Computer-Aided Design Speeds Development of Safe, Affordable, and Efficient Batteries

Storage - Mar 22, 2017

NREL energy storage researchers Kandler Smith and Chuanbo Yang, along with postdoctoral researcher Francois Usseglio-Viretta, evaluate a simulation still of Li-ion battery components in the Energy Systems Integration Facility’s 3-D visualization room. Photo by Dennis Schroeder, NREL 41705

Computer-aided simulation speeds the development cycle for just about any technology, from robot-controlled manufacturing to landing a spacecraft on the surface of Mars. Computer-aided simulation often offers the most productive and affordable way to prototype an evolving technology, including safe, affordable, and efficient plug-in electric vehicle (PEV) batteries.

Related Research on
Battery Recycling Market - Global Forecast to 2022

Over the past seven years, the Energy Department's Vehicle Technologies Office (VTO) has sponsored the National Renewable Energy Laboratory (NREL), four other national laboratories, seven industry partners, and four research institutions to develop cutting-edge battery simulation tools under the Computer-Aided Engineering for Electric-Drive Vehicle Batteries (CAEBAT) project. CAEBAT's adaptable battery simulation models have been added to commercial software packages and are already in use by many battery developers and automakers.

Limited driving range and high costs are two of the key market challenges that have prevented more widespread adoption of PEVs. CAEBAT tools help improve the performance, driving range, safety, and lifespan of PEV batteries, while reducing cost and the number of build-break-test cycles required for design improvement.

Lithium-ion (Li-ion) batteries, the industry standard due to their light weight and high storage capacity, are widely used in applications that range from everyday electronics to airplanes, spacecraft, and PEVs. Yet as recent recalls of products such as hoverboard toys and cell phones show, Li-ion batteries can, on rare occasions, overheat and lead to safety hazards. Considering this and the complexity of significantly larger PEV batteries, understanding the interactions among electrochemical, thermal, and mechanical physics across a range of length scales is crucial for better Li-ion batteries. That process, however, can be expensive—and predicting such interactions makes one of the most dynamic cases for accurate computer simulation. Project work under CAEBAT addresses these very issues. 

"The CAEBAT tools we've developed rely on a range of multi-physics engineering approaches to optimize battery electrical, chemical, thermal, and mechanical response for the real world," said Kandler Smith, CAEBAT task leader with NREL's Transportation and Hydrogen Systems Center. "Models developed at NREL are being incorporated into research and industry software tools to help simplify and speed the design and validation process for batteries."

NREL's energy storage team has led CAEBAT efforts since the project's inception in 2010. Through a number of competitive partnerships, CAEBAT teams—including industry partners ANSYS, CD-adapco, and EC Power—have developed the software tools necessary to help battery designers and manufacturers create affordable, high-performance Li-ion batteries for next-generation PEVs. Their software packages are now used by most major battery and automotive companies throughout the world for electrochemical and thermal design of battery cells and packs.

"Ten years ago, there were only basic physical models for batteries that could predict one aspect of battery behavior at a time," explained Smith, "and these were for very small battery cells and packs."

NREL's battery modeling team started looking at ways to create a more flexible model. They were interested in predicting battery behavior at larger scales, under a wider variety of performance and abuse circumstances, for the safe and reliable integration of Li-ion batteries.

NREL's development of the multi-scale multi-domain (MSMD) model—a modular, multi-physics framework that has since served as the foundational underpinning of CAEBAT—provided a basis for commercial software companies to use and incorporate into programs for battery design. Coupled with industry's interest in faster methods to develop vehicle batteries, the MSMD model launched what would be an ongoing multi-laboratory, multi-industry, and multi-year project.

CAEBAT's success is driven by the collaborative nature of the work. Many players, including the NREL-led consortium of partners , have contributed to and benefited from CAEBAT. Argonne National Laboratory (ANL) and Sandia National Laboratories are working with NREL to provide experimental data and collaborate on simulations. Industry partners Ford and General Motors have extended the models to their vehicle prototyping processes. ANSYS, CD-adapco, and EC Power developed adaptable tools that many battery and vehicle companies are using today. Additionally, universities such as the Massachusetts Institute of Technology, Pennsylvania State University, Texas A&M University, and University College London, along with the U.S. Council for Automotive Research, have partnered with the laboratory on a variety of CAEBAT projects. Oak Ridge National Laboratory, which is leading its own CAEBAT project, is also collaborating with NREL.

"With laboratory, industry, and university partners, NREL has contributed its technical expertise in battery mechanics and materials, as well as computation and modeling to effectively achieve the VTO vision of developing computer-aided engineering tools that accelerate the design of PEV batteries," said Ahmad Pesaran, NREL manager for the CAEBAT program, currently on detail with the DOE's Vehicle Technologies Office "At the same time, we've been able to tap others to help extend and validate the models."

Ahmad Pesaran leads the CAEBAT-III kickoff meeting with partners from industry, academia, national laboratories, and other research institutions. Photo by Dennis Schroeder, NREL 35959.

CAEBAT is now in its third phase, and all of the work leading up to today has made it possible for researchers to achieve highly nuanced capabilities for accurate battery simulation. While CAEBAT-I focused on harnessing the MSMD model to create software tools for cell and battery design, CAEBAT-II centered on safety modeling while improving the MSMD model's speed and efficiency—which led to the development of the advanced GH-MSMD model. The GH-MSMD model speeds computational time by a factor of 100 to 1,000, providing the ability to run a 20-minute driving profile simulation in less than a few seconds. The GH-MSMD model is now being incorporated into ANSYS's Fluent 17 software package.

Today, with a robust suite of advanced electrochemical modeling tools and extremely fast simulation capabilities, project work under CAEBAT is expanding to the microstructure level to better understand the impact that material recipe and manufacturing controls have on battery electrodes. As this body of work evolves, the energy storage team is working closely with NREL's materials science researchers for insight on electrodes, such as the Li-ion silicon anode, while collaborating with the laboratory's computational science researchers to continually develop new algorithms and codes for particle-scale simulations of active battery materials.

CAEBAT microstructure applications involve the simulation of physics at the electrode microstructure level. Seen above, researchers calculated the pore size of an electrode's microstructure geometry (right), as well as the lithium displacement within an electrode to evaluate the difficulty of movement (left), both scaled with color. Images created by Francois Usseglio-Viretta, NREL

"Our microstructure modeling efforts are exploring the interaction of battery physics with electrode morphology—how all the different structures and particle shapes are held together in a conductive matrix, and how these geometries affect battery performance so that manufacturers can improve designs from the material recipe up," said Smith.

In collaboration with Texas A&M, NREL prototyped software with a scalable code to simulate battery electrochemical reactions on microstructure particles. ANL is validating the models by providing electrode samples and electrochemical measurements—and scanning their 3-D geometries with its Advanced Photon Source, which uses X-rays to develop a 3-D map of the electrode and pore structure. The computationally complex simulations are conducted using NREL's high-performance supercomputer, Peregrine.

Advancements in crash test simulations from CAEBAT are already providing automakers with faster access to information when any changes are made to a vehicle's components or battery design. By offering the NREL models to industry partners, researchers have received valuable data to continually improve and validate their safety models.

"One of the goals of our safety modeling work is to enable the industry to reduce the number of physical crash tests done for batteries and vehicles," said Shriram Santhanagopalan, NREL senior energy storage engineer whose work centers on battery safety modeling. He explained that, 20 years ago, automakers used to test multiple vehicles for every test condition—crashing 10 or 20 replicates of the vehicle before approving it for commercial use. "Now with crash simulation software, such as LS-Dyna, carmakers probably test three or four vehicles," he said. "That level of confidence, and significant savings in cost, time, and energy, came from years of testing and simulating the crash response of vehicles."

NREL's battery safety modeling work involves the coupling of CAEBAT-developed electrochemical, thermal, and chemical models with LS Dyna mechanical crash simulations in order to develop highly accurate models that simulate the impact of a vehicle crash on a vehicle's battery—and how that battery reacts. These kinds of simulations for PEV prototyping save a significant amount of time and resources for automakers, which in turn translates into reduced costs for consumers.

"Understanding the interplay of a battery's chemical and mechanical reactions in a crash event is crucial in order to control such reactions. The models provide much greater engineering precision beyond the prevention of a mechanical breach to the battery pack's hardware," said Santhanagopalan.

Interns Julia Hartig (left) and Zenan Wu (right) set up mechanical and electrochemical characterization tests of battery cells in order to validate and improve existing battery predictive models under CAEBAT. Photo by Dennis Schroeder, NREL 41483

CAEBAT teams are now working to efficiently integrate mechanical-electro-chemical-thermal models, predict the behavior of battery packs and cells that experience overheating from mechanical crush, and evaluate the impact of microstructure geometries on battery electrode behavior. Such developments are bridging CAEBAT's existing work on performance, lifespan, and safety modeling to develop a more complete suite of tools. This will help manufacturers draw on predictions at the microstructure level to influence predictions at the vehicle level, and vice versa.

"We've reached a point where we're getting nanoscale profiles of silicon in a battery's material recipe that, when integrated into models, can lead to better predictions of PEV driving range, or how a vehicle's battery will respond in a crash," said Matthew Keyser, NREL senior engineer and acting group manager of the energy storage team. "As we gain more and more data through CAEBAT collaborations, the models will only become more comprehensive and accurate."

Moving forward, CAEBAT teams plan to conduct microstructure evaluation, combine multi-domain models, and continue addressing today's key PEV-market issues from a battery's electrode makeup to the vehicle it powers.

Source : National Renewable Energy Laboratory (NREL)

Published on Global Energy World: Mar 22, 2017