CFD-Simulation of an Industrial Furnace in the Hot-dip Galvanization

Process

Christoph Triebl a,*, Christoph Spijker a, Harald Raupenstrauch a, Alexander Jarosik b, Gerhard Angeli b a Chair of Thermal Processing Technology, Montanuniversitaet Leoben, Leoben, Austria

b voestalpine Stahl GmbH, Linz, Austria

* Corresponding author: Christoph Triebl, Franz-Josef-Strasse 18, 8700 Leoben, Austria
E-Mail: christoph.triebl@unileoben.ac.at

Abstract

Zinc coatings are widely spread for protecting steel against corrosion in many different applications. The commercially most important method for producing zinc coatings is hot-dip galvanization. State of the art in continuous processes is invented by Sendzimir, at which pre-heating of the strip is implemented prior dipping into the liquid zinc bath. The surface of the steel strip is first cleaned by removing oil residues in a direct-fired furnace by flash evaporation due to the fast temperature increasing. After degreasing the strip is passed through an annealing furnace for reducing the surface to eliminate oxides and improve the surface quality for the galvanization. Although the Sendzimir- process is more than 70 years old, detailed conditions in the pre-heating furnaces are not determined and potential for optimization still exists. [1, 2]
To investigate the operation conditions of a direct-fired furnace in a hot-dip galvanizing process, the furnace was modeled by using computational fluid dynamics (CFD). The commercial software ANSYS Fluent 15.0 was chosen to model the turbulent combustion in the direct-fired furnace. Since free jets have to be simulated, the realizable-k-ε-model is used for modeling turbulence. In the furnace, non- premixed burners are installed, which are fed by pre-heated air and natural gas. In modeling non- premixed combustions, convective and diffusive time scales are of the same magnitude, but the chemical time scale is smaller by several orders of magnitude. For this reason, fast chemistry and subsequently local chemical equilibrium are assumed [3, 4]. Simplifying the model by introducing fast chemistry, equal diffusivities, isobaric, and adiabatic conditions, the Equilibrium model is chosen for simulating combustion in the direct-fired furnace [5]. Due to the fast chemistry, reaction takes place in asymptotically thin layers, which leads to a one-dimensional combustion normal to these layers, at which only the transport equation for the mixture fraction Z has to be solved [3, 4]. The fluctuation of Z caused by turbulence is taken into account by a probability density function (PDF) [5]. For simulating radiation, the discrete ordinate method (DOM) is used, which models scattering, diffusive reflection, as well as refraction and can be adopted in complex geometries, such as furnaces [6].
The results of the simulation show that the complex flow conditions and high turbulence in the direct fired furnace are caused by the asymmetric arrangement of the burners. The path lines of the burners are influenced by the inert gas flow and the shape of the ceiling of the furnace. Different inert gas flow rates above and below the steel strip results in different concentrations of the components of the flue gases in the upper and the lower part of the direct-fired furnace. Due to the high temperature of more than 1500 K, the steel strip heats up immediately after entering the furnace via radiation, which plays a major role in the heat treatment of the strip.

References

1. A. R. Marder, The metallurgy of zinc-coated steel, Progress in Materials Science, Volume 45, Issue 3, 2000, p. 191-271
2. T. Sendzimir, Process for coating metallic objects with layers of other metals, 1938, U.S.
Patent No. 2,110,893
3. N. Peters, Laminar diffusion flamelet models in non-premixed turbulent combustion, Progress in Energy and Combustion Science, Volume 10, Issue 3, 1984, p. 319-339
4. N. Peters, Laminar flamelet concepts in turbulent combustion, Symposium (International) on
Combustion, Volume 21, Issue 1, 1988, p. 1231-1250
5. G. M. Goldin, A priori investigation of the constructed PDF model, Proceedings of the
Combustion Institute, Volume 30, Issue 1, 2005, p. 785-792
6. S. T. Thynell, Discrete-ordinates method in radiative heat transfer, International Journal of
Engineering Science, Volume 36, Issues 12â??14, 1998, p. 1651-1675