RT info:eu-repo/semantics/article
T1 Quantitative analysis of the prediction performance of a Convolutional Neural Network evaluating the surface elastic energy of a strained film
A1 Martín Encinar, Luis
A1 Lanzoni, Daniele
A1 Fantasia, Andrea
A1 Rovaris, Fabrizio
A1 Bergamaschini, Roberto
A1 Montalenti, Francesco
K1 Neural Network
K1 Energy density
K1 Red neuronal
K1 Energía elástica
K1 3312 Tecnología de Materiales
AB A Deep Learning approach is devised to estimate the elastic energy density at the free surface of an undulated stressed film. About 190000 arbitrary surface profiles are randomly generated by Perlin noise and paired with the corresponding elastic energy density profiles , computed by a semi-analytical Green’s function approximation, suitable for small-slope morphologies. The resulting dataset and smaller subsets of it are used for the training of a Fully Convolutional Neural Network. The trained models are shown to return quantitative predictions of , not only in terms of convergence of the loss function during training, but also in validation and testing, with better results in the case of the larger dataset. Extensive tests are performed to assess the generalization capability of the Neural Network model when applied to profiles with localized features or assigned geometries not included in the original dataset. Moreover, its possible exploitation on domain sizes beyond the one used in the training is also analyzed in-depth. The conditions providing a one-to-one reproduction of the “ground-truth” profiles computed by the Green’s approximation are highlighted along with critical cases. The accuracy and robustness of the deep-learned are further demonstrated in the time-integration of surface evolution problems described by simple partial differential equations of evaporation/condensation and surface diffusion.
PB Elsevier
SN 0927-0256
YR 2025
FD 2025
LK https://uvadoc.uva.es/handle/10324/77889
UL https://uvadoc.uva.es/handle/10324/77889
LA eng
NO Computational Materials Science, Volume 249, 2025, 113657
NO Producción Científica
DS UVaDOC
RD 09-oct-2025