DECEMBER 2002 21

The Science and Technology of Steelmaking– Measurements, Models andManufacturingA. McLeanUniversity of Toronto

This J.K. Brimacombe Memorial Lecturewas presented at the ISS 60th ElectricFurnace Conference in November in SanAntonio, Texas.

In our efforts to charac-terize and improve theperformance of an exist-ing steelmaking process,or in our quest to gener-ate useful knowledge asa basis to develop newmanufacturing routes,measurements andmodels should be con-sidered as two interdependent require-ments. Without measurements, ourmodels are incomplete and unsatisfac-tory. Without models, we fail to realize,or perhaps even comprehend, the po-tential significance of our measurements.Sometimes in our enthusiasm, we con-struct sophisticated, elegant models andforget the reality of the actual manufac-turing process. In this computer age, weneed to remember the importance of ob-servations and accurate measurements.In addition, as engineers and applied sci-entists, we have an obligation and a re-sponsibility to facilitate the transfer of newknowledge into the realm of operating

practice. During this process of genera-tion, evaluation and communication ofnew knowledge, knowledge exchange isperhaps the most difficult step. In thiscontext, the preeminent aim of collabo-rative activities between our educationalinstitutions, industrial organizations, gov-ernment funding agencies and profes-sional societies, is to ensure the availabil-ity of high quality people who not onlyunderstand the fundamental aspects andpractical implications of their discipline,but also are fully equipped with the es-sential skills of communication that willenable them to participate throughouttheir career in this most challenging andsatisfying activity, the science and technol-ogy of steelmaking.

INTRODUCTION“A never-ending challenge to the com-petitiveness of the steel industry is the

application of knowledge on the shopfloor where, finally, productivity andquality are realized.”1993 Howe Memorial Lecture, The Ironand Steel Society of AIME1

During the past century, the steel indus-try has been transformed by major tech-nological changes. Coupled with theimplementation of innovative productionsystems, there have been increasing de-mands for improved steel quality. Thesedemands have been met by advances inour knowledge and understanding of thechemical, physical and thermal interac-tions between steel, gas, slag and refrac-tory phases, which occur within indi-vidual reaction vessels and also duringtransfer operations between the primaryfurnace and the casting mold. In a typi-cal steel plant, the reactors include con-verters or arc furnaces, ladle furnaces,vacuum vessels, tundishes and castingmolds. In recent years, transfer opera-tions between one reactor and the nextmust be controlled precisely. Otherwise,these transfer steps become destroyersof quality.

As we move into the next century, thereis an increasing need for new processcontrol strategies to achieve consistentquality products with maximum yield,

The exchange of knowledgebetween researchers, users andthe resource community isrequired to transform a funda-mental scientific concept into areliable operating system.

22 DECEMBER 2002

optimum cost and minimum impact onthe environment. Successful new strate-gies are most likely to evolve from aknowledge-based foundation of mea-surements, models and a solid under-standing of the fundamental aspects ofthe manufacturing process (Figure 1).

At an International Conference in To-kyo on Computer-Assisted Materials De-sign and Process Simulation, in a paperentitled “Computer Simulation of Solidi-fication and Casting Processes,” J. KeithBrimacombe and his co-authors empha-sized the importance of “balancing theapplication of sophisticated computermodels with painstaking measure-ments.” While acknowledging that “thecomputer has transformed the ability toanalyze casting and solidification proc-esses quantitatively with mathematicalmodels,” they went on to state “withoutcareful companion measurements, bothin the laboratory and in-plant, to char-acterize thermophysical properties andcomplicated boundary conditions, aswell as to validate the mathematical mod-els, the power of the computer is severelyblunted.”2 Although these commentswere made with respect to casting proc-esses, they are equally valid within thegeneral context of this article.

In the following sections, examplesare provided of some laboratory dataand plant studies, together with processmodels and innovative technologies thathave been developed for monitoring and

control of specific aspects of the steel-making operation. These technologicalconsiderations are discussed within thephilosophical context of several themes,which are in many ways a reflection ofsome of the ideas and concepts ex-pressed so eloquently by Brimacombein memorial lectures to professional so-cieties, in keynote presentations at inter-national conferences and in commen-taries published during his years aspresident of The Minerals, Metals andMaterials Society, and later, ISS. They arebut a few threads drawn from themulticolored tapestry that was the fabricof Keith’s career.

OXYGEN CONTROLIN MOLTEN STEEL“When you can measure what youare speaking about and express it innumbers, you know somethingabout it. When you cannot measureit, your knowledge is meager and un-satisfactory.”Lord Kelvin

In discussions of steel quality, interac-tions between oxygen and elements suchas carbon, manganese, chromium, sili-con, aluminum and calcium are particu-larly interesting. The same is true withrespect to reactions involving the ele-ments sulfur, phosphorus, hydrogen andnitrogen. Control of oxygen potentialthroughout steelmaking operations isessential. Without oxygen control, con-trolling the behavior of the other ele-ments is impossible. In this context, con-trol of the composition, properties andperformance of steelmaking slags andrefractories is also a mandatory require-ment for producing quality steel.

A major thrust in current steelmak-ing technology is the production of steelwith lower residual concentrations ofoxygen and sulfur. These elements havea profound influence on the quality ofthe final steel product because of theireffect on inclusion formation. Since

these reaction products have a directbearing on the properties of steel in ser-vice, considerable attention has beendirected toward developing a thoroughunderstanding of the factors that influ-ence their formation and properties, aswell as methods for minimizing their pres-ence in solid steel. Factors that influencetheir formation include: interactionswithin the molten steel itself, interactionsbetween molten steel and air duringtransfer operations from one vessel toanother, interactions between steel andmolten slag, and finally, interactions be-tween steel and refractory materials.

In steelmaking processes, molten steelis inevitably brought into direct contactwith air. Little et al. have shown that largeinclusions are primarily due toreoxidation of the metal during transferoperations.3 An exposed metal meniscusin the ladle or tundish, and a moltenstream or turbulence of the metal in themold cavity, provide ample opportunityfor liquid steel to react with air, which isan unlimited supplier of oxygen.4 In ad-dition, these conditions can also contrib-ute to the absorption of nitrogen, as wellas hydrogen, particularly when the hu-midity content of the atmosphere is high.

Reactions between molten steel andthe surrounding atmosphere have beeninvestigated using levitation melting anda falling droplet technique (Figure 2).5, 6

The solutes oxygen and sulfur and theinterrelationships that exist betweenthese solutes and the elements nitrogenand hydrogen have been studied. Whenpresent in molten steel, oxygen and sul-fur behave as surface-active elements.Experiments have shown that both oxy-gen and sulfur drastically decrease thenitrogen absorption rate (Figures 3 and4). The same behavior has been ob-served with respect to hydrogen absorp-tion. On the other hand, in the case ofoxygen absorption, the rate is not af-fected by the sulfur content of the steel,but rather by diffusion in the gas phase(Figure 5). Thus, if steel is produced

Figure 1 The knowledge exchangeprocess.

DECEMBER 2002 23

with very low oxygen and sulfur contents,then molten steel is more susceptible tohydrogen and nitrogen absorption. Forexample, during ladle desulfurizationtreatments, nitrogen absorption may beabout 15 ppm when the final sulfur levelis 90 ppm. However, when the final sul-fur level is 20 ppm, nitrogen absorptionmay be as high as 50 ppm.7 This impliesthat while problems associated with thepresence of sulfides may be avoided byproducing low sulfur steels, new prob-lems may be introduced due to high ni-trogen contents and possibly even high

hydrogen contents. For this reason,there are advantages to producing steelwith final sulfur concentrations that areat the higher end of the sulfur specifica-tion rather than at the lower end. In thisway, nitrogen and hydrogen absorptioncan be minimized. By the same logic,delaying the addition of deoxidizers and/or desulfurizing treatments will also de-crease opportunities for nitrogen andhydrogen absorption.

BEHAVIOR ANDCHARACTERIZATION OFSTEELMAKING SLAGS“We must see what we can learn fromeach other. The power of knowledgestems not just from its depth, but alsofrom its breadth.”Presidential Perspective: The Power ofKnowledge8

To a large extent, oxygen control insteelmaking is affected by the compo-sition and properties of the slag phase.Without oxygen control, the behaviorof the other elements cannot be man-aged. Steelmaking slag systems involveat least eight important components:CaO, MgO, MnO, FeO, SiO

2, Al

2O

3, P

2O

5,

CaF2, and often a number of others such

as Na2O, TiO

2 and Cr

2O

3. One of the

greatest obstacles to the application

of physical chemistry principles to theelucidation of slag-metal and slag-gasreactions is a lack of knowledge aboutactivities of the reacting species. Sinceiron oxide in slag participates in nu-merous reactions between metal, slagand gas phases, and important steel-making reactions such as desulfuriza-tion and dephosphorization are relatedto the ferrous oxide activity, evaluationof the activity of this component is par-ticularly important.

In addition, molten slag can contrib-ute to reoxidation of the steel. For ex-ample, consider the reoxidation of alu-minum by ferrous oxide in molten slags:

2

33

Al(in Fe) + 3 FeO(in slag)

= Al Fe

. . . . . (1)2O s( ) +

Equation (1) will proceed to the right-hand side with an increase in FeO activ-ity. Hence, it is required to lower the FeOactivity in molten slag especially in sec-ondary refining vessels. For this purpose,aluminum or calcium carbide is spread

Figure 2 Levitation melting and fallingdroplet facility for the investigation ofreactions between molten steel and gasatmospheres.5, 6

Figure 3 Absorption of nitrogen by steeldroplets after falling for various timesthrough different gas atmospheres.5

Figure 4 Decreased in nitrogenabsorption rate by steel droplets with anincrease in sulfur content.5

Figure 5 Absorption of oxygen by steeldroplets containing different amounts ofsulfur, aluminum (0.1 percent) andtitanium (0.1 percent) after falling forvarious times through air.6

24 DECEMBER 2002

over the slag. Currently, control strate-gies for such practices are based onsamples taken from the molten slag dur-ing ladle treatment.

Based on practical experience, plantoperators pay particular attention to theconcentrations of “total” Fe plus MnO,(%T.Fe) + (%MnO). Manganese oxidecan also reoxidize aluminum although(%MnO) < (%T.Fe). For example, Haraet al. have reported this effect where, asshown in Figure 6, the total oxygen con-centration in steel after RH degassingincreased with an increase in (%FeO)+ (%MnO) in the slag phase.9 However,chemical analysis for FeO and MnO inslags requires at least 20 minutes, whileladle treatment is generally completedwithin 15 to 20 minutes. Hence, aknowledge of (%T.Fe) + (%MnO)based on chemical analysis cannot re-ally assist plant operators to control themolten slag of a particular heat duringactual production.

In view of the industrial significanceof ferrous oxide activity, numerous ex-perimental determinations have beenconducted by many laboratories dur-ing the past five decades. Nevertheless,our knowledge of the activity of FeO inmulticomponent industrial slags is stillfar from satisfactory. The conventionalexperimental technique for determin-ing ferrous oxide activity consists ofequilibrating molten slags with liquid

iron. Chemical analysis for oxygenmade on samples taken from liquid ironyields values of ferrous oxide activities.Alternatively, molten slags contained inan iron crucible are brought into equi-librium under a stream of CO-CO

2 or H

2-

H2O. These techniques however, require

a relatively complex experimental setupand long duration.

Rapid Determination of FerrousOxide ActivitiesOgura et al. have developed a “shopfloor” activity determinator for ferrousoxide based on the principles of zirco-nia sensors.10 This equipment usessamples taken from the slag phase.However, knowledge of sample weightis not required, and activity measure-ments can be completed within fiveminutes. This electrochemical tech-nique based on magnesia-stabilized zir-conia permits rapid determination offerrous oxide activity within homoge-neous and heterogeneous slag sys-tems.11 Using this technique, FeO activ-ity has been measured for manydifferent slag systems by Iwase et al.12, 13

This research led to the manufac-ture of an analytical instrument to de-termine FeO activities.10, 14 A diagram ofthe automatic activity determinator isshown in Figure 7. To operate this sys-tem, a solid state sensor (F) is attachedto the elevator mechanism (E), and aniron crucible (K) is placed on the steelpedestal (O). The electrical contact tothe outer electrode of the zirconia cellis made via this pedestal. Pure silver(M) is premelted within the iron cru-cible. A sample of slag (L), 1 to 3 g, isplaced in the iron crucible. The ironcrucible is moved upward via an eleva-tor mechanism (P) into a transparentsilica reaction tube (H). The furnace issubsequently heated to a temperatureof 1,750 degrees Kelvin under a streamof argon within 2.5 minutes accordingto a computer program. The electro-chemical cell is lowered until it contacts

both the molten silver and the slag.Open-circuit cell voltages generated aremonitored on the LCD of the microcom-puter. The cell potential is then convertedto the FeO activity and displayed on theLCD. In this way, a single activity mea-surement is completed within five min-utes. This determinator, which is nowused by several steel companies in dif-ferent parts of the world, has permittednew control strategies to be developedfor the behavior of phosphorus and sul-fur in liquid iron and for reducing steel-making slag volume.

Applications of the ActivityDeterminator for Optimizationof Steelmaking OperationsAn example of the industrial applica-tion of the activity determinator byHamm et al.15 relates to control of thedesulfurization reaction, which can beexpressed as:

Figure 6 Relationship between totaloxygen contents in steel after degassingand slag chemistry.9

Figure 7 Illustration of automaticactivity determinator: (a) Mo rod; (b)zirconia cement; (c) Mo + MoO

2 reference

electrode; (d) reference electrode; (e)elevator mechanism; (f) zirconia cell; (g)water-cooled brass flange; (h) transparentsilica tube; (i) tungsten filament; (j) Arinlet; (k) iron crucible; (l) slag sample;(m) silver; (n) Pt-PtRh 13 thermocouple;(o) stainless steel pedestal; and (p)elevator mechanism.11

DECEMBER 2002 25

CaO S Fe CaS FeOb g b g b g+ + = +

.....( )2

Equation (2) implies that if the activ-ity of CaO (i.e., the basicity of the slag) iskept constant, a logarithmic plot of thesulfur partition ratio against FeO activityshould be a straight line with a slope of–1. The relationship obtained by Hammet al. is shown in Figure 8. Although theslope is different from the theoreticalvalue, there is a good correlation be-tween the two parameters, indicating thatthe sulfur partition ratio can be moni-tored, and therefore controlled, throughFeO activity measurements.

Figure 9 shows the relationship be-tween a

FeO, which was determined by the

activity determinator and (%T.Fe) +(%MnO) obtained by chemical analysisfor industrial slags taken from differentladle steelmaking operations. As shownin this figure, the empirical parameter,(%T.Fe) + (%MnO), is proportional tothe FeO activities. However, FeO activitiesof slags from one melt shop are differentfrom those at another even though theempirical parameter, (%T.Fe) + (MnO),is the same.

From a thermochemical point-of-view, this behavior is quite understand-able if differences in slag basicity areaccounted for. In other words, we canstate that the empirical parameter is not

an appropriate model for close controlof FeO activity. Oh-nuki et al. have re-ported that improved slag reduction,together with increased steel cleanliness,was achieved at the ladle furnace withthe aid of measurements obtained by theFeO activity determinator.16

Manganese control during steel-making is governed by the equilibriumreaction:

Mn FeO MnO Fe+ = +b g b g.. .. .( )3

Where (i) and [i] indicate component(i) in slag and molten steel, respectively.The partition ratio is expressed as:

log % / %

log

. . . . . ( )

Mn Mnb gm rb g= + a constantFeO

4

Where, (%Mn) and [%Mn] are theweight-percent concentrations of man-ganese in the slag and metal phases.According to Equation (4), a logarith-mic plot between (%Mn)/[%Mn] andthe a

FeO should yield a straight line with a

slope of unity. Hamm et al. have shownsuch a relationship based on actual datafrom practical steelmaking operations(Figure 10).15 These authors also re-ported that a

FeO measurements can be

used to calculate alloying efficiency with

a higher degree of precision, thus re-sulting in significant cost reduction andmore accurate heat chemistries.

Characterizing Slag BehaviorIn discussing the refining capacities ofslags, the ratio of basic oxides to acidicoxides expressed in weight percent, orsometimes as a molar ratio, is tradition-ally used as a measure of slag basicity.However, the simple (CaO/SiO

2) ratio ig-

nores the effects of other oxides. Theratio (CaO + MgO)/(Al

2O

3 + SiO

2) im-

plies that lime and magnesia behave asequivalent basic oxides and that aluminaand silica have the same degree of acid-ity, neither of which is the case. To somedegree, this can be offset by using em-pirical coefficients although the modi-fied ratios are somewhat restricted intheir applicability to specific applications.

Optical basicity provides a good foun-dation for a better understanding of thebehavior of molten slags than the con-ventional basicity ratios. The concept ofoptical basicity was developed by glassscientists Duffy and Ingram17 and intro-duced to the metallurgical communityby Duffy, Ingram and Sommerville in thelate 1970s.18 This approach has provedto be a valuable tool for designing slagsor fluxes that will have the required char-acteristics with respect to behavior of,for example, sulfur, phosphorus, hydro-gen, magnesia and alkali oxides.19-25 The

Figure 8 Relationship between sulfurpartition ratio and FeO activity.15

Figure 9 Relationship between FeOactivity determined by the activity sensorand (%T.Fe) + (%MnO).15

Figure 10 Plant data showing therelationship between manganese partitionratio and FeO activity.15

26 DECEMBER 2002

relationships between sulfide capacity,water vapor capacity and optical basic-ity are illustrated in Figure 11.26 In thecase of slag compositions for ladle des-ulfurizing treatments, a highly basic slagis generally used. While such slags havea high capacity for absorbing sulfur, theyalso have a strong ability to transfer hy-drogen from moisture-containing atmo-spheres into molten steel at a faster ratethan would be the case with a more acidicslag. This has important implications withrespect to ladle slags, as well as the se-lection of tundish fluxes and continu-ous casting mold powders. The opticalbasicity of molten slag can be calculatedusing the following relationships:

Λ Λ=

=

=

=

∑

∑

i ii

n

ii Oi

i Oii

n

N

NXn

Xn

1

1

Where:Λ = Optical basicity of the slagΛ

i= Optical basicity value of compo-

nent “i”N

i= Compositional fraction

Xi

= Mole fraction of component “i”in the slag

nOi

= Number of oxygen atoms in com-ponent “i”

PREVENTION/DETECTION OFDETRIMENTAL VARIATIONSDURING STEEL PROCESSING“The very pliancy of the steelmakingprocesses demands increased watch-fulness to prevent and detectunsought variation.”Henry Marion Howe27

As previously mentioned, the steelmak-ing system can be described in terms of aseries of metallurgical reactors linked bya number of transfer steps. Within this sys-tem, the furnace and ladle are batch re-actors, while the tundish and molds arecontinuous reactors. This difference has

important implications with respect tothe influence of transfer operations onsteel quality. In the quest for steel qual-ity, we must not only prevent detrimen-tal variations during processing, butalso develop sensor technologies thatwill permit us to continuously monitorand detect, at an early stage, the occur-rence of any detrimental variations dueto perturbations or imperfectionswithin the processing system. There isgood evidence from plant observationsthat quality achieved within one reactorcan be lost during transfer to the next. Inaddition, short, intermittent variationsduring the transfer process can have rela-tively long-time, adverse effects on behav-ior within the next reactor. These ad-verse effects relate to changes in chemistry,temperature and fluid flow, the formationof inclusions and ultimately, their loca-tion and distribution within the cast prod-uct. They can also affect surface quality,segregation and crack formation. Im-provements in our understanding of theimportance of these phenomena have ledto developments in processing technolo-gies, which are having a pronounced ef-fect on steel quality. Several examples ofthese aspects of steel processing are de-scribed in the following sections.

Optimization of the Electric ArcFurnace ProcessIn the latter half of the 1990s, research-ers at the University of Toronto partici-pated in a collaborative research effortbetween Goodfellow Technologies Inc.,of Mississauga, Ontario, Canada and Co-Steel Lasco of Whitby, Ontario, Canada.The group examined the potential forreplacing electrical energy with chemi-cal energy through post combustion inan electric arc furnace.28 By monitoringthe chemical composition of the gasesleaving the furnace and integrating theresults within the control system for thefurnace (Figure 12), the operating con-ditions can be adjusted to take greateradvantage of the chemical energy avail-able within the furnace, which otherwisewould be lost.29 Up to 30 percent of thetotal energy input can be lost in the exitgases, and much of this is in the form ofunused chemical energy due to incom-plete combustion within the furnace.Expert Furnace System OptimizationProcess (EFSOP) is also a useful tool forevaluating the design and operationalaspects of the furnace fume system,which, in turn, has beneficial implica-tions with respect to safety and environ-mental regulations.30

Figure 11 Relationships betweensulfide capacity (broken line), watervapor capacity (solid line) and opticalbasicity.26

Figure 12 Diagram of the gasmonitoring system and instrumentaionfacilities for dynamic control of theelectric arc furnace operation.

DECEMBER 2002 27

While other systems are available foranalyzing exit gases, particularly in thecase of integrated steel plants using oxy-gen converters, this system has severalunique features in terms of the combi-nation of hardware and software usedfor the dynamic control of an electricarc furnace operation.

After an extensive development pro-gram, the new technology has beenimplemented in regular production.Substantial energy savings of more than$1 million/year and productivity im-provements of more than 5 percent havebeen achieved.31 The main advantagesof this new processing technology canbe summarized as follows: reduced en-ergy costs, increased productivity, opti-mized furnace practice, improved op-erational safety and environmentalcontrol strategies.

Presently, EFSOP installations are inregular production operation withinthe United States, Canada, Mexico andthe U.K. This is an auspicious start forthis new technology and augers well forthe future.

Furnace to Ladle TransferDuring furnace tapping, an uncontrolledand variable amount of air is entrainedby the tapping steam and carried intothe ladle. This air entrainment has a di-rect bearing on fluid flow patterns in theladle, which in turn, determine whetheran alloy additive dissolves in the steel oris carried up to the surface to be lost byoxidation due to contact with the atmo-sphere. For this reason, the timing, lo-cation and method of deoxidizer andother alloy additions all influence alloyrecovery and final steel quality. As previ-ously mentioned, in addition toreoxidation, steel can pick up nitrogenand hydrogen from a humid atmosphereduring this transfer operation.

Improvements in furnace tappingoperations to minimize slag carryoverinto the ladle and to promote the for-mation of a more compact tapping

stream include the use of slidegate valvesand bottom tapping systems. With morecompact streams, protecting the steelfrom contamination with air, by gasshrouding the stream with a protectiveatmosphere should be easier.

When elements with a high affinity foroxygen and sulfur, such as calcium orrare earth elements, are added to theladle, the effectiveness of the added ele-ment is strongly influenced by the quan-tity and composition of the slag carriedover from the furnace. This furnace slagcan constitute a major source of oxygenthat then transfers to the steel duringladle processing. This transfer of oxy-gen from slag to metal is enhanced dur-ing any gas stirring operation that bringsfresh metal into contact with the slag. Inaddition, as the steel is poured from theladle, the interaction that occurs betweenthe descending slag layer and the refrac-tory ladle walls can produce a reactionproduct or glaze, which serves as a res-ervoir for oxygen and perhaps sulfur.These elements can then transfer intothe next batch of steel during ladle proc-essing and can adversely affect steel qual-ity. If there is substantial carryover of slagfrom the furnace, the iron oxide andmanganese oxide content, which is likelyto be high, then a reducing agent, maybe added to minimize the use of expen-sive deoxidizing agent or loss of alloyingelements that may be added to the steel.If phosphorus oxide is present in the slagcarried over from the steelmaking vessel,then during slag reduction or steel deoxi-dation, phosphorus may revert to the steelfrom the slag phase. Again, this is a de-stroyer of quality.

Continuous Monitoring of LadleMetallurgy OperationsIn recent years, acoustic technology andvibration analysis to monitor and controlsteelmaking processes have found in-creasing application. The sound emittedduring certain metallurgical operationsreflects the corresponding operating con-

ditions. Thus, the acoustic response maybe used to monitor and control the pro-cess. Accelerometers have been used todetect vibrations generated by gas-solidinteractions within the blast furnace, intorpedo cars during hot metal treatment,within steelmaking converters during de-carburization and slag foaming, and inladles, tundishes and submerged nozzlesduring steel processing. These applica-tions have been reviewed in more detailelsewhere.32

Studies conducted within the Hungar-ian steel industry have shown that withthe aid of a remote microphone, the des-ulfurization rate of hot metal or moltensteel during injection of appropriate re-agents can be measured by the change insound caused by the generation ofbubbles.33 We can also continuouslymonitor the desulfurization process bymeasuring, with an accelerometer, the vi-brations caused by gas injection. This tech-nology could also be used as an onlinedetection system to continuously moni-tor changes in the concentration of dis-solved oxygen during processing of mol-ten steel. An investigation of factors suchas temperature and surface tension,which influence the sound emitted bybubbles when gas is injected into a liq-uid, has been conducted by Zhang.34 Formore than three decades, the term “ladlemetallurgy” has been used to signify anumber of processes, which include: steelreheating, deoxidation, alloying, desulfu-rization and modification of inclusionsand various techniques for promotingthermal and chemical uniformity withinthe steel by injecting argon or nitrogenthrough porous refractory units locatedwithin the base of the ladle or throughsubmerged lances. The application ofvibration monitoring to control argonstirring during ladle processing hasbeen described by Kemeny,35 and byMinion et al.36

During treatment of steel in a ladle arcfurnace, minimizing the refractory attackcaused by arc flare is important. This is

28 DECEMBER 2002

achieved by ensuring that sufficient slagis present with appropriate foaming char-acteristics to enclose the arc discharge. Aunique technology has been developedby Kemeny et al. to control this processby monitoring the vibrations generatednot only by the rising gas bubbles, butalso by the arc flare (Figure 13).

The gas bubbling process is oftenhighly variable and frequently excessive.This can lead to slag entrainment, steelreoxidation and accelerated refractoryattack. These two separate processes –gas bubbling and arc flare – create pres-sure waves that can be independentlymonitored by measuring the vibrationson the outside steel shell of the ladle.Separate characteristic frequency rangescan be identified, which relate to the ris-ing gas bubbles and the arc flare. If thereis insufficient slag present, this will be in-dicated by a high reading on the arc flareindicator. As synthetic flux is added to theladle, the arc flare value will progressivelydecrease until an acceptable level of fluxaddition is achieved. Using this nonintru-sive, continuous monitoring technologyto independently control both the bub-bling and arc heating processes, the fol-lowing benefits have been achieved: con-sistent stirring practice, predictable alloyrecovery, less aluminum fade, predictabledesulfurization, clean steel with low total

oxygen content, less nozzle clogging, lessarc flare and slag overheating, improvedrefractory performance and reduced ar-gon consumption.

Several techniques have been used todetect slag carryover. These include elec-tromagnetic devices, weight measure-ments and vibration sensors. Because ofthe significant difference in density be-tween metal and slag, there is a changein momentum transfer to a submergedrefractory nozzle when slag is entrainedby the stream. This change in momen-tum is reflected in a change in vibrationalmovement of the nozzle and also thetundish.37-39 By continuously sensingthese vibrational signals with acceler-ometers, monitoring the transfer op-eration and detecting the onset ofslag carryover is possible. In some in-stances, detecting the onset of vortexformation before slag has left the ladleor tundish may be possible. This wouldpermit the operator to implement cor-rective measures before slag carryoveractually occurs.

Ladle to Tundish TransferRecognizing that the tundish, like theladle, is a metallurgical reactor involv-ing interactions between molten steel,slag and refractory phases, as well as theatmosphere, is important. It is now ac-cepted that the tundish is not only a dis-tributing vessel, but also a continuousreactor. This is particularly significantbecause the tundish links together theladle that is a batch reactor with the cast-ing molds, which are themselves con-tinuous reactors. A major challenge forthe steelmaker is to ensure that thetundish is in fact a continuous refinerand not a continuous contaminator. Inthis context, tundish slags play an im-portant role. There are three require-ments for a good tundish flux:

♦ Thermal insulation of the molten steel♦ Protection of the steel from reaction

with the atmosphere

♦ Absorption of nonmetallic inclusionsfrom the molten steel

Thermal insulation is best obtainedwith a solid powder layer. On the otherhand, prevention of reoxidation, nitro-gen and hydrogen absorption from theatmosphere by the molten steel, as wellas absorption of nonmetallic inclusionsfrom the steel, are all best achieved bythe presence of a liquid layer. Viscosityof the tundish flux is also important forproducing clean steel. If viscosity is toohigh, this will limit the ability of the flux toabsorb inclusions. If viscosity is too low,then flux may be drawn down into themold, particularly during ladle transferoperations, when conditions are un-stable. Tundish slags can now be de-signed using the concept of optical ba-sicity to obtain a flux that has therequired characteristics in terms ofchemical properties, as well as viscosity,to ensure the tundish will function as arefining vessel and not a contaminator.24

Methods of preventing reoxidationduring metal transfer between ladle andtundish and tundish and mold by ex-tended refractory nozzles and/or gas-shrouding devices are now well-estab-lished practices. However, significantcontamination can still occur duringladle-change operations, when exposureof steel to air for even a few seconds canlead to the casting of inferior steel overthe next several minutes. An increase inthe nitrogen content of the steel by a fewppm is a valid indicator that reoxidationhas occurred. Since the rate of oxygendissolution is about 200 or 300 times fasterthan that of nitrogen,6 analysis of samplesfor total oxygen can yield values of severalhundred ppm. This represents massivereoxidation and the generation of verylarge inclusions. Much of this reoxidationmaterial will separate in the tundish andcollect in the tundish flux. In so doing,iron oxide and manganese oxide con-tents of the tundish flux may be mark-edly increased.

Figure 13 Vibration monitoring of gasrinsing and arc flare during ladle-furnaceprocessing.

DECEMBER 2002 29

The influence of thermal effects on liq-uid steel flow in tundishes during a ladlechange operation and during steady stateconditions have been reported by Lowryand Sahai based on results from an el-egant one-third scale water modelingstudy, together with in-plant tracermeasurements, on a six-strand bloomcaster.40 From residence time distribu-tion diagrams pertaining to the actualsteel casting tundish, we can concludethat steel flow following a ladle change isradically different from that under almostisothermal, steady state conditions dur-ing the latter half of a ladle pour. Duringa ladle change, the new steel enteringthe tundish is at a higher temperaturethan the steel already present in thetundish. The new, higher temperaturesteel actually reaches the nozzles at theend of the tundish before it reaches thenozzles closest to the incoming ladlestream. This behavior is the reverse ofthat observed under steady state condi-tions. Steady state flow is reestablishedwhen the new steel displaces the previ-ous steel. This requires almost 2.5 timesthe mean residence time for the tundish.This time period could amount to 25percent or more of the ladle casting timedepending on tundish capacity and cast-ing rate. This is a significant portion ofthe casting time during which the metalflow is in a reverse mode. These effectswill influence factors such as alloy addi-tions, inclusion flotation and inter-grademixing and should be included in anyconsideration and analysis of tundishdesign for improvement in steel quality.

During the casting operation, ladleslag can be drawn by vortex formationinto the tundish as the metal level in theladle decreases. Again, this is an ex-ample in which slag that is a refiningagent in one vessel becomes a contami-nating agent in the next. Some of thisentrained slag may enter the tundishflux and consequently alter its behav-ior, while other portions may be car-ried over into the mold and end up

adversely affecting the performance ofthe mold flux or generating defects inthe final product. From flotation consid-erations, there may also be a higher con-tent of inclusions in the last portion ofsteel to leave the ladle. In a typical ladlechange operation, total oxygen contentin the tundish can change quite drasti-cally, both at the entrance to the tundishclose to the incoming ladle stream andat the exit above the mold. In addition,the weight of metal in a tundish understable operating conditions may dropby about half during the ladle changeoperation. In turn, this can adverselyaffect the quality of the steel enteringthe casting molds. While the ladlechange may require approximately twominutes, the effect on the quality of thesteel leaving the tundish could extendfor more than 15 minutes.41

Tundish to Mold TransferDuring the past decade, studies byBrimacombe et al. have generated in-creasing awareness of the critical impor-tance of “meniscus control” because ofthe direct correlation that has been es-tablished between fluid flow behavior atthe meniscus and surface quality of thecast products.42 When the metal deliverysystem involves the use of submerged-entry nozzles (SEN), the behavior of fluidflow in the mold, as well as at the menis-cus, is extremely important. The designof the nozzle, as well as the depth of sub-mergence, directly influences fluid flowwithin the casting mold, behavior at themeniscus, vortex formation and distri-bution and location of inclusions in thefinal product.

To minimize nozzle blockage andenhance the flotation of nonmetallic in-clusions, argon gas is sometimes in-jected into the SEN. This can have a pro-nounced effect on fluid flow behavior. Ifargon injection is excessive, mold slagcan be drawn into the molten steel andend up as defects in the product. Evenin the absence of argon injection, if the

submerged entry nozzle has not beenproperly fired and is porous, or if cracksare present in the nozzle wall caused bydamage during handling, or if cracksform during use, air can be drawn intothe nozzle. The oxygen from this willcause reoxidation, while the nitrogen willbehave essentially like argon and causeconsiderable disturbance of the moldpowder at the meniscus. In addition, ifthe formation of reoxidation material isexcessive, the composition of the moldflux can change, and this can adverselyaffect subsequent performance.

The behavior of synthetic slags in thecasting mold is a critical parameter forthe production of quality steel. This isthe last slag with which the molten steelhas contact before solidification. All ofthe previous work that has been con-ducted in the furnace, ladle and tundishto generate quality steel can be destroyedin the casting mold. For this reason, themold can be considered the “heart” ofquality, and the mold slag must be de-signed to perform a range of functions,which is considerably greater than thatrequired of any of the previous slags inthe furnace, ladle or tundish. These func-tions include prevention of reoxidation,provision of thermal insulation, availabil-ity as a flux to dissolve impurities, per-formance as a lubricant and possessionof good heat transfer characteristics toenhance solidification. The interactionof the mold flux with refractory nozzles,and the behavior with respect to the ab-sorption of hydrogen from the atmo-sphere and subsequent pickup of hy-drogen by the steel, are also importantconsiderations when designing the mostappropriate composition of the flux fora particular application.

In the case of small billets, reoxida-tion of mold streams by direct contactwith air, will result in the formation ofoxide slag within the mold. If this oxidematerial is drawn into the solidifying steel,it can produce large centerline defects.If it is entrapped as a solid crust between

30 DECEMBER 2002

the solidifying steel shell and the moldwall, then large surface slag defects canbe generated. While open streams canbe protected by gas shrouding systems,differences in stream character be-tween strands, or even instability at dif-ferent times on the same strand, canadversely affect meniscus behavior,fluid flow patterns, surface quality, pin-holes and the location of inclusions inthe final product.

With multistrand billet casting and acentral ladle stream, smooth streams areusually found at the ends of the tundish.Such streams entrain very little air andthe steel meniscus is quiescent with a slowmotion from the outside of the mold to-ward the stream. In this case, mold scumwill float on the surface but could be drivendeep within the steel pool by the incom-ing tundish stream. Rough streams, suchas those issuing from tundish nozzles lo-cated nearest to the ladle stream, have

not only a higher surface to volume ratiothan a smooth stream, but will also en-train more air, causing increased turbu-lence in the mold and excessive splash-ing at the meniscus (Figure 14). Withina turbulent mold pool, new steel iscontinuously brought to the surface,where further contact with the air canoccur. Turbulent pools provide little op-portunity for proper separation ofreoxidation products. These oxides, inthe form of mold scum, are thrown tothe outside of the mold, where they canbe entrapped in the freezing interfaceat the surface of the billet and give riseto surface slag defects.

In addition to differences in streamcharacteristics between one strand andanother, a major concern with respect tosteel quality is the variability in characterof a specific stream during the castingoperation caused by fluid flow changeswithin the tundish. From experiments

conducted with various tundish nozzledesigns, it has been found that steelstreams of consistent quality could beobtained with regular nozzles possess-ing a logarithmic-type entry modifiedto include eight deep slots or castellationsin the upper rim. With these castellatednozzles the appearance of the tundishstreams remained unchanged by con-ditions in the tundish.(4)

Based on recent advances in the un-derstanding of electromagnetic forcefields and their effects on liquid metalflow, together with parallel developmentsin induction power systems, Beitelmanand co-workers have successfully pio-neered a new system for steel flow con-trol in the mold during billet casting. Thetechnology is applicable for open streampouring between the tundish and moldand also in cases where submerged en-try nozzles are used to protect the steelfrom reoxidation.43

This novel in-mold electromagneticstirring system (M-EMS), incorporates twosets of stirring coils with independent con-trol of each of the electromagnetic fields(Figure 15). The stirring coil in the lowerregion of the mold (M-EMS) provides themain stirring action while the coil in themeniscus region has a more localized ef-fect and is termed the auxiliary coil-stir-ring modifier (AC-SM). Depending on the

Figure 14 Smooth tundish streams with little air entrainment, quiescent conditions at themeniscus, no splashing and deep penetration in the mold pool compared with rough tundishstreams, massive air entrainment, turbulent conditions at the meniscus, considerablesplashing, droplet formation and shallow penetration within the mold pool.

Figure 15 Dual-coil electromagneticstirring system for billet casting.44

DECEMBER 2002 31

casting requirements, the stirring inten-sity at the meniscus can be enhanced,diminished or essentially eliminated byadjusting the current input to the AC-SM and controlling the direction of themagnetic field relative to that of the M-EMS.44 Since solidification of the steelskin begins at the meniscus, excessivemotion at the meniscus will produceripple marks on the solidified billet sur-face, which serve as “finger prints” or“witness marks” reflecting conditionsat the meniscus at the time solidifica-tion was initiated.

Chaotic movement at the meniscuswill produce billets with rough irregularwave marks that are frequently associ-ated with surface cracks and in the caseof submerged pouring practice, accel-erated wear of the SEN, mold powderentrapment and surface defects. On theother hand, a calm meniscus will pro-duce a smoother surface on the solidi-fied steel billet and reduced refractorywear of the nozzle.

Since this system provides indepen-dent control of liquid steel flow at themeniscus and within the lower region ofthe mold, electromagnetic stirring con-ditions can be optimized to achieve maxi-mum benefits in each of these locationsfor both open stream and submergedpouring practices.

Operating experience at Ispat Sidbecin Canada, as well as several other plants,has confirmed that the dual-coil systemhas major beneficial implications withrespect to the quality of continuously caststeel, as well as productivity.44

This innovative casting technologypromotes good surface quality, reducesthe incidence of pinholes, decreasessegregation of elements such as car-bon, sulfur and phosphorus and dimin-ishes crack formation and center linedefects. In addition, by preventingoverstirring at the meniscus, mold pow-der entrapment, surface defects andaccelerated wear of the SEN can beavoided.45

INNOVATIVE DEVELOPMENTSTHROUGH SYNERGISTICPARTNERSHIPS“There must be builders of bridges be-tween the creators of knowledge andthe creators of wealth.”1996 Campbell Memorial Lecture ASMInternational42

On our technological journey throughthe 21st century, we will cross a numberof bridges that will span our knowledgegaps, bringing a new understanding ofexisting processes and permit the devel-opment of new processing routes, newdiagnostic systems, new products andnew partnerships between people.

In this context of collaborative inter-actions, basic studies should be per-formed by the research community inclose cooperation with the user com-munity and the resource community, theprovider of equipment and materials(Figure 16).

We must recognize and acknowledgethat these three communities representquite distinct cultures, with very differentobjectives, and collaboration is not alwaysan easy task. Nevertheless, when effectivecommunication is established betweenthe members of this triumvirate, there is apowerful driving force for the synergisticdevelopment of innovative technologies.

We cannot overemphasize that con-siderable research, together with in-plant evaluation, is required to transform

a fundamental scientific concept into areliable operating system.

In this collaborative process, the ex-change of knowledge between the dif-ferent communities is perhaps the mostdifficult step since it not only requires awillingness to share knowledge, butalso an ability and a willingness to re-ceive knowledge.

In the knowledge exchange process,the receptor capacity of individuals is fre-quently the rate-limiting step.

To promote and encourage the ex-change of knowledge and experiencebetween the laboratory and the work-place and facilitate development and useof new products and processes, well-trained people will always constitute thecritical link (Figure 17).

In the end, the preeminent aim of col-laborative activities between our educa-tional institutions, industrial organiza-tions, government funding agencies andprofessional societies, must be to ensurethe availability of high quality people.

ROLE OF THE PROFESSIONALSOCIETY“We must also continue to buildbridges with other societies with com-mon goals, all for the benefit of ourmembers, our industries, and the glo-bal village. Breaking down the wallsbetween solitudes is never easy but itis the only way.”A Letter to the Membership46

Figure 16 Triumvirate for collaborativepartnerships.

Figure 17 High quality people andinnovative developments.

32 DECEMBER 2002

For many years, the ISS has played apivotal role in the organization of inter-national meetings. The promotion ofthis kind of interaction is perhaps oneof the most worthwhile endeavors withwhich a professional society can be in-volved. It is certainly an activity that wasstrongly endorsed and encouraged byBrimacombe. Throughout his career,and particularly during his tenure aspresident of three professional societ-ies, Brimacombe emphasized theimportance of good relationships andcommunication between different pro-fessional groups within North America,as well as those located in countries over-seas. In the spirit of technical exchange,he played a major role in organizing jointconferences, international symposia andintensive courses designed specificallyfor industrial personnel (Figure 18). Ameasure of the success he achieved inthese endeavors may be deduced fromthe many honors and distinctions he re-ceived from professional institutionsthroughout the world.

In addition to his scientific accom-plishments, and they were many,Brimacombe had a tremendous inter-est in and concern for educational is-sues as confirmed by his presidency ofthe TMS Foundation, his presidency ofthe Canadian Foundation for Innovationand his encouragement of the forma-tion of joint student chapters. During al-most three decades at UBC, Keith greatly

Figure 18 Activities of the professionalsociety.

influenced the training of numerousyoung engineers, as well as visiting re-searchers from universities and indus-trial organizations, many of whom nowoccupy prominent positions withinacademia and industry around theglobe. In the final analysis, perhaps ourgreatest challenge is to ensure the avail-ability of men and women with a soundunderstanding of the fundamental as-pects and practical implications of theirdiscipline and who are fully equippedwith the behavioral attributes and com-municative skills that will enable them toapply their knowledge with wisdom andintegrity throughout their professionalcareers within this most challenging andsatisfying field of activity, the science andtechnology of steelmaking.

JAMES KEITH BRIMACOMBE– BRIDGE BUILDER PAREXCELLENCE“Passing the torch to the next genera-tion of materials professionals beginswith lighting the candle of dreams andall that is possible in our talentedyoung people.”Presidential Perspective; Passing theTorch47

Throughout his career, Brimacombe wasa great builder of bridges:

♦ Bridges of education to enhanceknowledge exchange

♦ Bridges of innovation to transform in-dustrial operations

♦ Bridges of friendship to facilitate com-munication between different cultures

Regardless of whether we work in in-dustry, government or academia, all ofus have the great privilege of participat-ing in this bridge-building process.Therefore, let us build with vision, en-thusiasm, integrity and wisdom.

In this way, the torch will be passed onfrom this generation to the next, and to-gether we can ensure that the shiningflame of Brimacombe’s scientific andphilosophical legacy will brighten theroad that leads to knowledge, enlightenthe pathway called wisdom and set ablazethe imagination and minds of the youngengineers and teachers of tomorrow ontheir journey through the 21st century.By these endeavors, we will honor andcommemorate the highest standards ofachievement and professionalism so wellexemplified by Brimacombe throughouthis distinguished career.

References1. J.K. Brimacombe, “Empowerment with Knowledge

– Toward the Intelligent Mold for the ContinuousCasting of Steel Billets,” 1993 Howe Memorial Lec-ture, Metallurgical Transactions, Vol. 24B, 1993,pp. 917-935.

2. J.K. Brimacombe, I.V. Samarasekera and S.L.Cockcroft, “Computer Simulation of Solidificationand Casting Processes,” Proceedings of the Inter-national Conference on Computer-Assisted Materi-als Design and Process Simulation (COMMP ‘93),ISIJ, 1993, pp. 102-110.

3. J. Little, M. Van Oosten and A. McLean, “Factors Af-fecting the Reoxidation of Molten Steel During Con-tinuous Casting,” Canadian Metallurgical Quar-terly, Vol. 7, 1968, pp. 235-246.

4. W.J. Maddever et al., “An Investigation of CastingStreams,” Canadian Metallurgical Quarterly, Vol.12, 1973, pp. 79-88.

5. L.A. Greenberg and A. McLean, “Nitrogen Pickup inLow Sulfur Steel,” Steelmaking Conference Pro-ceedings, ISS, Vol. 63, 1980, pp. 201-208.

6. L.A. Greenberg and A. McLean, “The Kinetics of Oxy-gen Dissolution in Liquid Iron and Liquid Iron AlloyDroplets,” Transactions of the Iron and Steel In-stitute of Japan, Vol. 14, 1974, pp. 395-403.

7. J.D. Young, A. McLean and I.D. Sommerville, “Gas-Slag-Metal Reactions in the Ladle, Tundish andMold,” Japan-Canada Seminar on Secondary Steel-making, CSIRA-ISIJ, Proceedings Volume, 1985, pp.C4, 1-15.

8. J.K. Brimacombe, “Presidential Perspective: ThePower of Knowledge,” Journal of Metals, Vol. 45,No. 11, 1993, p. 74.

9. Y. Hara et al., “Development of Reoxidation Preven-tion Method to Produce Clean Steel,” Current Ad-

Figure 19 Bridging the gap.

DECEMBER 2002 33

vances in Materials and Processes, ISIJ, Vol. 4, 1991, pp.1,202-1,205.

10. T. Ogura et al., “Activity Determinator for the AutomaticMeasurements of the Chemical Potentials of FeO in Met-allurgical Slags,” Metallurgical Transactions, Vol. 23B,1992, pp. 459-466.

11. M. Iwase et al., “Measurements of the Ferrous OxideActivities in Slags to Reduce Steelmaking Slag Volume,”High Temperature Materials and Proc-esses, Vol. 20, Nos. 3-4, 2001, pp. 269-277.

12. V. Espejo and M. Iwase, “A Thermodynamic Study of CaO+ Al

2O

3 + FexO at 1,673 Degrees Kelvin,” Metallurgical

Transactions, Vol. 26B, 1995, pp. 257-264.13. H. Hoshino and M. Iwase, “Chemical Potentials of Com-

ponents in the System CaO + P2O

5 + FexO,” Metallurgi-

cal Transactions, Vol. 27B, 1996, pp. 595-603.14. M. Iwase, T. Ogura and R. Tsujino, “Automatic FeO Activity

Determinator for Slag Control,” Steel Research, Vol. 65,pp. 90-93.

15. A. Hamm et al., “The Use of Slag Iron Oxide ActivityMeasurements for Steelmaking Process Control,” Steel-making Conference Proceedings, ISS, Vol. 77, 1994, pp.153-156.

16. K. Oh-nuki, T. Isono and K. Umezawa, “Effect of Slag Com-position on the Activity Coefficient of FeO,” Current Ad-vances in Materials and Processes, ISIJ, Vol. 6, No. 1, 1993,p. 220.

17. J.A. Duffy and M.D. Ingram, “Establishment of an OpticalScale for Lewis Basicity in Inorganic Oxyacids, MoltenSalts and Glasses,” J. Am. Ceram. Soc., Vol. 93, 1971, pp.6,448-6,454.

18. J.A. Duffy, M.D. Ingram and I.D. Sommerville, “Acid/Basic Properties of Molten Oxides and Metallurgical Slags,”J. Chem. Soc. Faraday Trans., Vol. 74, No. 6, 1978, pp.1,410-1,419.

19. J.A. Duffy, “Use of Refractivity Data for Obtaining OpticalBasicities of Transition Metal Oxides,” Ironmaking andSteelmaking, Vol. 19, No. 6, 1989, pp. 426-428.

20. T. Nakamura, Y. Ueda and J.M. Toguri, “A New Develop-ment of the Optical Basicity,” Trans. JIM, Vol. 50, 1986, pp.456-461.

21. D.J. Sosinsky and I.D. Sommerville, “The Composition

and Temperature Dependence of the Sulfide Capacity ofMetallurgical Slags,” Metallurgical Transactions, Vol.17B, 1986, pp. 331-337.

22. A. Bergman, “Some Aspects on MgO Solubility in Com-plex Slags,” Steel Research, Vol. 60, No. 5, 1989, pp. 191-195.

23. R.W. Young et al., “Use of Optical Basicity Concept forDetermining Phosphorus and Sulfur Slag-Metal Parti-tions,” Ironmaking and Steelmaking, Vol. 19, No. 3,1992, pp. 201-219.

24. Y.D. Yang et al., “Water Vapor Solubility in Tundish Slags,”J. Iron & Steel Res. Int., Vol. 5, No. 2, 1998, pp. 7-14. Also,Second International Conference on ContinuousCasting of Steel, Chinese Society for Metals, 1997, pp.113-120.

25. Y.D. Yang et al., “The Correlation of Alkali Capacity withOptical Basicity of Blast Furnace Slags,” Iron & Steelmaker,ISS, Vol. 27, No. 10, 2000, pp. 103-111.

26. E. Wei et al., “The Fundamentals of Using Custom-De-signed Slags to Remove Impurities from Liquid Steel,”Electric Furnace Conference Proceedings, ISS, Vol. 60,2002, pp. 761-771.

27. H.M. Howe, “The Metallurgy of Steel,” Scientific Publish-ing Co., New York, 1890.

28. N.G. Kournetas et al, “Effects of Post-Combustion Validatedby Steel Mill Demonstration,” Electric Furnace Confer-ence Proceedings, ISS, Vol. 55, 1997, pp. 373-380.

29. S.H. Anderson, E.J. Evenson and N.G. Kournetas, “EAFOptimization Through Continuous Offgas Analysis,” SteelTimes International, No. 5, 1997.

30. S.H. Anderson, “New Technology Developed for Optimi-zation of Electric Arc Furnace Fume Systems,” Environ-mental Science and Engineering Magazine,” No. 7,1997, pp. 56-57.

31. E.J. Evenson, H.D. Goodfellow and J. Guerard, “EnergyOptimization and Continuous Fume Analysis at Co-SteelLasco,” Electric Furnace Conference Proceedings,”ISS, Vol. 55, 1997, pp. 435-453.

32. X.F. Zhang, A. McLean and I.D. Sommerville, “PotentialApplications of Acoustic Techniques in Iron and Steelmak-ing,” Steelmaking Conference Proceedings, ISS, Vol.74, 1991, pp. 659-662.

33. G. Pal et al., U.S. Patent No. 4,617,830 (issued Oct. 21,1986).

34. X.F. Zhang, Ph.D. Thesis, University of Toronto, 1991.35. F.L. Kemeny, “Tundish Nozzle Clogging – Measurement

and Prevention,” Alex McLean Symposium, Proc., ISS,1998, pp. 103-110.

36. R.L. Minion et al, “Improved Ladle Stirring Using Vibra-tion Technology at Stelco Hilton Works,” Iron &Steelmaker, Vol. 25, No. 7, 1998, pp. 25-31.

37. T Itoh et al., Transactions of the Iron and Steel Insti-tute of Japan, Vol. 22, 1982, B90.

38. F. Fitzgerald, Steel Research, British Steel Corp., June1983.

39. M. Downey et al., Proceedings of the 6th InternationalConference on Refining Processes, Scaninject VI, Luleå,Sweden, 1992, pp. 489-500.

40. M.I. Lowery and Y. Sahai, “Thermal Effects on the Flow ofLiquid Steel in Continuous Casting Tundishes,” Iron &Steelmaker, Vol. 19, No. 3, 1992, pp. 81-86.

41. H. Kimura, “Current Advances in Materials and Proc-esses,” ISIJ, Vol. 4, 1991, pp. 1,198-1,201.

42. J.K. Brimacombe, “The Challenge of Quality in ContinuousCasting Processes,” 1996 Edward DeMille CampbellMemorial Lecture ASM International, Metallurgical andMaterials Transactions, Vol. 30B, 1999, pp. 553-566.

43. L. Beitelman, “Continuous Casting of Steel Billets with In-Mold Dual-Coil Electromagnetic Stirring System,” Elec-tromagnetic Processing of Materials, InternationalCongress, Paris-La Defense, Vol. 2, 1997, pp. 335-341.

44. J. Limoges and L. Beitelman, “Continuous Casting of HighCarbon and Alloy Steel Billets with In-Mold Dual-Coil Elec-tromagnetic Stirring System, Iron & Steelmaker, ISS, Vol.24, No. 11, 1997, pp. 49-57.

45. L. Beitelman, “Flexible Control of Stirring in a Continu-ous Casting Mold with a Dual-Coil System,” ElectricFurnace Conference Proceedings, ISS, Vol. 55,1997, pp. 333-340.

46. J.K. Brimacombe, “A Letter to the Membership,” Jour-nal of Metals, Vol. 44, No. 12, 1992, pp. 41-42.

47. J.K. Brimacombe, “Presidential Perspective: Passingthe Torch,” Journal of Metals, Vol. 45, No. 9, 1993,p. 66.