Nonlinear supports in FEA: Gap Elements (NX Nastran) [Q&A #2]

Recently I got a great question from Ardalan (thanks for writing mate!) about defining nonlinear supports in FEA. I think he does a better job of defining the problem:

First how to constraint an entity with in tolerance rather than completely fixing it? i.e. how to define a certain range of motion for an entity beyond which no movement would be possible?

Second how to fix an entity only in one direction of an axis but not the opposite one, e.g. fixing a node in +Y yet allow it to move in -Y?

Both of those questions can be answered with one element type from NX Nastran library: Gap Element

*If you are an impatient person, you can scroll down – there is a video there in which I explain how to define Gap Elements at the very bottom of this post 🙂

Why may this be useful?

There are many instances where gap elements will excel. My personal favorite is the support ineffective in one direction. Some time ago I discussed how to avoid contact in the analysis. Supports ineffective in one direction would fall into this category I would say. I can understand that there is an argument that this solution is not “far away” from contact. It uses nonlinear approach, needs to be iterative etc. I won’t argue here, especially since I hate discussions about classifications as such. I will put it this way:

Gap elements have many great uses, especially when you want to quickly and relatively easy calculate supports that are ineffective in one direction.

They are also useful if you need to have a sliding connection between 2 elements. As the name suggests you can define the “gap” that needs to be closed before forces are transmitted!

Why can support be ineffective in one direction?

I will stop here for a second, to explain why supports can be ineffective, and what does it really mean. Usually, this is explained in terms of tension/compression problems.

Take a look at this (beautiful) flower:

Nonlinear support: innefective in tension (nonlinear FEA)

It simply stands on the floor, and the fact it is not going “deeper” into the floor means this is a support. In fact, this is a very strong support. If you would apply a compressive load on the flower (who would love to squish such a beautiful thing ?!?) it would resist anyway. It is actually astonishing how much load you can apply on such support!

But if you would apply tension to the flower (gently and by the pot!) you would simply lift the flower up! In such case, there is no support there at all!

Nonlinear support: innefective in tension (nonlinear FEA)

This is basically what an ineffective in one direction support is. I would say that support ineffective in tension like this one is the most popular type, but there are plenty more. You know I can’t resist making bad sketches so here we go!

Different types of nonlinear supports ineffective in one direction

I have already shown you the support ineffective in tension (aka “flower support”). But of course, there are more.

You can easily imagine support ineffective in compression. Just imagine a lamp that hangs from a ceiling on a cord:

Nonlinear support: innefective in compression (nonlinear FEA)

If you apply enough vertical force upward you will “lift” the lamp. This means that the cord is a support that only works in tension!

Nonlinear support: innefective in compression (nonlinear FEA)

But of course, this is not all! Similar things can be made for bending and this can sometimes catch you off-guard.

Imagine a connection that looks like this:

Nonlinear support: innefective in certain bending direction (nonlinear FEA)

We already discussed how such connections work here. In short, the moment is divided into a pair of forces. Tension goes into the wall thanks to the bolts, while compression is simply carried by contact between the end plate and the wall at the bottom.

The fun starts when we switch the bending direction. In such case, the connection is hopeless, as there are no bolts to carry tension at the bottom. Such support is ineffective in “upward” bending direction!

Nonlinear support: innefective in certain bending direction (nonlinear FEA)

There are plenty more possibilities, but those are the main ones.

Supports with gaps

The second part of the question was about gaps in supports. Simply put sometimes support can start working only after a certain deformation takes place, as I described here.

The simplest example of such support would be a cantilever with additional support. If fact I will use this example in Femap to show you how things are done!

Nonlinear support: Support with a gap

In such example, you can see that firstly the cantilever will deform (maximal deformation at the end). But later, when the gap will be closed, this will start to work as a simply supported beam. In such case, deflection will increase in the center of the beam, rather than on the rightmost edge (it will already be supported at that stage!).

Note also, that I draw the “additional support” only at the bottom of the cantilever – if the force will be upward there is no support at the end at all!

Gap elements for the rescue!

The above example is a perfect case for gap elements in NX Nastran. Describing how to define it would take some time, so instead below you can see a video I made with all the instructions needed to do this task 🙂

Free FEA course!

I have a free nonlinear FEA course just for you! Subscribe below to get it!

If you enjoyed the post you can share it with friends – that would be a great help! If you have a spare 15 seconds write a comment with your thoughts on the matter or any questions you might have. I have a good history of replying to each and every comment 🙂

Free nonlinear FEA course!


  1. Mike Beaumont September 12, 2017 at 4:18 pm - Reply

    Always happy to see a new post, thanks Łukasz 🙂

    • Łukasz Skotny September 12, 2017 at 5:53 pm - Reply

      Hey, Mike!

      I’m very glad that you like it 🙂

      All the best

  2. Silviu September 12, 2017 at 7:40 pm - Reply

    Hi, Lukasz!

    I just want to say that I think what you are doing here is great. I’m always extremely captivated when reading your posts.
    Keep up the good work! Please! 🙂


    • Łukasz Skotny September 12, 2017 at 8:05 pm - Reply

      Hey, Silviu!

      Thank you so much for the encouragement – this means a lot!
      I’m very glad that you like my blog 🙂

      All the best

  3. Goyo September 14, 2017 at 2:17 am - Reply

    Very good!
    Bravo Master FEA

    • Łukasz Skotny October 24, 2017 at 9:31 pm - Reply

      Hey Goyo!

      I’m so happy that you like it 🙂

      All the best!

  4. Asmita March 23, 2018 at 5:52 am - Reply

    Your way of explaining the topic is excellent. Thanks for the blog.

    • Łukasz Skotny March 23, 2018 at 7:05 am - Reply

      Hey Asmita!

      I’m really glad that you like it 🙂

      All the best

  5. Josh May 22, 2018 at 4:38 pm - Reply

    Nice post! I am trying to create a small program in Microsoft Excel (for optimization purposes), in which I need to use something like a gap element. Do you know how I could do this? I’m struggling to understand the degrees of freedom and how to build a gap element into my stiffness matrix.

    Thanks again,

    • Łukasz Skotny May 22, 2018 at 5:17 pm - Reply

      Hey Josh!

      I’m really happy that you like it! Sadly I’m not sure how to help you. In truth I have never tried to write my own FEA code – so I have no experience here. I really do hope you will find some help somewhere! Good luck!

  6. Boris Jakimov July 19, 2018 at 4:46 pm - Reply

    Great stuff as always Lukasz!

    • Łukasz Skotny July 19, 2018 at 5:07 pm - Reply

      Thanks Boris!

      I’m super happy that you like it!

Leave A Comment

Do NOT follow this link or you will be banned from the site!