A

T A

Volume 60

A T

2015

Issue 3

DOI: 10.1515/amm-2015-0336

M. Rywotycki*,#, A. Szajding*, Z. Malinowski*, T. Telejko*, A. Godasz*, M. Bene**

The influence of burner locations in the heating furnace on the charge temperature field

Wpyw usytuowania palnikw w piecu grzewczym na pole temperatury nagrzewanego wsadu do

przerbki plastycznej

Charge heating in industrial furnaces is a difficult and complex process. There are many physical phenomena which
influence heat transfer. At the charge surface heat transfer takes place by radiation and convection. In order to ensure correct
operation of the technological system, it is necessary to achieve the required charge temperature in the whole volume and
ensure its uniformity.
The influence of selected burner locations inside the furnace on the charge temperature has been analysed. The
temperature field and its uniformity in the round charge made of steel for hot open die forging have been analysed. The model
and numerical calculations were performed with the ANSYS-Fluent 14.5 package.
Keywords: industrial furnaces, heat transfer, numerical modelling,
Nagrzewanie wsadu w piecach przemysowych jest trudnym i zoonym procesem. Celem zapewnienia prawidowej pracy
cigu technologicznego konieczne jest osignicie przez wsad wymaganej temperatury w caej objtoci, oraz zapewnienie
odpowiedniej rwnomiernoci nagrzewania.
W pracy okrelono wpyw sposobu nagrzewania wsadu w piecu komorowym dla wybranych wariantw usytuowania
palnikw grzewczych. Analizie poddano pole temperatury i jego jednorodno w nagrzewanym wsadzie stalowym
przeznaczonym do przerbki plastycznej. Model i obliczenia wykonano pakietem numerycznym ANSYS-Fluent 14.5.

1. Introduction
Charge heating for open die forging is among the most
important stages of the final product manufacturing. The
problem of auniform temperature field in the ingot that is
heated is important in the production process. The heating
in the furnace is adifficult and complex process. There are
many physical phenomena which influence heat exchange.
Heat transfer occurs by radiation and convection between the
furnace and the surface of the charge. In order to ensure the
correct operation of the technological system, it is necessary
to achieve the required temperature in the whole volume of the
charge and ensure its uniformity. Modelling of the temperature
field and other phenomena as regards the process of steel
production has been analysed by several authors, with both
commercial software and original formulations having been
applied [1-8].
The performance of heating furnaces can be controlled
by the proper definition of the heating time and the furnace
temperature profile. The temperature distribution in the furnace
is limited by the design features of the furnace, the shape and
dimensions of the charge and the physical properties of the

heated material. An incorrect selection of these parameters

may cause incorrect heating of the charge, disrupt the operation
of the furnace or prolong the heating time which, in turn,
can contribute to an increase in the energy consumption and
areduction of the furnace efficiency [9-15].
The influence of selected configurations of the heating
burners in the furnace on the charge temperature has been
analysed. The temperature field and its uniformity in the heated
round steel charge for hot open die forging have been analysed.
The model and numerical calculations were performed with
the ANSYS-Fluent 14.5 package.
2. Heat transfer models
The determination of the charge temperature field
requires solving the heat transfer equation of the charge
volume for the prescribed heat transfer boundary conditions
resulting from the furnace atmosphere and the furnace
walls. The heat transfer between the charge surface and the
furnace atmosphere occurs by radiation and convection. In
order to fully describe the heating furnace operation model,

*AGH University of Science and Technology, AL. A.MICKIEWICZa 30, 30-059 KRAKOW, POLAND
**czech technical university in prague, trojanowa 13, 120 00 prague 2, czech republic
#

Correspondimg author: [email protected]

1982
it is necessary to describe the heat transfer by its structural
components - the walls and fluid composed of combustion
gases. Such acomprehensive approach to the model
construction makes the results of the numerical calculations
accurate, taking into account the most important factors
which influence the charge heating in the chamber furnace.
Neglecting convection in the furnace chamber leads to an
underestimation of the heat flux on the surface of the heated
charge [16].
In order to accelerate the computations, atwo dimensional
(2D) model was applied, allowing the temperature in the
cross-section of the charge placed in achamber furnace to be
determined [17].

2 T

+ qv = c p
2+
y 2

(1)

where:
cp specific heat, [J/(kg K)]
qv efficiency of the internal heat source, [W/m3]
x,y Cartesian coordinates, [m]
heat transfer coefficient of the charge, [W/(m K)]
charge density, [kg/m3]
time [s]
The solution of the equation (1) is atime dependent
temperature field T (x, y, ), which should meet the
boundary conditions at the surface of the material heated.
The boundary conditions were determined on the basis of
the adopted heat transfer mechanisms: prevailing radiation
and supplementary convection related to the movement of
hot combustion gases in the furnace chamber. The model
of Conjugate Heat Transfer for Fluid/Solid interfaces [18]
has been used in the calculation. The Discrete Ordinates
(DO) [18] model of radiation heat transfer implemented in
the numerical package was used in the calculations, whose
Angular Discretization takes on the following values: Theta
Divisions 2, Phi Divisions 2, Theta Pixels 1, Phi Pixels
-1. Other values of those parameters have been considered
during the research. The constant value of the emissivity
of the charge surface and furnace walls amounting to 0.83
and 0.3 respectively has been taken. The combustion gases
flow was described by the standard turbulence model k-e
using Standard Wall Function [18]. The following constant
values have been assumed for this model:: Cmu = 0.09,
C1 =1.44, C2 = 1.92 [19]. The combustion of natural gas in
the burner was modelled using the combustion functions of
the ANSYS - Fluent package. The properties of the material
heated and the furnace structural materials were updated with
the temperature changes during heating [17]. This ensures
asignificant improvement of accuracy in the temperature
field determination.
The system of equations forming the mathematical
model of the heating charge in the furnace was solved by
the method of control volumes with second order upwind
scheme of discretization. The Coupled algorithm was used
for the description of the coupling of the pressure and
velocity fields in the model during the calculations. The
controlled level of residues was at the level of at least 10-3
for momentum and continuity and 10-6 for energy and DO
- intensity.

3. Experimental investigation
Models describing heat transfer in the chamber
furnaces typically assume that radiation is the prevailing
mechanism of heat transfer, and convection is neglected
without modelling the combustion gases flow through the
heating furnace chamber. Neglecting this mechanism in
the zones where burners operate may lead to an incorrect
simulations of the temperature field [16]. At the first stage
of the investigations, an experiment of the charge heating at
the laboratory chamber furnace operated by the Department
of Thermal Engineering and Environment Protection of
AGH University in Krakow was carried out. The charge
temperature at 3 points and the temperature of refractory
lining and insulation at 3 points were recorded. A scheme
of the furnace is shown in Fig. 1. The furnace length was
1800 mm, height 1000 mm and width 900 mm, and it was
equipped with one burner. The thickness of the insulating
layer was 220 mm. The charge dimensions were: diameter
of 80 mm and length of 337 mm. Temperature measurement
was carried out continuously at asampling rate of 10 Hz,
with type K thermocouples.
Ti1
Ti2
Ti3

insulation
outlet
charge
burner

Tc2

Tc1
Tc3

atmosphere

Fig. 1. Scheme of the research station (Ti - point temperature

measurement of insulation, Tc - temperature measurement charge
point)

The measurement also included recording of natural

gas consumption and acombustion gases analysis, which
was the basis for determining the combustion gases volume.
Measurement of the volume flow of the gas supplied to the
burner was made using arotameter. An analysis of the exhaust
flue duct was made using agas analyser, the PG 250 Portable
Multi-Gas Analyser. The results of the analysis were used to
determine the amount of combustion air (Table 1.).

Chemical composition of combustion gases

TABLE 1

CO2
O2
CxHy

Unit
%
%
%

32 min
10.02
1.54
0.04

65 min
9.34
1.35
0,03

NO2

ppm

NO
CO
N2

ppm
ppm
%
-

54
26
88.4
1.07

59
98
89.28
1.06

1983
4. Numerical computations and results of the laboratory
furnace
The results of numerical calculations were compared
with the measurement results. Numerical calculations were
performed and the steel charge heating process was analysed.
The heating time to the temperature of 1000C in the
charge axis was about 65 minutes. The calculations have
been started from the cold charge placed into the furnace.
The computing results were used to validate the developed
heat transfer model implemented in an ANSYS - Fluent
environment. In the case of the numerical model, after 65
minutes the temperature achieved in the middle of the
charge is 1012C, and so the difference in relation to the
experimental data is below 2%. Figure 2 shows the geometry
and the numerical mesh applied in the calculations.

Fig. 4. The temperature and velocity field after 65 minutes of heating

in the laboratory furnace

The temperature and velocity fields are typical for heating

processes in chamber furnaces. Heat transfer conditions differ
essentially in the furnace chamber. In the zone were burners
are installed, there is an intensive convective heat transfer
between hot gases resulting from natural gas burning and the
surface of the walls and the charge. However, in other zones
combustion gases move much faster. The calculation results
presented in Fig. 5 and Fig. 6 show an agreement of calculation
results and temperature measurements at the test rig.
a)

b)

1200

Fig. 2. The geometry and the numerical mesh applied in the laboratory
furnace model

800

Numerical solution
Measurement

800
Temperature, OC

Temperature, OC

The calculation results yield the temperature fields in the

charge and in the laboratory furnace walls. The temperature
and velocity fields have been presented for ahalf of the heating
time at the 32nd minute (Fig. 3) and at the end of the heating
process at the 65th minute (Fig. 4).

1200

Numerical solution
Measurement

400

400

20

40
Time, min.

60

80

20

40
Time, min.

60

80

Fig. 5. Comparison of the numerical calculations of the charge

temperature with measurements at the test strand a- at the charge
axis, b - at the charge surface

The major differences of the temperature curves at the

charge axis may result from neglecting the heat of phase
transformation in the formulated model.
1100

Numerical solution - 20 mm from furnace interior

Measurement - 20 mm from furnace interior
Numerical solution - 110 mm from furnace interior
Measurement - 110 mm from furnace interior
Numerical solution - 220 mm from furnace interior
Measurement - 220 mm from furnace interior

1000
900
800

Temperature, OC

700
600
500
400

Fig. 3. The temperature and velocity field after 32 minutes of heating

in the laboratory furnace

300
200
100
0
0

20

40
Time, min.

60

80

Fig. 6. Comparison of the numerical calculations of the furnace

temperature with measurements at the test station

1984
The temperature field in the charge cross-section after
heating is presented in Fig. 7. Atypical isotherm distribution
for heating in achamber furnace can be seen. At the end of
heating time the temperature at the charge centre obtained
from the measurement was 1003 oC, and from the calculations
was 1010 oC. At the same time, at the surface atemperature
of 1019 oC and 1012 oC from measurements and calculations
have been obtained, respectively. The highest temperature was
measured at the top corner at the burner side and it was about
1042 oC.

Fig. 8. Influence of mesh size on the calculation results

Fig. 7. The charge temperature field heated in the laboratory furnace

after 65 minutes

5. Numerical computations and results for industrial

chamber furnace
Two geometries of chamber furnace were created for the
same model settings, with burners installed at the side wall
and at the chamber roof. The heating process of the charge in
the form of 2 cylinders with adiameter of 800 mm and length
of 6000 mm was analysed. The cylinders were placed next to
one another at the furnace bottom using spacers (not shown
in the picture), which ensure auniform combustion gases
flow around the charge. 10 hours heating time was taken in
both cases. A 20 oC initial temperature of both the charge
and the furnace has been assumed. The dimensions of the
furnace chamber were: width 3200 mm, height 3000 mm
and length 8000 mm. The simple burner geometry was the
same in both cases.
The authors had no measurement data in the case of
industrial furnaces so the mesh sensitivity analysis was
performed on the mean temperature of the charge 1 and
2 (left and right side). The analysis started from the mesh
numbering of 2460 volumes and it gradually concentrated
until the average temperature of the charge did not change.
The results are shown in Figure 8. The average values of
temperature are presented in the form of adimensionless
ratio of the largest number of numerical mesh (12788 items)
to individual meshes. As can be seen from the Figure 8 graph,
for the mesh with 4465 volumes the differential average
surface temperature of the charge 1 is less than 2% compared
to the result obtained for the mesh with 12788 volumes. It
was therefore considered that the mesh number of 4465 is
sufficient for obtaining correct results. The accuracy of mass
and energy balance was less than 1%.

The temperature field of the charge and the furnace (Figs.

9 and 10) shows the influence of the location of the burners. The
change in the burner locations makes asignificant difference
in the heating of the charge located in the furnace chamber.
Placing the burners at the furnace roof without changing the
exhaust gases escape hole position results in aconsiderably
higher temperature within the chamber. It involves achange
in the heating conditions of the charge resting on the furnace
bottom. The comparison of temperature distributions at the
charge axes for chamber type I and II is presented in Fig. 11.
The temperature increase rate varies depending on the burner
locations and the position of the charge at the furnace bottom.

Fig. 9. The temperature and velocity fields for chamber furnace I after
10 hours

Fig. 10. The temperature and velocity fields for chamber furnace II
after 10 hours

1985
1600

Acknowledgments

Left charge - furnace 1

Right charge - furnace 1
Left charge - furnace 2
Right charge - furnace 2

Research work was supported by NCBiR, as a project

GEKON1/O2/213082/4/2014

Temperature, OC

1200

800

Literature
400

0
0

200

Time, min.

400

600

Fig. 11. The comparison of temperatures variations at the charge axis

heated in chamber type I and II

For burners mounted at the furnace roof it is necessary

to change the layout and system of exhaust gases flow from
the furnace chamber. Modifications should aim at auniform
discharge of gases flow from the chamber at two sides of the
furnace. This should ensure that the differences in the charge
temperature depending on its location decrease.
6. Conclusion
The two dimensional model of the chamber furnace
with various locations of gas burners have been developed
using the ANSYS - Fluent 14.5 package. The effect of
burner locations in the chamber furnace on the temperature
field of the charge for plastic forming processes have
been analysed. In order to verify the model assumptions,
amodel of the laboratory chamber furnace has first been
constructed. The numerical results have been compared
with the measurements. It follows from the analysis
that achange in the burner positions in the furnace has
asignificant effect on the temperature field of the charge.
A change in the burner position should be accompanied
by achange in the combustion gases discharge system. It
should allow auniform flow of the hot exhaust gases over the
charge surface. A further stage of research will be focused on
the development of athree dimensional model to facilitate
more accurate numerical calculations. The model verification
will be performed based on measurements conducted at an
existing industrial chamber furnace for the charge heating
prior to the forging process.
Received: 20 January 2015.

[1] M. Hojny, M. Gowacki, Steel Research International 79, 868874 (2008).

[2] A. Cwudziski, Archives of Metallurgy and Materials 58,
1077-1083 (2013).
[3] L. Sowa, A. Bokota, Archives of Metallurgy and Materials 56,
359-366 (2011).
[4] T. Telejko, Z. Malinowski, M. Rywotycki, Archives of
Metallurgy and Materials 54, 837844 (2009).
[5] Z.Malinowski, T.Telejko, B.Hadaa, Archives of Metallurgy
and Materials 57,325331 (2012).
[6] W. Piekarska, M. Kubiak, Z. Saternus, K. Rek, Archives of
Metallurgy and Materials 58, 1237-1242 (2013).
[7] W. Piekarska, M. Kubiak, Z. Saternus, Archives of Metallurgy
and Materials 57, 1219-1227 (2012).
[8] W. Piekarska, M. Kubiak, A. Bokota, Archives of Metallurgy
and Materials 56, 409-421 (2012).
[9] M. Kieloch, R. Wyczkowski, S. Wyczkowski, Gospodarka
Paliwami iEnergi 12, 8 11 (2000).
[10] Z. Malinowski, Hutnik Wiadomoci Hutnicze 4, 147 150
(1997).
[11] M. Kieloch, Energooszczdne imao zgorzelinowe nagrzewanie
wsadu stalowego. Wyd. WIPMiFS PCz, Czstochowa (2002).
[12] M. Kieloch: Racjonalizacja nagrzewania wsadu, Wyd.
WIPMiFS PCz, Czstochowa (2010).
[13] R. Straka, J. Makovika, M. Bene, Environment Protection
Engineering 37, 13-22 (2011).
[14] M. Kieloch, . Piechowicz, J. Boryca, A. Klos, Archives of
Metallurgy and Materials 55, 647-656 (2010).
[15] P. Mullinger, B. Jenkins, Industrial and process furnaces:
principles, design and operation, Elsevier, ButterworthHeinemann, (2008).
[16] Z. Malinowski, M. Rywotycki, Hutnik Wiadomoci Hutnicze
75, 601604 (2008).
[17] Z. Malinowski, Numeryczne modele wprzerbce plastycznej
iwymianie ciepa. AGH UWN-D, Krakw, (2005).
[18] ANSYS 14.5 User Manual.
[19] C.E. Baukal, Industrial Burners Handbook, CRC Press (2004).