Content uploaded by I. Ruiz-Bustinza
Author content
All content in this area was uploaded by I. Ruiz-Bustinza on Nov 07, 2016
Content may be subject to copyright.
123
METALLURGY AND FOUNDRY ENGINEERING Vol. 37, 2011, No. 2
* Ph.D.: School of Engineering. Universidad Panamericana. Augusto Rodín 498, 03920 México D.F. México;
e-mail: robglez@up.edu.mx
** M.Sc.; Prof., Ph.D., D.Sc.: Sid-Met-Mat Research Group. Universidad de Oviedo. ETSIMO. Indepen-
dencia 13, 33004 Oviedo. Spain; e-mail: siderurgia@etsimo.uniovi.es
*** Ph.D.: Centro Nacional de Investigaciones Metalúrgicas (CENIM/CSIC). Avda de Gregorio del Amo, 8.
28040-Madrid, Spain; e-mail: iruizbustinza@cenim.csic.es
**** Ph.D., D.Sc.; Ph.D.: AGH University of Science and Technology, Faculty of Metals Engineering and Indu-
strial Computer Science, Krakow, Poland; e-mail: mkarbow@agh.edu.pl
Roberto González* , Miguel A. Barbés**, Luis F. Verdeja**,
Iñigo Ruiz-Bustinza***, Javier Mochón***, Ramón M. Duarte***,
Miros³aw Karbowniczek****, Piotr Migas****
MECHANISMS KNOWLEDGE OF THE FLOW AND WEAR
IN THE BLAST FURNACE CRUCIBLE
WITH THE NODAL WEAR MODEL
1. INTRODUCTION
When describing the mechanisms of wear in the lining, and of metal flow around the
dead man (inactive coke) that may be present in the crucible of a blast furnace during its
work life, the following considerations and work methods aimed at tackling the problem
may be found in specialized references [1–3].
a) These are complex phenomena, in which a great number of variables take part, so
a quantitative estimation of wear suffered by the materials and of the density/intensity of
the current lines for the melt along the operation life of the crucible are difficult to find.
b) On the other hand, it is usual for different corrosion mechanisms to act in parallel,
making the evaluation of the participation and importance of each of one of them in the
wear process of the lining very difficult [4]. Furthermore, the fluid velocities in the
crucible are not homogenous as they depend on the occupation degree (geometry), and
characteristics (porosity) and quality of the dead man.
c) Attempts have been made to reproduce and simulate, at a laboratory scale, the opera-
tion conditions the hearth of the blast furnace may withstand (with fluid-refractory
relative velocities of 0.5 cm/s) and design the furnace lining in accordance to the
results obtained in samples in contact with corrosive fluids (pig iron and slag).
d) Finally, an attempt is being made to understand the movement of the melt inside the
crucible through the use of room temperature physical models.
124
These models are also accompanied by measurements using radioactive tracers to
study the characteristics of the pig iron flux in contact with the dead man in the hearth
below the nozzles level. Figure 1 shows a 3D finite element simulation of the melt flow,
where the complexity of the results makes it difficult to estimate wears.
Fig. 1. 3D finite element simulation of melt flow in a dead man hearth, showing mesh and element
distribution (a) and fluid velocity (b)
A fact that differentiates the NWM in respect to other practices aimed at understanding
this problem, is that both wear and scab growth, as well as dead man characteristics, circula-
tion velocities of the fluids (pig iron and slag), furnace operation (production) parameters,
geometric design characteristics and nature of the crucible materials, are not independent
variables and, the values each one of them may reach, slightly or strongly determine all the
others. For example, when trying to explain the behavior of refractory linings submitted to
corrosive action (pig iron and slag), different wear mechanisms or physical characteristics
or geometries the dead man may adopt inside the hearth are described, as if these results
were not a consequence of a determined operation (production) practice, under which the
blast furnace has been demanded to produce for specific periods of time.
The novelty of the NWM is to admit that wear, blast furnace production and dead man
physical and geometric characteristics are connected and dependant phenomena. On the other
hand, the variables that may explain the bidimensional (2D) behavior of a specific section of
the blast furnace crucible, according to the NWM, are reduced to the following four properties:
global heat transport coefficient between pig iron and lining /,
pig iron lining
gi
h−
−
nodal temperature at the pig iron-refractory interface Ti,
nodal position variation (increase or decrease) of the abscissa in the pig iron-refractory
interface Δxi,
nodal position variation (increase or decrease) of the ordinate in the pig iron-refrac-
tory interface Δyi.
While the temperature of the melted components inside the hearth of the blast furnace
(tapping temperature T∞) is an important parameter in order to control the furnace operation,
the temperature at the refractory-melt interface (nodal temperature Ti) or the global heat
transport coefficient between the pig iron and the lining /pig iron lining
gi
h−
− aren’t variables
= >
#
accessible to direct experimental determination [5]. Nevertheless, in all situations studied so
far, experimental data supports the hypothesis, theories and developments of the NWM.
However, the characteristics of problems laid out in blast furnace crucibles, bottom
and walls of electric furnaces, or copper metallurgy PS convertor linings, have only a few
points in common [6, 7].
As already mentioned, both Ti and /pig iron lining
gi
h−
− are not variables accessible through
direct experimentation. They may be, however, indirectly determined if the crucible is
equipped with a temperature measurement system (T ≤ 600 oC) and the corresponding algo-
rithm to determine the position of isotherms inside the lining throughout the furnace opera-
tion life. The flowchart and input and output variables that take part in the thermal 2D prob-
lem posed in the NWM are shown in Figure 2 and Table 1, respectively, in order to study
specific sections of the crucible in a blast furnace (Fig. 3). Particularly, under “zero wear”
conditions the values of both Δxi and Δyi would be null. Nevertheless, at the end of three, six
or nine months of operation, the situation would have changed and each node at the interface
would have experimented a negative variation (corrosion) or a positive one (growth). The
usual procedure would be to calculate the geometric and thermal parameters at the pig iron-
refractory interface nodes in a 2D section after 9 months, using as a starting point, in the
convergence iterations to be executed, the Δxi, Δyi, Ti and /pig iron lining
gi
h−
− values, obtained
for this same section of the crucible, for example, after a 6 months operation period (Fig. 2).
Fig. 2. Determination of nodal properties: /pig iron lining
gi
h−
− and Ti of different 2D sections of a blast
furnace hearth
126
Table 1. Input-output variables in the 2D section study of the NWM of the blast furnace hearth
2. CORROSION OF MATERIALS IN THE CRUCIBLE
OF THE BLAST FURNACE AND THE NWM
If the NWM is considered, the wear-growth of scabs in 2D sections of the materials in
contact with corrosive fluids at the blast furnace hearth, the profiles to be drafted after long
operation periods, may be considered dependant of the following variables:
global heat transport coefficient between pig iron and lining /,
pig iron lining
gi
h−
−
equivalent thermal conductivity of the materials at the lining ,
lining
λ
mean values of the global heat transport coefficient from the cold zone of the refracto-
ry towards the cooling exterior system ,
r
g
h and
nodal geometric variables of either wear or growth (Δxi and Δyi) referred to the initial
geometry of the pig iron-refractory interface at zero wear conditions.
Thus, according to the NWM, the behavior of a 2D section S1 (Fig. 3) of the crucible
along a period of time Δt, of the blast furnace campaign, would be defined in the following
way:
Model-2D-S1
()
/;;;;
pig iron lining r
lining g i i
gi
hhxy
−
−λΔΔ (1)
127
Fig. 3. Position of the twelve sections and tap-holes (TH-1, TH-2 and TH-3) for a blast furnace
hearth with 10.0 m of diameter at the tuyeres level. Porosity
ε
(%) of the dead man (inactive coke) in
the hearth is indicated
Depending on the operational characteristics of the blast furnace (load and gas move-
ment, pig iron/slag production and quality of raw materials), a value of the /pig iron lining
gi
h−
−
parameter will show that, besides being a nodal variable, it may also have considerable
variations along the melt-refractory interface (walls, bottom or corners of the crucible),
corresponding, for example, to the S1 section in Figure 3. In other words, the operation
characteristics of the furnace are related to the /pig iron lining
gi
h−
− value along the pig iron-
refractory interface.
On the other hand, the design of the crucible depends on the geometry, nature and
distribution of the materials used during construction. Besides physical, mechanical
and chemical characteristics, it may be stated that one of the critical factors (though not the
only one) in the lining design is the equivalent thermal conductivity .
lining
λ In this way, by
taking into account this physical property, the model includes (equation 1) a representation
of the variable that best represents the products used during construction.
The geometrical variables that play a role in the nodal wear (or growth), are repre-
sented in the 2D model by the Δxi and Δyi parameters.
Also, one of the variables with special consequence in the study of this process is
related, according to the NWM, with the cooling system characteristics acting on the cold
128
face of the refractory. Major differences would be observed when using forced cooling by
water, air or oil. The r
g
h parameter makes reference to the numerical value of the global heat
transport coefficient between the refractory and the cooling fluid being used.
Depending on the nature and thickness of materials used during the blast furnace con-
struction, cooling at the beginning of the campaign may be inefficient: the elevated capacity
to evacuate heat by the external cooling system may be baffled by the high thermal resis-
tance of the crucible refractories or by a production capacity well below the furnace nomi-
nal standard.
On the other hand, for some metallurgical installations experimental identification of
isotherms in the lining and variables product of the NWM may be calculated through one
single expression, just as indicated by equation 1. However, the calculations for the blast
furnace are more complex [8, 9].
Along the interface of the refractory with the melt in the hearth of the blast furnace,
there isn’t a unique value for /.
pig iron lining
gi
h−
− Generally, different values for the global heat
transport coefficient are reached at the bottom, corners and walls of the crucible. Being
more precise, not even at the characteristic length of both the corners nor the bottom, the
magnitude of /pig iron lining
gi
h−
− is a constant, due to the fact that the transition of nodal values
from a corner towards a wall, takes place not in an abrupt or discontinuous way, but in
a progressive one. Because of this, at least three similar expressions of equation 1 should be
used to define the 2D section of a blast furnace: one for the wall, one for the bottom and,
finally, one for the corners.
Finally, another aspect that complicates the definition of the crucible with an expres-
sion in accordance to equation 1 is the fact that during the campaign, changes in production
practices may cause disruptions in the value of /pig iron lining
gi
h−
− and, thus, from that specific
moment in the operation life of the furnace, modifications in the wear mechanisms of the
crucible may occur.
3. CHARACTERISTICS OF THE DEAD MAN (INACTIVE COKE)
IN THE CRUCIBLE OF THE BLAST FURNACE AND THE NWM
When production in the blast furnace is started and the blowing-in process takes place,
all of the available volume from the collecting-vat and hearth is occupied by coke. At the
tuyeres level, the furnace is slowly heated by the partial combustion of the coke. This
material at the mentioned level is consumed and substituted by the same material from the
upper inlet until reaching a threshold temperature at the hearth, that will allow the ferric
load to be added [2, 10].
Once the proportion of ferric-reduction load fed to the furnace reaches stationary con-
ditions, the situation of the coke after a three-month operation period won’t be very diffe-
rent to the one represented by Figure 4a: dead man resting over the base of the hearth.
Taking as a starting point the Ti and /pig iron lining
gi
h−
− values, the product of the NWM
analysis in one of the crucible sections (Fig. 2), Table 2 presents a briefing of expressions
used to calculate the most representative variables at different zones of the crucible
throughout the work campaign of the furnace (Fig. 4).
129
Fig. 4. Different possibilities in design and position of the inactive coke region (dead man and hearth
raceway) in a blast furnace
Table 2. Equations to calculate the porosity, mean value of coke particles size and superficial velocity
of fluids in the blast furnace hearth
Porosity at the center (0,0) of the static bed of inactive coke particles (Fig. 3) [11].
0.0981 1 1000
(
%
)
100
centro
E
A
dP
Sdt
⎡⎤
⎛⎞
+
⎜⎟
⎢⎥
⎝⎠
⎢⎥
ε=
⎛⎞
⎢⎥
⎜⎟
⎢⎥
⎝⎠
⎣⎦
(2)
Nodal superficial velocity, si
v− (m· s-1). Ranz–Marshall–Kitaiev equation [12–14].
()
1/1.30 1/1,30
0.5
20,5 0.33
20.60
0.60 Pr
2.0
gi i ii
si
i
i
h
vNu
CRMK T
−
−
−
⎡⎤
⎡⎤
ρμ
⎢⎥
=⎢⎥
−
⎢⎥
−⎢⎥
⎣⎦
⎣⎦
(3)
Mean value of coke particles size distribution at the static bed, .
p
D Kitaiev equation [12].
()
1/ 0.75
0.30
0.9
si i
p
gi
vT L
DCK
h
−
−
⎡⎤
=−
⎢⎥
⎢⎥
⎣⎦
(4)
Median superficial velocity, ,
si
v− and porosity, ,
ε at bed. Ergun equation [15–17].
() ()
22
23 3
11
150 1.75
isi isi
p
p
vv
P
LD
D
−−
−ε −ε
μρ
Δ=+
εε
(5)
130
The symbol ε in equation 2 of Table 2, corresponds to the porosity at the center of the
hearth and may be obtained from the production of pig-iron A (t⋅hr–1), slag production E
(kg⋅t pig-iron–1), crucible surface S (m) and pressure variation referred to a one hour time
unit
()
,dP dt consequence of the furnace operation (decrease during tapping and raise
when the tap-hole is sealed) [11].
The nodal surface velocity vsi of the fluid around coke particles in contact with the
bottom, corners or wall of the crucible are calculated from the following nodal properties of
the fluid:
nodal temperature Ti,
viscosity μi,
density ρi,
Prandtl number Pri,
Nusselt number Nui and
global heat transport coefficient between the melt and the refractory
/pig iron lining
gi
gi
hh
−−
−= in expression 3 of Table 2 [1214].
The mean size of coke particles p
D at the region of the dead man inside the hearth, is
a function of /
;;pig iron lining
si i gi
gi
vTh h
−
−−
−= and of the characteristic linear dimension of
the system analyzed L: bottom, corner or wall of the crucible in equation 4 of Table 2 [12].
Finally, the mean surface velocity si
v− and porosity ε at the bed of inactive coke parti-
cles, are related with the Ergun equation through the following variables: ,,,
pii
Dμρ L and
ΔP (pressure variation in the crucible due to tapping or sealing of the tap-hole) as indicated
by equation (5) in Table 2 [15–17].
Figure 4 shows some of the different options that the influence zones of the dead man
may present, along with regions fitted for the free circulation of the melt (raceway of the
hearth). In order to estimate the thickness of these regions where the melt will freely circu-
late, without the barriers imposed by the bed of coke particles, the following equation must
be considered:
()
()
ii i
TTeC
∞
⎡⎤
μ−μ Δ=
⎣⎦ (6)
Where μi(Ti) represents the viscosity of the melt in contact with the wall, corner or
bottom of the crucible, μi(T∞) is the viscosity at the interface of the dead man influence zone
and melt separation, Δe(Δx o Δy) is the thickness of the raceway of the hearth and C is
a constant which characterizes the region of the crucible being studied.
Finally, the following have been published recently: the first outputs of speed pig iron,
size of particles of coke and porosity of dead man in the hearth of blast furnace in the begin-
ning blowing-in [18, 19].
4. CONCLUSIONS
The nodal wear model represents an efficient complement to work with and interpret
experimental isotherms at the lining of any given 2D section of a blast furnace crucible, and
the possibility to obtain the following:
131
1. The resulting differential wear of the refractories that constitute the working lining of
the furnace crucible at any specific zone: walls, corners or bottom.
2. In the same way noticeable wear at corners or central part of the crucible may be de-
tected in a 2D section. Formation of protective scabs at the walls or bottom may also be
detected.
3. The erosion of the lining reflects in the values of Ti and of /,
pig iron lining
gi
h−
− indepen-
dently from the fact that the wear mechanism acting is of the chemical or mechanical
type.
4. Starting from the Ti and /pig iron lining
gi
h−
− values, in a determined 2D section of the cru-
cible, coke properties may be estimated, along with the major or minor influence of the
dead man region with respect to the zones where the melt freely circulates (raceway of
hearth).
5. Considering the results in different 2D sections of the crucible for a specific moment in
the operation life of the furnace, a 3D representation may be possible, including both
of the refractory lining profiles as well as those corresponding to the dead man influ-
ence zone.
Acknowledgements
The authors wish to thank the Ministry of Education and Science (MEC): MAT2003-
00502, the Ministry of International Affairs and Cooperation (MAEC): MAEC-AECI-B/
1629/04; B/2884/05; B/5814/06, B/7648/07 and the CSIC-Madrid (Spain) for making
the scientific and technological cooperation between CENIM/CSIC and Oviedo University
and AGH University of Science and Technology Krakow possible through the “Associated
Unit”.
REFERENCES
[1] Omori Y.: Blast furnace phenomena and modelling, Ed. Elsevier, England (1987) 345352, 395405
[2] Habashi F.: Handbook of Extractive Metallurgy, Metal Industry: Ferrous Metals. Vol. I. Ed. Wiley-VCH,
Germany (1997) 57104
[3] Bennett J. P. , Smith J.D.: American Ceramic Society, Ceramic Transactions, 125 (2001) 135154
[4] Verdeja L.F., Sancho J.P., Ballester A.: Materiales Refractarios y Cerámicos, Ed. Síntesis, Madrid (2008)
1922, 156176
[5] Romero M.A., Jimenez J., Mochón J., Menéndez J.L., Formoso A., Bueno F.: Rev. Metal, Madrid, 36 (2000)
1, 4046
[6] Parra R., Verdeja L.F., Barbés M.F., Goñi C., Bazán V.: JOM, 57 (2005) 10, 2936
[7] Goñi C., Barbés M.F., Bazán V., Brandaleze E., Parra R., Verdeja L.F.: J. Ceram. Soc. Jpn, 114 (2006) 8,
665668
[8] Lüngen H.B., Rüther H.P., Clixby G., Cassella G.: Investigations on blast furnace wear phenomena espe-
cially in the heart, European Commission. Technical Steel Research. EUR 19347 EN (2000) 193
[9] Mülheims K., Pritchard W.D.N., Steiler J.M., Schulte M.: Wear blast furnace heart, European Commission.
Technical Steel Research. EUR 20109 EN (2002) 272
132
[10] Sancho J.P., Verdeja L.F., Ballester A.: Metalurgia Extractiva: Procesos de obtención, Ed. Síntesis, Madrid
(2000) 5574
[11] Zaïmi S.A., Venturni M.J., Sert D.: Rev. Metall-Paris, Journées Sidérurgiques ATS-2002, 99 (2002) 11,
1819
[12] Havelange O., Danloy G., Venturini J.M., Pierret H., Rüther H.P., Mielenz O., Köchner H., Alexander J.A.,
Post J.R., Clixby G.: Determination of coke bed voidage in the blast furnace hearth, European Commission.
Technical Steel Research. EUR 20942 EN (2004) 194
[13] Danloy G., Falzetti M., Formoso A
., Herfurth E., Vega J.: Modeling of gas and char flows at high PCI
through experimental and theoretical studies of the raceway and the dead man, European Commission.
Technical Steel Research. EUR 20094 EN/DE (2002) 224
[14] Alcock C.B.: Principles of pyrometallurgy, Academic Press, London (1976) 9899
[15] Ballester A., Verdeja L.F., Sancho J.P.: Metalurgia Extractiva: Fundamentos, Ed. Síntesis, Madrid (2000)
235238
[16] Poirier D.R., Geiger G.H.: Transport phenomena in materials processing, TMS, Pennsylvania (1994) 9398
[17] Jimenez J., Mochon J., Sainz de Ayala J.: ISIJ Int., 44 (2004) 3, 518526
[18] Martín R., Barbés M.A., Barbés M.F., Marinas E., Ayala N., Mochón J., Verdeja L.F., García F.: R. Metal,
Madrid, 45 (2009) 4, 295304
[19] Barbés M.F., Barbés M.A., Marinas E., Fernández B., Martín R., Mochón J., Verdeja L.F.: Bol. Soc. Esp.
Ceram, V., 48 (2009) 3, 153156
Received
November 2011