# Compression only constraint between two area element

Hi All,

My name is Flavio and I’m a beginner user of Sofistik. As shuch, I’m still learning the general philosofy and how the software works.
I created a simple model of an aluminium structure. The structure is a folded sheet supported on one edge on an aluminium angle.
My aim is to model a compression only constraint (could be either area constraint or line constraint) between the angle and the folded sheet.
In the image below there is the model I created so far. Everything is working fine, except the link which is now a bilateral link (i.e. link reacts in tension and compression).

How do I define a link which reacts only in compression?

Thanks you very much.

Hello

In this case you should work with spring elements. This has the advantage that you can define a non-linear spring for compression forces. You could have a look at an older post. There is the general feature explained.

For an elastic coupling you could have look at the TEDDY example:

TEDDY > File > Examples > sofimshc > english > area_models > elastic_interface_regions.dat

Furthermore I have created a short example (SOFiSTiK 2020) for you. It shows you a compression only connection between two area elements. For a rigid coupling you should increase the spring stiffness (SOFiMSHC -> SPRI CP… CT …)

Flächendruckkopplung.sofistik (39.5 KB)
Flächendruckkopplung.dwg (109.3 KB)

A line constrain is easier, because you can work with the standard SOFiPLUS(-X) commands.

• Select Structural Lines
• Choose “Material” as a Spring Typ
• Define non-linear material

Keep in mind that both methods require a nonlinear analysis.

Best regards
Frederik Höller

I’ll have a look at the tutorials and let you know.

Thanks.

I tried to understand the possible workflow for non linear springs but I cannot clearly understand how it works.
In my case, I used the line link to model a compression only element with a work law that has no stiffness in tension (you can see the line link from the images attached in my previous post).
Similarily, I modelled point links which are supposed to react only in tension with the appropiate work law.
I then exported the model to SDD and run the linear analysis.
After that I run the superpositioning (even though not specifically required) and then the “Combine Load” task to create the combinations for the non-linear analysis.
After that, I run the “Analysis of Combined Loads” and I receive an error about the “tangential stiffness”
Below the analysis log.

Parameter of System of Equation 3
(Sparse Direct Solver)
Number of unknowns 25248
Total entries 667567
Total entries after fill 3170856
Iteration 3 Residual 238.424 energy 0.303 Step 2-1 f= 0.161
Iteration 4 Residual 7053.952 energy 10.09 Step 3-1 f= 0.274
Iteration 5 Residual 226723216.0 energy10128. Step 4-1 f= 0.465
Iteration 6 Residual ************ energy****** Step 5-1 f= 0.791
Iteration 7 Residual NaN energy****** Step 6-1 f= 1.344
+++++ error no. 202 in program TVERS
Tangential stiffness gets negativ. Iteration stopped. 5001
+++++ error no. 197 in program VERS
Displacements or rotations are too big.
This error message can be switched to a warning with CTRL WARN 197.
+++++ warning no. 195 in program PHI
Node 9 rotates more than one time around its axis in LC 5001
+++++ warning no. 195 in program PHI
Node 10 rotates more than one time around its axis in LC 5001
+++++ warning no. 195 in program PHI
Node 11 rotates more than one time around its axis in LC 5001
+++++ warning no. 195 in program PHI
Node 12 rotates more than one time around its axis in LC 5001
+++++ warning no. 195 in program PHI
Node 13 rotates more than one time around its axis in LC 5001
+++++ warning no. 129 in program PHI
Further warnings of type no 195 are supressed
TIME NEEDED 4, TOTAL 4
DATE 2021-03-18, 16:13:31

``````+++++    2 errors and   12 warnings +++++
``````

+++++ DATA ERRONEOUS, PROGRAM ABORTED +++++

*** Stop : Program ASE : Analysis of Combined Loads
*** Process canceled, error detected!

It seems that there is some instability, but I fail to understand where in my model the instability is.
I would attach the model but as new user I’m not allowed.

Thanks.

I contacted the Sofistik Support and they suggested to run a TH3 analysis. Effectively, the model converges regularily with the TH3 analysis, but there is a convergence problem with the TH2 analysis.
I understand the difference between the two types of analysis, however I have some difficulties in understanding how to adjust my model to make it work with the TH2 analysis.

Any suggestion?

Have a look at my small example. The example shows how to implement a connection between two structural areas that only transfers compression forces.
Why are you working with theory II. and III. order? As far as I understand you, you only want to account a non-linear spring behaviour. So use “SYST PROB NONL” or set the checkbox “First Order”. Do you really need a geometric non-linear behaviour? The differences between the types of analysis can be found within the ASE manual.

SSD / TEDDY > Help > User Manuals > All Manuals… > ase > chapter 3.4 SYST - Global Control Parameters

If you have a call back to a support activity please send your question via the SOFiSTiK Application Manager directly to the support (attach your last ticket ID).

Frederik,

to be honest I’m new to Sofistik and it’s hard for me to read the text input. Although I understand the small code, I’m not sure on how to implement it in my model.
Anyway, I don’t need the geometric non linear behaviour (at this stage, the real structure may be sensible to geometric non-linearity, but I don’t want to analyse this), but as far as I understand this is the only way of doing a non-linear analysis.

You are sayint “set the checkbox First Order” does not this equate to a linear analysis?

Regards.

OK no problem .
There is a difference between a “linear elastic” and “First order” analysis in SOFiSTiK.

• linear elastic (LINE): linear “static” analysis
• First order (NONL): material non-linear springs, cable failure on compression, non-linear bedding,…
• 2nd order (TH2): First order analysis + geometric non-linear behaviour due to large deflections

As I understand you correctly, you only need a non-linear spring material. This means the option “First order” is in your case sufficient.

Fredrick,

thank you very much for clarifying this for me.
I assumed that “Linear” and “First Order” where the same analysis.
I’ll try to run my model with the “First Order” option and see if it works.

Out of curiosity (and also because I may need it in the future) is there any guideline on how to run a buckling (i.e. TH2 or TH3) analysis?

Regards.

1 Like

Hello Flavio

Yes we have a few verification benchmarks about buckling.
You can find the examples and their documentation on our website.

www.sofistik.com > Infocenter - Overview > Documentation - Verification Benchmarks

• design code benchmarks > filter: topic > steel
• mechanical benchmarks > filter: procedure > buckling

Best regards
Frederik