Hydrogen degassing in a vacuum arc degasser using a three-phase Eulerian method and discrete population balance model

Faris Karouni, Bradley P. Wynne*, Jesus Talamantes-Silva, Stephen Phillips

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

14 Citations (Scopus)

Abstract

A three-phase Eulerian model incorporating slag-steel interactions to predict the rate of hydrogen removal from molten steel in a full-scale industrial vacuum arc degasser (VAD) has been developed. The interfacial area for hydrogen transfer is calculated using a bubble population balance model. This accounts for bubble growth due to changes in hydrostatic pressure as well as coalescence and breakup. The predicted velocity field and bubble distribution are compared with experimental data in the literature. The bubble size predictions under atmospheric pressure conditions are sensitive to the initial bubble size, while under vacuum conditions they are relatively independent of the initial size. The omission of the slag layer from the model results in a 12% increase in the hydrogen removal rate. Variation in the slag eye diameter as a function of argon flowrate is simulated and compared with empirical correlations in the literature. Hydrogen measurements from a full-scale VAD unit at Sheffield Forgemasters International Ltd. steelworks are compared to the model predictions for a series of melts of varying initial hydrogen content. Based on the initial hydrogen content of the liquid steel, the model predicts the amount of hydrogen removed to within ±20% of final experimental measurements obtained from the melt.

Original languageEnglish
Article number1700550
Number of pages11
JournalSteel Research International
Volume89
Issue number5
Early online date6 Apr 2018
DOIs
Publication statusPublished - 3 May 2018

Keywords

  • CFD
  • degassing
  • eulerian
  • hydrogen
  • PBM
  • slag

Fingerprint

Dive into the research topics of 'Hydrogen degassing in a vacuum arc degasser using a three-phase Eulerian method and discrete population balance model'. Together they form a unique fingerprint.

Cite this