Control of fluid flow, heat transfer and inclusions in continuous casting: CFD and neural network studies

Petri Väyrynen*, Shenqiang Wang, Jukka Laine, Scppo Louhenkilpi

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review


Fluid flow and heat transfer calculations have been carried out in tundish and mould including different kind of submerged entry nozzles. The effect of different kind of tundish dams have been studied in a bloom tundish in steady state and transient conditions. Many different CFD parameters, like turbulence models and mesh density, were tested during the studies. CFD calculations were also carried out to study the effects of swirling flow inside the SEN as well as different kind of SEN nozzles on mould How phenomena. Different kinds of criteria for the ideal mould flow were derived. Neural network model was developed to predict and control the tundish temperature from process parameters and casting time. The model, which is based on the Bayesian MLP neural network, includes 13 inputs from ladle cycle and the output is the tundish temperature. The results were good; the mean error for temperature was 3.4°C.

Original languageEnglish
Title of host publicationJim Evans Honorary Symposium - Held During TMS 2010 Annual Meeting and Exhibition
Number of pages8
Publication statusPublished - 2010
MoE publication typeA4 Article in a conference publication
EventTMS Annual Meeting and Exhibition - Seattle, United States
Duration: 14 Feb 201018 Feb 2010
Conference number: 139


ConferenceTMS Annual Meeting and Exhibition
Abbreviated titleTMS
CountryUnited States


  • CFD
  • Continuous casting
  • Mathematical modelling
  • Neural network

Fingerprint Dive into the research topics of 'Control of fluid flow, heat transfer and inclusions in continuous casting: CFD and neural network studies'. Together they form a unique fingerprint.

Cite this