Introduction to FEA CD $59.99
|
Q: I have applied a force to my FE model of a sailing boat deck to
represent a cleat pull out load. The stresses I see are far bigger than the
deck material can take. From my experience with manufacturing these
boats it cannot be that bad?
The model is probably set up with a point force applied to the shell mesh of the deck. The
reason that you are getting very high local stresses is that the load is trying to act in the FE
model as a force at a finite point. The finite point is of indeterminate size, but in FE terms it
is very, very small. It is rather like a lady passenger walking on your deck with a stiletto heel.
You probably wouldn’t allow that!
This results in very high stresses as the elements immediately adjacent to the loaded point
load try to transmit the stresses out to their neighbours. The stress gradient – that is the rate
at which the stress drops off as you move away from the point, is very high. This means that
the results will be very inaccurate.
As an interesting exercise, if you refine the mesh at the loaded point the stresses will keep
getting higher and higher – you are in fact chasing a mathematical singularity. The area of
load application gets smaller and smaller and the pressure applied to maintain the net
force goes up and up. Or in other words, that lady passenger just keeps getting sharper
and sharper heels. Her weight stays the same, but the local stresses she causes are going
to just get bigger and bigger.
So how do you deal with this? The modeling I mean – not the passenger!
In practice the cleat will be attached to the deck at
several bolting points and be well tightened down. That
creates a much bigger footprint (excuse the pun) over
which the load will act.
Here is an example of a cleat, it happens to have two
pads that attach to the deck. Notice the footprints
shown below the cleat image.
If we can manage to mesh the deck structure so that
we map the footprint of the cleats reasonably well, then
we can apply a distributed load over the mesh.
Don't forget we said the cleats would be well tightened down, it would also help if they
were bonded to the deck.
If the cleats are not well tightened then we would have to consider the action of the bolts
as they connect to the deck, but even then there would be probably be washers or load
spreaders under the decking.
There are several other methods to distribute load, such as using rigid spider type
elements. This link will take you to descriptions of this type of element.
FETraining