|Return To Program Contents Page|
Session Chairperson: Einar K. Jensen, Elkem a/s Research, N-4602 Kristiansand S., Norway
MATHEMATICAL MODELLING OF THE METAL FLOW IN SIDEWELL FURNACES: Y.S. Kocaefe, R.T. Bui, D. Kocaefe, Department of Applied Sciences, University of Quebec at Chicoutimi 555, Boul. de l´Université, Chicoutimi, Quebec, Canada G7H 2B1
Many different types of furnaces are used in aluminum industry for melting and treating the metal. Sidewell furnaces are commonly used for scrap melting. Metal flow in the furnace is provided by different types of equipment such as pumps and impellers, and plays an important role in furnace performance. A three-dimensional model has been developed to simulate the metal flow in such furnaces. In this paper, the model will be described, and the results will be presented showing how various operating and design parameters such as impeller position and rotational speed, sidewell geometry and arch design affect the metal flow.
Na AND Ca PICK-UP FROM HALL BATH IN INGOT FURNACES: David H.DeYoung, Aluminum Company of America, Molten Metal Processing Center, Alcoa Technical Center, Alcoa Center, PA 15069
A series of laboratory tests were conducted to explore the reaction of bath with Al-Mg alloys, as may occur when Hall bath enters ingot furnaces in a smelter-supplied ingot plant. Results showed that Mg clearly reacts with bath and adds Na and Ca to the metal by the reactions: 2(Na3AlF6) + 3 Mg 6 Na + 2 AlF3 + 3 MgF2 (1) and CaF2 + Mg Ca + MgF2 (2) Temperature, Mg concentration in the alloy, the quantity of bath added to the metal, and the particle size of the bath added all influenced the reaction. At 700°C the Na and Ca pick-ups that occurred from these reactions were approximately 15 ppm while at 900°C the pick-ups were greater than 300 ppm. The pick-ups increased with increasing quantity of bath up to a limiting amount, beyond which no further pick-up occurred. The variation of the reaction with particle size indicated that the reaction is limited by the available bath-metal surface area.
EFFECTS OF SOLUTE INTERACTIONS ON GRAIN REFINEMENT OF COMMERCIAL ALUMINIUM ALLOYS: M.A. Kearns, P.S. Cooper, London & Scandinavian Metallurgical Co. Limited, Fullerton Road, Rotherham, South Yorkshire, S60 1DL, England
An analysis of the effects of alloy additions on the grain refinement in a series of model and commercial aluminium alloy compositions is reported. The data of Birch and Fisher published earlier describing grain refinement in 32 systems, including AA3004, AA5083, AA6063 and AA7050, are considered in terms of the supercooling effect of each alloy addition. A simple model describing grain size in terms of additive supercooling effects of individual alloy additions is proposed which fits the data reasonably well. Deviations from simple additive behaviour are evident in systems where strong intermetallic interactions occur. Interaction coefficients amongst solutes are invoked to explain these deviations. Individual alloying effects on grain refinement are treated in terms of constitutional supercooling parameters and behaviour is shown to be similar to that described in earlier studies. Zirconium is shown to have a general poisoning effect in a range of alloy compositions.
THE BUBBLE SIZE AND MASS TRANSFER MECHANISMS IN ROTOR STIRRED REACTORS: Stein Tore Johansen, Svend Grådahl, Per Ola Grøntvedt, Pål Tetlie, Rune Gammelsæter, SINTEF Materials Technology, N-7034 Trondheim, Norway; Karl Venås, Pal Skaret, Karl Venås a.s., Trondheim, Norway; Erling Myrbostad, Bjørn Rasch, Hydro Aluminium a.s R&D Centre, N-6600 Sunndalsøra, Norway
The performance characteristics of the HYCAST Rotor has been investigated in a water model. The removal rate of oxygen from water is measured for a large number of operational conditions. By combining different techniques the energy dissipation in the reactor, the mass transfer at the top surface and the bubble sizes may be determined experimentally. Investigations of the bubble dynamics by high speed video show that the bubbles in this type of reactors are very different from what is hitherto believed. Based on such a picture the paper demonstrates that the mass transfer and bubble sizes in the HYCAST Unit may be explained by a relatively simple theory which seems to explain all available experimental data. It is also demonstrated how we from this theory can scale from the oxygen removal in water model to hydrogen removal in liquid aluminum.
MODELLING OF SURFACE SEGREGATION DEVELOPMENT DURING D.C. CASTING OF ROLLING SLAB INGOTS: A. Mo, H.J. Thevik, SINTEF Materials Technology, Box 124 Blindern, N-0314 Oslo, Norway; B.R. Henriksen, E.K .Jensen, Elkem Research, Box 40, N-4602 Kristiansand, Norway
A two dimensional mathematical model for the development of macrosegregation at and close to the ingot surface during DC casting of rolling slabs is presented. The model accounts for macrosegregation caused by exudation of interdendritic melt and macrosegregation associated with solidification shrinkage. Equations for the conservation of energy, solute, momentum, and mass during the stationary phase of the process are solved numerically by a finite element method in a solution domain defined by a vertical cross section of the ingot. The main simplifications in the modelling concept are to assume the solidified part of the mushy zone to move with the casting speed, and to consider a binary alloy solidifying according to the lever rule. The thickness and solute concentration of the surface layer and the macrosegregation close to the surface are calculated, and modelling results are compared with measurements on real castings.
10:10 am BREAK
A MODEL TO PREDICT THE STEADY STATE PULL-IN DURING D.C.-CASTING OF ALUMINIUM SHEET INGOTS: Arild Håkonsen, Hydro Aluminium R&D Materials Technology, P.O.Box 219, N-6601 Sunndalsøra, Norway
By combining empirical results with a dimensional analysis, a model for the prediction of pull-in during DC-casting of aluminium sheet ingots is constructed. The model predicts the steady state pull-in along the rolling faces of an ingot. By using an alloy dependent constant, the nominal geometry and the casting speed, the model can easily calculate a near optimal mould shape. It is also possible to calculate the resulting shape of the rolling faces of an ingot for a given mould geometry, alloy constant and casting speed. The model requires only a spread sheet for these calculations. Comparisons with experiments show a very good agreement. The best fit values of the alloy constant for several commercial aluminium alloys are presented. The value of this constant varies within 12%.
THE MECHANISM OF PULL-IN DURING D.C.-CASTING OF ALUMINIUM SHEET INGOTS: Hallvard G.Fjær, Institute for Energy Technology, P.O. Box 40, N-2007 Kjeller, Norway; Arild Håkonsen, Hydro Aluminium R&D Materials Technology, P.O. Box 219, N-6601 Sunndalsøra, Norway
The pull-in phenomenon apparent in the DC-casting process of aluminium sheet ingots is investigated both analytically and by use of numerical models calculating temperatures, strains and stresses. The major part of the pull-in takes place above the bottom of the sump, but is mainly caused by thermal contractions and deformations in regions below. The strong and almost exactly linear dependency of the pull-in on the casting speed is explained. The effects on the pull-in of different cooling conditions and of various thermophysical and thermomechanical properties are investigated. 2D plain strain approximations have been successfully applied in calculating the pull-in for the centre part of the rolling face, but 3D effects are significant close to the narrow ingot surface. This is illustrated by comparison of results from 3D and 2D calculations. Modelling results are also compared with measurements of ingot thickness variations. A very good agreement is obtained.
WATER COOLING IN DIRECT CHILL CASTING: PART 1: BOILING THEORY AND CONTROL: John Grandfield, Comalco Research Centre, P.O.Box 316, Thomastown, Victoria 3074, Australia
The intensity of the water cooling in direct chill casting affects the process heat flow and the performance of the casting process. The intensity depends on the boiling behaviour on the ingot surface. The basic mechanisms of boiling, the different boiling modes (nucleate, unstable film and stable film boiling), the effect of water composition, temperature, flow and velocity on boiling behaviour and heat transfer are reviewed. A new method for measuring the heat transfer coefficients has been developed and some new results on the effect of water composition and temperature are presented together with results from inverse calculations based on insitu measurements during casting.
A COMPREHENSIVE APPROACH TO WATER COOLING IN DC CASTING: B. Magnin, L. Maenner, Y. Caratini, Pechiney Centre de Recherches de Voreppe, BP27, F-38340 Voreppe, France
The numerical simulation of DC casting requires a good knowledge of the heat transfer between ingot and cooling water. This is especially true for the starting phase, where most defects are initiated. An experimental set-up has been developed to characterize this heat exchange close to process operating conditions. The method is based on the measurement of the cooling of a preheated instrumented aluminium block quenched by a water curtain. A 2D transient inverse model allows to determine the distribution of heat transfer coefficient and temperature on the whole surface of the block. In this way, heat flux can be assessed both at the water impingement point and below. The results obtained with the experimental apparatus are in good agreement with heat transfer rates calculated from in-situ temperature measurements using cast-in thermocouples. This device is thus an efficient and economical method to determine the influence of cooling water characteristics on the law describing heat transfer coefficient as a function of surface temperature. These laws have been measured and introduced in a numerical simulation of DC casting in order to study the effect of water cooling and casting parameters on the thermal evolution of the ingot during the different phases of casting.
FUNDAMENTALS OF UBC DECOATING/DELAQUERING FOR EFFICIENT MELTING: Wesley Stevens, Francois Tremblay, Arvida Research and Development Centre, Alcan International Limited, 1955 Mellon Blvd., Jonquière, Québec, Canada G7S 4K8
Aluminum used beverage cans (UBC) should be cleaned to completely remove all organic and residues to maximize metal recovery on remelting. This process must also minimize oxide formation. The aluminum industry uses three principal decoating technologies. The most used is the rotary kiln. The other processes are the belt decoater and the fluidized bed decoater. The presentation will cover the fundamental science for decoating shredded UBC to minimize metal loss on remelting. Laboratory work and plant trials have shown the requirements for complete decoating. These are: correct temperature range for the coatings to be pyrolized, good process gas to shred contact for all the scrap, sufficient oxygen being present to complete the oxidation of the carbon compounds present. The basic principles of the three decoating technologies used for the decoating of UBC will be discussed in relation to how they correspond to the fundamental process requirements.
|Search||Technical Program Contents||1997 Annual Meeting Page||TMS Meetings Page||TMS OnLine|