Abstract
Porous transport layers, sintered from sphere-based titanium material, can be used in electrolyzers. A sedimentation-based stochastic sphere model was used to create many versions of a sphere-based micro-structure. Aided by Lattice-Boltzmann simulations, the permeability was calculated based on the artificial micro-structures. These data were then used to train a convolutional neural network in order to characterize material on the basis of its micro-structure. Binary image series from the real material of porous transport layers were characterized using a deep learning model trained with artificial data. A reasonable degree of accuracy was achieved by the neural network in the prediction of the through-plane permeability of the real material, when the samples were manufactured from spherical particles.Graphic
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call