20 / 352
16
cooling, which is much slower than heating, the process is reversed, but the stresses are of
a lower intensity due to the slower temperature change.
Furthermore, one should also consider the influence of residual stresses which are
caused by braking in a similar manner as the thermal shock stresses due to uneven heating
and cooling, only that they have the quasi-static and not the cyclic character. The residual
stresses developed in this manner are tensile, also relatively high, in the tread surface
layer and compressive, relatively low, in the rest of the wheel, the fatigue crack zone
included. Combined action of all mentioned stresses (mechanical, thermal and residual) is
obviously such that there is a resulting cyclic-stress which can cause the initial fatigue
crack, located at the highest stress concentration/chuck mark, to develop to its critical
size. It is also obvious that there is the resulting quasi-static tensile stress at braking,
which can cause the crack of a critical size to propagate in unstable manner till brittle
fracture of the wheel.
For the next analysis several facts and data had to be considered.
Impact toughness testing indicated an extremely low resistance to crack growth, only
about 1 J, accompanied by a low resistance to crack initiation (about 15 J). However,
impact loads are not typical for railroad wheels, but they can be involved.
Fracture toughness, obtained using 30 mm thick CT specimens, is
K
I
= 104.9 MPa
√
m,
but the thickness requirement for valid
K
Ic
was not satisfied. Since thickness requirement
would be satisfied for fracture toughness value 64.2 MPa
√
m, it is accepted that the real
value of plane strain fracture toughness is in the range 64.2
<
K
Ic
<
104.9 MPa
√
m /4/
.
Fatigue tests indicated very small resistance to crack growth corresponding to only
20,000 cycles for crack propagation before fracture, where the stress intensity factor
range (
Δ
K
≈ 20 MPa
√
m) is much lower than its critical value (
K
Ic
= 64.2 MPa
√
m) and the
corresponding stress range (
Δ
σ
= 87 MPa) is much lower than the yield stress, being for
this steel class
R
eh
= 500 MPa.
Using these data it is possible to evaluate the tensile quasi-static stress for the brittle
fracture of the wheel having a fatigue crack of approximately semi-elliptic form,
modelled in Fig. 13, of crack length
c
= 14 mm and crack depth
a
= 35 mm.
It was calculated that the value of remote tensile stress is
σ
= 282.5 MPa, sufficient to
produce brittle fracture. The tensile stress of this magnitude may appear during wheel
braking by the brake system used in the considered case.
3.5. Assessment of the possibility of crack detection before fracture
According to fatigue tests, the number of cycles during fatigue crack propagation till
fracture is about 20,000. The corresponding range of stress-intensity factor applied was
Δ
K
≈ (18 to 20) MPa
√
m, which corresponds to the stress amplitude
σ
= 79.2 – 88 MPa
that can be achieved only during braking process, since in normal service stress in this
wheel is much lower /4/. For that, the applicable cycle number for crack growth
monitoring is determined by braking application. However, since the given number of
cycles is very small, the likelihood of crack detection before fracture is negligible,
because frequent wheel examination by non-destructive methods should be required. It is
reasonable to conclude that material having such a high rate of crack growth, used in
condition when brake shoe is acting upon wheel tread surface, does not allow crack
monitoring before fracture in a satisfactory manner.