Assessment of Euler–Lagrange models for fluid-induced cohesive particle deagglomeration using particle-resolved DNS
Publication date
2026-05-12
Document type
Forschungsartikel
Author
Organisational unit
Publisher
Elsevier
Series or journal
Particuology
ISSN
Periodical volume
114
First page
408
Last page
430
Peer-reviewed
✅
Part of the university bibliography
✅
Language
English
Keyword
Particle-laden flows
Particle-resolved DNS
Euler-Lagrange
Effective sphere
Homogeneous isotropic turbulence
Deagglomeration
Abstract
Accurate modeling of agglomerate breakup is essential for predictive simulations of particle-laden turbulent flows. Breakup models applied in the context of agglomerates represented by single spheres often rely on simplifying assumptions about the agglomerate structure and relevant stress mechanisms, raising questions about their fidelity. In this work, the fluid-induced breakup model by Breuer and Khalifa [Powder Technology 348, 105–125 (2019); Computers & Fluids 194, 104315 (2019)] developed for compact, nearly spherical agglomerates is systematically assessed within the Euler–Lagrange framework using high-fidelity reference data from particle-resolved direct numerical simulations coupled with the discrete element method. A single agglomerate composed of 500 primary particles is released into homogeneous isotropic turbulence, with Reynolds number, Hamaker constant, and particle size systematically varied to generate 18 different application cases. Comparisons between the two approaches demonstrate that fragmentation ratios and breakup rates reasonably agree in many cases, both confirming that breakup is augmented by increasing turbulence intensity and is hindered by stronger cohesion. A stress analysis further reveals that the turbulent stress dominates the breakup of large agglomerates, the drag stress acts on intermediate sizes, and the rotary stress mainly disrupts the smallest particle clusters. While the breakup model in the Euler–Lagrange method exhibits characteristic deviations from the resolved data, sudden size drops due to symmetric binary breakup and reduced reagglomeration compared to the gradual erosion in PR-DNS, the overall breakup rates collapse onto a common scaling based on the adhesion number and Reynolds number across both methods. These results highlight that, despite structural simplifications, the Euler–Lagrange breakup model reproduces the essential breakup dynamics observed in PR-DNS at a fraction of the computational costs, making it a practical framework for large-scale simulations of cohesive particle-laden flows.
Description
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Version
Published version
Access right on openHSU
Metadata only access