Endplate Steel Connection - Experimental X Ansys Analysis

  • Last Post 27 November 2018
Fabricio.Urquhart posted this 23 November 2018


I am trying to do a benchmark, analysing to add in my master thesis. But I am in front of some problems. The epxerimental analysis shows higher stiffness than the FEA Analysis.

I think that I am doing something wrong, about the value of stiffness factor. I think that is necessary to calibrate this coefficient.

What do you think about? See the graph below.

Attached Files

Order By: Standard | Newest | Votes
peteroznewman posted this 23 November 2018

Hello Fabrico,

Please explain what you mean by stiffness factor.  Is that in the contact definition?


Please insert an image of the model showing loads and boundary conditions.

Was the material used in the experiment tested to obtain stress-strain data?

What was the measured thickness of the parts in the experiment?

Please show an overlay of the experimental material stress-strain curve and the FEA material stress-strain curves.

I assume the experimental data actually measured force and deflection. If you plot what was measured and overlay the same measurement from the FEA, that is one less source of error.


Fabricio.Urquhart posted this 27 November 2018

Hello Peter.

Yes the stiffness factor in the contact definition. "LIMITE P/RIGIDA" is the graph plotte by Eurocode-3 (2010), I showed it, only to  have the idea, if the connection is rigid or flexible.

The load applied and the boundary condition you can see in the pictures below:


The experiment/PHD thesis and the previous work (master thesis) are attached. The material stress-strain curve you can see in page 187 of PHD thesis. The comparation you can see in the master thesis page 172.

Attached Files

peteroznewman posted this 27 November 2018

I recommend you perform sensitivity studies on all the factors that you believe have an effect on the curves.

For example, solve a model with a 20% change in:

  • Normal Stiffness Factor in the Contact definitions
  • Stress values in the Mulitlinear material table

I anticipate the curve will move very little for the 20% change in Normal Stiffness Factor and it will move much more for the 20% change in the material stress.

Fabricio.Urquhart posted this 27 November 2018

Thank you Peter, I will do it.

I tried to use the values below, recommended ofr Maggi,2004. But with these values, the model does not converge.