Benedikt Alkin †, *, 1, 2 , Tobias Kronlachner †, *, 1, 3, Samuele Papa †, 1, 4, 5,
Stefan Pirker 3, Thomas Lichtenegger 1, 3, Johannes Brandstetter @, 1, 2
† core contributor, * equal contribution 1 NXAI GmbH, Linz, Austria 2 ELLIS Unit Linz, Institute for Machine Learning, JKU Linz, Austria 3 Department of Particulate Flow Modelling, JKU Linz, Austria 4 University of Amsterdam, Amsterdam, Netherlands 5 The Netherlands Cancer Institute, Amsterdam, Netherlands @ Correspondence to: johannes.brandstetter@nx-ai.com
Advancements in computing power have made it possible to numerically simulate large-scale fluid-mechanical and/or particulate systems, many of which are integral to core industrial processes. Among the different numerical methods available, the discrete element method (DEM) provides one of the most accurate representations of a wide range of physical systems involving granular and discontinuous materials. Consequently, DEM has become a widely accepted approach for tackling engineering problems connected to granular flows and powder mechanics. Additionally, DEM can be integrated with grid-based computational fluid dynamics (CFD) methods, enabling the simulation of chemical processes taking place, e.g., in fluidized beds. However, DEM is computationally intensive because of the intrinsic multiscale nature of particulate systems, restricting either the duration of simulations or the number of particles that can be simulated. Moreover, the non-trivial relationship between microscopic DEM and macroscopic material parameters necessitates extensive calibration procedures. Towards this end, NeuralDEM presents a first end-to-end approach to replace slow and computationally demanding numerical DEM routines with fast, adaptable deep learning surrogates. NeuralDEM is capable of picturing long-term transport processes across different regimes using macroscopic observables without any reference to microscopic model parameters. First, NeuralDEM treats the Lagrangian discretization of DEM as an underlying continuous field, while simultaneously modeling macroscopic behavior directly as additional auxiliary fields. Second, NeuralDEM introduces multi-branch neural operators scalable to real-time modeling of industrially-sized scenarios — from slow and pseudo-steady to fast and transient. Such scenarios have previously posed insurmountable challenges for deep learning models. Notably, our largest NeuralDEM model is able to faithfully model coupled CFD-DEM fluidized bed reactors of 160k CFD cells and 500k DEM particles for trajectories of 28s, which amounts to 2800 machine learning timesteps. NeuralDEM will open many new doors to advanced engineering and much faster process cycles.