Learning Forum

[ChristophMaier]

Hello everybody,

For an analysis of some measurements we made, I tried to simulate our setup using Ansys Workbench. In general, it is a simple three-point bending setup (see figure below). Since we want to include this simulation in a paper of ours, I tried to stay as close as possible to the real setup. At first I tried to use rigid bodies for both the stamp and the support structure, but I ran into problems with the contact. For this reason, I used only mild steel, which should work since the tested material is very soft compared to steel. On the tested material there are three line bodies, which are particularly interesting for us, and I connected them to the solid via bonded contacts. The stamp is moving down to a (hopefully later) variable negative z coordinate, but right now it's only 1 mm.

My problem is the solution of the simulation (see figure below with a scaling factor of 15) does not behave as I expect. The two supports seem to absorb the moment (but with this type of support the moment should be zero, at least for a simple beam structure). On the far left and far right of the structure, there should be a constant angle, which is not the case.

For all three relevant contacts (between beam and supports and between beam and stamp) I choose frictionless contacts with asymmetric behavior (see figure below). Large deflection is switched on, as is the quasi-static solution.  

 

I have seen some other simulations, also on this website here. Most people used frictionless contacts. Some used a remote displacement, which I would also use, but in this case I wanted to stick as close as possible to the real design. I also didn't use symmetry, but that shouldn't be the solution. Does anyone have any ideas why it doesn't behave like a normal, simply supported beam and where my error lies?

Thank you in advance for any advice.

With kind regards

Chris

 

[Armin_A]

Hi Chris,

Could you describe what is the material of the solid body (plate) where the line bodies are connected to? I suggest using rigid behavior for the tooling as the resulting deformation of the supports appears undesired to you. You can at least increase the Young's modulus of the current tooling to evaluate whether the response is moving into the right direction. I assume you employed linear elastic behavior for all materials.

Also, try using a frictional contact type as it is better in line with reality. This is particularly important as you seem to be planning to compare your simulation results with actual experiments.

[ChristophMaier]

Thank you for your reply!
The solid consists of a linear orthotropic material, which should represent a long-fiber composite. 
Increasing the modulus of elasticity is definitely a good idea. As soon as the result corresponds to the expected value, I will try to use rigid bodies.