Advanced Computing in the Age of AI|Wednesday, September 30, 2020
  • Subscribe to EnterpriseAI Weekly Updates:  Subscribe by email

Tackling the Complex and Challenging World of Automotive Battery Design 

<img style="float: left;" src="" alt="" width="95" height="53" />Researchers developing low-cost, long-lived and safe lithium-ion batteries for hybrid and electric vehicles are enlisting high performance computing (HPC) modeling and simulation capabilities to unravel the complexities of this important technology. Dr. Robert Spotnitz brings us up to date.

Designing automotive batteries is not for the faint of heart. This is a complex business and only in its infancy – especially when it comes to using advanced digital manufacturing processes such as modeling and simulation to explore the ins and outs of battery behavior.

One of the leaders in the drive to create better batteries for the automotive industry is Dr. Robert Spotnitz, president of Battery Design LLC, a leader in battery design and modeling software. Spotnitz has been at this for several decades and is deep into the complexities that make battery design so challenging. In fact, "challenge" is a word he uses frequently; it's one of the defining characteristics of this manufacturing sub-sector.

In this case, the challenge is to develop low-cost, long-lived, safe, reliable batteries – primarily based on lithium-ion – for hybrid, plug-in hybrids, and electric vehicles (EVs). Judging between the various battery models under development requires a fundamental understanding of their operation and the response of the battery's component parts to varying operating and environmental conditions, as well as the subtle differences occasioned by chemical or geometry design changes.

In the business of batteries, high performance computing and modeling and simulation are just being introduced. Historically, Spotnitz says, battery companies have experts on staff who build and test physical prototypes repeatedly. This is fine for small batteries, but not practical when it comes to the larger automotive versions.

"It may take several years of testing on expensive equipment in order to evaluate battery performance," he notes. "As a result, each OEM tends to keep that information secret."

For example, say GM qualifies a LG Chem cell for the Volt and now Ford wants to use that same cell in one of their vehicles. But GM will not share its proprietary test results, so Ford will have to repeat those same costly and arduous tests to have confidence that the cell will work in its vehicles.

However, if Ford had simulation software that accurately predicted battery performance under a wide variety of conditions, including the capture of internal gradients within a simulated battery cell, the company could not only see immediately if a particular design works, but what happens as the engineers tweak the design parameters.

This is precisely what is beginning to occur. Battery Design Studio, developed by Spotnitz and distributed by CD-adapco, is a unique tool that allows engineers to design lithium-ion batteries, simulate their performance, and analyze the data. It includes both finite element method (FVM) and computational fluid design (CFD) capabilities.

No Two Batteries are Alike
It also addresses another of the many challenges confronting battery designers – each battery is different and has a lifecycle all its own. Spotnitz explains that the materials stored in the cell – for example, a pack of batteries – can be exposed to different stressors such as high temperature. The battery will "remember" that experience, which makes it different from all other cells until it dies. "These are not like other devices where, once you know their state, you can predict what happens next," he says. "With each battery, you have to know their history."
The challenges are worth the prize. Batteries have the potential to deliver up to six kilowatt-hours per kilogram, making them competitive with gasoline. However, most of today's batteries only deliver 250 watt-hours per kilogram which is nowhere near what lithium-ion technology is capable of. Another big hurdle – range. Unless the battery can deliver 200-300 miles, it won't be competitive. Again, the technology will support that kind of mileage; what's needed, says Spotnitz is the innovative design to achieve those goals.

Today, with the application of the modeling and simulation software, the OEMs are able to design batteries with built in controls that prevent overcharge and overdischarge. The development of even more complex models designed to extend battery life are being hampered by a lack of understanding of what causes batteries to fade over time. "This is a case where our modeling capabilities have exceeded our understanding of the physics," Spotnitz says.

Simulation that is moving ahead is exemplified by work at MIT where researchers are simulating the effect of crush on big batteries. This approach, which is a lot more cost effective than testing physical prototypes, uses CFD to see how heat is dissipated and FEA to understand the mechanical deformations involved.

Thermal management of the battery packs is one of the major thrusts of battery design today. Questions such as how much coolant is required, the design of coolant plates, spacing of the cells, the use of heat spreaders, etc., are all being simulated to design the optimum battery pack. This within a variety of cell designs including cylindrical, prismatic and stacked plate cells. To make life even more complicated, the automakers have issued their own challenge to the battery designers – they want the simulations accurate enough so they can predict battery performance within two degrees.

Small is Beautiful
One of the exciting frontiers of battery design now being explored is occurring at the molecular level. All manner of different particles are being considered to improve the energy density of the batteries, including nanoparticulates. "This is an example of where chemistry has surpassed our modeling capability," comments Spotnitz.

The chemists can now perform atomic layer deposition of thin films that are stable at high voltages, as well as construct core-shell structures, and use nanoparticles to mae highly intricate agglomerates of particles. The challenge confronting Spotnitz and other researchers is how to devise better modeling algorithms to predict how these particles will behave when they are subjected to charge and discharge.

Another material of great interest is nanostructured silicon which holds promise for advances in high capacity anodes in lithium batteries. However, silicon tends to crack when cycled. Predicting silicon product morphology upon cracking is another in the long list of jobs to tackle using HPC capabilities – in this case, employing FEA to develop the simulations.

In addition, Spotnitz and other researchers are looking into various aspects of material design, such as the atomic orientation of crystals. The way a crystal is exposed to an electrolyte will cause it to charge and discharge at different rates. He notes, "Being able to account for the crystal's orientation is important so we can predict how the crystal will behave in a battery. The real challenge in battery technology is to connect the molecular level with the macro level. Fortunately we have a lot of infrastructure in place to accomplish this."

This includes models based on object oriented programs that can take lower level battery physics and incorporate it into higher level physics that provide insights into battery behavior that were previously unavailable. In addition, CD-adapco has created full-vehicle simulations that can accept various battery pack models – when an improvement is introduced into the pack model, the simulations shows the impact of the change on the entire vehicle.

Spotnitz says that CAE tools are available today for: unit cell design at the particle level; cell design including the impact of such factors as current collectors and heat generation on functionality; and module/pack design that includes its impact on system power and thermal behavior.

In the near future he and his colleagues will be simulating cell pack abuse including crush in order to improve overall vehicle and passenger safety. Also coming up are more investigations into particle design to increase the energy, power and life of tomorrow's automotive batteries.

For Spotnitz and other researchers involved in the complex world of lithium ion batteries, there is no shortage of challenges.

Add a Comment

Do NOT follow this link or you will be banned from the site!
Share This