UK uses machine learning to study microstructures to design better batteries -Lithium - Ion Battery Equipment
Researchers at Imperial College London have shown how machine learning can help design better-performing lithium-ion and fuel-powered lithium-ion batteries,media reported. The new machine learning algorithm allows researchers to explore possible designs for the microstructure of fuel-powered lithium and lithium-ion batteries, and then run 3D simulation models to help researchers make changes that improve battery performance.
Performance improvements include faster charging of smartphones, extended range of electric vehicles, and increased power of lithium-ion batteries powered by hydrogen fuel in data centers.
Fuel-powered lithium batteries can harness clean hydrogen fuel from sources such as wind and solar to generate heat and electricity, while lithium-ion batteries in smartphones, laptops and electric cars are also a popular form of energy storage. The performance of both is closely related to its microstructure: the shape and arrangement of the small holes inside the battery will affect the energy generated by the fuel-powered lithium battery and the charging and discharging speed of the battery.(Lithium - Ion Battery Equipment)
However, because the pores are so small, on the micron scale, it can be difficult to study the specific shape and size of such pores at a resolution high enough to correlate them with the overall performance of the battery .
Now, researchers at Imperial College have used machine learning techniques to help them virtually explore such pores and run 3D simulation models to predict the battery's performance, based on its microstructure.
Using a novel machine learning technique called "deep convolutional generative adversarial networks" (DC-GANs), based on training data obtained from nanoscale imaging at a synchrotron, a particle accelerator the size of a football field, the researchers Learn to generate 3D image data of battery microstructures.
Andrea Gayon-Lombardo, lead author of the study, from Imperial's Department of Earth Science and Engineering, said: "Our technique allows us to zoom in on batteries and cells to see what characteristics affect overall performance. Developing this image-based machine learning technique It could provide new ways to analyze images of this size."
When running 3D simulation models to predict battery performance, researchers want a data volume large enough to be statistically representative of the entire battery. At present, it is difficult to obtain image data of a large number of microstructures with a high resolution. However, the researchers found that the code could be trained to generate larger datasets with the same properties, or to intentionally generate structures that lead to models with better-performing batteries.
By limiting their algorithm to results that can be produced today, the researchers hope to be able to apply the technique to battery manufacturing to design optimized electrodes for next-generation batteries.
Performance improvements include faster charging of smartphones, extended range of electric vehicles, and increased power of lithium-ion batteries powered by hydrogen fuel in data centers.
Fuel-powered lithium batteries can harness clean hydrogen fuel from sources such as wind and solar to generate heat and electricity, while lithium-ion batteries in smartphones, laptops and electric cars are also a popular form of energy storage. The performance of both is closely related to its microstructure: the shape and arrangement of the small holes inside the battery will affect the energy generated by the fuel-powered lithium battery and the charging and discharging speed of the battery.(Lithium - Ion Battery Equipment)
However, because the pores are so small, on the micron scale, it can be difficult to study the specific shape and size of such pores at a resolution high enough to correlate them with the overall performance of the battery .
Now, researchers at Imperial College have used machine learning techniques to help them virtually explore such pores and run 3D simulation models to predict the battery's performance, based on its microstructure.
Using a novel machine learning technique called "deep convolutional generative adversarial networks" (DC-GANs), based on training data obtained from nanoscale imaging at a synchrotron, a particle accelerator the size of a football field, the researchers Learn to generate 3D image data of battery microstructures.
Andrea Gayon-Lombardo, lead author of the study, from Imperial's Department of Earth Science and Engineering, said: "Our technique allows us to zoom in on batteries and cells to see what characteristics affect overall performance. Developing this image-based machine learning technique It could provide new ways to analyze images of this size."
When running 3D simulation models to predict battery performance, researchers want a data volume large enough to be statistically representative of the entire battery. At present, it is difficult to obtain image data of a large number of microstructures with a high resolution. However, the researchers found that the code could be trained to generate larger datasets with the same properties, or to intentionally generate structures that lead to models with better-performing batteries.
By limiting their algorithm to results that can be produced today, the researchers hope to be able to apply the technique to battery manufacturing to design optimized electrodes for next-generation batteries.
