Search Documents

By A. A. Hejja

"Liquidus temperature and electrical conductivity data measured on synthetic slags were modelled by the use of neural nets (NN). In this work the applied multilayer feedforward NNs were trained by a conjugate-gradient optimisation for which a three layer formulation was used. Very good fits were obtained for both the liquidus temperature and the conductivity data. The synthetic slags were prepared from pure oxides to represent a wide range of compositions likely to be encountered in ferromanganese and silicomanganese smelting. The slag constituents were in the following range: Mn0;5-30%, Ca0;20-35%, Mg0;5-15%, Si02;27-58%, Al20 3;5%. Liquidus temperatures varied from 1300° to 1380° and increased with increasing basicity ratio. The electrical resistivity of slags decreased with the increase of basicity ratio from 0.55 to 1.1 but above 1.1 basicity ratio the resistivity tended to increase depending upon the MnO content.Neural NetsVarious neural net (NN) topologies such as Hopfied nets, Hamming nets, Carpenter/Grossberg classifiers have been discussed in some detail by Lipmann (1). In this work, the applied multilayer feedforward NNs are not trained by backpropagation (BP) using the generalized delta rule (2) (GDR), but by a conjugate-gradient optimisation [3,4] - a procedure similar to that proposed by Barnard and Cole [5]. A neural net is capable ·of learning the non-linear functional relationship existing between a set of inputs and outputs [ 6]. In this capacity a neural net is used to determine the functional relationship that exists . between N dependent and independent variables, which characterize the system under consideration For a three layer neural net [1] the following function may be formulated, which relates X; inputs to the output Y 0 from node o:"

Jan 1, 1997

By Bo Zhou

During secondary aluminum recovery, complex aluminum scraps are melted and refined often in a rotary melting furnace and the aluminum metal has to pass through a molten salt layer and melt down in a bottom aluminum bath. Previous experimental study and industrial observations have indicated that salt shell formation and re-melting on the metal solids in the early stage plays a critical role in the overall melting process of the scrap. In the present paper, mathematical models are developed to simulate the process of salt shell formation and its re-melting, as well as aluminum melting in the molten melt. Influence of the salt layer properties, the residence time of the metal particles in the salt layer, the particle size, the temperature of bulk melt were investigated. Reasonable agreement with the measurement data was reached. The developed melting model will be used to construct a population balance model of the scrap melting behavior for a rotary melting furnace, taking into account the distributed characteristics of the scrap. Finally, an overall process model based on Computational Fluid Dynamics (CFD) for the rotary furnace will be constructed, coupling with the population balance model for the scrap melting.

Jan 1, 2003

By Bo Zhou

During secondary aluminum recovery, complex aluminum scraps are melted and refined often in a rotary melting furnace and the aluminum metal has to pass through a molten salt layer and melt down in a bottom aluminum bath. Previous experimental study and industrial observations have indicated that salt shell formation and re-melting on the metal solids in the early stage plays a critical role in the overall melting process of the scrap. In the present paper, mathematical models are developed to simulate the process of salt shell formation and its re melting, as well as aluminum melting in the molten melt. Influence of the salt layer properties, the residence time of the metal particles in the salt layer, the particle size, the temperature of bulk melt were investigated. Reasonable agreement with the measurement data was reached. The developed melting model will be used to construct a population balance model of the scrap melting behavior for a rotary melting furnace, taking into account the distributed characteristics of the scrap. Finally, an overall process model based on Computational Fluid Dynamics (CFD) for the rotary furnace will be constructed, coupling with the population balance model for the scrap melting.

Jan 1, 2003

By G. Brooks, J. Naser, M. A. Sattar

CFD models have been developed and numerical simulations have been carried out to predict the formation of foam in oxygen steelmaking. Foam was considered as a separate phase comprising a mixture of gas and liquid. Bubble break up and coalescence models have also been incorporated in a CFD model to predict the number density of individual bubble classes. A population balance equation was used to track the number density of each bubble class. Decarburization with heat generation from chemical reactions was integrated in the process. User subroutines were written in FORTRAN to incorporate the foam formation, the bubble break up and coalescence rate and decarburisation in the main program. The model predicted the foam height, bubble number density, velocity of phases, decarburization, and turbulence. The result from the model has been compared with available data from literature and found to be in reasonable agreement with the experimental and plant data.

Jan 1, 2014

The entrainment of copper matte in the slag during the slag skimming stage of a Peirce-Smith converter was studied with a dam break model. Experiments were carried out with a range of fluid pairs where the lighter fluid was allowed to overflow a weir and the amount of entrained heavier fluid measured. The resulting entrainment is measured as the ratio of the heavy to light fluid withdrawn. The fluids include oil-water, water- glycerol and salt solutions-water pairs to cover a wide range of densities and viscosities. It was found that the Richardson number, Ri, correlates the entrainment well. For the case of the liquid-liquid interface initially at the same height as the weir, the entrainment was found to vary with Ri-1/3. The transient flow was simulated with a volume of fluid (VOF) code written in-house and used a stairstep approach to represent flow obstacles. The code was validated against the experimental results with good agreement for the variation of the heights of the two fluid with time. The simulation showed that the prevailing mechanism for entrainment is due to a pressure differential created across the liquid in the converter during pouring resulting in the heavier liquid being rotated upwards and over the weir.

Jan 1, 1999

By J. L. Liow

Studies have been carried out on the causes and effects of slopping in an annular cylindrical geometry such as the top submerged combustion reactor. Properties of standing waves formed such as the wave profile, period, pressure distribution and velocity have been derived analytically. The wear profile due to wave erosion was calculated based on the assumption of linear motion and that the wear is a dissolution process. The use of the j, factor showed that the wear was proportional to the square root of the wave velocity at any point. Using the dimensions of the lsasmelt reactor given by Errington et al. (1987)[11, the wear profile due to wave erosion was calculated using the analytical solution. The wear profile calculated was shown to have a maximum just below the initial bath surface. The wear was found to be concentrated in the region where the first asymmetric standing wave was most active. The results showed that the refractory near the bath surface will be affected by wave erosion, more so than any other parts of the reactor. The contribution of wave erosion on wear must be accounted for in assessing the total refractory wear of a vessel in which slopping exists.

Jan 1, 1991

By P. Myllykoski, J. Nylander

"A commercial program package called Crownoff for the simulation of temper rolling in four high mills has been verified against measured data. The agreement between measured and calculated transversal strain distributions of the strip was found to be good. Effects of the strip crown and work roll bending force on the strain distribution were simulated. The results showed that when a single ground camber of work rolls is used for all strip widths; the strip crown should be parabolic, excluding the edge drop. An increase or decrease of the strip crown by 30 % during a coil can be compensated by a 2 to 5 t change in the work roll bending force. The ability to control the strip crown by work roll bending was insensitive to strip width. IntroductionCold rolled steel strips are usually temper rolled after recrystallization annealing, which follows cold rolling. The purpose of temper rolling is to suppress the yield point, finish the strip surface and correct possible flatness errors. The presence of a yield point might otherwise cause problems during subsequent forming processes. The in homogenous yielding in non-temper rolled steels leads to the formation of Lüders bands in the strip surface.The rolling force is influenced by the rolling speed and changes in frictional conditions in the roll bite. Therefore it is very important to plan carefully the mill settings and control algorithms during rolling. The purpose of mill control algorithms is to ensure constant reduction and the best possible strip flatness during the entire coil.Correctly predicting the rolling force is important for determining the mill presettings. Torque, forward slip and power requirements are not usually critical, if the reductions are small. However, the large elastic deformations occurring in the roll bite make the physical description of the process difficult. Physical rolling models developed for cold rolling contain assumptions that lose their validity as the reductions decrease to the level of one per cent."

Jan 1, 1994

By R. Nilsson

In the present work, an assessment of the sulphide capacities of a number of binary and ternary slag systems have been carried out in the light of a sulphide capacity model. The model is based on the expression: {CS- exp (-A G °/R T) ? ao2 / fs2-} where AGO is the standard Gibbs energy change for the above reaction, ao2 is the activity of oxygen ions and f, is the activity coefficient of the sulphide ion in the slag. In the present model, the ratio (ao2/fs2) is optimized for the experimental data available in literature. The systems assessed are the binaries MnO-SiO2, CaO-SiO2, Al2O3-CaO, as well as the ternaries CaO-Al2O3-SiO2 and CaO-MnO-SiO2. In the case of all these systems, the experimental data available in literature were complemented by measurements carried out at the Division of Theoretical Metallurgy. The sulphide capacities of some FeO- containing binary, ternary and quaternary slags are also examined. The potentiality of the model to estimate the sulphide capacities of complex slag systems from binary and ternary data is demonstrated.

Jan 1, 1996

By Jean-Christophe Gebelin, Mark Jolly, Samantha Jones

"The de-waxing process in investment casting is truly a black art which is carried out in a black hole - the autoclave. The process involves exposing wax, coated by a porous ceramic shell, to steam at a pressure of 10 bar and a temperature of about 180°C. The processes which then occur are truly multi-media and multi-physics. Steam condenses in the porous ceramic, wax melts and flows, the shells expand and sometimes crack. Why the shells crack is one of the most difficult questions asked in the industry.Modelling and validating the de-waxing process is not simple. The IRC at the University of Birmingham is attempting to model this process using Physica+ a code developed in the UK. This paper will present the latest results in this attempt to model one of the casting industries most intractable problems.IntroductionThe de-waxing process using an autoclave involves the following physical phenomena:· Flow of the steam in the autoclave;· Heat transfer from the steam to the wax, possibly through the shell;· Melting and flow of the wax.At the same time, other phenomena may occur, such as the flow of the steam through the shell, the condensation of the steam on the shell surface and on the wax pattern surface, the flow of wax the through the shell, etc ...This shows how complex the de-waxing process is from a physical point of view. The main defect that appears during this process is the cracking of the shell. Previous modelling work carried out by Snow [1] has shown that it is possible to predict the stress in the shell, making numerous assumptions. In this paper, the modelling to predict the state of stress in the shell that is currently in development at the University of Birmingham is presented together with the presentation ofthe physical phenomena to model and the assumptions used. The parameters that are needed in such models will also be discussed, and some experimental work undertaken at the moment presented."

Jan 1, 2001

By Anna Franzkowiak, Dean Gregurek, Alfred Spanring, Angelika Ressler

In order to determine the most suitable refractory products and improve the lining lifetime for the diverse furnaces used in the nonferrous metal industry, corrosion tests are performed at RHI’s Technology Center. The practical facilities include the cup test, induction furnace, rotary kiln, and drip slag test described in this paper, which enable a comprehensive understanding of the chemo-thermal brick wear on a pilot scale. The corrosion trials are performed with actual slags generated during operations at a customer’s plant. To determine the highest influencing wear parameter, every single test is combined with a detailed mineralogical investigation and thermochemical calculations performed using FactSage. Based on the results, tailored refractory solutions for the nonferrous metal industry can be provided in combination with trials conducted at the customer’s site.

Jan 1, 2014

By K. Th. Mavrommatis

Cleanness of steel is justified, measured and estimated in steelmaking process through different concepts which are mainly based on operational, empirical or semiempirical parameters and describe ?cleanness as a state?. In this work a theoretical concept is developed, considering ?cleanness as a process? and introducing a dynamic aspect into the quantification of cleanness of a steel melt, linking operational parameters with physicochemical variables, which are based on first principles. This concept and the here derived and developed variables and parameters are based on the earlier defined agglomeration index (AI) of a steel melt, the ratio of the mean inclusions diameter (MID) to the mean inter-particle distance (MIPD), which is shown to be a measure of the agglomeration tendency of the melt after deoxidation, and can serve both, to monitoring cleanness during steelmaking operation and to be useful as a basis for mathematical modelling of cleanness and of the metallurgical measures in plant, associated with production of clean steels.

Jan 1, 2006

By Mike Dry

While McCabe-Thiele plots are useful for designing single-component solvent extraction systems, not least because of their simplicity, they are less useful for multi-component solvent extraction. This paper presents a relatively simple method, compared to other approaches, for modelling solvent extraction as chemical equilibrium reactions between the various aqueous and organic species. The essence of this approach is how values for the equilibrium constants can be extracted from standard laboratory data. Synergistic solvent extraction of nickel is used to illustrate the method.

By Y. Houbaert, L. Kestens, R. Decocker

"A magneto-elastic model is presented to calculate the orientation dependence of the magnetostrictive strain, observed at saturated magnetisation in ferromagnetic materials with a cubic crystal structure and an arbitrary crystallographic texture. The formula of Becker and Doring is used to express the anisotropy of magnetostriction for a single crystal. In order to simulate the macroscopic average magnetostriction of a polycrystalline aggregate (with an arbitrary texture) the Reuss assumption of the elasticity theory was applied. According to this assumption the various orientations of the polycrystal can defonn without constraints, producing local strain incompatibilities in the microstructure but observing total stress equilibrium. The macroscopic strain is calculated as the weighted average of the individual strains of all orientations composing the polycrystal. The weight factors are determined by the volume fractions of the corresponding orientations of the given texture, which can be measured by standard X-ray diffraction techniques. The model is applied to simulate the variation of magnetostriction (at saturation) with respect to the rolling direction for a standard grade of Goss-oriented electrical steel. A brief comparison with experimental data allows validating the basic model assumptions. IntroductionBecause of tougher environmental constraints, the noise reduction in transformer cores has become increasingly important. It is generally accepted that the main source of noise is magnetostriction, which is known to be caused during the magnetisation process by the movement of 90° magnetic domain walls and the rotation of the magnetisation vector out of the easy directions. Therefore, it is obvious that when a ferromagnetic body is magnetised to saturation, the resulting magnetostriction will be dependent on the distribution of magnetic domains in the initial state.In the literature two coefficients are used to describe the resulting magnetostriction at a saturating magnetic field. The total strain, measured from an arbitrary demagnetised state to technical saturation, is called the saturation magnetostriction, and is denoted by (~)s. The tenn magnetostriction coefficient will be employed to designate the strain measured from the ideal demagnetised state (i.e. one in which no preferred orientation of domains exists) to technical saturation. Failure to recognise the difference between these two quantities has led to considerable confusion in the literature.The magnetostriction coefficient for a single crystal is a well-known property for ferromagnetic cubic materials. In 1939 Becker and Doring [1] have derived the following relation, describing the magnetostriction coefficient for single crystals with a cubic symmetry:"

Jan 1, 2001

By M. Patel, A. W. Piper, R. A. Forsdick

"Lead bullion is refined in kettles to remove the contained impurities such as copper, antimony, silver and bismuth. These impurities are removed sequentially by the addition of one, or more, reagents to the bath. Impurity elements react selectively with the reagents to form a dross layer on the surface that is subsequently removed for further processing. Mixing the lead bath using specific impeller designs, which form a vortex to ensure that good contact takes place, carries out the reactions. It is important that the mixing is controlled to achieve the optimum bath conditions and hence good dross grades.This paper presents initial modelling work on the kettle refining process using both physical modelling (Water Models) and mathematical modelling (Computational Fluid Dynamics - CFD). Comparisons between measured plant data and simulation results are given.IntroductionThe refining of lead bullion takes place in hemispherical kettles of varying sizes. It is normal practice to remove the impurity elements sequentially, by the addition of reagents, which selectively react with the impurity element. The impurity element forms a dross layer on the surface, which can be removed for subsequent processing at the end of the reaction period.All the refining stages in the production of lead and lead alloys involve mixing the lead bath, although different conditions are required for each stage. At Britannia Refined Metals mixing is achieved using either 3 or 4 angled or flat blade impellers, which generate a vortex in the central region of the kettle. However, throughout the industry a wide range of mixer types are used which are all trying to achieve the same result."

Jan 1, 1999

By Mark Jolly

"Over the last ten years or so many man hours have been dedicated to developing methods for simulating shape casting processes. The majority of the methods developed have been devised by a combination of computer, mathematics and materials specialists often with limited knowledge of foundries and foundrymen. As a result of this it would appear that although there are many software packages for the foundryman to use there still exist fundamental misunderstandings as to their usefulness and suitability within the foundry.The paper aims to identify how well current software packages perform. The relationships between physical phenomena, practical defects and software capabilities are presented. Some discussion of the use of criteria functions is also presented. Finally issues arising from the postprocessing of results are discussed as the presentation of the results to the lay-person in simulation techniques is possibly one of the most important aspects in influencing the adoption of this type of software in the foundry community."

Jan 1, 1999

By Pengfu Tan

The control of converter slag chemistry and slag viscosity has played an important role in optimizing converter operations. A thermodynamic model of copper P-S converter and a viscosity model of converter slag have been used to predict the chemistry and viscosity of converter slag. The effect of oxygen potential, SiO2/Fe in slag, detection of slag blow endpoint, and temperature on slag viscosity, liquidus temperature of converter slag and skimming operation has been discussed. Some applications of the industrial operations have been presented.

Jan 1, 2006

By A. Broussaud

An original methodology has been developed for steady state simulation of selective grinding and mineral liberation in rod and ball mills. Three fundamental aspects of the methodology are presented successively : modeling ore texture, modeling grinding and its selectivity, and modeling mineral liberation. The methodology for modeling ore texture includes the conversion of 2D data into 3D reconstructed information and the use of the BILCO coherent material balance software. The models for grinding Simulation are classified into three levels, corresponding to three levels of available experimental data. All models account for grinding selectivity and for variations of engineering para-meters and operating conditions. An algorithm has been introduced to forecast mineral liberation, compatible with accurate simulation of selective grinding, which generates reasonable results.

Jan 1, 1987

By Andreas Siegmund

During the decade of the nineties the conventional method of primary lead smelting (sinter machine/blast furnace) was successfully challenged by the introduction of novel direct and continuous smelting processes. In combination with a significant shift in market structure, as well as more stringent government regulations these modem technological innovations like the QSL, Kivcet, and Isa/Ausmelt processes became more than major competitors. They have proven to be economically and environmentally viable. The intensification of the metallurgical reaction by applying oxygen enriched or pure oxygen bath or flash smelting principles resulted in cost savings and a higher flexibility with respect to raw materials and additives usage. By synchronizing individual auxiliary plant sections with the smelting process a virtually waste-free production and low-emission mode of operation with nearly optimum energy exploitation is achieved. The primary lead industry will gradually substitute the sinter -blast furnace operation with novel high-energy efficient technologies in the new millennium.

Jan 1, 2003

By Andreas Siegmund

During the decade of the nineties the conventional method of primary lead smelting (sinter machine/blast furnace) was successfully challenged by the introduction of novel direct and continuous smelting processes. In combination with a significant shift in market structure, as well as more stringent government regulations these modem technological innovations like the QSL, Kivcet, and Isa/Ausmelt processes became more than major competitors. They have proven to be economically and environmentally viable. The intensification of the metallurgical reaction by applying oxygen enriched or pure oxygen bath or flash smelting principles resulted in cost savings and a higher flexibility with respect to raw materials and additives usage. By synchronizing individual auxiliary plant sections with the smelting process a virtually waste- free production and low-emission mode of operation with nearly optimum energy exploitation is achieved. The primary lead industry will gradually substitute the sinter -blast furnace operation with novel high-energy efficient technologies in the new millennium.

Jan 1, 2003

By Edgar E. Vidal, Keith Smith, James A. Yurko

Beryllium, like many other reactive metals, is extracted from a halide (generally BeF2) using magnesiothermic reduction, also known as the Kroll process. The beryllium is produced in a molten state, which when solidified resemble "pebbles". Although beryllium is used in limited quantities compared to other metals (< 100 Mt/annual), it is a critical material for the nuclear, defense/aerospace, and large science markets because of its unique nuclear and mechanical properties. Beryllium processing has unique environmental, health, and safety concerns which must be handled appropriately in the 21st century from a worker and environmental safety perspective, and must be designed with this first in mind. In 2011, Materion began operations of its new Kroll plant – informally known as the “Pebbles Plant” - in Elmore, Ohio. The plant was majority funded by a U.S. Department of Defense Production Act (DPA) Title III grant, which is a unique government/industry partnership program to, "create assured, affordable, and commercially viable production capabilities and capacities for items essential for national defense." After a twelve year hiatus from primary beryllium production, Materion has reentered its extraction business and has relearned how to make pebbles. The purpose of this presentation is to give an overview on beryllium extraction and discuss some of the unique aspects of building and operating a safe beryllium facility during the production scale-up phase.

Jan 1, 2015