If the stress applied to the material is not
relieved by its failure then the defect is caused
to increase in size at a fast speed which rapidly
increases to the speed of compressive waves in
the steel. This fast ductile failure continues
until the supply of energy maintaining the
applied stress becomes exhausted at which point
'crack arrest' is said to occur. The properties
of the steel at places where permanent
deformation or tearing are happening change
continuously, its hardness increasing and its
remaining ductility decreasing.
back to index
The phenomenon in which ductile steel pipes,
pressurized by gas or vapour pressure, have
ruptured by longitudinal through-wall crack
propagation of fast ductile failure is far from
unknown.
Ref.1
The mechanism of failure is that
the speed of propagation of the crack along the
pipe is greater than the velocity of sound in the
fluid within the pipe, so that the stress applied
to the pipe at the instantaneous location of the
crack-tip is not relieved by escape of the
pressure. Such failures have been arrested by
changes in the structure of the pipe, e.g. at
joints and flanges, and by change in the
properties of the 'over-burden' within which the
pipes had been buried. A critical, unstable crack
at the belt-line of the Sizewell 'B' reactor
pressure vessel, will commence to grow as a shear
crack at 45 degrees to the surface, at a linear
speed of 3.24 kilometres per second, and after
the crack tip has travelled a short distance, the
mode will change to cleavage, perpendicular to
the vessel surface, and with a linear speed of
5.89 kilometres per second, which is the velocity
of compressive waves in A533B steel. The velocity
of sound in the reactor coolant at the inlet
temperature of 225 is 1.2 kilometres per second.
Once the crack has become unstable, and started
to propagate in either ductile mode, then it will
not stop running until the vessel has been
effectively dismantled (Ref. 2). The time for the
crack tip to traverse the circumference of the
vessel at the belt-line is about 1.5
milliseconds. The mechanical energy absorbed by
the creation of crack surfaces as the crack
progressed once round the vessel is only about
one percent of the energy available in the hot
water. Once the structural strength has been
removed by fast ductile crack propagation, the
acceleration of the vessel fragments due the
continuing exertion of the internal pressure is
about one hundred times the acceleration due to
gravity. The lateral gap between the pressure
vessel and the concrete housing is only a few
inches. The direction vertically upwards is
virtually unrestrained. The mechanical energy
equivalent stored in the hot water in the reactor
pressure vessel, and which would be released
before the pressure had fallen from 2000 to 100
psi, is about the same as four tonnes of TNT. If
a one third part of the reactor vessel, weighing
150 tonnes, absorbed one third part of this
pressure energy, then it would have sufficient
upwards momentum to rise freely more then 300
metres above the point of projection. Loss of
energy penetrating the containment building would
not be significant. It is quite possible that
such an event would be accompanied by a reactor
power surge generating more than the usual heat
output, as any control rods remaining attached to
the upper-vessel moved upwards with it. The
Sizewell 'B' nuclear power station is provided
with what the Nuclear Installations Inspectorate
calls 'defence in depth', and 'multiple
independent barriers' to prevent the escape of
fission products from the reactor fuel into the
environment. All of this is swept away and made
absolutely ineffective by this event. The
consequences would be Chernobyl transformed to
Western Europe.
back to index?