Engineering Failure Analysis 45 (2014) 449455

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

Using three-dimensional nite element analysis for simulation

of residual stresses in railway wheels
Reza Masoudi Nejad
School of Mechanical Engineering, Sharif University of Technology, Tehran, Iran

a r t i c l e

i n f o

Article history:
Received 24 March 2014
Received in revised form 11 July 2014
Accepted 26 July 2014
Available online 7 August 2014
Keywords:
Residual stresses
Stress eld
Heat treatment process
Finite element analysis
Thermal loads

a b s t r a c t
One of the most important issues in railway wheels is residual stresses. It is desirable to
produce less residual stresses when possible and to decrease the remaining residual stresses in the wheels. The objective of this paper is to provide an estimation of the residual
stresses in the rail wheel caused by the stress eld from heat treatment process of a railway
wheel. A three-dimensional nonlinear stress analysis model has been applied to estimate
stress elds of the railway mono-block wheel in heat treatment process. After forging or
casting, railway wheels are heat-treated to induce the desirable circumferential compressive residual stress in the upper rim. Finite element analysis model is presented applying
the elasticplastic nite element analysis for the rail wheel under variable thermal loads.
Calculative analysis applying a nite element method (FEM) has been used to predict residual stresses. The quenching and annealing segments of the wheel manufacturing process
are simulated using a decoupled heat transfer and stress analysis. Three-dimensional nite
element analysis results obtained show good agreement with those achieved in eld
measurements.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
The railroad wheel has the initial residual stress created by the manufacturing process, and this residual stress changes
due to the mechanical stress caused by service conditions. The residual stresses of railroad wheels are inuenced by the heat
treatment during manufacture processing. In recent decades, numerous studies have attempted to estimate the residual
stresses. Residual stresses play an important role in mechanical member [1] or destroy its surface material. Farrahi et al.
[2,3] investigated the effects of the residual stress eld and relation to fatigue crack closure and crack growth behavior.
Residual stresses induced by the manufacturing process and this residual stress changes due to the thermal stress induced
by braking, are not apparent visually. The object of several investigations on manufacturing processes is to show a layer of
compressive residual stress on the surface of parts to inhibit propagation of cracks. The effects of the residual stress and
metal removal on the contact fatigue life have been estimated by seo et al. [4,5]. Okagata et al. [6] evaluated the fatigue
strength of Japanese railway wheel and presented the fatigue design method of high speed railway wheel by considering
the effect of manufacturing conditions on the fatigue strength of the material. In the literature [713], some of these issues
are studied using some experimental observation, analytical calculations and FEA calculations within various contexts.
Most of the previous studies described above have estimated the residual stresses using numerical simulations and nite
element method in rail. Unfortunately, existing techniques for estimation residual stresses of the rail wheel problems are

E-mail address: reza.masoudinejad@gmail.com

http://dx.doi.org/10.1016/j.engfailanal.2014.07.018
1350-6307/ 2014 Elsevier Ltd. All rights reserved.

450

R. Masoudi Nejad / Engineering Failure Analysis 45 (2014) 449455

Fig. 1. Mechanical material data for railway wheel.

Fig. 2. Finite element modeling of wheel.

Fig. 3. Thermal load conditions due to the heat treatment process.

R. Masoudi Nejad / Engineering Failure Analysis 45 (2014) 449455

451

Fig. 4. Contour plots of the temperature distribution (in C) due to the heat treatment process: (a) rst phase, (b) second phase, (c) third phase, and (d) forth
phase.

452

R. Masoudi Nejad / Engineering Failure Analysis 45 (2014) 449455

simple model. Finite element models of this nature generally need a ne mesh to achieve accurate results. The simple model
cannot obtain accurate stress eld results of railway wheel under the inuence of thermal loads.
In this paper, a three-dimensional elasticplastic nite element method using the true geometry of a railway wheel has
been used to model and accurately predict the stress distribution due to heat treatment process. The manufacturing process
simulation consisted of two parts, a nonlinear transient thermal analysis and a nonlinear static structural analysis. The heat
treatment process cools the rim of the wheel much faster than the plate of the wheel. The rim of the wheel is sprayed with
water. The plate of the wheel has not been cooled as rapidly.
2. Finite element modeling and residual stress
Manufacturing process used in forming railway wheels induced a wide variety of residual stresses. In the railway wheel
manufacturing process, a heat treatment is used to decrease the risk of rim crack initiation by increasing the surface hardness. Under service condition, the thermal brake loading on rim wheel develops higher tensile residual hoop stress. However,
this process generates higher tensile hoop stress that could contribute to help the formation of rim fatigue cracks. They
reduce fatigue strength and exacerbate the effect of cracks and material defects. The full wheel with true prole geometry
has been constrained to build the nite element model. For modeling wheel we simulate its diameter 920 mm. The stress
analysis in railway rim wheel is used three-dimensional elasticplastic nite element method. In Fig. 1 shows the Elastoplastic behavior at different temperatures using a bilinear isotropic hardening model. In this simulation for residual stress
two parts of analysis used; non-linear thermal analysis and non-linear static structural analysis. The non-linear thermal
analysis determined the temperature distribution of the wheel that varies over time, was used as an input load for the
non-linear static structural analysis. In nite element software, we have this ability to use thermal analysis of model for
structural solution. Fig. 2 shows nite element modeling of wheel in manufacture process for residual stress analysis. The
model mesh has 37,653 eight-noded elements with three translational degrees of freedom at each node with quarter point

Fig. 5. Baseline temperature-time history of two nodes during heat treatment processing.

Fig. 6. Stress distribution (in MPa) in wheel rim during quench processing.

R. Masoudi Nejad / Engineering Failure Analysis 45 (2014) 449455

453

node locations. The heat thermal analysis consist of four phases; rst is high temperature step with water spray on the tread
surface of rim (about 3 min) then, put in room temperature (about 6 min), the elevated temperature draw (about 6 h) and
nally, time for cooling in room temperature (about 7 h). Fig. 3 shows the thermal load conditions due to the heat treatment
process in the thermal model.

Fig. 7. Contour plots of the von Mises stress due to the heat treatment process of railway wheel analysis: (a) rst phase, (b) second phase, (c) third phase,
and (d) forth phase.

454

R. Masoudi Nejad / Engineering Failure Analysis 45 (2014) 449455

The heat treatment process cools the rim of the wheel much faster than the plate of the wheel. The rim-quenched wheels
strengthen the steel, induce the desirable residual compressive stresses in the upper rim, and improve the wear resistance.
The benecial residual compressive stresses are known to help prevent the formation of fatigue cracks at the tread surface of
rim and therefore they are signicant to the trains safety. The input parameters included material properties, temperatures,
times, and boundary conditions, which were thought to affect the residual stress eld. The parameters required for the analysis of heat transfer, including thermal conductivity, and specic heat. Thermal conductivity k, the ability of the material to
conduct heat energy is described by a transfer of property is usually expressed in units of W/m C. This value varies with
temperature as an input to the application is given. Since the heat transfer analysis, including free expansion, so the specic
heat at constant pressure of the material (cp) is used. The heat capacity is dened as the product of cp and the material density resulting in units of J/m3 C. In this paper, the density is constant and equal to 7860 kg/m3 were considered. The heat
capacity varies with temperature as input has been dened for the software. Convection occurs from all wheel surfaces during the quenching process. The heat transfer coefcient for wheel to air is 27 W/m2 C and the heat transfer coefcient for the
portion of the wheel tread which is exposed to the water spray during the quenching process is 3042 W/m2 C. Radiation
from all surfaces of the wheel is permitted during the heat transfer analysis. For this purpose, two parameters are used to
determine the radiative heat transfer. The StefanBoltzmann constant, r is 5.67  10 8 W/m2 K4 and the surface emissivity,
e is 0.96.
One of the most important factors in residual stress forming is a type of constraint in the structure it means that two different movement boundary condition show different residual stresses. In this analysis, the center of wheel in either direction
and outer wheel rim in one direction (perpendicular to the axis) is bounded by using displacement boundary condition.

3. Result and discussion

The non-linear transient thermal analysis yields the time-dependent temperature eld which causes the estimation of the
residual stresses. Since the temperature varies with location and time. As already mentioned in the analysis of air and water
heat transfer coefcient is assumed constant. Fig. 4 shows the temperature distribution due to the heat treatment process in
the thermal model. According to Fig. 4 the effects of water spray on the wheel rim cooling are visible. Also, at the end of the
quench, the minimum temperature in the wheel rim is 326 C and the maximum is 802 C, reduced little from the initial
temperature of 869 C.
Fig. 5. shows the temperature-time histories of two nodes in the model (one on the tread surface of rim and the other at
the plate of the rim wheel), beginning at the initial temperature (869 C), through the quench, annealing, and nally to room
temperature at the end of the cool down. The effect of latent heat release is shown in this gure. Latent heat is released during the transformation from the austenite to pearlite.
The variation of stress distribution of node on the tread surface of rim during quench processing as a result of the thermal
gradient is plotted in Fig. 6. The vertical axis in Fig. 5 again denotes time (s) and the horizontal axis is stress (MPa). Figs. 7 and
8 illustrates contour plots of the Von Mises stresses and von Mises stress according to wheel tread, respectively. For railroad
mono-block wheel, the magnitude of von Mises stress appears to be 553 MPa after this heat-treatment and the tendency of
von Mises stress is not symmetry at upper and down side. The magnitude of this Stress higher than the yield strength of the
steel wheel, that the wheel is a plastic deformation. This Figure clearly indicates the compressive region near the tread and
the tension which exists below the compressive layer. Thus, this analysis shows that the fatigue cracks initiate at a depth
45 mm below the tread surface for railway mono-block wheel. This results show good agreement with those achieved in
eld measurements [14,15].

Fig. 8. Von Mises stress according to wheel tread.

R. Masoudi Nejad / Engineering Failure Analysis 45 (2014) 449455

455

4. Conclusions
Three-dimensional nite element analysis for simulation of residual stresses in railway mono-block wheels is developed
in this paper, which is based on the residual stresses from heat treatment process. For this purpose, Non-linear 3-D nite
element analysis is used for residual stress. The following conclusions can be made:
(1) The nite element analysis results show the very signicant stress distribution. The resultant stress eld is high value
and their effects are not negligible in the crack initiation.
(2) The results revealed that stress eld is highly sensitive to the variable thermal loads. Therefore, this factor signicantly
affects the stress eld of railway wheels during heat treatment process.
(3) The results conrmed the magnitude of the circumferential stress in the wheel since this has been identied as a reasonable means of assessing the possibility of fatigue cracks to initiate.
(4) The railway mono-block wheels are rim-quenched using a water spray to induce benecial hoop residual compressive
stress at the tread surface.
(5) Results of the baseline analysis suggest the presence of a 45 mm thick residual compressive layer with stresses as high
as 553 MPa in a new mono-block wheel. Therefore, the residual stresses have a signicant effect on fatigue life.
In this paper, the residual stresses in the rail wheel caused by the stress eld from heat treatment process of a railway
wheel have been investigated. Future research needs to consider interactive effects of those parameters. Also, other effects,
such as the process parameters, boundary conditions, and material properties need to be included in railway wheel.
References
[1] Farrahi GH, Tirehdast M, Masoumi Khalil Abad E, Parsa S, Motakefpoor M. Failure analysis of a gas turbine compressor. Eng Fail Anal
2011;18(1):47484.
[2] Farrahi GH, Majzoobi GH, Hosseinzadeh F, Harati SM. Experimental evaluation of the effect of residual stress eld on crack growth behaviour in C(T)
specimen. Fract Mech Anal 2006;73(13):177282.
[3] Smith DJ, Farrahi GH, Zhu WX, McMahon CA. Experimental measurement and nite element simulation of the interaction between residual stresses
and mechanical loading. Int J Fatigue 2001;23(4):293302.
[4] Seo JW, Goo BC, Choi JB, Kim YJ. Effect of removal and residual stress on the contact fatigue life of railway wheels. Int J Fatigue 2008;30(1011):11521.
[5] Seo JW, Kwon SJ, Jun HK, Lee DH. Effects of residual stress and shape of web plate on the fatigue life of railway wheels. Eng Fail Anal
2009;16:2493507.
[6] Okagata Y, Kiriyama K, Kato T. Fatigue strength evaluation of the Japanese railway wheel. Fatigue Fract Eng Mater Struct 2007;30:35671.
[7] Lingamanaik Siva N, Chen Bernard K. Prediction of residual stresses in low carbon bainiticmartensitic railway wheels using heat transfer coefcients
derived from quenching experiments. Comput Mater Sci 2013;77:15360.
[8] Lingamanaik Siva N, Chen Bernard K. Thermo-mechanical modelling of residual stresses induced by martensitic phase transformation and cooling
during quenching of railway wheels. J Mater Process Technol 2011;211(9):154752.
[9] Lin KY, Huang JS. Analysis of residual stresses in railroad car wheels based on destructive test measurements. Theoret Appl Fract Mech
1989;12(1):7386.
[10] Moyar GJ. An analysis of the thermal contributions to railway wheel shelling. Wear 1991;144(12):11738.
[11] Shen XH, Yan J, Zhang L, Gao L, Zhang J. Austenite grain size evolution in railway wheel during multi-stage forging processes. J Iron Steel Res Int
2013;20(3):5765.
[12] Seo JW, Kwon SJ, Jun HK, Lee DH. Effects of residual stress and shape of web plate on the fatigue life of railway wheels. Eng Fail Anal
2009;16(7):2493507.
[13] Tawk D, Mutton PJ, Chiu WK. Experimental and numerical investigations: alleviating tensile residual stresses in ash-butt welds by localised rapid
post-weld heat treatment. J Mater Process Technol 2008;196(13):27991.
[14] Lonsdale C, Demilly F, Fabbro VD. Wheel rim residual stress measurements. <www.standardsteel.com>.
[15] Hlavcek P, Brumek J, Horsk L. Using of abrasive water jet for measurement of residual stress in railway wheels. Tehnicki vjesnik 2012;19(2):38790.