E. Ranjbarnodeh et al, Journal of Advanced Materials and Processing, Vol. 1, No.

2, 2013, 29-37 29

Heat Transfer Modelling in Wide Gap Rail Thermite Welding

A. Mohassela, A. H. Kokabi a, P. Davamia, E. Ranjbarnodeh b*, M. Movahedia

a
Department of Materials Science and Engineering, Sharif University of Technology, Tehran, Iran
b
Young Researchers Club, Department of Engineering, Islamic Azad University, East Tehran Branch Tehran,
Iran

ARTICLE INFO ABSTRACT

Article history: In the current study, a thermal model was used to investigate the
Received 24 Aug 2012 effects of welding parameters on the distribution of temperature in
Accepted 5 jan 2012 wide gap aluminothermic rail welds. To solve the governing
Available online 20 May 2013 thermal equation, SUTCAST, a finite difference program (FDM)
was employed while different aspects such as phase change and
convective heat transfer in weld pool were taken into account in the
Keywords: numerical solution. To validate the predictions, the modeling
Heat transfer results were compared with the experiments and despite some
Aluminothermic rail welds utilized simplifications, a reasonable agreement was found between
FDM numerical and experimental results. The effects of welding
Preheating temperature
Initial liquid temperature variables such as preheating; initial liquid temperature and weld
Weld gap. gap on temperature distribution were studied. The results of the
modeling showed that the magnitude of weld gap had the greatest
influence on thermal behavior of the joint and the highest weld gap
results in most stable results.

1. Introduction rail segment with a 4.5 to 6 m length and

employing two narrow gap size welds.
Thermite welding is a process in which heat Although this operation can eliminate one
released from aluminothermic reaction is used defect but creates two intrinsic imperfections
for melting large industrial components [1,2].
which will be active at certain conditions in
Effortlessness, cost-effective and ease of future. Use of one wide gap weld instead of
transportation are the main reasons for the two narrow gap welds is another solution
extensive use of the process for welding and
which can decrease the number of welds in
repairing the rails in railway industries [3,4]. railroads and cause saving in time and cost so
Weld gap is one of the most important factors a decrease in cost about 900$ per each weld
affecting the weld properties, so thermite can be attainable [5].
welding is divided into three groups based on Similar to all fusion welding processes, final
the gap size. In narrow gap and wide gap quality of thermite welds depends on thermal
thermite welds, the gap size is smaller than conditions introduced during welding
1inch and greater than 3 inches, respectively operation.
and the gap between 1 to 3 inches is named
medium gap size. Repairing the faulty rail can
be done by cutting defective section, using a

*Corresponding author. Tel.: +989122140219

E-mail address:islam_ranjbar@yahoo.com(E. Ranjbarnodeh).
E. Ranjbarnodeh et al, Journal of Advanced Materials and Processing, Vol. 1, No. 2, 2013, 29-37 30

Macro and micro-structures, welding defects important parameter that influences the
and residual stresses are some of the welding- thermal behavior of narrow gap welds and
related events affected by weldment thermal welds with greater gap size exhibit low
history. In spite of thermal cycle importance, sensitivity to variation of other parameters, so
a few investigations have been carried out on that stable thermal conditions can be attained
heat transfer analysis of thermite welded in welds with greater gap size. According to
structures. Schroeder [6] and Jha[7] studied the authors research, no comprehensive study
thermal distribution in base rails and heat has been conducted in wide-gap thermite rail
treatment after welding but ignored to study welds.
the thermal conditions in weld metal and in In this study, heat transfer and thermal
base rail immediately adjacent to weld metal. distribution in wide gap thermite rail welds
Chen et al. [8] have simulated heat transfer in and the effects of initial melt temperature,
narrow gap size thermite welds by a finite- preheating temperature and gap size are
element method using FIDAP package from investigated. Furthermore, preheating
fluent but disregarded studying thermal operation was taken into the account through
conditions appeared in wide gap thermite measurable preheating temperature which is
welds regardless of thermal variations which more applicable than preheating heat-flux
occur when gap size range changes. They used by Chen.
included heat conduction in modeling
explicitly and took into the account other 2. Model development
thermally physical phenomena such as
Figs. 1 and 2 show different parts of physical
solidification, melting, heat conduction and
radiation through material properties, initial model used for wide gap thermite rail welds.
and boundary conditions. They also employed The external segment of model including two
halves of silica form mould and the welded
uneasily measurable preheating heat-flux to
describe preheating conditions obtained in rails are seen in Fig. 1 and the rails with the
parent rail while this could be performed with molten filler metal are shown in Fig. 2. The
dimensions of the mould is 381 mm153 mm
temperature as a measurable parameter. Their
results show that gap size is the most 138 mm.

Fig.1.Physical model including two halves of form mould and rails.

E. Ranjbarnodeh et al, Journal of Advanced Materials and Processing, Vol. 1, No. 2, 2013, 29-37 31

Fig. 2.Rails and weld metal.

The model is symmetric on the longitudinal Where k is the thermal conductivity, T the
plane, so one half of the model was meshed temperature in Kelvin,Qo the heat generation
and studied. A heat transfer model was rate, p the density, Cp the specific heat andt
introduced to analyze thermal behavior of this the time. From the other thermally physical
process using SUTCAST, the finite-difference process, heat convection was included
package from Razi Metallurgical Research through material properties and radiation was
Center. This software has originally been ignored.
developed to simulate heat transfer and fluid A convection type heat transfer was applied
flow in casting parts. It is necessary to delete by the boundary condition summarized in
the previously defined air gap at rail-melt Table 1 and illustrated in Fig. 3 along with the
interface, to allow melt penetration and hence initial and boundary conditions shown in the
the occurrence of welding. It can happen if a same Fig. The (0, 0, 0) was selected as
high heat convection coefficient is applied at reference point and a mesh with 2 mm length
the rail-melt interface. Because of intricate in all directions was chosen for one half of the
character of the process, it is difficult to physical model. Tables 2 and 3 show the
incorporate all the involving parameters in the chemical compositions of steel rail and weld
model. According to Chen's study, conduction metal, respectively.Weld metal has relatively
as prevailing heat transfer method was the same composition as steel rail does, so the
explicitly considered in the model as follows: same thermal properties were selected for
both of them. Fig.4 shows variation of
T thermal conductivity and specific heat as a

k
T  Q Cp 1 function of temperature, respectively.
t

Table 1.Boundary conditions

Boundary zone back Front Right Left Bottom Top

Environment type Convection Insulation Convection Convection Convection Convection
E. Ranjbarnodeh et al, Journal of Advanced Materials and Processing, Vol. 1, No. 2, 2013, 29-37 32

Fig. 3.Initial and boundary condition.

Table 2.Chemical composition of steel rail (wt%)

C Si Mn P S Fe
0.554 0.240 0.880 0.016 0.024 Bulk

Table 3.Chemical composition of weld metal (wt%)

C Si Mn P S Fe
0.577 0.44 0.600 0.012 0.011 Bulk

Fig. 4.Thermal conductivity and specific heat of rail and weld metal as a founction of temperature.
E. Ranjbarnodeh et al, Journal of Advanced Materials and Processing, Vol. 1, No. 2, 2013, 29-37 33

High heat transfer by convection in molten shown in Fig. 2. The thermal result of sample
metal at temperatures above melting point S5 was used for verification of thermal
was taken into account by applying a factor of model.The prepared joint was cut and
3 in thermal conductivity of steel rail. Latent sectioned longitudinaly and then,
heat generated or absorbed during solid-state macroetched with 17% HCl solution in
transformations and melting or solidification molten state for 30 min. Penetration depth
phenomena were also included through the observed in experimental specimen was
specific heat term in equation 1. compared with simulation result. The
simulation result was in good agreement with
3. Validation experiment. Further details are mentioned in
In order to validate the developed model, two the next section. After validation, in order to
segment rails with standard specification investigate the effects of initial liquid
UIC60-Grade700A, and composition temperature, weld gap and preheating
tabulated in Table2 and welding conditions temperature, welding conditions given by
listed in Table 4 were joined together with Table 5 were designed and the results are
molten metal from reaction of thermite discussed in the following section.
powder in a silica mould enclosing the rails as
Table 4.Experimental welding variables

variable Weld gap (mm) Preheating temperature (oC) Initial liquid temperature (oC)

Value 70 850 2250

Table 5.Welding conditions used in simulation

Sample Weld gap (mm) Preheating temperature(oC) Initial liquid temperature (oC)

S1 60 850 1850

S2 60 850 2050

S3 60 850 2250

S4 50 850 2250

S5 70 850 2250

S6 60 350 2250

S7 60 600 2250

4. Results and discussion to rail center and then, was decreased from
Fig. 5 compares the amount of melt penetration center to corner. Fig. 7 shows the effect of
obtained from experimental measurement melt temperature on temperature distribution
with simulation result at S5 condition. Despite at S1, S2 and S3 specimens. As can be seen, in
the assumed simplifications, only 0.17% addition to melt penetration, the temperature
difference existed and hence, there was a of all points in rail head, web and base
good agreement. Temperature variations at S1 increased with the increase in melt
condition and in two different time ranges are temperature. This was an expected
shown in Fig. 6. As can be seen, the phenomenon, because of the increasing heat
temperature was first raised from rail corner input with increasing melt temperature.
E. Ranjbarnodeh et al, Journal of Advanced Materials and Processing, Vol. 1, No. 2, 2013, 29-37 34

Fig. 5.Comparison between amounts of melt penetration obtained from the experimental measurement with the
simulation result at S5 condition.

(a)

(b)
Fig. 6.Isothermal profile at S1 condition: (a) Time 9.11 sec.-45.57 sec(b) Time: 54.69sec.-91.14 sec.
E. Ranjbarnodeh et al, Journal of Advanced Materials and Processing, Vol. 1, No. 2, 2013, 29-37 35

S1 S2 S

Fig. 7.Effect of initial melt temperature on melt penetration.

The effect of weld gap on melt penetration is maximum temperature of rail points, with
shown in Fig. 8. The extent of high increasing weld gap. The amount of heat input
temperature areas increased with the increase has increased with the increase in weld
in weld volume as well as the increase in volume, as expected.

Fig. 8.Effect of gap size on melt penetration.

The preheating temperature is another penetration.The larger preheating temperature

parameter which can affect the rail causes the greater extent of melt penetration.
temperature distribution. Fig. 9 shows the Heat sink which is introduced by rail bulk
effect of preheating temperature on melt caused temperature of rail points to increase
E. Ranjbarnodeh et al, Journal of Advanced Materials and Processing, Vol. 1, No. 2, 2013, 29-37 36

in addition to melt penetration. Fig. 10 briefly with maximum possible weld gap, the effect
shows the effect of the discussed parameters of other parameters will be negligible. In this
on melt penetration. As can be seen, the effect situation, the results will be less sensitive to
of weld gap on weld penetration was greater the variation of other factors and stable weld
than the others. This is in agreement with penetration can be obtained.
Chen's results. So, if weld metal is established

Fig. 9.Effect of preheating temperature on melt penetration

Fig. 10.Effect of initial melt temperature, gap size and preheating temperature on melt penetration.

5. Conclusions 2- Increasing melt temperature as well as

weld gap and preheating temperature increase
1- The model can predict the amount of weld
the temperature of all rail points. It seems that
penetration in a relatively good agreement
the effect of weld gap is greater.
with experimental data.
E. Ranjbarnodeh et al, Journal of Advanced Materials and Processing, Vol. 1, No. 2, 2013, 29-37 37

3- To have stable weld penetration, maximum welded rail steel", Word Academy of
possible weld gap is recommended. Science, Engineering and Technology,
Vol. 60,2009, pp.558-562.
Acknowledgment [5] K. Sawley, J. Sun, D.R. Poirier, "Wide-
The authors are grateful to Railway Research gap welds offer railroads potential
Center and Razi Metallurgical Research savings",Railways Track and Structures,
Center for supporting this project. Vol. 4, 1999, pp. 12-14.
[6] L.C. Schroeder, D.R. Poirier, "The
References mechanical properties of thermite welds in
[1] R.N. Hart, "Thermite Welding premium alloy rails",Materials Science
Process", 2rd ed., Lindsay Publications, Engineering, Vol. 63, 1984, pp. 1-8.
Bradley, 1987, p123. [7] B. Iha, "Thermite Welding of Rail Steel",
[2] J.E. Brumbaugh, "Welders Guide", 3rd Ph.D. Thesis, Illinois Institute of
ed., Howard W. Sams& Co., Indianapolis, Technology, Illinois, 1989.
1975, pp. 19. [8] Y. Chen, F.V. Lawrence, C.P.L. Barkan,
[3] J. Myers, G.H.Geiger, D.R. Poirier, J.A. Dantzig, "Heat transfer modeling of
"Structure and properties of thermite welds rail thermite welding",Proceeding of
in rails",Welding Journal, Vol. 258, IMechE Part F: J. Rail and Rapid Transit,
1982, pp. 8-19. Vol. 220, 2006, pp. 207-217.
[4]S. Rajanna, H.K. Shivanand,
"Improvement in mechanical behavior of
expulsion with heat treated thermite
Journal of Advanced Materials and Processing, Vol. 1, No. 2, 2013 38