Dissertations / Theses on the topic 'Ladle metallurgy'

Extensive computer predictions have been carried out by the author to study flow, addition dispersion and particle motion during central gas injection into cylindrical vessels. In conjunction with numerical computations, experiments were conducted in a 0.30 scale water model of a 150 ton steel processing ladle, using a Froude number scaling criterion. Two typical gas injection configurations (i.e., conventional central injection and C.A.S. alloy addition procedure) were investigated.
Flow visualization studies were carried out using a suspended network of silken threads, mean velocity vectors and overall flow patterns were determined by video recording techniques, while mean velocity vectors and associated turbulence level were also measured with laser doppler velocimetry. These measurements show very reasonable agreement with equivalent numerical predictions.
To simulate the subsurface motion of additions, spherical wooden balls of various densities were dropped from typical heights, and their subsurface trajectories, immersion times, etc., recorded by means of a video recorder. Frame by frame analysis of the video tapes showed trends which are in good accord with computed trajectories.
Mixing times of simulated molten additions were measured by the conductivity measurement technique. These were compared with prediction from an equivalent tracer dispersion model and excellent agreement achieved.
For industrial application, flow, particle motion, and mixing times in a 150 ton steel processing ladle have been predicted and their technological significance discussed.

The performance characteristics of a tundish, such as the flotation of inclusions and of slag entrainment, are largely influenced by the fluid flow phenomena. Physical and mathematical modeling in water is widely used to understand fluid flows in a tundish, and as a tool to improve, control and design procedures for high quality steel processing operations. These approaches were used to study the performance of fluid flow for a new design of ladle shroud. The new "Dissipative Ladle Shroud" (DLS) was studied, using a one third scale, delta shaped, four-strand tundish. The results were compared with those achieved with the standard ladle shroud. The standard k-ε turbulence model, together with the continuity and momentum equations, were employed in order to analyze, and visualize, the velocity fields generated within the tundish operating at steady state. Different cases have been analyzed, including a conventional ladle shroud (LS) with a bare tundish, and a tundish furnished with an impact pad. Similarly, the new design of the shroud (DLS) was studied under equivalent conditions. The physical experiments included the use of Particle Image Velocimetry (PIV) and conductivity tracer techniques. The PIV measured the instantaneous velocities at the outlet of the DLS and the LS at different flow rates, showing the detailed jetting characteristics of water leaving the two types of ladle shroud. Residence Time Distribution (RTD) curves were also obtained for the different flow arrangements previously mentioned, and the dispersion of a colored dye tracer was observed at different intervals of time during tundish operation, and analyzed using the video visualization technique.
Les caractéristiques de performance d'un répartiteur telles la flottaison des inclusions et de la scorie sont largement influencées par les phénomènes d'écoulement des fluides. Les modèles physiques et mathématiques sont communément utilisés pour comprendre la dynamique des fluides dans un répartiteur et pour améliorer, contrôler et concevoir les procédés de traitement d'acier de haute qualité. Cette approche a été utilisée pour étudier la performance de la dynamique des fluides avec un nouveau design de poche de couleé. La nouvelle "Dissipative Ladle Shroud" (DLS) a été étudiée en utilisant une échelle d'un tiers de répartiteur, en forme delta et possédant quatre drains de couleé. Les résultats ont été comparés avec ceux obtenues avec le poche de couleé normal. Le modèle de turbulence k-ε, avec les équations de conservation des masses et de quantité de mouvement, a été utilisé pour analyser et visualiser les champs de vitesse générés par l'opération du répartiteur en conditions stables. Différents cas ont été analysés, incluant un "conventional ladle shroud" (CLS) avec un répartiteur vide, et un répartiteur muni d'un bloc d'impact. De même, le nouveau design de poche de couleé (DLS) a été étudié sous des conditions équivalentes. Les expériences physiques incluent l'utilisation du "Particle Image Velocimetry" (PIV) et de techniques de traceur de conductivité. Le PIV mesure la vitesse instantanée à l'issue du DLS et du CLS à différents taux de flux, en montrant les caractéristiques détaillées du jet d'eau quittant les deux types de poche de couleé. Des courbes de Distribution des Temps de Résidence (RTD) ont également été obtenues pour les différents arrangements de flux précédemment mentionnés, et la dispersion d'un traceur de teinture colorée a été observée et analysé à différent intervalles de temps pendant l'opération du répartiteur en utilisant une vidéo de visualisation technique.

The onset of a 'late' rotating vortex over an off-centre drain nozzle at 2/3 radius was studied in an 1160-mm diameter tank. It was found that using a sloped bottom ladle could be beneficial in terms of steel yield, provided the exit nozzle is located 'centrically'.
Miner modification of the nozzle (skewed nozzle) to impart a radial component of velocity to the spinning vortex core was found to be effective in making AMEPA system sensitive to early slag entrainment phenomena by diverting the core away from the central vertical axis of the nozzle.
A 0.75 scale water model was constructed to simulate the flow of liquid steel through a ladle shroud in the presence of gas infiltration. It was found that the ladle shroud slag detector could be temporarily 'blinded' by gas bubbles or permanently blinded by a standing submerged gas jet.

Cleanness and uniformity in steel properties are important for high quality steel. Physical and/or mathematical models can be used in order to achieve optimum conditions for the clean steel during steelmaking processes. In the present study, important metallurgical transport phenomena in steelmaking ladles and tundishes have been investigated using both mathematical and physical (water) models.
Through appropriate solutions of the Navier-Stokes equations, the intermixing of fluid within gas-stirred ladles can be modelled quite satisfactorily. It is shown that off-centered bubbling gives the most consistent results in terms of minimising mixing times, since angular velocity components intermix fluid across the width of a ladle. Comparisons between mathematical and experimental data are presented.
Fluid flow, heat transfer and inclusion flotation have been modelled mathematically for testing the behaviour of several tundish designs. Computations are presented to illustrate the importance of thermal natural convection currents in mixing the upper and lower layers of steel. Particle removal rates are also experimentally studied with the aid of the novel E.S.Z. (Electric Sensing Zone) system, and compared with computational results.

The present work was to study the role of ladle glaze as apotential supplier of inclusions to the steel melt during theladle refining process. In this study, the total number ofinclusions at the beginning and at the end of the ladletreatment process was found to be increasing with ladle age,which is the number of heats, the ladle being used. Asubstantial increase in inclusion population was noticed aftera certain ladle age.

Totally four types of inclusions named as; type-1 (MgO),type-2 (spinel), type-3 (an oxide solution) and type-4 (spinelin the center surrounded by the oxide solution of type-3) wereobserved in the beginning of the ladle refining process.Thermodynamic calculation revealed that the type-3 and type-4inclusions were generated by the reactions between EAF slag andladle glaze. Even a part of inclusions of type-2 (spinel phase)could be formed by these reactions. Three types of inclusionswere found before casting, viz. type-5 (oxide solution with lowcontents of MgO and SiO2), type-6 (small MgO islands embeddedin an oxide solution) and type-7 (spinel in the centersurrounded by the oxide solution of type-5). Inclusions of bothtype-5 and type-7 were the products of the reaction betweeninclusions of type-2 and the liquid metal. On the other hand,the occurrence of pieces of MgO having sharp edges in the oxidesolution suggested that the type-6 inclusions were generated bythe ladle glaze.

A preliminary examination on the possibility of inclusionseparation by bubble floatation, experiments using cold modelswere also carried out. De-ionised water and silicon oil wereused as the bulk phase. Charcoal particles of different sizeranges were employed as the dispersed phase. The examination ofcharcoal-water-gas system indicated that the positivefloatation coefficient is not a sufficient condition for theinclusion separation. The experimental results were found to bein contradiction with the prediction of a typical model thatconsiders interfacial energies. The omitting of the drag forcewas believed to be the reason causing the failure of the modelprediction in the charcoal-water-gas system. The failure of themodel prediction suggested a need of a new model taking intoaccount interfacial energies, drag force, buoyancy force andgravity force.

Key words:oxide inclusions, ladle metallurgy, ladleglaze, inclusion population, ladle age, interfacial tension,inclusion separation

The present study was carried out to investigate theformation and chemical development of non-metallic inclusionsduring ladle treatment of steel.

To begin with, an investigation of the deoxidation processand the impact of aluminium addition was carried out. For thispurpose, a new experimental setup was constructed. The setupallowed the examination of the deoxidation process as afunction of time by using a quenching technique. Preliminaryexperiments showed that homogeneous nucleation of alumina tookplace in the areas supersaturated with aluminium. Theseexperiments also showed that agglomeration of alumina particleswas a very rapid process, even when the convection in the meltwas negligible.

In order to examine whether aluminium supersaturation occursduring industrial practice, the deoxidation of liquid steelwith aluminium wire injection in a gas-stirred ladle wassimulated by mathematical modeling using a Computational FluidDynamics (CFD) approach. The results showed that theconcentration of aluminium in the vicinity of the aluminiumwire injection was high enough to generate homogeneousnucleation of alumina.

Aiming at an understanding of the inclusion chemistry duringthe ladle process, an industrial study was performed atUddeholm Tooling AB. The impact of slag-lining reactions andladle glaze as a source of inclusions in the melt was alsostudied. The experimental results were analysed from athermodynamic viewpoint to gain an insight into the origins ofthe inclusions and their changes along the process of the ladletreatment. Six types of inclusions were found in the steel. Thetypes of inclusions present varied along the history of theladle treatment. Three types of inclusions were found in theliquid steel before deoxidation, namely type A (a liquidinclusion with high SiO2 concentration), type B (spinel) andtype C (a combination of type A and type B). Thermodynamicanalysis indicated that these types of inclusions could begenerated by the reaction between the Electric Arc Furnace(EAF) slag and the ladle glaze, during the filling of theladle. The addition of aluminium resulted in the formation ofalumina inclusions (type E), which agglomerated and separatedfrom the steel very fast. The spinel inclusions of type B werefound to be unstable at low oxygen potentials. The inclusionsof this type would react with the liquid metal forming theinclusions of type F (spinel in the centre surrounded by anoxide solution containing Al2O3, CaO and MgO). Further,reaction between the liquid metal and inclusions of type Fwould result in the inclusions of type G, an oxide solutionmostly consisting of Al2O3 and CaO with small amounts of MgOand SiO2. The inclusions of type G were the only inclusionsfound in the steel before casting.

Ç
EMTAS Steel Plant produces alloy steels used principally by the automotive industry. Demands of customers for cleaner steel with &ldquo
fewer, smaller and homogenously distributed inclusions&rdquo
are getting tighter with time. The tight demands of the customers are checked by ultrasonic testing and faulty regions are cut off and scrapped which results in production losses in steel plants. Decreasing production losses due to scrapping by producing cleaner steels is therefore a major concern. The objective of this study is to investigate factors affecting the cleanliness of the steels produced by Ç
EMTAS. With this objective, the steelmaking practice used was thoroughly investigated and the possible origins of inclusions in the final product was tried to be found. Based on the findings, process parameters like temperature, time, quantities of additives to steel, intensity of stirring etc. were varied with the objective of determining the optimum production practice.

The aim of this research was mainly to investigate desulfurization behaviour of ladle furnace slags of a low-sulfur steel. First, the change of the amount of unstable oxides (FetO and MnO) in slags was studied from the converter till the end of ladle furnace treatment for different steel quality groups. It was found that the change in the amount of unstable oxides could be examined best with low-C steel qualities. Then, the change of activity of iron oxide was investigated for a member of low-C steel quality group, namely 4937K, a low-S steel. The relation between the degree of desulfurization at the ladle furnace, % DeS (measured), and the change in activity of iron oxide, % Decrease in aFetO, was examined first using an empirical expression obtained from literature. Then, this relationship was studied using steel oxygen activity measurements conducted at the ladle furnace. With this method, a regression formula was proposed for 4937K slags relating activity coefficient of iron oxide, gama(FetO), with the major slag components. The results obtained from steel oxygen activities were found to be representative for 4937K steel-slag system showing similar behaviour with the ones from the empirical expression obtained from literature. Similar results were found using Temkin equation and the polymeric anion model. In relation to formation of less harmful inclusions to suppress adverse effects of sulfur, morphological studies of some steels produced in ERDEMiR plant were performed. D-type globular inclusions with severity values of 1-2 were observed.

The ladle furnace process is an important process in the steel manufacturing industry. The purpose of this process is to optimize the composition of the elements in the melt as well as to homogenize the temperature in the liquid. It is common practice to model this process using smaller water models. In order to accurately scale these models a variety of criteria and scaling factors are needed. The central phenomenon which all else is derived from is the two-phase gas plume dominating the fluid flows. The plume, and its dependant parameters are difficult to define. Which ones ought to be used and how to use them has not been standardized. Concerns have been raised whether the most common method of scaling is even applicable in ladle metallurgy. This report gives an account for studies concerning these variables and their effect on the subject. The objective of this report is to highlight ways to improve these simulations with respect to debated parameters. The conclusion of this study points out the reasons for why these variables may be of importance for the modeling of the ladle furnace process. It also specifically mentions future work that should be conducted in order to provide deeper knowledge of thedifferent parameters affecting the method of modeling.
Skänkmetallurgin är en viktigt process inom stålindustrin. Syftet med denna process är att optimera den kemiska sammansättningen i smältan och att homogenisera temperaturen i vätskan. Det är vanligt att modelera denna process med hjälp av vattenmodeller. För att träffsäkert skala dessa modeller krävs en mängd kriterier och skal-faktorer. Det mest centrala fenomenet, utifrån vilket allt annat kan härledas, är två-fas gasplymen som dominerar flödena i skänken. Plymen och dess beroende parametrar är svåra att definera. Vilka som bör användas och hur de används har inte standardiserats. Oro har väckts över om den vanligaste skalningsmetoden ens går att använda i skänkmetallurgi. Denna rapport redogör för studier rörande dessa variabler och deras påverkan på ämnet. Syftet med denna rapport är att belysa olika tillvägagångssätt till att förbättra dessa simulationer med hänsyn till debaterade parametrar. Slutsatsen för denna studie lyfter fram anledningarna till varför dessa variabler är av vikt för modellering för skänkmetallurgin. Även framtida arbete som bör utföras föratt ge djupare förståelse för de olika parametrarna belyses.

In this study it was aimed to analyse refining processes such as decreasing reducible oxide content of ladle slag with affecting parameters in low carbon aluminum killed grades and for the research Erdemir low carbon steel grades 7112K and 7110K are selected. There was a negative correlation between reducible oxide amount in ladle slag and desulphurization capacity of ladle slag with metal-slag reaction and steel internal cleanliness. To refine these properties of slag, aluminium dross, which was aluminium production discard and has a metallic content around %30-35 was used and after ladle treatment operation, decrease in reducible oxides such as FetO, MnO, SiO2, P2O5 was analysed. After the study it was observed that 653 kg. of converter slag leaked during tapping of steel and SiO2 ve P2O5 content of ladle slag had negligible change after ladle treatment. According to the results, it is observed that initial %10-12 (FetO + MnO) content was reduced to % 4.5-5.0 (FetO + MnO) after ladle treatment with use of aluminium dross. Beside of this, in order to see the effect of this slag reduction on steel cleanliness, low carbon aluminium killed grades were compared with ultra-low carbon aluminium killed grades having (FetO + MnO) content of %16-17 in slag. It was seen that reoxidation of aluminium (loss of dissolved aluminium) during continuous casting for ULC (ultra-low carbon) gradesis 144 ppm but for LC grades it was 94 ppm and it was being expected that ULC steel group would have higher inclusion content after casting.

The purpose of this thesis was to analyze and understand the functions of the current temperature model used at the steel mill in at SSAB in Oxelösund. Furthermore, to try to optimize the model by creating a better description of the ladles thermal status. A copy of the current model was made using Matlab and its performance was compared to actual temperature measurements. Along with this, additional process data was used to find a correlation between these data and temperature alterations in the liquid steel. These correlations were then used to create a new description of the ladles thermal status in a modified temperature model. The results show that the thermal status of the ladle has a big influence on the steel temperature throughout the steel making process. Also, the thermal status is greatly affected by the how the ladles are handled when being empty. The results also show that a model using a unique thermal status calculated for each heat, based on correlations derived from stored process data, provides a better result than the current model which is based on pre-determined heat loss curves.
Syftet med det här projektet var att kartlägga och förstå funktionerna i den befintliga temperaturmodellen som används i stålverket i Oxelösund samt att försöka optimera modellen genom att på ett bättre sätt beskriva skänkens värmestatus. En kopia av den nuvarade modellen gjordes med hjälp av Matlab och jämfördes med verkliga temperaturmätningar. Sedan användes ytterliggare processdata för att hitta samband mellan dessa och temperaturförändringar i stålet. Dessa samband änvändes sedan för att skapa en ny beskrivning av skänkens värmestatus i en modiferad temperaturmodell. Projektets resultat visar på att skänkens värmestatus har en väldigt stor påverkan på temperatur-förändringar i stålet och att det som påverkar värmestatusen mest är hur skänken hanteras när den är tom. Resultaten visar också att en modell med en unik värmestatus för varje skänk uträknad med samband härledda från processparameterar ger ett bättre resultat än   resultaten   från   dagens   modell   där   förbestämda   värmeförlustkurvor  bestämmer skänkens värmestatus.

In present thesis, the fundamental studies on the reaction between MgO based refractories and slag were undertaken for the development of a carbon-free bonding MgO lining material. Alumina was selected as a potential binder material. Due to MgO-Al2O3 chemical reaction, the developed refractory was bonded by MgO·Al2O3 spinel phase. To begin with, an investigation of the dissolution process of dense MgO and MgO·Al2O3 spinel in liquid slag was carried out. To obtain reliable information for dissolution study, a new experimental method was therefore developed. In this method, a cylinder was rotating centrally in a special designed container with a quatrefoil profile. This method also showed a good reliability in revealing the dissolution mechanism by quenching the whole reaction system. The experimental results showed that the dissolution process of MgO and spinel was controlled by both mass transfer and chemical reaction. It was found that the rapid dissolution of spinel was mainly because of its larger driving force. To improve the resistance against slag penetration, two aspects were studied to develop carbon-free MgO refractory. First, colloidal alumina was used and the effect of its addition into MgO matrix was investigated. The use of colloidal alumina was to form bonding products in the grain boundary of MgO. The results showed that the alumina addition greatly improved the resistance of MgO based refractory against slag penetration in comparison with the decarburized MgO-carbon refractory. It was found that the improvement of resistance was mainly related to the spinel-slag reaction products of CaO·Al2O3 and CaO·MgO·Al2O3 solid phases at the grain boundaries. Second, the effect of particle size distribution on the penetration resistance of MgO was investigated. The most profound improvement against the slag penetration was obtained by using a proper particle size distribution. The results highlighted the importance of considering the refractory structure. Experiments were undertaken to investigate the dissolution mechanism of different types of MgO based refractories in liquid slag. It was observed that the dissolution of spinel bonded MgO refractory was much slower than the decarburized MgO-carbon refractory. The primary dissolution in spinel bonded MgO refractory occurred at the slag-penetrated layer, and the removal of this layer by peeling off enhanced the dissolution rate rapidly.

QC 20170918

European RFCS LEANSTORY project

Gas stirring is currently the most commonly used method of homogenizing liquid steel in commercial processes. However, due to the harsh environment during the process, physical models built out of e.g. plexiglass have been used to easier understand the complex phenomenon occurring in the process. The models are also used to optimize stirring conditions and estimate mixing times. Instead of liquid steel, water has been used for modelling, to increase safety and reduce costs. The water models are usually scaled down to sizes that are easier to handle. Scaling correctly requires fulfilling commonly used criteria and dimensionless numbers. This report investigated the accuracy of these dimensionless numbers and the relations commonly used for scaling. Existing studies and relations were evaluated, and the theoretically best suitable scaling equations were tested. Three bottom blown vessels were built, in order to test the existing relations. By applying scaling criteria and calculating gas flow rates accordingly, the correlation between theoretical mixing time and measured mixing time could be investigated. This thesis concluded that the correlation between the measured mixing times was not significant, however, by implementing the scale factor λ1/2 a better approximation seems to be given.
Gasomrörning är för tillfället den vanligaste metoden som används för att homogenisera flytande stål i kommersiella processer. På grund av de svåra förhållandena under processen har fysiska modeller, byggda av exempelvis plexiglas, använts. Detta för att enklare förstå de komplexa fenomen som uppstår under processen. Modellerna används också för att optimera omrörningsförhållandena och för att uppskatta blandningstider. Istället för flytande stål har vatten använts vid modellering för att öka säkerheten och minska kostnaderna. Vattenmodellerna är vanligtvis nedskalade till storlekar som är lättare att hantera. En korrekt skalning kräver att vanliga kriterier och dimensionslösa tal uppfylls. Denna rapport undersökte noggrannheten för dessa dimensionslösa tal samt relationer som vanligtvis används vid skalning. Befintliga studier och relationer utvärderades och de teoretiskt mest lämpliga skalningsekvationerna testades. Tre stycken kärl med bottenblåsning byggdes för att testa dessa relationer. Genom att tillämpa skalningskriterier och beräkna gasflödeshastigheterna, kunde korrelationen mellan de teoretiska och uppmätta blandningstiderna undersökas. Denna avhandling drog slutsatsen att korrelationen mellan de uppmätta blandningstiderna inte var signifikant. Dock verkade en bättre approximation fås när skalfaktorn λ1/2 implementerades.

A numerical study was carried out to predict three dimensional turbulent flow heat and mass transfer in RH-ladles. To model the turbulence in the system, the popular Launder and Sharma (7) version of the low Reynolds number k-$ varepsilon$ model was employed. The governing equations for fluid flow and heat transfer were solved using a control volume-based finite difference method. The buoyancy effect in the axial momentum equation was incorporated through the Boussinesq approximation. From the fluid flow and turbulence characteristics predicted by the numerical model, the mass transfer occurring from the dissolution of steel rods immersed in the liquid steel was computed using the Sherwood number correlation recently proposed by Mazumdar et al. (25). The parameters studied include: the flow rate of steel, the heat flux from the solid walls and various radial positions of the vertically immersed steel rod. The fluid flow results are presented as vector plots. The heat transfer aspects are shown through temperature contours at various vertical planes and the mass transfer results are graphically presented in plots showing variance of mass transfer coefficients with key parameters.
The flow field results were found to be in qualitative agreement with previously reported 3D numerical studies for similar systems. Due to the lack of any experimental or numerical results related to heat and/or mass transfer in RH-ladles, the heat and mass transfer results obtained in the present study could not be compared and verified.

It is well known that inclusions affect the properties of steels and alloys. Therefore, it is important to understand what type of inclusions that exist and how they behave and especially with a focus on large size inclusions. Thus, the large size non-metallic inclusions in ferroalloy FeTi70R were investigated in two dimensions (2D) by using scanning electron microscopy (SEM) in combination with an energy dispersive technique (EDS). It was found that the FeTi70R ferroalloy contain complex oxide inclusions consisting mostly of CaO, SiO2 and TiOx. Furthermore, experimental trials were performed to investigate how these inclusions behaved when entering a melt. More specifically, a comparison between pure Fe and an Alloy 825 grade were made. These results determined the parameters effect on the transformation of the inclusions in the melt. The large size non-metallic inclusions in Alloy 825 during the ladle treatment were investigated during industrial trials by using both two dimensional (2D) and three dimensional (3D) microscopic investigations. The results showed that inclusions consisted of spherical oxides and clusters made up of oxides and nitrides. Further investigations found that the spherical inclusions were transformed from existing NMI in the FeTi70R ferroalloy and slag particles. As for the clusters, they originate from deoxidation products. Furthermore, small inclusions precipitated in the local zones around the added FeTi70R ferroalloy and titanium nitrides. Investigations also found that only Al2O3-MgO and TiN clusters exist after casting. Industrial trials were performed during the last period of the ladle treatment and using a combined electromagnetic (EMS) and gas (GS) stirring. The purpose to investigate the effect of different EMS directions on the agglomeration and on the removal of Al2O3-MgO and TiN clusters. The investigations were then performed in 3D after an electrolytic extraction of the metal samples. The results show that electromagnetic stirring in the upwards direction is best for the agglomeration of the Al2O3-MgO and TiN clusters. However, electromagnetic stirring in the downwards direction is more effective to remove clusters from the melt. This is in agreement with the theoretical predictions based on Stokes’, Brownian and Turbulent collisions. Also, the calculations showed that for Al2O3-MgO clusters with sizes <20 μm the Turbulent collision is the defining factor for agglomeration. However, both Stokes’ and Turbulent collisions are dominant for larger inclusions. For the TiN clusters, turbulent collisions is the dominant factor. Further investigations with more heats and stirring modes were done by using 2D microscopic investigations. More specifically, the number, size, composition and morphology of different inclusions were determined by using SEM in combination with EDS and Inca Feature analyses. The results show that the EMS in downwards direction with a 0.04 m3 min-1 gas flow rate promotes a general removal of Al2O3-MgO and TiN inclusions. Furthermore, that the upwards EMS direction promotes a drastically increase of inclusions having an equivalent size smaller than 11.2 μm. Moreover, the stirring with a 0.02 m3 min-1 gas flow rate has a better removal rate for both downwards and upwards stirring directions compared to the stirring with a 0.04 m3 min-1 gas flow rate. However, no influence on the inclusion composition and morphology could be seen from the different stirring modes.
Det är välkänt att inneslutningar påverkar egenskaperna i stål och legeringar. Därför är det viktigt att förstå vilken sorts inneslutningar som existerar samt hur de beter sig, särskilt stora inneslutningar. Således undersöktes de stora icke-metalliska inneslutningarna i ferrolegeringen FeTi70R i två dimensioner (2D) genom att använda svepelektronmikroskopi (SEM) i kombination med energidispersionsteknik (EDS). Det visade sig att ferrolegeringen FeTi70R innehåller komplexa oxidinneslutningar bestående huvudsakligen av CaO, Si02 och TiOx. Vidare utfördes laborativa försök för att undersöka hur dessa inneslutningar uppträder när de kommer ner i det smälta stålet. Mer specifikt så gjordes en jämförelse mellan rent Fe och Alloy 825 legeringen. Dessa resultat fastställde hur olika parametrar påverkade inneslutningarnas omvandling i smältan. De stora icke-metalliska inneslutningarna i Alloy 825 legeringen under skänkugnsbehandlingen undersöktes vid industriella försök med hjälp av både tvådimensionella (2D) och tredimensionella (3D) mikroskopiska undersökningar. Resultaten visade att inneslutningarna bestod av sfäriska oxider och kluster uppbyggda av oxider och nitrider. Ytterligare undersökningar visade att de sfäriska inneslutningarna omvandlades från befintliga icke-metalliska inneslutningar i ferrolegeringen FeTi70R och slaggpartiklar. När det gäller kluster härrör de från desoxidationsprodukter. Vidare så bildades små inneslutningar i de lokala zonerna kring den tillsatta FeTi70R legeringen och kring titanitriderna. Undersökningarna fann också att endast Al2O3-MgO och TiN-kluster existerar efter gjutning. De industriella försöken utfördes under den sista perioden av skänkugnsbehandlingen och med en kombinerad elektromagnetisk (EMS) och gas (GS) omröring. Syftet med detta var att undersöka effekten av olika riktningar på EMS omrörningen på agglomerationen och avskiljningen av Al2O3-MgO- och TiN klustren. Undersökningarna utfördes sedan i 3D efter en elektrolytisk extraktion av metallproverna. Resultaten visar att en uppåtriktad elektromagnetisk omröring är bäst för agglomereringen av Al2O3-MgO inneslutningar och TiN kluster. Emellertid så är en nedåtriktad elektromagnetisk omröring effektivare för att avlägsna kluster från smältan. Dessa resultat överensstämmer med de teoretiska förutsägelserna baserade på Stokes, Brownian och Turbulent kollisionsvolymer. Dessutom visade beräkningarna att för Al2O3-MgO kluster med storlekar <20 μm är det turbulenta kollisioner som är den avgörande faktorn för agglomerering. Dock är det både Stokes och de Turbulenta kollisionerna som dominerar för större inneslutningar. För TiN klustren är det de turbulenta kollisionerna som är den dominerande faktorn. Ytterligare undersökningar med fler charger av Alloy 825 och olika omrörningsmetoder gjordes baserade på 2D mikroskopiska undersökningar. Mer specifikt så bestämdes antalet, storleken, sammansättningen och morfologin för olika inneslutningar med användning av SEM i kombination med EDS och Inca Feature analyser. Resultaten visar att en nedåtriktad EMS i kombination med ett gasflöde på 0,04 m3 min-1 främjar ett generellt avlägsnande av Al2O3-MgO och TiN inneslutningar. Vidare främjar den uppåtriktade EMS omrörningen en drastisk ökning av inneslutningar med en ekvivalent storlek av 11,2 μm eller mindre. Vidare har omröringen med ett gasflöde på 0,02 m3 min-1 en bättre avskiljningshastighet för inneslutningar i både nedåt och uppåtgående omrörningsriktningar jämfört med omrörningen med ett gasflöde på 0,04 m3 min-1. Emellertid kunde inget inflytande på sammansättningen och morfologin ses från de olika omrörningssätten.

QC 20171122