3-D CFD Analysis for Blast Furnace Hearth Wear

David Roldan1, Yu Zhang1, Rohit Deshpande1, D. (Frank) Huang2, Pinakin Chaubal2, and Chenn Q. Zhou1
1
Purdue University Calumet, Department of Mechanical Engineering, Hammond, IN 46321, USA
2
Mittal Steel Company, USA R&D Center, East Chicago, IN 46312, USA

Tel: 219-989-2665
Email: [email protected]

Key Words: Blast Furnace Hearth, Erosion, CFD, Heat Transfer

ABSTRACT

Distributions of liquid iron flow and refractory temperatures have significant influences on the blast furnace hearth wear. Conditions
inside the furnace are generally inferred from the measurements of thermocouples/heat flux sensors embedded in the refractory. A 3-D
computational fluid dynamics (CFD) model has been developed to simulate hot metal velocities and temperatures as well as refractory
temperatures. The CFD model has been validated using both laboratory measurements and industrial blast furnace on-line data. In this
research, a new methodology is developed which uses the 3-D CFD model and the 1-D heat transfer model to predict the blast furnace
hearth erosion and inner profile based on the detailed hearth refractory temperature records of a blast furnace. Its application to the
Mittal Steel IH7 blast furnace has demonstrated the effectiveness of this methodology.

INTRODUCTION

Blast furnaces will remain the predominate iron-producing equipment continuously in the foreseeable future. Because the capital cost
are intensive for building a blast furnace system, to prolong the campaign life will be one of keys for reducing the ironmaking cost. A
longer campaign life can significantly lower costs and increase productivity because less blast furnace downtime and financial cost
would be needed for repairs and refractory relining.

At present time, most of the existing hearth erosion models are based on the heat transfer through refractory lining only [1, 2, 3, 4] and
some of them considered the effect of hot metal flow [5]. The traditional approach to define the carbon hearth erosion is based on the
calculation of the isothermal lines of 1150°C [1, 2, 3, 4, 5], which is the solidification point of the molten iron saturated with carbon. The
theory of using 1150°C isothermal line as the “erosion” line is that under this temperature, the hot metal either has been solidified or
has been saturated with carbon and no longer to absorb any more carbon from the carbon brick. The 1150°C isothermal line might be a
good divider to separate the good lining and bad lining, but, it would be neither the erosion line nor the inner profile. The embedded
consumable sensors [6] may monitor the erosion of a blast furnace hearth lining locally and in-one-way, straight forward, which can not
predict the skull build-up or the inner profile. The inner profile of a hearth will be the key to the prediction of the hearth erosion, the
maintenance of the hearth, and the operations of casting.

In order to determine the hearth erosion and hearth inner profile, the hearth side wall and bottom temperatures are generally needed. In
theory, there are three more sets of information should be also obtained: 1). the detailed initial hearth design and all thermal
conductivities hearth lining materials, 2). the skull properties, and 3). the hot face temperatures. In this paper, a new methodology is
presented which uses the thermocouple data, the 1-D heat transfer model and the 3-D CFD results to predict the hearth erosion and
inner profiles and requires an iteration procedure. The details of the methodology along with an example to demonstrate the
methodology are described in the following sections.

MATHEMATICAL MODEL

Governing Equations

The 3-D CFD software numerically solves the governing equations of the flow properties on a computational grid with specified
boundary and initial conditions. In this work, the 3-D computational domain includes the shell, ram, refractory materials, hot metal,
and the deadman. Some assumptions have been employed to formulate the governing equations. The hearth is considered to be at
steady state. The free surface of the hot metal is assumed flat and horizontal. The governing equations of flow properties are derived
from fundamental conservation laws and relevant state relations. The equations of mass, momentum, enthalpy, and species are all

AISTech 2006 Proceedings - Volume 1 167

 elliptic-type partial differential equations. For convenience in numerical formulation, these equations are arranged in a common form,
Equation (1).

3
∂ ∂ξ (1)
(ρu i ξ − Γ )=S
i =1 ∂x i ξ ∂x i ξ

in which ξ is a general flow property, xi, i=1,3 are coordinates, ui, i=1,3 are velocity components, Γ is effective diffusivity, and Sξ is
the sum of source terms. For the region with permeability between 0 and 1, it is treated as porous medium requiring special treatment
in the governing equations.

Boundary conditions are needed to solve the governing equations. The free surface of hot metal is an inlet flow boundary with given
velocities and temperature. The inlet velocity of the hot metal based on the production rate is assumed uniform at the inlet. Non-slip
boundary conditions exist on all solid walls. The upper surface of the refractory walls is set as adiabatic. At the taphole exit, pressure is
kept at atmospheric pressure and a mass flow boundary condition is implemented to ensure mass balance. Side and bottom wall
surfaces are cooled with water at a given temperature. Heat transfer coefficients are given on the outer surfaces based on the
temperature measurements. More details on the governing equations used in the CFD model can be found in a previous publication[7].

Methodology for Predicting the Erosion and Inner Profile of the Hearth

There are a few models to predict the erosion and inner profiles and almost all of the models use 1150° isothermal line as the erosion
line. Heat conduction models based on thermocouple readings in the lining have been reported in literatures [1, 2, 3, 4]. A more accurate
and straight forward approach to determine the inner profile of the blast furnace hearth is to directly use the 3-D CFD hearth model and
thermocouple measurement results[7]. In this model, a routine was developed to adjust the inner profile by adding the skull on to the
previous hot face or reducing the previous lining layer (including the skull) thickness until there is a good match between the measured
and calculated temperatures. This method is simple and valid but is very inefficient algorithm. In order to provide a more accurate and
efficient method for predicting the erosion and inner profiles, a new methodology has been developed in this research that uses 1-D
heat transfer model and thermocouple data to estimate the primary inner profile and then uses the 3-D CFD model to refine it. Based
on the CFD results shown in Figure 1 which is a plot of heat flux distributions, the 1-D heat conduction assumption is reasonable in
most of the regions in the refractories except at or near the corner area. The details of the 1-D heat transfer model [10] can be found in
Appendix.

Figure 1 Heat flux paths

The hot face is defined as the interface between the hot metal and the skull or refractory. In reality, the hot face temperatures vary at
different locations. It is almost impossible to make measurements of these temperatures. However, the 3-D CFD model, which predicts
the distributions of hot metal velocity and temperature inside the hearth, the skull, and the refractory, can provide detailed hot face
temperatures at every location. Based on the 1-D heat transfer model, the thermocouple data, and the 3-D CFD model, the hearth
erosion and inner profiles can be estimated in the following iteration procedure, which is also illustrated in Figure 2:

1. Using the 3-D CFD model to simulate a base case which has the original hearth geometry without skull
2. Using the thermocouple data and the 1-D heat transfer model to estimate the primary inner profile with the hot face
temperatures from the base case in Step 1.
3. Using the 3-D CFD model to compute the hot metal flow field and all the temperature distributions based on the primarily
estimated inner profile
4. Obtaining the hot face temperatures from Step 2 and re-estimating the inner profile from the thermocouple data and the 1-D
heat transfer model

168 AISTech 2006 Proceedings - Volume 1

 5. Repeat steps 2 and 3 till the hot face temperature is converged
6. Tuning the inner profile till all the calculated temperatures match with those measured with a required accuracy.

Use 3-D CFD to simulate the base case

Obtain the hot face temperatures (Thf, i)

Use 1-D heat transfer model to calculate

the primary inner profile

Thf, i=Thf, i+1

Run 3-D CFD model i=i+1

Obtain the hot face temperatures (Thf, i+1)

no
Check: convergence
yes
Finalize inner profile

Figure 2 Methodology for Predicting the Erosion and Inner Profile of the Hearth

RESULTS AND DISCUSSION

The new methodology has been used to estimate the erosion and inner profiles for the Mittal Steel IH7 blast furnace hearth for
different time periods. The following shows an example to demonstrate this methodology. In this example, the first month of the new
hearth operation in October 2003 was considered. The production rates and all the thermocouple data were averaged during this time
period. The detailed properties of the refractories for the blast furnace hearth linings and structure as well as the thermal conductivities
of the sidewall skull and bottom skull were reported in a previous paper [8]. The non-uniform thermal connections [ 9], i.e., the actual
thermal conductivity of the ram layer between the cooling device and the sidewall carbon brick were also considered.

Calculated Inner Profiles

The base geometry in a vertical cross section through the taphole is shown in Figure 3a. The base geometry has the original thickness
of all the refractories and has no skull and erosion. The hearth is divided into three zones: the dead zone, coke zone, and coke free zone
which have the “permeable porosities” of 0.0, 0.5, and 0.8, respectively. The primary and finalized inner profiles are showed in Figures
3b and 3c, respectively. In both figures, there is skull formation on both the side and bottom walls of the hearth. The finalized inner
profile has a slightly thicker skull formation than the primary inner profile. The results indicate that the erosion is very minimal during
the first month of the operation.

Figure 4 illustrates the average hot face temperature as a function of height for different iterations. The figure shows that the hot face
temperature values converge after three iterations. The average hot face temperature increases with each iteration. Therefore, the skull
thickness should increase since the hot face temperature is proportional to the skull thickness (See Appendix for the details of the 1-D
Heat Transfer Model). In each iteration, the result shows that the average hot face temperature decreases as the height increases. This is
because as the hot metal flows to the bottom more energy is being lost through the side walls which leads to the temperature decrease.
The results show that it is important to consider the non-uniform distributions of the hot face temperatures.

AISTech 2006 Proceedings - Volume 1 169

 Coke zone Dead zone Coke free
Carbon =0.5 =0 zone
Brick =0.8

Fire Brick

Mullite Brick
Al-C Brick
Graphite Beam

(a) Base geometry

Skull

(b) Primary inner profile

Skull

(c) Finalized inner profile

Figure 3 The Inner Profiles

1600
Avg. hot face temp. (C)

1500

1400

1300

Base
1200 1st iteration
2nd iteration
1100
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Elevation (m)

Figure 4 Average hot face temperatures

Comparisons between Thermocouple Data and CFD Results

During each iteration procedure, thermocouple data is used to estimate the inner profile and then compared with CFD results. At the
beginning, the difference between the measured and calculated refractory temperatures is large and this difference reduces with more

170 AISTech 2006 Proceedings - Volume 1

iterations. Figure 5 is a parity plot. The agreement between the measurements and CFD results is well even at the third iteration. The
average difference is 16%. This difference can be reduced by the final tuning of the skull thickness locally.

200
180
160
140

Meeasured T (C)
120
100
80
60
40
20
0
0 20 40 60 80 100 120 140 160 180 200
CFD Results (C)
Figure 5 Measured and Calculated Temperatures at the Third Iteration

Flow Characteristics

The CFD software can be used to analyze velocity and temperature distributions in a blast furnace hearth. Figure 6 displays a 3-D view
of temperatures and velocity vectors including a vertical plane and a horizontal plane through the tap hole. The black lines outline the
dead zone and coke zone, in addition to the thickness of the skull and the fire brick. Figures 7 and 8 show the temperatures and
velocity vectors in a side and a top view respectively. It is noted that in the figure x is along the vertical direction, z is along the
horizontal direction toward the tap hole and y is along another horizontal direction perpendicular to x and z directions. This x-y-z
coordinate direction is applicable to all the plots in this manuscript. Because the dead zone is not permeable to the fluid, the hot metal
is not able to flow through the deadman directly to the tap hole. Therefore, the fluid can only travel through the more porous zone.
Velocity of hot metal near the side wall under the tap hole is large, being sucked up into the tap hole for discharge. In contrast,
velocities of hot metal flow are very low near the deadman region and at the corner opposite to the tap hole. At the plane near the tap
hole level, peripheral flow velocity is high and a ring shaped channel flow formed, bypassing the dead zone at the center.

The hot metal temperature distribution of is rather uniform especially in the upper part. The temperature corresponds well to the
molten iron velocity. Namely the temperature is high at the side wall and the corner of bottom under the tap hole where the velocity is
high. The temperature distribution in the skull and the refractory is quite non-uniform due to water cooling effects. In the skull, the
temperature decreases from the top to the bottom vertically, which is consistent with the hot face temperatures shown in Figure 4.

Figure 6 3-D Flow Characteristics

AISTech 2006 Proceedings - Volume 1 171

 (a) Temperature contours

(b) Velocity vectors

Figure 7 2-D Side View

(a) Temperature contours

172 AISTech 2006 Proceedings - Volume 1

 (b) Velocity vectors

Figure 8 2-D Top View

SUMMARY

A new methodology has been developed in this research which integrates the 1-D heat transfer model with the 3-D CFD model to
predict the hearth erosion and inner profiles. The 1-D heat transfer model estimates the refractory and skull thicknesses locally based
on the thermocouple data and hot face temperature calculated by the CFD model. The 3-D CFD model calculates the velocities and
temperatures of the hot metal as well as temperatures in the refractories and skulls based on the inner profile estimated by the 1-D heat
transfer model. An iteration procedure has been developed to perform the calculations efficiently and accurately. The application of
this methodology to the Mittal Steel IH 7 blast furnace has demonstrated its effectiveness.

ACKNOWLEDGEMENTS

Special thanks to Indiana 21st Century for supporting this research and to Dr. Fang Yan for helping in the software development.

REFERENCES

1. Shibata, K., Kimura, Y., Shimizu, M., and Inaba, S., “Dynamics of Dead-man Coke and Hot Metal Flow in a Blast Furnace
Hearth,” ISIJ International, 1990, Vol. 30, No. 3, pp. 208-215
2. Torrkulla, J. and Saxen, H., “Model for Estimation of Blast Furnace Hearth Conditions,” IRONMAKING CONFERENCE
PROCEEDINGS, 1999, Vol. 58, pp. 649-655
3. Torrkulla, J., and Saxen, H., “Model of the State of the Blast Furnace Hearth,” ISIJ International, Vol. 40 (2000), No.5, pp. 438-
447
4. Torrkulla, J. Brannbacks, J., and Saxen, H., “Indicators of the Internal State of the Blast Furnace Hearth,” ISIJ International, Vol.
42 (2002), No.5, pp. 504-511
5. Takatani, K., Inada, T., and Takata, K., 2001, “Mathematical Model for Transient Erosion Process of Blast Furnace Hearth,” ISIJ
International, Vol. 41, No. 10, pp. 1139-1145.
6. Stokman, R., Stein Callenfels, E., and Laar, R., “Blast Furnace Lining and Cooling Technology: Experiences at Corus IJmuiden,”
Iron and Steel Technology, 2004, pp. 21-28
7. Yan, F., C.Q. Zhou, D. Huang, P. Chaubal , and A. Zhao, "3-D Computational Modeling of a Blast Furnace Hearth", Iron and
Steel technology, Vol. 2, No.1, pp. 48-58, 2005
8. D. Huang, P. Chaubal, H. Abramowitz, and C. Zhou, “Hearth Skulls and Hearth Wear Investigation of ISPAT Inland’s #7 Blast
Furnace”, AIST 2005 Proceedings, Vol. 1, pp. 101-112, 2005

AISTech 2006 Proceedings - Volume 1 173

 9. D. Huang, P. Chaubal, F. Yan, and C. Zhou; “The Use of a CFD Model for Understanding the Internal Conditions in a Blast
Furnace Hearth”, 2005, the 5th European Coke and Ironmaking Congress, June 12-15, 2005, Stockholm, Sweden, pp.1-15.
10. D. Huang, “Hearth Wearing Models (Design & Pseudo Codes)”, October 22, 2003, Internal Document of USA R&D Center,
Mittal Steel Company (formerly, R&D Center, Ispat Inland Inc.)

APPENDIX – 1-D HEAT TRANSFER MODEL[10]

a. Side wall
With this assumption a one dimensional Fourier’s Law is used to estimate the erosion and skull in side the hearth. Figure A-1 shows
the temperature, radius and thermal conductivity of the structure side wall in the hearth. It is composed of three materials or layers
which are the carbon brick (CB), fire brick (FB), and skull. The skull layer is assumed so that it is incorporated in the formula. THF and
rHF are the temperature and radius of the hot face, respectively. In addition, TS and TD are the shallow and deep thermocouple
temperatures, respectively.

q4 q3 q2 q1

TS TD TCB TFB THF

Center line
kskull

rHF

kCB kFB rFB

rCB
rD

rS

Figure A-1 Side wall structure

The one dimensional steady-state Fourier’s Law in cylindrical coordinates, with no heat generation, is shown in Equation (A-1);

dT
q = −kA , where A = 2πrL (A-1)
dr
The side wall structure can be divided up into four sections, q1 thru q4 which is shown in Figure 1. Equation (A-2a) shows the formula
for q1 which is obtained by separating the variables in Equation (A-1) and integrating both sides to get

2 π L ( T HF − T FB ) (A-2a)
q1 =
1 r FB
ln
k skull r HF

Note that q2, q3, and q4 can be found similarly,

2 π L ( T FB − T CB )
q2 = (A-2b)
1 rCB
ln
k FB r FB

174 AISTech 2006 Proceedings - Volume 1

 2 π L ( T CB − T D )
q3 = (A-2c)
1 rD
ln
k CB rCB

2 π L (T D − T S )
q4 = (A-2d)
1 r
ln S
k CB rD

In this 1-D heat transfer model, it is assumed that the heat transfer only along the radius out forward, based on the heat balance, we
have

q1 = q 2 = q 3 = q 4 (A-3)

To solve above equations it gives:

THF − TD T D − Ts
= (A-4)
1 r 1 r 1 r 1 r
ln( CB ) + ln( D ) + ln FB ln( s )
k FB rHF k CB rCB k skull rHF k CB rD

Solve for rHF in equation (4) to get

rFB (A-5)
rHF =
1 rS
(T HF − T D ) ln
k CB rD 1 rCB 1 rD
exp − ln − ln k skull
(T D − T S ) k FB rFB k CB rCB

Equation (5) is only valid if all three layers exist. If there is no skull formation, for example erosion occurs but the fire brick is still
intact, then only q2, q3, and q4 are incorporated in the formula and q1 is not included. Follow Equations (A-1) through (A-3) and solve
for the radius of the hot face to get,

rCB
rHF = rFB = (A-6)
T − T D k FB r k r
exp ( HF ) ln( s ) − FB ln( D )
T D − T s k CB rD k CB rCB

Again Equation (A-6) is only valid if the carbon brick and fire brick exist. Similarly, the fire brick can be completely dissolve and only
the carbon brick and skull exist, then radius of the hot face can be found by,

rCB (A-7)
rHF =
1 rS
(T HF − T D ) ln
k CB rD 1 rD
exp − ln k skull
(T D − T S ) k CB rCB

Finally if only the carbon brick exist the radius of the hot face can be found by,

AISTech 2006 Proceedings - Volume 1 175

 rD (A-8)
rHF = rCB =
r T HF − T D
exp ln S
rD T D − T S
b. Bottom wall
Now all hot face radii can be found on the sidewall. Next, is to calculate the skull thickness on the bottom wall, which can be found in
a similar method. Figure A-2 shows the bottom structure of the furnace. Z represents the distance and Tup and Tdw are the upper and
lower thermocouples, respectively.

Hot Metal (THM)

Skull (kskull)

Z3 Z4 Z5 Z6
Fire Brick (kFB)
Mullite Brick (kMB)
Al-C Brick (kAl-C)

Graphite Beam (kGB)

Z2
Thermocouples

TUP
TD

Z1
Figure A-2 Bottom wall structure

One dimensional Fourier’s Law is still used for the bottom skull thickness but using Cartesian coordinates instead of Cylindrical
coordinates. The skull thickness calculation is shown in Equation (A-9),
z1 T hf − TUP z2 z z z
z 6 = z skull = − − AlC − MB − FB k skull (A-9)
k GB TUP − T DW k GB k AlC k MB k FB

Equation (A-9) is only valid if all the bottom refractories and skull exists. If there is no skull and all the refractories still exist then the
thickness of the fire brick can be found by,

z1 T hf − TUP z2 z z
z 5 = z FB = − − AlC − MB k FB (A-10)
k GB TUP − T DW k GB k AlC k MB

Equation (A-11) is used to calculate the skull thickness if the fire brick is completely dissolved and Equation (A-12) is used to
calculate the thickness of the mullite brick if no skull and fire brick exists.

z1 T hf − TUP z2 z z
z skull = − − AlC − MB k skull (A-11)
k GB TUP − T DW k GB k AlC k MB

z1 T hf − TUP z2 z
z 4 = z MB = − − AlC k MB (A-12)
k GB TUP − T DW k GB k AlC

Equations (A-5) through (A-12) can be used to determine the skull and erosion profile inside the whole furnace.

176 AISTech 2006 Proceedings - Volume 1