Page 137 - Introduction to Naval Architecture

STRENGTH 123

STRESSES

Whilst stress can be used as the yardstick by which to judge some aspects
of failure it is not adequate for all. However it is appropriate to consider
the stresses in a hull and how realistic calculations of them might be.
The first problem encountered is that due to the manufacturing
process the ship has built-in stresses. In particular the rolling of the basic
structural elements and their subsequent welding into the hull will
induce strain and stress. The welding process can also introduce
imperfections which act as discontinuities and cause stress concentra-
tions, hence the importance of radiographic examination of welds to
identify significant defects for remedial action. The resulting stresses
can be high enough to cause local yielding of the material causing a
redistribution of load locally. There will remain certain strains which
are an unknown quantity but which will add to the probability of failure
under extra applied loads, particularly in fatigue. Modern welding
methods and greater accuracy of build geometry can reduce the levels
of built-in strain but they do remain.
The next problem is the sheer complexity of the loading patterns
and of the ship structure. Whilst modern research and computer
methods provide an ability to deal with more and more complexity,
some simplification of the load and structure is still needed. A simple
example will illustrate this. Finite element analysis, which is discussed in
outline later, is a very powerful tool but the finer the mesh used in way
of a discontinuity, say the tip of a crack, the higher the stress obtained
by calculation. In the limit it becomes infinite. Clearly some yielding
will take place but the naval architect is left with the task of deciding
what is acceptable. This can be determined by comparing theory with
model or full scale experiments.
Traditionally the naval architect has treated the probler,! of overall
hull strength as an equivalent static one, making fairly gross simplifica-
tions and then relying upon a comparison with the results of
corresponding calculations for previously successful ships. This had the
merit that although the stresses derived were nominal, and might bear
no relation to the actual stresses, the new ship was likely to be
satisfactory in service provided it did not differ significantly from the
ships with which it was compared. The big drawback of the method was
that it was a 'play safe' one. It could not tell the designer whether the
new ship was grossly overdesigned or close to the limit of what was
acceptable. The growing importance of ensuring structural weight is
kept to a minimum has driven the naval architect to adopt more
realistic design methods as they have become available. Even these,
however, must be used with some caution because they cannot yet take
account of every factor affecting the problem.