International Journal of Engineering Research ISSN:2319-6890)(online),2347-5013(print)

Volume No.4, Issue No.1, pp : 4 - 8 01 Jan. 2015

Stress Analysis on Behaviour of Rails

M.Kameswara Reddy, Dr.K.V.S.Srinadh, T.V.Ravi Teja, Rafiuzzama Shaik

Abstract : It is axiomatic to say that the rails used nowadays KittisakKuntiyawichai, Enrico Spacone and Minho Kwon, in
have been subjected to heavier axle loads and high operating their paper[2] presented an alternative way to derive the exact
speeds. This may lead to derailment and eventually loss of element stiffness matrix for a beam on Winkler foundation and
life. In this paper Transient analysis (Dynamic analysis) for the fixed-end force vector due to a linearly distributed load. The
rail by using ANSYS 12.1 Software is discussed. In this element flexibility matrix is derived first and forms the core of
analysis deflection and stress are compared with wheel the exact element stiffness matrix. The governing differential
diameter, load, speed and range of distribution of contact compatibility of the problem is derived using the virtual force
load. Result is analyzed by the effect of variation of these principle and solved to obtain the exact moment interpolation
parameters. functions. The matrix virtual force equation is employed to
Keywords: Stress on Rails, Analyzing rails stresses obtain the exact element flexibility matrix using the exact
moment interpolation functions. The so-called “natural”
1.Introduction element stiffness matrix is obtained by inverting the exact
In October 2000 a high-speed train derailed less than one km element flexibility matrix.
south of Hatfield station near London in UK. Four passengers Ulf Olofsson and Roger Lewis, in their book on
were killed and more than seventy people were injured. The Tribology of wheel rail contact[1], have focussed on the
cause of the accident was fracture and subsequent friction, wear, and lubrication of the tiny contact zone (roughly
fragmentation of the outer rail on a curved section. The 1 cm2), where steel wheel meets steel rail, from a mechanical
investigations revealed numerous fatigue cracks at the running engineer’s viewpoint
corner. When one of these penetrated into the web and foot of Ma Weihua and LuoShihui, in their Seminar given in
the rail it resulted in a knock-on effect leading to the extension the Conference of Intelligent Computation Technology and
of adjacent cracks and, thereby, to the fracture of a complete Computation on Analyses of Wheel/Rail Unsymmetrical
rail section. Usually rail breakage will not have such drastic Problem[3], have discussed about wheel surface damage. They
consequences. Nevertheless it is a paramount objective of every discussed three types of surface damages: Wheel Tread sapling,
railway company worldwide to avoid rail breakages. Smith, Out of Round Wheel and Wheel flange wear. They presented
collecting data from various sources, provides a comparison of that the unsymmetrical wheel/rail contact was an important
the frequency of fracture events of various railway components reason which can leads to the wheel surface damage problem.
in Great Britain at the end of the 19th and the end of the They suggested that to improve the dynamic performance and
20thcentury. Although the list is certainly subject to some the stability of the vehicle, the strong wheel/rail unsymmetrical
uncertainties, he is able to conclude: “…it is clear that whilst contact must be avoided.
failures of wheels and axles have been reduced by a factor of Zerbst, U., Lundén, R., Edel, K.O. and Smith, R.A in
20 over the last century, failures of rails per train km have their book on Introduction to the Damage Tolerance Behavior
actually increased by a factor of more than 2”. Reasons behind of Railway Rails[4] introduce the most important questions
this trend are heavier axle loads, increased volumes of traffic regarding crack propagation and fracture of rails. These include
and axial tensile stresses at low temperatures due to the loading conditions: contact forces from the wheel and
continuously welded rails. thermal stresses due to restrained elongation of continuously
Worldwide there were comparable numbers of rail breakages. welded rails together with residual stresses from manufacturing
Systems for heavy haul freight with their high axle loads are and welding in the field. An overview of crack-type rail defects
significantly more affected than systems with predominantly and potential failure scenarios has been provided.
passenger transport, the authors [1] give an approximate figure P HosseiniTehrani, M Saket in their Seminar in
of the economic costs of rail fracture and its avoidance: € 2000 Conference Modern Practice in Stress and Vibration
million per year in the European Union alone. A positive trend Analysis[5] they have given study of multi-axial high-cycle
found with respect to the mentioned British network was that, fatigue initiation life prediction for railroad is done in this
whilst the detection rate of damaged rails which then had to be paper. Using ANSYS 11.0 software three dimensional elasto-
removed, has been increased continually throughout the period plastic finite element model of rail/wheel contact was
under consideration, the number of breakages was virtually constructed and fine mesh technique in contact region is used to
constant. This fact shows that countermeasures, such as non- achieve both computational efficiency and accuracy. Stress
destructive inspection and periodical grinding, have brought a analysis is performed and fatigue damage in railroad is
significant improvement towards failure prevention. On the evaluated numerically using multi-axial fatigue crack initiation
other hand, the requirements on the networks such as an model.
increased volume of traffic and higher axle loads etc. are 2.Sress Analysis on Rails
permanently increasing. Therefore, fatigue crack propagation in Dealing with a rail system loaded dynamically by
rails remains an important issue with respect to both the assuming the load is being transmitted through the rail span as
quantitative understanding of the Mechanisms and the three progressive line contacts. The base result of the analysis
development of analysis routines for practical application. includes the values of deflection and stress plotted against

IJER@2015 Page 4
 International Journal of Engineering Research ISSN:2319-6890)(online),2347-5013(print)
Volume No.4, Issue No.1, pp : 4 - 8 01 Jan. 2015

different loads, speeds and range of distribution of the contact The above figure shows the rail model used for the
load. The range of distribution of the contact load is analysis the report. The I-section is created as an area entity,
demonstrated by transforming the three progressive line which is then extruded to a dimension of 0.6 m (600 mm). The
contacts loading as three sets of increased number of line cross-section details are listed below.
contacts.  Web length=100mm;
2.1 Modeling in ANSYS software:  Web thickness=20mm;
The modeling of the rail is kept very simple, by  Base length= 200mm;
creating an I-section as cross section (entity-area). The area  Base thickness=20mm;
entity is then extruded to create rail (entity-volume). The length  Head length=100mm; and
of the extrusion is considered as 0.6 mt, which is a standard for  Head thickness=20mm.
Indian rail span [1]. 2.5 Meshing of the model generated:
2.2 Element used: The meshing of the model is done using hexagonal
Element used for meshing the model is “Solid45”. sweep. Element size used is 10 mm. The meshed elemental plot
SOLID45 is used for the 3-D modeling of solid structures. The of the volume generated (along with the boundary conditions)
element is defined by eight nodes and orthotropic material is shown in the figure below:
properties having three degrees of freedom at each node:
translations in the nodal x, y, and z directions. The element has
plasticity, creep, swelling, stress stiffening, large deflection,
and large strain capabilities.

Figure3: element plot (with boundary conditions) of the model

2.6 Stress distribution along the rail span model:
Figure1: element solid 45 The stress distribution and transmission for the rail
The solution output associated with the element is in two model with the transmitting load along the length of the span is
forms: generated. These results are shown in the figures below in
 Nodal displacements included in the overall nodal which the load is transfered along the span in three steps.
solution
 Additional element output (stresses as the chapter
uses)
2.3 Assumptions and Restrictions:
 Zero volume elements are not allowed.
 The element may not be twisted such that the element
has two separate volumes. This occurs most frequently
when the elements are not numbered properly.
 All elements must have eight nodes.
2.4 Model of the rail generated:
Figure 4:stress along the span at load step 1

Figure 2: model used in ANSYS

Figure 5:stress along the span at load step 3

IJER@2015 Page 5
 International Journal of Engineering Research ISSN:2319-6890)(online),2347-5013(print)
Volume No.4, Issue No.1, pp : 4 - 8 01 Jan. 2015

Figure 9: deflection along the span at load step 3

2.8 Load vs. stress curves:

Figure 6:stress along the span at load step 2

2.7 Deflection along the span:
The variation of the deflection along the span with the
transmission of the load (through the steps) is plotted in the
figures below.

Figure 10:variation of stress along the span with varying load

In the above fig. the graph shows variation of the

stress along the span for different loading conditions. The graph
gives a clear idea about how the curves are offset with the
varying load without actually changing the nature of the
response.
Figure 7: deflection along the span at load step 1

Figure 11: variation of the stress at a point with varying load

In the above fig depicts a graph between increase of

Figure 8: deflection along the span at middle load (x-axis) and stress (y-axis) at a given node and as expected
stress also increases with increase of load.
2.9 Speed vs stress curves:

IJER@2015 Page 6
 International Journal of Engineering Research ISSN:2319-6890)(online),2347-5013(print)
Volume No.4, Issue No.1, pp : 4 - 8 01 Jan. 2015

Figure 13: range of the distribution of the contact load

(increased distribution)
The stepped loading along the three steps is carried out
with the same number of the steps but the distribution of the
loading is increased (by increasing the number of the lines
Figure 12: graph showing independence of the stress with the through which the nodes used for point loading is increased), as
speed (timing of loading) mentioned in the above mentioned passage.The results show
that with the increase in the distribution of the loading the
In the above fig. a graph is drawn between length of nature of the stress response hardly changes, even the
the rail span (x-axis) and stress (y-axis), for varying speeds magnitude remains mostly constant, except for the region
(transmission of the loads). From this curve it can be where the peak amplitude comes into the picture, the change is
pragmatically observed that for same loading condition with predominantly a preferred change, as the peak of the stress
change of speed there is no change of stress distribution (both value gets reduced, i.e.; the maximum value of the stress along
in nature and magnitude). But it is a general notion that with the the span of the rail gets reduced. With this the report concludes
increase in speed the stresses in the rail do increase, not only a that with the increase in range of the distribution of the load the
notion but also proven in many test results. The only difference response plot does not alter for most of the span of the rail, but
between the practical results and the results shown above is that the peak definitely becomes low.
absence of the traction force. With this the thesis comes to a
conclusion that the variation in stress with the variation in the
speed is only associated with the traction force between the
wheel and the rail, as the modeling of the rail in this thesis does
not include the contact detail of the wheel and the rail. To
highlight the effect of the speed, the speed variations are made
highly exaggerated (0.072 to 720 Kmph).

2.10 Range of distribution of contact load vs. stress:

The distribution of the contact load is increased by increasing

the number of the lines of the nodes included in point loading
(not by making the loading as the pressure loading). Care must Figure 141: variation of the stress along the span of the rail
be taken regarding making the point load magnitude only a with increased distribution of the contact load
fraction as the range of the distribution increases (to maintain 2.11 Load vs. maximum deflection:
the total load constant). The figure below shows the increased
distribution of the load (in comparison to the loading of the
nodes along a single line). This can be associated to the
situations like, with the decrease in the stiffness of the wheel
the distribution of the load does increase (due to increased
deformation), increase in diameter of the wheel also increases
the distribution of the loading (as the line contact gradually
tends to be area contact).

Figure 15: variation of maximum deflection in span with

variation in load

IJER@2015 Page 7
 International Journal of Engineering Research ISSN:2319-6890)(online),2347-5013(print)
Volume No.4, Issue No.1, pp : 4 - 8 01 Jan. 2015

In the above figure, a graph is drawn between load (x- increase of range of distribution of contact load, the decrement
axis) and maximum deflection along the span of the rail (y- in the deflection value is getting saturated.
axis), the observation shows that as expected, with increase of
load the maximum deflection also increases. The increment 4 Conclusion:
obtained is linear.
Variation of the rail speed is the interesting result we have
obtained. As, both the results of the variation with the stress
2.12 Speed vs. deflection: and maximum deflection get the thesis to a common conclusion
that when the analysis does not include contact details,
physically referring to tractive details between wheel and rail,
along with the details of thermal and residual stresses, the
variation is independent of the wheel speed. With this the thesis
concludes that the reason for the variation in practical results
can be due to any one of the above mentioned causes or even
all of them intact.
1. The results of the variation of the stress and maximum
deflection with the variation of the load magnitude is mostly as
expected that with the increase in the magnitude of the load,
both the parameters do increase. But, the heart of the point of
the load variation is that the increase of the deflection is linear.
2. The results using the variation in the range of
distribution of the contact load are unique from most analysis
Figure 16: variation of maximum deflection with variation of
of the others, as the variationin its value can be assigned to the
wheel speed
variation in any of the following:
a. Variation in diameter of the wheel;
The graph shown above is a plot between speed of the
b. Variation in deformation at the point of contact; or
wheel (x-axis) and maximum deflection the span of the rail (y-
c. Due to any of the reasons resulting in the variation
axis). This result clearly supports the conclusion from the graph
contact.
shown in figure 12, which states that the stress is unaffected by
These results conclude that the stress is mostly unaffected by
the speed of the wheel (unless the contact is defined). The
the variation of the range of the distribution of the contact load,
above statement will hold true only if the deflection also
but the peak is definitely reducing. At the same time the
behaves unaffected by the changing rate of wheel speed. Hence,
variation of the maximum deflection gets to saturation after
the above result clarifies the point that deflection is unaffected
some increment in the range of the distribution of the contact
by the changing wheel speed.
force (before which the magnitude of the maximum deflection
does decrease).
2.13 Range of distribution of contact load vs. deflection:
REFERENCES:

i. A book by Ulf Olofsson and Roger Lewis, “Hand book of

railway vehicle dynamics” Published in 2006 by CRC Press.
ii. A paper by SuchartLimkatanyu, KittisakKuntiyawichai,
Enrico Spacone and Minho Kwon,”Natural stiffness matrix for beams
on Winkler foundation: exact force-based derivation”.Structural
Engineering and Mechanics, Vol. 42, No. 1 (2012) 39-53.
iii. A paper by Ma WeihuaLuoShihui, Song Rongrong,”Analyses
of Wheel/rail Unsymmetrical Contact Problem”. International
Conference on Intelligent Computation Technology and Automation
2008.
iv. A paper by Zerbst, U.; Lunden, R.; Edel, K.-O.; Smith,
R.A.”Introduction to the damage tolerance behaviour of railway
Figure 17: variation of maximum deflection with the variation rails”.In Engineering Fracture Mechanics (2009) Elsevier.
v. A paper by P HosseiniTehrani, M Saket.” Fatigue crack
of distribution in the range of the contact load
initiation life prediction of railroad”. 7th International Conference on
In the above figure a graph between range of Modern Practice in Stress and Vibration Analysis. Journal of Physics:
distribution of contact load and deflection is drawn, the Conference Series 181 (2009) 012038.
observation shows that with increase in the range of distribution
of contact load, the deflection value decreases, but with further

IJER@2015 Page 8