

References:

1. E. Sahraei, R. Hill, T. Wierzbicki, J. Power Sources 201 (2012) 307-321.

2. L. Greve, C. Fehrenbach, J. Power Sources 214 (2012) 377-385.

3. C. Zhang, S. Santhanagopalan, M.A. Sprague, A. Pesaran, J. Power Sources (2015), http://dx.doi.org/10.1016/j.jpowsour.2015.04.162

4. W.F. Fang, P. Ramadass, Z.M. Zhang, J. Power Sources 248 (2014) 1090-1098.