06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

Gilmar <[email protected]>

[FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification
1 mensagem

Łukasz from Enterfea <[email protected]> 6 de fevereiro de 2024 às 10:28

Para: Gilmar <[email protected]>

Hey Gilmar!

Oh, I have a treat for you today! We will do a Nonlinear FEA analysis of a pin
connection and check its capacity for various failure modes. Then we will design
the same connection with EN 1993-1-8, and see what the comparison will show!

This will be so much fun!

Analyzed Pin Connection

We will analyze a typical pin connection. However, since I want to show you
different failure modes, I will have to cheat a bit.

Nonlinear FEA always shows failure associated with the lowest capacity.
This means, that if I would like to show you all possible failure modes, I will have
to modify the model a bit to make the given failure mechanism the decisive one.

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:179015634793… 1/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

The base geometry of our model is given below, although I will have to change it
a bit toward the end:

As you can imagine from the dimensions, we will use a 100mm pin. What is less
obvious is, that the external plates are 30mm thick.

Pin failure in shear

I'm willing to guess that the first failure mode that comes to mind would be shear
failure of the pin. So let's start there.

To make sure, that the pin is the weakest link in the connection, it was assumed
to be made from steel grade 4.8 (this means that the yield strength of the pin is
320MPa). The plates, on the other hand, were made from steel S460 (so the
yield strength is 460MPa).

Firstly, let's do the Nonlinear FEA analysis of the connection:

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:179015634793… 2/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

As illustrated by the strain plot in the model cross-section above, the shear failure
in the pin is the obvious failure mechanism here. But I think it is ever better
demonstrated when you show the pin only, with magnified deformations:

I simply love such images from FEA outcomes :)

But as you already know, while the plots like that look great, it's the equilibrium
path that is the real outcome here:

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:179015634793… 3/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

It's way easier to read the max value from the chart when you have it in Excel, of
course. Here, the plateau starts when the force is 2895kN. Stress stare in our
model (shown in the cross-section) in the last step of analysis looked like this:

There are 2 figures because we have different yield stress values. The scale on
the left ends on 460MPa (showing the yield state in plates). The scale on the right
ends with 320MPa showing how the pin is yielded.

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:179015634793… 4/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

And now the cool part: code comparison!

This is a simple task, since there is a code procedure described in EN 1993-1-8

that allows to calculate the pin shear capacity. Truth be told, it's brutally simple:

And now, the kicker. The above is the capacity of a single shear plane, and our
pin has to fail in 2 shear planes at once to fail. This means that the actual code
capacity is 2 x 1451 = 2902kN which is almost equal to the capacity from
Nonlinear FEA (2895kN). The difference is only 0.2%.

Of course, in the “nice” material failures, this is to be expected, but it's always
cool to see how those things align :)

Following verification won't be as simple, though.

Contact capacity of the plates

Of course, its not always the pin that fails. Plates that are used in the connection
can fail as well. There are 2 mechanisms in which a plate in a pin connection may
fail, and here we will analyze too high contact pressure from the pin.

Luckily, to get this failure in Nonlinear FEA, a basic modification of the model is
needed. All we need to do is to reduce the yield strength of the plates to 235MPa
(so the steel grade would be a typical S235).

With such changes, plastic strains obtained in Nonlinear Analysis look like this:

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:179015634793… 5/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

It's easy to note that now it's the plates that get the punishment. Of course, to
check the capacity, we will need the equilibrium path again:

Again, having this chart in Excel makes it easy to read the capacity. In this case,
equal to 2143kN (measuring the peak on the curve). However, if we would
consider the 5% limit of plastic strains, the capacity would be slightly lower
(2026kN).
https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:179015634793… 6/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

Of course, the thing I like the most in this type of failure is how the model looks
like when you increase the deformation scale!

Code verification part!

This failure type is well described and recognized, so there is a simple equation
in EN 1993-1-8 that allows to calculate how much the plates can take:

Again, we have 2 openings, so the capacity is 1058 x 2 = 2116kN. Assuming we

would follow the 5% plastic strain limit, the capacity from FEA is 2026kN (4.4%
difference on the safe side). I don't know about you, but I would be happy with
such accuracy (assuming, of course, that it's the code outcome that is correct...).

Shear failure of the plates

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:179015634793… 7/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

This one is a bit more tricky. Plates in such connections can realistically fail in 2
ways. The first one is due to pin contact pressure. We already analyzed this
above. The second failure mode would be in shear.

To avoid this, codes give you guidelines on how much steel you need to have for
this failure mode NOT to happen (we want to contact to be the governing case).

In order to make our plates fail in this way, we actually have to change the
geometry of the model... changing the steel grade would still lead to failure due to
contact pressure in the base geometry. The changed geometry is:

If you ever designed any pin connections, you will notice that there is suspiciously
little steel around the opening - that is the point!

The rest of the model stays the same, and the plastic strains from Nonlinear FEA
look like this:

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:179015634793… 8/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

Of course, as in the previous cases, it's the equilibrium path that interests us
most:

In this case, the capacity on the chart is 1462kN, but it will be slightly reduced
again by the 5% allowed plastic strains (to 1442kN). And of course, again, the
coolest view comes from the model with increased deformation:

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:179015634793… 9/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

Looking at this failure, you can notice some differences from the previous one.
But we will get to this in a second. Before we should discuss...

Code comparison

And this becomes awkward. You see, there is no procedure that would allow us
to calculate the capacity of our case. The code simply gives us rules of how much
material we need “around the opening”. Of course, we didn't follow those rules,
which resulted in the above failure mode.

The thing is, that if you would end up in a situation where you simply can't follow
the code (due to geometry limitations or whatever), you simply wouldn't be able to
design this in accordance with EN 1993-1-8!

Even in such simple cases, sometimes you just can't rely on codes to
design things. This is where FEA really shines.

But let's not leave it like that. I usually like to at least try to estimate the capacity
to give me a sense if the FEA outcomes make sense.

The failure mode we are investigating could have 2 different types, depending on
geometry:

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:17901563479… 10/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

Since in our case, there is the same amount of material on the top and to the
sides, shear will be weaker. So, this is what we will investigate.

For the plate to fail, 2 shear planes have to form. The biggest thing we do not
know is how long those will be. An optimistic approach would be to assume that
the material will shear on the width of the pin (since the pin has to “fly through”
the plate). This is marked above on the left, and the length of the shear plane is
the marked 109mm.

On the other hand, the pessimistic approach would be, that it's only safe to use
as much material as we have on the very top of the plate (so 70mm as seen on
the geometry figure). Let's see where those assumptions would lead us:

Of course, in both cases, for a single plate to fail, 2 shear planes would have to
form, and both plates have to fail for the model to fail. This means that total of 4
shear planes will happen at failure.

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:17901563479… 11/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

The capacity of the model in such a case will be somewhere between 4x 284.9 =
1140kN and 4x 443.7 = 1775kN. If you asked me, a 1140 - 1775kN range is a big
range to choose from, but the FEA outcome we received (1442kN) nicely fits into
that range.

Understanding the difference in the failure modes

What I really like about FEA is, that it makes it easy to understand various failure
modes... if you know what you should look at.

I'm first to admit, that if you only look at stress distribution at failure of both plate
failure modes, it might be difficult to see the difference:

Of course, knowing everything I described above, you can spot, that the model
on the left has more steel, so it's the one failing due to pin contact. But honestly,
I'm not super sure if I would spot, based on the figure above, that those are 2
different failure modes!

Heck, even looking ad the plates with increased deformation scale could not help
me. While there are a few things that could help here, I would have to really pay
attention to notice those (like the necking in the plates on the right):

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:17901563479… 12/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

However, this doesn't mean that you have to know the answer beforehand, to be
able to interpret outcomes. I mean, if that would be the case, doing FEA would be
pointless!

The trick is in realizing, that when material yields, the stress starts to be constant
(in plastic plateau of structural steel). But this doesn't mean that it's all the same
whether you have 0.1% or 2% of plastic strains!

And here, as the model fails, it's the plastic strains we should look at:

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:17901563479… 13/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

I think that now it's clear, that on the left plastic strains focus around the contact
area, where the failure happens. While in the figure on the right, you can see
plastic strains starting to form a shear plane - indicating a different failure (just as
they did in the pin shear, we started with!).

Cool, wouldn't you agree?

This also means, that FEA can save you occasionally. If you did a proper model
hoping to find the failure mechanism that you think will govern the design, you
may find a different one that is even worse! Even if you were unaware of it at the
start.

It's of course always best to be able to predict how your models will fail, and you
really should be able to imagine failure of the component before the analysis. But
we are all humans in the end, and it makes me feel a bit more relaxed to know,
that if I blunder and fail to notice something... FEA will find it for me, as long as I
do the model well!

A few practical conclusions

Preparing such case studies is a massive effort, as you can imagine. So let's use
it well, and form a few practical conclusions that you can remember from this:

When you design a simple pin connection, following the code seems
to be the easiest way to design it. As you could see, the differences
between Nonlinear FEA, and code procedures were minimal in this case.

Code procedures often allow you to only design the “simple cases”. This
means that quite frequently you will find yourself outside what is
allowed by the code. In situations where following the code is impossible,
all that remains are estimates... and FEA. And as you could see here, FEA
is super accurate, and allows you to reliably assess capacity of more
complex models!

If you build a correct FEA model of whatever you are designing, FEA will
find the governing failure mode... even if you were unaware that such
failure could happen! This is a considerable benefit, and can save you in
difficult cases. Please remember that you need to be aware of a given failure
mode to check it by the code... so this is a huge FEA advantage.

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:17901563479… 14/15
06/02/2024, 13:13 Gmail - [FEA] Case Study: Pin Connection in Nonlinear FEA and Code Verification

And in the end... FEA is just spectacular! I don't know about you, but there
is something that simply speaks to me in those outcomes, and makes me
feel “in control”. Being able to analyze super complex problems and being
certain that the outcomes you have are simply correct... is a wonderful thing!

One last thing

I really hope that you like this case-study. And, it's likely, that you know someone
who would benefit from reading my newsletter as well. Please forward them this
email, so we can all share knowledge with each other!

And if you received this email from a friend, you can sign up to my newsletter
here: https://enterfea.com/fea-newsletter/ This way, you will receive weekly
emails like this one directly to your email!

Have a remarkable day, and see you next week!

Ł

You are receiving this email because your email was added to my newsletter of the enterfea community. If you don't want
to receive any more emails from me, you can unsubscribe from all mailings (even if you are already a customer) by
clicking here. In such a case, I will really appreciate your feedback.

The administrator of your personal data is Łukasz Skotny - Enterfea, ul.Skrzydlata 1/7, 54-129 Wrocław, POLAND. You
have the right to cancel your consent, request to see, change, delete or limit the use of your personal data. More
information on how your personal data is processed can be found here: https://enterfea.com/gdpr/

https://mail.google.com/mail/u/0/?ik=3be59fa6b0&view=pt&search=all&permthid=thread-f:1790156347931415677&simpl=msg-f:17901563479… 15/15