Search for: All records

The quantification of microstructural properties to optimize battery design and performance, to maintain product quality, or to track the degradation of LIBs remains expensive and slow when performed through currently used characterization approaches. In this paper, a convolution neural network-based deep learning approach (CNN) is reported to infer electrode microstructural properties from the inexpensive, easy to measure cell voltage versus capacity data. The developed framework combines two CNN models to balance the bias and variance of the overall predictions. As an example application, the method was demonstrated against porous electrode theory-generated voltage versus capacity plots. For the graphite|LiMn$$_2$$${}_2$O$$_4$$${}_4$chemistry, each voltage curve was parameterized as a function of the cathode microstructure tortuosity and area density, delivering CNN predictions of Bruggeman’s exponent and shape factor with 0.97$$R^2$$$R^2$score within 2 s each, enabling to distinguish between different types of particle morphologies, anisotropies, and particle alignments. The developed neural network model can readily accelerate the processing-properties-performance and degradation characteristics of the existing and emerging LIB chemistries.

 

The microstructural optimization of porous lithium ion battery electrodes has traditionally been driven by experimental trial and error efforts, based on anecdotal understanding and intuition, leading to the development of useful but qualitative rules of thumb to guide the design of porous energy storage technology. In this paper, an advanced data-driven framework is presented wherein the effect of experimentally accessible microstructural parameters such as active particle morphology and spacial arrangement, underlying porosity, cell thickness, etc. , on the corresponding macroscopic power and energy density is systematically assessed. For the Li x C 6 | LMO chemistry, an analysis performed on 53 356 battery architectures reported in the literature revealed that for commercial microstructures based on oblate-shaped particles, lightly textured samples deliver higher power and energy density responses as compared to highly textured samples, which suffer from large polarization losses. In contrast, high aspect ratio prolate-shaped particles deliver the highest energy and power density, particularly in the limit of wire-like morphologies. Polyhedra-based colloidal microstructures demonstrate high area densities, and low tortuosities, but provide no appreciable power and energy density benefit over currently manufactured particle morphologies. The developed framework enables to establish general microstructure design guidelines and propose optimal electrode microstructures based on the intended application, given an anode and cathode chemistry.