Destressing of Long Welded Rails – A Rational Approach

By

G S Yadav, Professor Bridge, IRICEN

Synopsis: During service the pattern of longitudinal thermal stresses in LWR may undergo significant
changes due to the effect of tractive and/or braking forces, change in vertical profile of track due to
settlement of track or maintenance operations like lifting /lowering of track, repairs of rail fractures
or replacement of defective rails etc. If the level of thermal stresses vary significantly from the
theoretical profile of thermal stresses, the track may become vulnerable to buckling and/or
fractures. Therefore, it becomes necessary to carry out the de-stressing of LWRs to restore the
longitudinal thermal stresses to desirable level. Indian Railway’s LWR manual prescribes certain
situations the occurrence of which warrants the de-stressing of LWRs. The requirement of de-
stressing due to SEJ gaps falling beyond permissible limits and due to repairs of rail fractures or
replacement of defective rails has been critically examined in this paper and a rational approach has
been suggested.

1.0 Introduction:

Destressing of LWR (as per para 6.4 of LWR manual) shall be undertaken when abnormal behaviour
of LWR gets manifested in one or more of the following:

(i) When the gap observed at SEJ

(a) Differs beyond specified limits
(b) Exceeds the maximum designed gap at SEJ
(c) When tip of tongue rail/corner of stock rail crosses the reference line
(ii) After specified special maintenance operations
(iii) After restoration of track following the unusual occurrences mentioned below
(a) Rail fractures or replacement of defective rail/glued joint
(b) Damage to SEJ/buffer rails
(c) Bucking or tendency towards buckling
(d) Factors causing disturbance to LWR such as accidents, breaches etc.
(iv) If number of locations where temporary repairs have been done exceed three per km

some of the reasons for de-stressing like (i) gaps at SEJ, (ii) damage to SEJ/buffer rails and (iii)
number of locations where temporary repairs have been done exceed three per km; appear to be
technically arbitrary. A rational approach based on effect of incidents on the stability of LWRs need
to be evolved.

The Stress pattern in central portion of LWR, sufficiently away from end of BL, has no bearing on the
gaps at SEJ. Condition of ballast and its packing (which determines the longitudinal ballast
resistance) and condition of fittings in BL determine the actual length of BL. The amount of
expansion/contraction of Rail at SEJ will actually depend upon actual breathing length in field
conditions.

The level of maximum thermal stresses in LWR depends upon variation of Rail temperature from the
stress free temperature. The thermal force diagram in LWR is shown in figure 1

AEαt αt
Maximum movement m at the end of LWR is given by (
m = (Lb αt)/2 ¿ R x 2 )
AEαt
The Breathing Length (BL) is given by Lb ¿ R

Maximum force in Central portion of LWR P= AEαt

Where, A = cross sectional area of Rail

E = Modulus of elasticity of Rail steel

α=Coefficient of thermal expansion of Rail Steel

t = variation of rail temperature from stress free temperature

2.0 Behaviour of Breathing length and its impact on gap at SEJ:

Gap observed at SEJ forms the main reason for destressing of LWRs during maintenance. However,
this doesn’t stand in technical scrutiny.

SEJ gap basically reflects the behaviour of breathing length (BL) and /or creep of LWR. In a
reasonably well maintained LWR with adequate toe load of fittings, there won’t be any creep. In
such a LWR the gap at SEJ will reflect only the behaviour of Breathing Length.
As shown in figure 2, the fittings in BL, especially GRPs, are subjected to not only wheel loads but
also longitudinal shearing force (equal to ballast resistance of one sleeper which for 60 kg rail is =
13.74 X 60 = 824 kg (assuming sleeper spacing as 60 cm)). GRPs in central portion of LWR are not
subjected to longitudinal shearing force as longitudinal ballast resistance is fully mobilised at end of
BL.

This results in accelerated wear and tear of GRPs due to crushing and bulging. On heavy density
routes the GRPs lose their efficacy within 2 to 3 years. The Toe load of ERCs is considerably reduced,
resulting into ineffective clamping action between rail and sleeper. This leads to relative movement
of rail with respect to sleeper, at least at some of the sleepers. A longer length of track comes in to
play to develop full longitudinal ballast resistance thus increasing the BL. This resultant increase in BL
leads to movement of Rail at SEJ more than the theoretically calculated values.

Fig2. Vertical load and longitudinal shearing force in GRP in Breathing Length of LWR

2.1 Effect of SEJ fittings on gap at SEJ:

The gap at SEJ is also influenced by the arrangement of fittings and fixtures at SEJ (figure 3). As per
standard arrangement of sleepers and fittings for SEJ, two number of angles are fixed with SEJ
sleepers (6 numbers in conventional design).

These angles are meant to ensure that SEJ sleepers do not move from their position to avoid
loosening of ballast packing and to avoid sleepers becoming out of square. In such a situation the
fittings on SEJ sleepers should be rail free fastenings. But standard drawing provides for normal ER
clips. This makes relative movement between rail and SEJ sleepers impossible. In such a situation the
SEJ sleepers (except two central sleepers) try to move with rail, but the tie angles fixed to sleepers
resist this movement. Thus a rigid frame resisting longitudinal movement of Rail at SEJ is created . In
such a situation the movement of Rail at SEJ will not take place as per LWR theory.

2.2 Effect of Change in Pattern of Thermal Stresses in the Central Portion of LWR:

If the pattern of thermal stresses in some part of central portion of LWR is changed due to
maintenance operations like rail replacement etc., it will not in any way affect the movement of Rail
at SEJ as the movement purely depends upon behaviour of BL (Creep in LWR is ignored presently,
and will be dealt with separately in para 4.1)

For example if Rail renewal is done at a location 1 (Figure 4), it will affect the force in track for a
length of two temporary BLs created due to discontinuity in track.

Therefore, as movement of Rail at SEJ depends on condition of track in BL, destressing of LWR on the
basis of movement of Rail at SEJ is technically unnecessary.

3.0 Analysis of Effect of Repairs:

Now let us try to understand the effect on thermal stresses in LWR due to repairs to track by way of
rail replacements in central portion of LWR.

If there is a fracture or rail replacement on any other account at location (1) in figure 4, the stresses
for a length of 2 temporary BLs (corresponding to variation of temperature at the time of fracture)
will be altered. If repairs are done in such a manner that the length of rail removed from track is
exactly same as the length of rail inserted (including welding) and stresses after welding are
equalised post repairs for a length of at least two temporary BLs (one temporary BL on either side of
the affected location), then the forces in LWR will get restored to original level (effect of hysteresis is
ignored in this analysis). However, in practice this is seldom achievable. We invariably end up either
with longer or shorter rail inserted in track as compared to rail removed from track. This alters the
pattern of thermal stresses in the affected portion of LWR permanently. The level of such stresses
(forces) can be calculated from the basic theory of LWR.

Fig. 4 Change is stress pattern in central portion due to repairs to fracture

If length of rail inserted in track varies by δL from the length of rail removed from track, and
equalisation of stresses is done, say over 250 meters (125 metres on either side), then the force
change in LWR would be

F = (δL / 250) * EA (Hooke’s Law)

The equivalent temperature which would cause this change in force can be calculated by equating
above value with AEαt

Thus, δt = δL / (250 * α)

The δL can be calculated as under:

If repairs are done, say at Td – 30, then the length of rail inserted in track would be more than the
length of rail removed by an amount of contraction of LWR over two temporary BLs corresponding
to Td-30

Lb = AEαt / R

Using standard values for 60 kg rail and 60cm sleeper spacing,

Lb= 76.86∗2.15∗10 6 * 1.152∗10−5 * 30 / 13.74 = 4156 cm = 41.56 m

Total contraction of rail (for both temporary BLs)

δL = 2 *(Lb αt)/2

= (2 * 41.56 *100 * 1.152∗10−5 * 30) /2

= 1.44 cm = 14.4 mm

Therefore, rail replaced at Td-30 will result in insertion of extra 14.4 mm rail in track. This extra
length may slightly vary on either side due to effect of hysteresis effects. Now the equivalent
temperature corresponding to 14.4 mm extra length of rail, assuming equalisation of stresses is done
for 125 metre of track on either side of affected location, can be calculated by formula:

δt = δL / (250 m * α)

= 14.4 / (250* 1000 * 1.152∗10−5 )

= 144 / (25 * 1.152) = 5 degrees

So, if repairs are done at Td-30 then the effective stress free temperature will get lowered by 5
degrees. If two such overlapping repairs are done, then the effective lowering of Td will be about 10
degrees in the overlapping portion. Such situation makes track vulnerable to buckling. When repairs
are done at higher temperature than Td, the corresponding stress free temperature in the affected
portion of track would become higher in a similar fashion; and track would become vulnerable to
fractures. If the effect of repairs do not overlap with each other, then excessive thermal forces will
not build up in the LWR.
From above analysis it emerges that the affected portion of LWR should be destressed if two or
more repairs are done over track length which makes the affected portions overlapping with each
other. As the stresses are equalised over 250 meters (i.e. 125 meters on either side of location of
repairs) any repairs within 250 metres of each other can be classified as overlapping repairs. Thus
the maximum length of the affected portion of track would be 250 metres plus 2 * 125 metres = 500
meters. This length only requires de-stressing. This conclusion is on the basis of assumption that
repairs are done at a temperature 30 degrees lower than destressing temperature.

4.1 Effect of Creep in LWR at SEJ gaps:

If LWR undergoes creep of rails in central portion, it indicates considerable weakening of fittings.
The creep may take place in the entire LWR or in some particular portion. If creep occurs in the
entire LWR then it may not pose immediate danger of buckling as sufficient compressive stresses are
not building up in the rails. However, there will be excessive movement at SEJs. Normally the gap at
one SEJ will be reduced and GAP at other side SEJ will be increased. If the creep is only in part of
central portion of LWR , then it poses serious danger of buckling as heavy build-up of compressive
stresses will take place at junction of creeping portion and non-creeping portion. This type of
behaviour will, however, not get captured by gap at SEJ.

Therefore, creep movement of LWRs in central portion should be monitored, say at every 500
meters, and any portion of track showing movement of more than 20mm should be thoroughly
investigated for loss of toe load of ERCs. Such affected portion should be destressed along with
rectification of cause of creep.

4.2 Damage to SEJ/buffer rails:

LWR manual also prescribes that destressing should be carried out if SEJ/buffer rails have been
damaged. This is totally unnecessary as damage to SEJ /buffer rails has no bearing on the health of
LWR. This provision should be done away with.

5.0 Conclusion:

The current stipulations in LWR manual regarding destressing on the basis of (i) gaps at SEJs, (ii)
temporary repairs to rail, and (iii) damage to SEJ/buffer rails exceeding three per km should be
modified as below:

(i) If two or more repairs, by way of replacement of rails, have been done in the track
within 125 meters of each other, the affected portion of LWR should be destressed
along with 125 metres of track on either side.
(ii) Gap at SEJ should be monitored during routine inspections. If gap at SEJ goes beyond the
permissible limit then the LWR should be investigated for misbehaviour of Breathing
Lengths and Creep in the entire portion of LWR. If the problem is limited to BLs then the
health of track should be restored in the Breathing Lengths. If there is a problem of
creep in the entire LWR, then immediate corrective action such as TFR should be taken
along with destressing of entire LWR.
(iii) The creep in the central portion should be monitored by erecting reference pegs at
about every 500 metres or preferably by using OHE masts as reference. If creep exceeds
20mm in a portion of LWR then the affected portion of LWR should be destressed. The
cause of creep should be investigated and corrective measures shall be undertaken
along with destressing of the affected portion.