Numerical Investigation of Slag Flow Through a Coke Funnel Analog and Packed Bed

Dong, Xue Feng; Jayasekara, Apsara S.; Sert, D.; Ferreira, Rui B.; Gardin, Pascal; Monaghan, Brian J; Chew, Sheng; Pinson, David; Zulli, Paul

doi:10.1007/s11663-021-02260-0

Cited by 4 publications

(1 citation statement)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Jing et al [ 16 ] predicted an optimum hydrogen enrichment at a minimum coke rate with the support of a 3D process blast furnace model. Similarly, a CFD approach was applied by Dong et al [ 17 ] to investigate the flow‐ability of liquid slag through the voids of a packed bed and were able to give some guidance on the understanding of flowing slag dependent on wettability, contact angle, and surface tension.…”

Section: Introductionmentioning

confidence: 99%

The First Unified and Transient Modeling Platform to Build a Digital Twin of Blast Furnaces Based on the Extended Discrete Element Method

Peters

2023

steel research int.

View full text Add to dashboard Cite

The current research addresses a unified and integrated Euler–Lagrange simulation platform based on the extended discrete element method that advances significantly the digital twin technology applied to blast furnaces. It includes a particulate phase characterized by its dynamic and thermodynamic state that is coupled to multiphase computational fluid dynamics (CFD) by heat, mass, and momentum transfer. Thus, the framework deals with transient and reacting 3D granular and multiphase flows. The latter includes all relevant processes taking place in a blast furnace: formation of a reducing agent, e.g., carbon monoxide, reduction of iron‐bearing material, and softening and melting with the subsequent transfer into the corresponding phase of the multiphase CFD solver and thus, identifying unambiguously the cohesive zone. Validation is carried out on different length scales from a particle level to the global dimensions of a blast furnace and shows a good agreement between predictions and experimental data. Applying fast and innovative algorithms allows for moderate computational times, i.e., 1000 s real‐time takes ≈2.3 d on an Intel(R) Xenon(R) Silver 4114 CPU @ 2.2 GHz processor with 40 cores and the entire domain divides into eight partitions.

show abstract

Section: Introductionmentioning

confidence: 99%