Influence of Ladle-slag Additions on BOF-Process Parameters

Influence of Ladle-slag Additions on BOF-Process Parameters

Anders Dahlin

Licentiate Thesis

Stockholm 2011

Department of Materials Science and Engineering Division of Applied Process Metallurgy

Royal Institute of Technology SE-100 44 Stockholm

Sweden

Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan i Stockholm, framlägges för offentlig granskning för avläggande av Teknologie Licentiatexamen, onsdagen den 4 maj 2011, kl. 10.00 i sal B1, Brinellvägen 23, Kungliga Tekniska Högskolan, Stockholm

ISRN KTH/MSE--11/09--SE+APRMETU/AVH

ISBN 978-91-7415-961-5

ii

Anders Dahlin, Influence of Ladle-slag Additions on BOF-Process Parameters

KTH School of Industrial Engineering and Management Division of Applied Process Metallurgy

Royal Institute of Technology SE-100 44 Stockholm

Sweden

ISRN KTH/MSE--11/09--SE+APRMETU/AVH ISBN 978-91-7415-961-5

© The Author

iii

“An expert is a person who has made all the mistakes that can be made in a very narrow field”

-Niels Bohr

iv

v

A ^BSTRACT

The influence of ladle-slag additions on the BOF-process performance were investigated in plant trials. The aim of the study was to recycle ladle slag from secondary steelmaking to the LD-converter to save lime and improve the slag formation. More specifically, two plant trial campaigns covering in total 83 heats, whereof 47 with ladle-slag additions and 36 without ladle-slag additions, were performed.

Slag and steel sampling of the process were performed at tapping as well as during blowing at 15, 35, and 65% of the total blowing time. During the first campaign, ladle slag was added through the chute and lime reductions were made manually to correct for the ladle-slag addition.

In the second campaign, a development of the approach was made to suite a normal production practice. More specifically, the ladle slag was added through the weight-hopper system and implemented in the process-control system. In this way, the lime additions were reduced automatically by approximately 260 kg per heat.

Moreover, the heat balance was compensated with a reduction in the iron-ore consumption. Additionally, the lance program was modified and the lance was lowered in the initial stages of the blow.

On the positive side, it was found that no demerits in the metallurgical performance of the process occur when ladle slag is recycled to the BOF-process.

Furthermore, only slight affections on the slag composition were found, mainly with respect to the Al 2 O 3 and FeO-content. In addition, the ladle slag was shown to melt during the initial stages of the blow. This contributed to an increased slag weight both during the blow and at tapping. However, a negative effect on the blowing time was experienced during the trials. Although, this effect was more pronounced during the first campaign and could be reduced with a controlled heat balance during the second campaign.

Key words: BOF, slag formation, ladle slag, recycling, steel, composition, energy

balance

vi

vii

A CKNOWLEDGMENTS

I would like to express my gratitude to Professor Pär Jönsson and Dr. Anders Tilliander for your encouragements and guidance throughout this project. I would also like to express my sincere appreciation, for all the help and valuable discussions, to M.Sc. Johan Eriksson at Swerea MEFOS.

For their help and engagement during the plant trials at SSAB I would like to thank Leif Nilsson, Dag Bergqvist and Magnus Andersson, as well as all the operators at the BOF and the laboratory at SSAB for showing such a curiosity towards research.

Financial supports from the Swedish Energy Agency, Jernkontoret – The Swedish Steel Producers’ Association and the committe JK21065, as well as grants from Axel Hultgrens foundation, are greatly acknowledged.

I would also like to thank Dr. Sharif Jahanshahi and Dr. Shouyi Sun at the CSIRO in Melbourne for giving me the opportunity to gain valuable experience from down under.

Many thanks are also expressed to my friends and colleagues at KTH. For my own health and wellness, as well as hard competitions and challenges on the ski-tracks, I would like to thank Fredrik Engström and Anders Bennitz at Luleå Technical University for many fun times during my years in Luleå.

The ones at home, who always wonder what I´m doing and when I will move to Borlänge again, my family, your support are truly precious to me.

The final thanks are expressed, from the bottom of my heart, to its best friend Mirja. Thank you for your guidance over the mountains of life.

Yours,

Anders Dahlin, Stockholm, April, 2011

viii

ix

S ^UPPLEMENTS

The present thesis is based on the following supplements:

Supplement 1: “Influence of ladle-slag additions on the BOF-process performance”

A. Dahlin, A. Tilliander, J. Eriksson and P.G. Jönsson Submitted to Ironmaking and Steelmaking for publication

Supplement 2: “Influence of ladle-slag additions on the BOF process under production conditions”

A. Dahlin, J. Eriksson, A. Tilliander and P.G. Jönsson Supplement 3: “A theoretical study of the effect of ladle-slag additions on

the BOF process”

A. Dahlin and P.G. Jönsson

The contribution by the author to the different supplements of the thesis:

1. Literature survey, experimental work, major part of the writing.

2. Literature survey, experimental work, major part of the writing.

3. Literature survey, modeling work, major part of the writing.

x

xi

C ^ONTENTS

I NTRODUCTION 1  

1.01   B ASIC O XYGEN S TEELMAKING 2  

1.02   T HESIS OUTLINE 3  

T HEORETICAL B ACKGROUND 5  

2.01   S LAG FORMATION IN THE BOF 5   2.02   R ECYCLING OF SLAG TO THE BOF 6   2.03   M ASS BALANCE EQUATIONS 6  

E XPERIMENTAL P ROCEDURE 9  

3.01   P LANT TRIALS 9  

3.02   C HEMICAL ANALYSIS 10  

3.03   M ODELING 11  

R ESULTS AND D ISCUSSION 13   4.01   E FFECT ON SLAG FORMATION 13   4.02   E FFECT ON ENERGY BALANCE 17   4.03   E FFECT ON STEEL COMPOSITION 20   4.04   R EDUCED LIME CONSUMPTION 22  

F INAL D ISCUSSION 23  

C ONCLUSIONS 25  

F UTURE W ORK 27  

R EFERENCES 28  

xii

1

Chapter 1

I NTRODUCTION

Since the discovery of the Bessemer process by Henry Bessemer in 1856 [1], slag formation in the Basic Oxygen Furnace (BOF) converter has been subjected to extensive research. In primary steelmaking, a rapid slag formation during the initial stages of the blow is crucial for the performance of the process. This is mainly to enable transfer of sulphur and phosphorus from the steel to the slag, but also to protect the refractory from severe erosion. The slag is created by fluxing the first formed FeO-SiO ₂ -rich slag with the addition of slag formers. More specifically, burnt lime with high contents of CaO and MgO-containing dolomitic lime. Thus, the rate of formation of a good metallurgical slag, with high contents of CaO and a MgO-content close to saturation, is governed by the dissolution rate of the slag formers.

During 2008, more than 1.3 Mt of slag was produced from the iron- and steelworks in Sweden [2]. The majority of this by-product is produced from the ironmaking Blast Furnaces (BF) and primary steelmaking LD-converters (LD). Furthermore, these slags are re-used externally or recycled internally. The slag from the LD is recycled and used as a slag former in the BF. In addition, the BF slags are used as construction material in roads and cement.

20 thousand ton of slag is annually produced from secondary steelmaking at SSAB

EMEA in Luleå. This slag has a high content of CaO and thus possesses the ability

to be recycled and used as a slag former to the BOF. The hypothesis is also that

this ladle slag, with a low melting point, has the possibility to improve the slag

formation. Furthermore, to increase the amount of liquid slag available during the

initial stages of the blow in the BOF. Additionally, if ladle slag is recycled to the

BOF, the amount of burnt lime added can be reduced thanks to the CaO-content

in the ladle slag. Furthermore, the recycling of ladle slag to the BOF will induce a

three time utilisation of the lime added in secondary steelmaking. First during ladle

refining, secondly as recycled slag to the BOF and thirdly as BOF slag recycled to

the BF.

Chapter 1

2

1.01 Basic Oxygen Steelmaking

Basic Oxygen Steelmaking (BOS) is the general name for the procedure of using high purity oxygen to convert the carbon-containing hot metal to liquid steel. The reactor is called a BOF. In Sweden, the reactor type used is called LD-converters from the name of the two cities, Linz and Donawitz where they were first developed. A schematic picture of a LD-converter is shown in Figure 1. This converter has porous stones in the bottom refractory and is therefore, specifically called a LD-LBE (Lance Bubbling Equilibrium) converter.

Primary steelmaking in an ore-based steel plant like SSAB EMEA, studied in the present work, starts with the hot metal produced from a BF. A typical chemical analysis of the hot metal after desulphurisation is 4.4% C, 0.4% Si, 0.4% Mn, 0.01%

S and 0.03% P. By injecting oxygen gas through a lance towards the metal bath, most of the impurities in the hot metal is oxidised and a liquid steel with typically 0.05%C, 0.15% Mn, <0.01% S, and <0.01% P is produced. To start with, steel scrap is charged into the magnesia-based refractory lined vessel. The scrap act as a coolant in the process and excess heat from the exothermal reactions in the process is used to melt the scrap and increase the yield. In addition to the scrap, iron ore can be used as a coolant as well.

The hot metal is poured on top of the scrap and the oxygen lance is inserted into the furnace. The blowing is started and high purity oxygen is injected at super-sonic speed at a flow rate of 340 Nm ³ /min until sufficient amounts of carbon has been oxidised and the aimed carbon content is reached in the liquid steel. During the process, slag-former additions are made continuously from above, to the converter through a weight-hopper system.

The oxidised impurities from the hot metal form, together with oxidised iron and added slag formers, a slag with a composition at tapping of typically 20% FeO, 40% CaO, 10% SiO2, 4% MnO, 12% MgO, <1% P2O5 and <0,1% S together with small contents of other oxides. Gas bubbles and metal droplets ejected into the slag form large reaction surfaces and create a foaming slag during the process.

This contributes to the high reaction rates experienced during oxygen steelmaking [3-4]. However, it may also cause the slag to exit the reaction vessel, which results in iron losses and a decreased yield of the process. This phenomenon is called slopping. A more detailed description of the slag formation will be given in the next chapter.

After the blow-end and the oxygen lance have been raised, the steel is tapped

through the tap-hole by tilting the furnace. The liquid steel now also contains a

considerable amount of dissolved oxygen, which is removed by adding elements as

Si or Al. After the BOS process, the liquid steel is brought to secondary

steelmaking units for alloying before being continuously cast into slabs.

Introduction

3

Figure 1 A schematic picture of and LD-LBE converter with the oxygen lance in blowing position.

1.02 Thesis outline

The main objective of this thesis was to study the effect on the BOF process when ladle slag is used as a slag former. The main purpose of the ladle-slag recycling to the BOF was to reduce the lime consumption and to improve the early slag formation in the process. To enable this, two plant trial studies of an industrial LD- converter have been made. A theoretical study of the effect of ladle-slag additions on the BOF process has also been made. Important measures, as steel composition, slag weight and composition have been studied. In addition, other process parameters as blowing time and slopping have also been carefully considered.

The thesis is based upon three supplements that cover different aspects of the effect of ladle-slag additions on the BOF process. A schematic figure of the outline can be studied in Figure 2.

In the first supplement, the initial attempt to speed up the slag formation in the

LD-converter with the addition of ladle slag was performed. In addition, effects of

the ladle slag addition on other process parameters including the steel composition

were studied as well. The results from Supplement 1 lead to the decision that the

ladle slag had to be implemented in the process control system of the converter. In

addition, it was decided to improve the method of adding the ladle slag to the

Chapter 1

4

converter. Supplement 1 enlightened the potential of ladle slag as a replacement for burnt lime. However, in order to improve the industrial applicability of the method, practices more closely to normal production needed to be used. This was studied in Supplement 2.

Figure 2 A schematic outline of the thesis and contents of the three supplements (S1-S3)

In Supplement 2, the effect of recycling of ladle slag to the LD-converter using a normal production practice was studied. In addition, the objective of Supplement 2 was to study the effect on the lime dissolution when recycling of ladle slag was performed. Furthermore, results from Supplement 1 suggested that some adjustments had to be made regarding the positioning of the lance. This was made during the campaign of plant trials in Supplement 2.

The result found in Supplement 1 and 2 of a disturbed heat balance due to the

recycling of ladle slag was studied more carefully in Supplement 3. This was made

with thermodynamic modeling using a commercial modeling package. The

objective was to quantify the energy penalty and study different options to obtain

an energy balance in the system. It was also of interest to use the thermodynamic

model to describe the effect of ladle slag addition on the final equilibrium

composition of the steel and to compare the result with the plant trial data.

5

Chapter 2

T HEORETICAL B ^ACKGROUND

2.01 Slag formation in the BOF

Important work during the 60s and 70s [4-8] investigated the slag formation in the BOF. More specifically, the evolution of the slag composition was studied using intermittent sampling of the steel and slag during oxygen blowing. These studies are the basis of the knowledge today and can be summarised as follows. When the blowing of the BOF is started, and the lance is positioned high above the metal bath, mainly iron and silicon is being oxidised. This slag, with high contents of FeO and SiO 2 is detrimental for the furnace lining, but important to ensure a high dissolution rate of lime. The added slag formers of burnt lime and dolomitic lime starts to dissolve and the CaO-content of the slag increases throughout the blow.

The silicon available in the hot metal is completely oxidised during the first stage of the blow. As the rate of decarburisation increases, less FeO is formed and the concentration of SiO 2 decreases with the increase in slag weight. The FeO-content becomes stagnant during the main decarburisation period, but increases again towards the end of the blow as the availability of carbon in the hot metal decreases.

In addition, this influences the slag weight inside the converter that also is stagnant during the middle part of the blow, but increases as the iron oxidation increases.

Figure 3 Evolution of metal and slag composition during blowing [9].

Chapter 2

6

2.02 Recycling of slag to the BOF

The method of recycling slag to the BOF to promote a rapid slag formation has been investigated earlier. The hypothesis is that the addition of a pre-melted slag would increase the weight of the liquid slag during the initial stages of the blow and thus, improve the slag formation. At Corus steel plant at Ijmuiden [10], recycled BOF slag was used to improve the slag formation and reduce the dust generation from the process. The work also showed that the metallurgical performance and lining wear in the process were unaffected by the slag recycling.

The use of recycled converter slag was also studied by Donayo et al. [11]. More specifically, the effect of recycling 13 kg slag per ton of steel and simultaneously reducing the lime additions with 4.1 kg per ton was investigated. The authors concluded that the method could be performed with an improved slag formation and a maintained phosphorus capacity in the process.

Another study was made by Lee et al. at Posco Pohang works [12]. Here, ladle slag from secondary steelmaking was recycled to the LD-converter to an extent of 10- 30 kg per ton of steel. The total amount of CaO added to the process was reduced, but no demerits were found in the metallurgical reactions. More specifically, similar phosphorus levels were shown at the blow end when ladle slag was used to replace burnt lime.

In the present work, the methodology of previous work [4-8], [10-12] is combined to study the effect of the ladle-slag addition on the slag formation and the performance of the BOF process.

2.03 Mass balance equations

The slag-weight calculations in this work are based on the conservation-of-mass principle. The components considered in the mass-balances are Si/SiO ₂ , Ti/TiO ₂ , and CaO. An assumption is made that the losses of each component can be neglected. Thus, the mass-balance equation for each component becomes:

Si/SiO 2 -balance

  SL  m  HM   m   SC  m   AD  m   LS  m   LO  ⁰ 

m _{Si Si Si Si Si Si} (1)

Ti/TiO ₂ -balance

  SL  m  HM   m   AD  m   LS  m   SC  m   LO  0 

m _{Ti Ti Ti Ti Ti Ti} (2)

CaO-balance

  ^{SL m}   ^AD

m _CaO  _CaO (3)

Theoretical Background

7

where m is the mass of component x in the slag (SL), hot metal (HM), scrap (SC), additives (AD), liquid steel (LS) and losses (LO).

Calculation of the slag weight can be made with the following equation:

 

X O X O

X SLAG X

M a M W

SL

m m

b

a

  

 100

(4) where W is the concentration of the oxide in the slag in weight-percent and M is

the molar weight of the oxide (X _a O _b ) and the metal (X). For CaO-balance, the last term of equation (4) is excluded.

The slag weight at tapping is approximated as the average value of the slag weight calculations with equations (1), (2) and (3) inserted into equation (4). This assumes that the slag formers are completely dissolved at tapping.

To estimate the slag weight during blowing, the slag weight is calculated using the

average value of the slag weight calculation with equations (1) and (2) inserted into

equation (4). This is because there is an unknown amount of undissolved slag

formers available inside the converter at this time. As a result, the term m _X (AD) is

neglected. In addition, the scrap is assumed to not have been melted until the time

of interruption. Thus, the term m _X (SC) is neglected as well.

Chapter 2

8

9

Chapter 3

E XPERIMENTAL P ^ROCEDURE

3.01 Plant trials

Two plant trial campaigns were performed to study the effect of ladle-slag additions on the BOF process. Throughout the trials, the number of factors affecting the slag formation process was minimised and only clean scrap, slabs and sheet metal, was used in the process. In addition, to decrease the effect of fluctuations in the hot metal analysis, and be able to compare every experimental heat with a reference of similar process conditions, it was tried to blow every other or at least every third heat as a reference heat without any addition of ladle slag.

The ladle slag, with the composition presented in Table 1, was added in different ways in the first and second plant trial.

During the first plant trials, the ladle slag was added in the chute and charged to the converter together with the scrap. In the second campaign, it was decided to add the ladle slag under normal production conditions. Hence, the ladle slag was added as other slag-former additions through the weight-hopper system located above the converter. Additionally, the ladle slag was implemented in the process-control system. This made it possible to compensate for the negative effect on the heat balance by the ladle-slag addition. Furthermore, for the second campaign, it was decided to change the lance schedule for heats with ladle slag and have a lower positioning of the lance during the initial stages of the blow. This was made to reduce the FeO-content of the slag and prevent slopping during the blow.

Table 1 The average composition of the ladle slag used during the two campaigns of plant trials.

All analyses are given in wt-percent.

  Fe  CaO  SiO ₂   MnO  P ₂ O ₅   Al ₂ O ₃   MgO  V ₂ O ₅   TiO ₂   S  Supplement 1   9.5  39.8  8.9  6.1  0.2  24.8  6.4  1.1  1.2  0.1  Supplement 2  9.7  41.4  7.1  5.3  0.2  24.5  6.7  1.6  1.1  0.1 

1000 kg of ladle slag was recycled to the process in each experimental heat during

the two plant trials. In addition, reference heats without ladle-slag additions were

also studied as explained earlier. Compensation concerning the CaO-amount

Chapter 3

10

available in the ladle slag was made in the following way. For ladle-slag heats during the first plant trial campaign the burnt lime addition was reduced with 400 kg per ton added ladle slag. In the second plant trial campaign, the lime additions were reduced automatically by the process-control system. The lime reductions were on average 260 kg per ton of ladle slag. It should also be noted that the total charge weight of the converter had been increased from 114 to 130 ton between the performance of trial campaign No.1 and 2.

Sampling of the process was made at the blow-end by normal practice using a sub- lance with a dual-thickness steel sampler. This included a steel sample together with a temperature and oxygen activity measurement. In addition, a slag sample was taken manually during tapping with a slag scoop.

To study the effect of the ladle-slag additions on the slag-formation process, sampling was also made during blowing. During the first campaign of plant trials, this was made at 35% and 65% of the total blowing time. In the second plant trial the focus was to take samples at the point where 1000 Nm ³ had been blown, i.e. at approximately 15% of the total blowing time. Steel sampling was made manually using a dual-thickness sampler and slag samples were withdrawn by hand using a slag scoop. At the point of sampling, the blowing was stopped and the converter tilted to be able to reach the metal bath and slag layer inside the vessel.

In total, 35 heats were studied during the first plant trials and 48 heats during the second plant trials. In 22 (first campaign) and 25 (second campaign) of them, ladle slag was added as a slag former.

3.02 Chemical analysis

The chemical analysis of the steel and slag samples was carried out at the laboratory at SSAB EMEA in Luleå. Before chemical analysis, some preparation of the slag samples was necessary. Slag samples taken at tapping were crushed and magnetically separated from metal droplets and then grinded into powder using a ring mill. Obvious metal droplets and pieces of undissolved lime in the slag samples taken at interruption of the process were removed before crushing. Afterwards, the slag samples were passed through a 0.25 mm grate after a short mill grinding procedure of 10s. This was believed to separate the slag from the metal droplets in the sample. The slag samples were analysed with the X-Ray Fluorescence (XRF) technique using an ARL 8680 instrument.

The steel samples taken at tapping was analysed using the Optical Emission

Spectroscopy (OES) technique with an ARL 4460 instrument. Metal samples taken

during blowing were analysed using XRF. In addition, the carbon and sulphur

content of the metal was determined with a LECO CS200 instrument. The spread

and accuracy of the analysis procedure is shown in Table 2.

Experimental Procedure

11

Table 2 The spread and analysis accuracy for the different methods used in the plant trials. S ₁ : After desulfurisation. S ₂ : Before desulphurization.

  Slag sample  

(XRF) 

  Hot metal sample (XRF)  Liquid steel sample (OES)    Content 

[wt‐%] 

Spread (1σ)  [wt‐%] 

  Content  [wt‐%] 

Spread (1σ)  [wt‐%] 

Content  [wt‐%] 

Accuracy  [%] 

Fe  20  ±0.3  C  4.8  ±0.026  0.025  ±10 

CaO  41  ±0.5  Si  0.31  ±0.007  0.024  ±21 

SiO2  9  ±0.2  Mn  0.36  ±0.003  0.15  ±6 

MnO  3.8  ±0.04  P  0.036  ±0.0005  0.004  ±13 

P2O5  0.6  ±0.015  Ti  0.087  ±0.001  0.006  ±15 

S  0.05  ±0.005  S 1 0.004  ±0.0002  0.009  ±17 

Al2O3  1.6  ±0.03  S ₂ 0.05  ±0.0011     

MgO  10.5  ±0.8           

TiO2  1.7  ±0.035           

3.03 Modeling

The theoretical study was performed in a commercial multi-phase equilibrium calculation package called the MPE. Among other commercial thermodynamic packages, the MPE has been developed especially to have practical relevance for pyrometallurgical processes. A detailed description of the equilibrium calculation software can be found elsewhere [13].

In brief, the equilibrium calculation program utilises different models and databases of thermodynamic data in a calculation routine to minimise the total Gibbs free energy of the system. The databases consist of critically assessed experimental data in binary and ternary systems. The Kapoor and Frohberg [14] model is used to describe the multi-component slag phase. Additionally, other solution models [13]

describe the solid solutions and the metal phase in the system.

To study the effect of the ladle slag addition on the BOF process, a reference simulation was made of a heat with similar process conditions as heats in the plant trials. First, the reference simulation was made without any ladle slag additions to determine the heat balance of the converter. Secondly, the same heat was simulated once again but now with the addition of ladle slag. In addition, a number of different options to compensate for the disturbed energy balance, when adding ladle slag, were studied in five case studies. These are presented in Table 3.

The hot-metal silicon and carbon contents were 0.4 and 4.6 wt-percent respectively.

Additionally, the aimed carbon content and tapping temperature was 0.046 wt-

percent and 1650°C. The temperature of the slag and gas was assumed to be the

same as the temperature of the steel. Furthermore, the hot metal temperature was

1350°C in every simulation. The principle of the energy- and mass balance is

showed schematically in Figure 4.

Chapter 3

12

Table 3 Summary of the different simulations made in the theoretical study.

Simulation  Type of heat 

A  Reference heat 

B  With ladle slag 

C  With ladle slag, reduced lime addition 

Case 1  With ladle slag, lower tapping temperature 

Case 2  With ladle slag, lower iron ore addition 

Case 3  With ladle slag, increased oxygen 

Case 4  With ladle slag, lower scrap weight 

Case 5  With ladle slag, FeSi addition 

The effect of the ladle-slag additions on the heat balance could be evaluated by comparing the heat balance of the different simulations to the reference heat (Sim.

A) In addition, the effect on the steel composition was also evaluated to see if the ladle-slag addition has positive effects on the thermodynamic equilibrium of the process.

Figure 4 Materials and gases considered in the heat- and mass balance calculations in the

simulation.

13

Chapter 4

R ESULTS AND D ^ISCUSSION

In this section, the main discoveries from the three supplements regarding the effect of the ladle-slag addition on the BOF process will be presented and discussed. First, the effect on the slag formation will be treated and later the effect on the metallurgical performance. Furthermore, the effect on the energy balance of the converter will also be shown and brought to discussion. An estimation of the lime savings attainable by recycling of ladle slag will also be presented in this section.

4.01 Effect on slag formation

Through analysis of the samples taken at interruption of the blowing in the first and second plant trials it was able to investigate the slag formation procedure inside the converter. In Figure 5, the alumina (Al ₂ O ₃ ) content of the slag is plotted as a function of the blowing time elapsed until sampling. It can be seen that there is an obvious rise in alumina content for heats where ladle slag has been used. This can be explained by the alumina content of the ladle slag, see Table 1. In addition, these results show that the ladle slag is melted during the initial stages of the blow. As the blowing proceeds, the alumina content of the slag is diluted and lowered. However, the average alumina content is 1.1 wt-percent higher at tapping in heats where ladle slag has been used compared to the reference heats.

During the second plant trials, the effect of the ladle slag on the lime dissolution

was studied. This was made by adding the same amount of burnt lime to heats with

and without ladle slag. Furthermore, by taking samples after 1000 Nm ³ of oxygen

gas had been blown. The dissolved amount of CaO could afterwards be estimated

through mass-balance calculations. The procedure is explained in detail in supplement

2. The results are shown in Figure 6 where the amount of dissolved CaO in the slag

is plotted as a function of the amount of CaO added to the process at the time of

sampling.

Chapter 4

14

Figure 5 The alumina (Al ₂ O ₃ ) content of the slag during blowing. At tapping, the average alumina content is 2.2 wt-percent for heats with ladle slag and 1.1 wt-percent for heats without ladle slag.

Figure 6 The amount of dissolved CaO in the slag at 15% of the total blowing time.

Results and Discussion

15

Figure 6 shows that despite a higher amount of added CaO, the heats with ladle slag did not show any increase in the amount of dissolved CaO to the slag. This indicates that there are more undissolved slag formers available inside the converter for heats with ladle slag in comparison to the reference heats. This is confirmed by studying the photographs that were taken at the time of sampling. Four typical examples of the visual appearance of the slag layer are shown in Figure 7. The appearances could be divided into four categories:

A. A well dissolved slag layer

B. A well dissolved slag layer with visible pile(s) of slag formers C. Less dissolved slag layer

D. Less dissolved slag layer with visible pile(s) of slag formers

Figure 7 Different visual appearances of the slag layer at sampling at 1000 Nm ³ of blown oxygen.

As can be seen in Figure 7 the slag layer in picture C and D are darker and hence, contain less dissolved slag formers. Additionally, the heats without ladle slag addition where mostly represented by the appearance of category A and B.

However, the heats with ladle-slag additions could be described by the slag layer in category C and D (See Figure 6).

From the analysis of the slag it was also possible to estimate the slag mass available

inside the converter. The procedure of doing this is explained in detail in Supplement

1 and Supplement 2 and will not be discussed here. The sampling procedure of tilting

Chapter 4

16

the converter brings the liquid part of the slag closest to the converter opening.

This, in turn, makes it possible to take samples from the liquid part of the slag.

Thus, it is governed that it is the weight of the liquid part of the slag that is being calculated.

The slag weight at the different points of sampling during the two plant trials is plotted in Figure 8. The figure shows that the slag weight has been increased for heats with ladle slag at times approximately representing 35%, 60-65% and 100%

(tapping) of the total blowing time. The average slag weight at tapping was 112 kg per ton hot metal for heats with ladle-slag additions and 93 kg for heats without ladle-slag additions. However, the slag weight was somewhat lower during the second campaign. Moreover, the difference in slag weight between a ladle-slag heat and a reference heat was lowered as well. During the first plant trial, ladle-slag heats had an average increased slag weight of 21 kg per ton hot metal compared to the reference heats. This was lowered during the second plant trials to 14 kg per ton hot metal.

Figure 8 Slag weight evolution during blowing expressed as kg per ton hot metal.

The main difference in the end point slag composition when adding ladle slag to

the process is found in the FeO-, CaO-, SiO 2 - and Al 2 O 3 -contents. This can be

seen in Figure 9 where the slag composition at tapping is shown for heats with and

Results and Discussion

17

without ladle slag during plant trial No.2. The increase in Al ₂ O ₃ has been discussed earlier and can be explained by the Al 2 O 3 content of the ladle slag. In addition, the decrease in contents of CaO and SiO ₂ is partly due to the dilution effect of the ladle slag. Moreover, the increase in FeO-content for heats with ladle slag is another explanation. However, no major difference is seen in the basicity (B2) expressed as the CaO to SiO ₂ ratio. The higher FeO-content of the slag for heats with added ladle slag contributes to the increase in slag weight, as seen in Figure 8.

Figure 9 The slag composition at tapping for heats with and without ladle slag during the second campaign of plant trials.

The slopping behaviour of the slag was seen to increase for heats with ladle slag in the first campaign of plant trials. That, in combination with an increased FeO content of the slag during blowing lead to the decision to reduce the distance between the lance and the metal bath for heats with ladle slag during the second campaign. The resulting experience from plant trial No.2 was that this reduced the tendency for slopping when ladle slag was recycled to the process. However, the effect of the adjustment on the equilibrium steel composition had to be monitored.

4.02 Effect on energy balance

The effect of the ladle-slag addition on the energy balance of the converter was

studied in Supplement 3. The results are summarised in Figure 10. Recycling of 1000

Chapter 4

18

kg of ladle slag to the process is accompanied by an energy penalty of 2082 MJ (Sim. B). By reducing the lime consumption, which is enabled by the CaO-content in the ladle slag, the energy penalty is lowered to 1728 MJ (Sim. C). However, a number of different options are available to compensate for the increased energy need. The result showed that the tapping temperature of the steel can be reduced by 12°C to compensate for the energy penalty (Case 1). On the other hand, this might not be possible practically since the higher temperature might be needed during secondary steelmaking. However, there is also a possibility to reduce the scrap amount with 1.5t (Case 4) or add chemical energy in the form of FeSi (Case 5). Unfortunately, this will have a negative effect on the yield of the process which can be seen in Figure 10. Furthermore, the addition of FeSi will have to be accompanied by an increase in the addition of CaO since it otherwise will decrease the basicity of the slag. The effect on the equilibrium steel composition is discussed in section 4.03.

Figure 10 A summary of the results of the energy- and mass balance of the converter from the theoretical study. LS: Ladle slag, HB: Heat balance, DoLi: Dolomitic lime, SW: Slag weight, StW:

Steel weight.

During the first campaign of plant trials, where the ladle slag was added through

the chute and charged into the converter, the only possibility to compensate for the

energy balance was to increase the oxygen content. This was simulated in Case 3

(See Figure 10) during the theoretical study. The result showed that theoretically 96

Nm ³ of oxygen gas is needed to increase the temperature of the steel to its correct

tapping temperature. More specifically, to oxidise Fe to FeO. The corresponding

Results and Discussion

19

increase in blowing time during the plant trial was 232 Nm ³ , which approximately corresponds to 40 seconds. This is shown in Figure 11. However, it should be noted that this is an average value calculated from all the heats during the campaign of plant trials.

For the second campaign of plant trials, the ladle-slag addition was implemented in the process control system of the converter. This resulted in a compensation for the heat loss by a reduction in the iron-ore addition and increased oxygen consumption. This option was later studied in Supplement 3. It was shown that by reducing the iron-ore additions by 380 kg and increasing the oxygen-gas amount with 50 Nm ³ it is possible to reach the same energy level as for a reference heat (See Case 3, Figure 10). This however, has the demerit of a small increase (9s) in blowing time.

Figure 11 The average oxygen consumption in the process for heats with ladle slag (LS) and heats without ladle slag (REF) additions during the first (S1) and second (S2) plant trial together with the theoretical study (S3).

The increase in oxygen consumption during the second campaign was found to on

average be 66 Nm ³ for heats with ladle slag additions compared to heats without

ladle slag (See Figure 11). This is close to the result of Case 3 in Figure 10 and can

thus be explained theoretically. However, it should be mentioned that the operator

of the converter has the control of when to stop the blowing. Thus, the human

factor is also involved in the result of the oxygen consumption during the plant

trials.

Chapter 4

20

4.03 Effect on steel composition

The theoretical effect on the end-point steel composition was studied using a commercial thermodynamic modeling package. The result of the phosphorus distribution between the slag and the steel, together with the phosphorus and sulphur content of the steel is shown in Figure 12. The figure shows that the sulphur content of the steel is somewhat increased when the ladle slag is added.

From a theoretical point of view, the desulphurisation capacity of the BOF slag is poor mainly due to a high FeO-content. This limits the possibilities to obtain a low oxygen activity in the steel necessary for sulphur refining. Also, since the ladle slag contains some amount of sulphur, the majority will be transferred to the steel. As a result, the increase in the sulphur content of the steel is 5 ppm. However, no increase in sulphur content of the steel was detected during the plant trials.

In Figure 13, the sulphur content of the liquid steel at blow-end is plotted against the temperature of the steel. It can be seen that there are three distinct levels of the sulphur content of the liquid steel. In addition, there is no difference between heats with ladle slag and without ladle slag. Furthermore, the lance adjustment made on ladle-slag heats did not affect the metallurgical performance regarding the sulphur content. However, it should be mentioned, that the expected effect of the ladle slag addition (See Figure 12) of a 5 ppm increase in the sulphur content is smaller than the accuracy of the method used for analysis with OES (See Table 2).

Figure 12 The result of the effect on the sulphur and phosphorus content of the steel for the

different simulations in the theoretical study.

Results and Discussion

21

Figure 12 show that the best way to maintain the metallurgical performance of the process is to use Case 1 to compensate for the energy penalty. Thus, lower the tapping temperature of the steel with 12°C. The methods studied during the plant trials were Case 3 (increase the oxygen amount) and Case 2 (reduce iron ore and increase oxygen amount) in plant trial No.1 and 2 respectively. These two options offer a slightly decreased phosphorus content of the steel with 4 and 1ppm respectively. However, these differences are also too small to be detectable practically.

Figure 13 The sulphur content of the liquid steel for heats with and without ladle slag that was desulphurised below 0,005 wt-percent prior to BOF steelmaking. Heats with a lance adjustment are specifically marked in the figure.

During the plant trials, it was not possible to detect any effect of the ladle-slag

addition on the phosphorus content of the steel. This can be seen in Figure 14

where the phosphorus content of the steel for every heat during the plant trials is

plotted against the steel temperature. The lance adjustment, made for ladle-slag

heats during the second campaign of plant trials, did not have any effect on the

end- point steel composition.

Chapter 4

22

Figure 14 Phosphorus content of the liquid steel for all heats during the trials. Heats with lance adjustments are marked specifically in the figure.

4.04 Reduced lime consumption

One of the objectives of the ladle-slag recycling to the BOF process was to some extent replace burnt lime as a raw material for slag formation in the converter.

During the first trial, 400 kg of lime was reduced from the original slag former additions calculated by the process computer. However, in the end the operator has the control of adding more (or less) of this proposed value. With the new process control system that was implemented before campaign No.2, the operators became more confident with the performance of the system. Therefore, the slag-former additions became more controlled. During the second trial when the process control system compensated for the CaO content in the ladle slag, the reduction in lime consumption was on average 260 kg per ton added ladle slag.

The amount of 1000 kg of ladle-slag addition per heat was chosen because it

corresponds to an amount that lead to a recycling of all the ladle slag produced by

secondary steelmaking within the steel plant, if recycling is done on a daily basis.

23

Chapter 5

F ^INAL D ^ISCUSSION

In the present study, the effect of ladle-slag additions on the slag formation and BOF-process performance has been investigated. The main discoveries are summarised in Figure 15 and discussed below.

In the theoretical study (S3) it was shown that the ladle -slag addition results in an energy penalty for the process. This energy penalty was not compensated for in the first plant trial (S1). Thus, there was a need for an increased use of oxygen gas to oxidise FeO to be able to reach the correct tapping temperature. The effect of this option was shown in the theoretical study and the result agrees with the result from the first campaign of the plant trials. During the second campaign of plant trials (S2), the energy penalty of recycling ladle slag was compensated by the process- control system by decreasing the iron ore content and increasing the amount of oxygen gas to the process. As a result, the increase in blowing time was decreased compared to the first plant trial. In addition it was closer to the increase that was simulated in the theoretical study.

An agreement between the three studies was also shown regarding the change in steel composition when recycling of ladle slag is made. More specifically, it was generally shown that the effect is low and the metallurgical performance is maintained when ladle slag is used. However, a potential kinetic effect of an increased slag weight was not included in the theoretical approach. Although it was shown, that the change in the equilibrium state of the slag/metal reactions is less than what is experimentally detectable.

The main results, an increased slag formation and a maintained metallurgical performance, have been in agreement with previous studies [10-12]. However, the amount of added ladle slag has been lower in this work. The addition of 1000 kg per heat in this work corresponds to a value of 7.7-8.7 kg per ton liquid steel compared to 10-30 kg per ton of steel in previous work [10-12]. Mink and Dye [10]

also concluded that the lining wear was unaffected by recycling of BOF slag to the

converter. The affection on lining wear of slag recycling needs to be performed

Chapter 5

24

during longer plant trial periods and that was not possible during these shorter tests.

Figure 15 The contribution of each supplement to the final results of the effect of ladle slag recycling on the BOF process performance.

A small step towards a non-waste steel production has been investigated in this work. In addition, lime savings of at least 260 kg per heat has been shown to be viable without any implications in the performance of the process. If implemented, this will reduce the energy and CO 2 -emmisions from the production of burnt lime.

However, the total benefit of the ladle slag recycling has to be examined with an environmental and economical life-cycle-analysis.

The theoretical study has explained many of the effects of the ladle slag addition

discovered during the plant trials. Thus, it would have been beneficial to have

performed the theoretical study before the plant trials. In such case, more options

to compensate for the energy penalty would have been examined practically and

may have been able to shed more light on the effect of ladle-slag additions on the

BOF process. Unfortunately, this was not possible within the time frame of the

project so the theoretical study had to be postponed until the plant trials had been

performed.

25

Chapter 6

C ^ONCLUSIONS

The effect on the BOF performance of using ladle slag as a substitute for lime in the slag formation process has been studied during this work. The possible lime reduction thanks to the recycling has been shown to be at least 260 kg per ton of added ladle slag. The most important conclusions regarding the effect on the BOF- process parameters are summarised as follows:

 No demerits in the metallurgical performance of the BOF process are associated with the addition of recycled ladle slag. Similar sulphur and phosphorus levels in the final steel were seen for heats with and without the addition of ladle slag.

 No major difference in the slag composition occurred when ladle slag was recycled to the process.

 The increased tendency for slopping when recycling ladle slag to the process could be controlled by a modification of the lance position during the initial stages of the blow.

 Melting of the ladle slag occurred early in the process and increased the Al ₂ O ₃ content of the slag before 15% of the total blowing time compared to a reference heat.

 The slag weight was increased during the blow in heats where ladle slag was recycled. However, the lime dissolution was not affected by the addition of ladle slag.

 Because of an increased amount of added slag formers. The slag weight at tapping was increased with the addition of ladle slag.

 To not affect the steel composition, the best way to compensate for the disturbed energy balance by the recycling of ladle slag is to reduce the tapping temperature. If not possible, the recycling will demand a decrease in the yield or increase in the blowing time.

 The increase in blowing time, when using ladle slag and compensating for

the energy penalty with an adjustment of the iron ore additions is

approximately 10 seconds.

Chapter 6

26

27

Chapter 7

F ^UTURE W ^ORK

Ladle slag recycling to the BOF process has been shown to be a viable procedure

and possible during daily production and that lime savings of at least 260 kg per ton

of ladle slag is possible. This will save energy in the production of burnt lime and

reduce the raw material usage in the BOF process. However, it is suggested that an

environmental and economical life-cycle analysis should be performed before

implementation of ladle slag recycling to the BOF process in daily production. In

addition, the effect of the ladle slag on the lining wear of the converter needs to be

studied carefully. This needs to be made with trials with the length of a BOF

campaign. Thus, for a time period of more than 3000 heats.

28

R ^EFERENCES

[1] E. M. Michaelis. Basic oxygen steelmaking - a new technology emerges, pp. 1–10, Proceedings, The Metals Society, London, (4-5 May 1978).

[2] Jernkontoret, Stålindustrin gör mer än stål, Handbok för restprodukter 2009, Jernkontoret, Stockholm, 2010

[3] A. Chatterjee, N.-O. Lindfors and J. Å. Wester, Process metallurgy of LD steelmaking, Ironmaking Steelmaking, No.1, pp.21-32, (1976)

[4] K. Kawakami, Kinetics of blowing reaction in a Basic Oxygen Furnace, J. Met, 18, pp.836-845, (1966)

[5] R. K. Iyengar and F. C. Petrilli, Slagmaking reactions in the BOF process, Open Hearth Proc., Metal Soc. AIME, Cleveland, USA, pp.21-32, (April 1973)

[6] D. Cosma, P. Stubbs and B. A. Strathdee, B.O.F slag control studies at Dofasco, Mc Master symposium on The role of slag in basic oxygen steelmaking, Hamilton, Canada, p.4-1, (May 1976)

[7] A. I. Van Hoorn, J. T. van Konynenburg and P. J. Kreyger, Evolution of slag composition and weight during the blow, Mc Master symposium on The role of slag in basic oxygen steelmaking, Hamilton, Canada, p.2-1, (May 1976)

[8] C. Cicutti, M. Valdez, T. Pérez, J. Petroni, A. Gómez, R. Donayo and L. Ferro, Study of slag-metal reactions in an LD-LBE converter, 6 ^th Int. Conf on molten slags, fluxes and salts, Stockholm-Helsinki, Paper 367, (2000)A

[9] M. Andersson, T. Sjökvist and P. Jönsson, Processmetallurgins grunder, Educational compendium in Applied Process metallurgy, Department of Materials Science and Engineering, 2006

[10] P. Mink and S. Dye, Reducing Vessel Fume Losses and Recycling of BOF residuals, Steelmaking Conference Proceedings, pp.639-651, (2002) [11] R. Donayo, A. Gómez, E. Lagos, J. Pérez, A. Repetto, S. Alzari and C.

Cicutti. Improvements in the converter process at Siderar, AISTech 2004 – Iron and Steel Technology Conference Proceedings, Nashville, USA,

pp.775-782, (Sept 2004)

[12] T. S. Lee, I. S. Choi and W. Y. Song, The technology of recycling ladle slag, Stahl und Eisen, No.12, pp.113-117, (2003)

[13] L. Zhang et al., Development and Applications of Models for Pyrometalurgical Processes, Materials Forum, Vol. 25, 2001, pp.136-153

[14] M.L Kapoor and M.G. Frohberg, Theoretical treatment of Activities in

Silicate Melts, in Chemical Metallurgy of Iron and Steel, The Iron and

Steel Institute, London, 1971, pp.17-22