An optimisation system developed by researchers at the University of Michigan could cut the time for both simulation and physical testing of new lithium-ion batteries by about 75%.
Testing the longevity of new battery designs faster could provide a boost to battery developers searching for the right combination of materials and configurations to increase the range of electric vehicles.
Wei Lu, U-M professor of mechanical engineering and leader of the research team behind the framework, published his work in open access journal Patterns-Cell Press.
Testing configurations— from materials used to the thickness of the electrodes to the size of the particles in the electrode— can usually take several months of cycling the battery up to 1,000 times to mimic a decade of real-time use.
Lu said: “Our approach not only reduces testing time, but it automatically generates better designs.
“We use early feedback to discard unpromising battery configurations rather than cycling them till the end. This is not a simple task since a battery configuration performing mediocrely during early cycles may do well later on, or vice versa.
“We have formulated the early-stopping process systematically and enabled the system to learn from the accumulated data to yield new promising configurations.”
To get a sizable reduction in the time and cost, U-M engineers harnessed the latest in machine learning to create a system halts cycling tests that don’t show promise using mathematical techniques known as Asynchronous Successive Halving Algorithm and Hyperband.
It also takes data from previous tests and suggests new sets of promising parameters to investigate using Tree of Parzen Estimators.
In addition the system generates multiple battery configurations to be tested at the same time, known as asynchronous parallelization— with the algorithm immediately calculating a new configuration following successful or discarded tests without the need to wait for the results of other tests.
U-M’s framework is effective in testing designs of all battery types.
Lu said: “By significantly reducing the testing time, we hope our system can help speed up the development of better batteries, accelerate the adoption or certification of batteries for various applications, and expedite the quantification of model parameters for battery management systems.”
The research was funded by LG Energy Solution.