Digital twin generation for adsorption in porous materials using Stochastic MorphoDeep

Adsorption dynamics in complex porous materials are vital in fields like catalysis and environmental engineering, yet their modeling is hindered by intricate pore morphology and network geometry. Here, we introduce Stochastic MorphoDeep, a digital twin generator that employs stochastic modeling to represent complex microstructures, mathematical morphology to mimic adsorption dynamics, and deep learning to accelerate simulations. By requiring only basic porosity parameters —such as pore volume and surface area—as inputs, Stochastic MorphoDeep establishes a robust framework for generating digital twin microstructures, enabling accurate predictions of adsorption behavior across diverse materials. This model has been applied to platelet-shaped structures but is generalizable to other types of microstructures, provided that a realistic microstructure generation model exists. The model’s performance is validated against experimental data obtained from tailored materials, demonstrating good accuracy in capturing the dynamic and heterogeneous nature of adsorption processes. Adsorption in complex porous materials is critical for applications like catalysis, but modeling it is hindered by structural heterogeneity and computational cost. Here, a digital twin framework called Stochastic MorphoDeep combines stochastic geometry and deep learning to simulate adsorption efficiently.