What is the maximum charge speed that a battery can withstand without starting to degrade? What is the best cooling strategy to prevent the localized heating that causes batteries to age faster? How can thermal runaway be better prevented? At the moment, it is difficult to provide answers to these questions, which are crucial to achieving widespread electrification.
Simulations that can be run on a large industrial PC
Tests can be difficult to interpret: there are simply too many interactions between temperature, currents, states of charge, lithiation, aging, etc. While P2D modeling is commonly used in industry, it only covers fragments of the battery. Very expensive computational resources are needed to run P4D (pseudo-4-dimensional) models that can simulate the entire battery. So, these questions continue to go unanswered...
Liten has achieved a major breakthrough by developing a P4D model of a complete 18650 cell, which can reproduce the first five minutes of a 4C fast charge using 16 computing cores over 24 hours. “If you're aiming to conduct an engineering study, a large industrial engineering PC will do the job," says Benoit Mathieu from Liten.
A customized model, specifically designed for batteries
The model was designed with the support of a Paris-based team from CEA's Energy Division. The model runs on the TRUST simulation platform, which uses numerical methods that can faithfully reproduce the coupling of physical effects specific to batteries. “It's custom-designed, in contrast to off-the-shelf commercial software platforms."
The tool is currently in the prototype stage and can produce qualitative results, however, it has already demonstrated impressive capacities. For example, it can determine the electrical potential in the collector of two wound electrodes, at any time during charging. It can also predict the formation and assess the size of lithium deposits on graphite (“lithium plating"). The cell's temperature field is clearly displayed, from the base to the top and from the edges to the center. “In order to validate these models, we need to examine the core of the battery during operation, using cutting-edge Synchrotron X-Ray Tomography techniques," explains Benoit Mathieu.
“Explore a wide range of manufacturing parameters and charge control algorithms"
The model is already sufficiently advanced to address industrial problems with cells. What's more, the team already has plenty of ideas on how it can be improved. Notably, funding has been raised for a thesis on finer mesh resolution and work on it is scheduled to begin in September.
In particular, the plan is to produce a realistic cell charging model, a quantitative simulation of lithium plating, and a specific mesh for the temperature field, as well as to couple the Trust code with a mechanics code to model cell stresses caused by the “breathing" of electrode materials. “We're also going to use supercomputers to produce more accurate simulations, and perform more calculations to explore a wide range of manufacturing parameters and charge control algorithms for batteries," explains Benoit Mathieu.