Mechanical Technology February 2016

16 / 40

14 Mechanical Technology — February 2016
⎪ Proactive maintenance, lubrication and contamination management ⎪
column is based, is trickier. Probably
the best known method is by using the
relationships developed by Cahoon and
further studied by Auerkari [7], who
derived a relationship between hardness
and ultimate tensile strength, though
this relationship relies on knowledge of
n, the work-hardening exponent for the
material, which is not readily determin-
able non-destructively.
Using an algorithm derived from
published hardness tensile relationships
[8], the tensile strength of the column
material was obtained.
From ASTM A515, however, it is
known that the yield strength of Grade
60 is approximately 50% of its UTS,
and this value has been used. A hard-
ness of 93.4 HVEQ corresponds to a
tensile strength of 330 MPa (47.85 ksi),
and thus the yield strength observed,
115 MPa (23.93 ksi), is unacceptably
below the minimum requirement of
≥221 MPa specified for Grade 60 ma-
terial, even allowing for probable errors
in hardness testing and the derivation
of the value.
Discussion
That the column was subjected to the
effects of a fire is without doubt. Several
adjacent items of plant were sufficiently
damaged to be discarded on the basis of
visual examination alone – and a great
deal of the piping and support structures
were destroyed. The effects of the fire on
this column, however, present a number of
anomalies in terms of the damage observed
and properties determined.
The effects of exposure of the column
to a fire under the circumstances that are
known to have occurred in this case were
expected to be more severe at the base of
the column, which was closest to the fire
epicentre, and were expected to decrease
with height and therefore distance from the
conflagration itself. In addition, the thermal
lagging on the vessel exterior was expected
to offer a large degree of protection from
external heating.
Heating of the lower part of the column
as the result of exposure to fire was expected
to cause a partial collapse under its own
weight and the weight of the product con-
tained within the vessel. This was expected
to restrict buckling to the part of the column
directly affected by the fire, in this case the
lower portion. While some evidence of this
was observed, the deformation was minor
and the overall deformation of the column
was in the form of a fairly uniform curve,
with two areas of ovality, approximately
equidistant between the mid-point of the
column and the ends. This cannot be ex-
plained in terms of localised fire damage
to an erected self-supporting column. It
can, however, be attributed to the effects of
thermal stress-relief whilst in a horizontal
position, particularly if the stress-relief
temperature was high.
The reduction in mechanical properties,
as revealed by hardness testing, indicates
a uniform heating over the whole length
of the column, to a temperature at which
significant degradation of the pearlite phase
would be expected to occur – in the region
of 650-700 °C.
The uniformity of loss of mechanical
properties also gives cause for concern. This
loss is the result of microstructural changes,
principally the degradation of pearlite as the
result of exposure to elevated temperatures.
As with the deformation observation, such
damage would be expected to become more
pronounced as the epicentre of the fire was
approached, and the thermal lagging on the
vessel would be expected to offer a high
degree of protection, giving further tem-
perature differences and therefore greater
differences in the degradation properties.
As with the deformation observed, however,
the uniformity of the drop in mechanical
properties can readily be explained if it had