I’m working on a configuration similar to a valve where my initial geometry has its two branches interpenetrating each other which is not the “real” configuration.
The real configuration is the moment where the valve is in closed position with both branches in contact at their tips. I can’t run a simulation with contact activated if the master and slave surfaces are interpenetrated so my idea is to run a first simulation where I open the valve by applying pressure.
Then I save the mesh in open position with the stress associated to create a new simulation where I turn off the pressure and activate the contact to let the branches go down in dynamic
and create my close configuration.
My idea was to use the RESTART function to activate contact but apparently it’s not possible to change the setting during a restart.
Also, I tried to use in cgx the command send all abq to save the mesh in open position, but then I loose all the entities I created so I can’t define anymore my contact…
I appreciate any hints.
You could start your simulation with the two parts separated enough to have the tips in free state condition without touching, then do an initial step were you move your parts to the initial real position, so the tips will get in contact and compute the stress at the base of the tips. Then you could do a second step with the actual loads on the parts. The trick is making a multistep simulation to keep the stress/deformation, and start with the part not in real position, but one that leave you the tips in real free state.
@irphe, I would have done it the same way as @SergioP describes here.
Now I’m in the case where I can’t move and replace the two objects because they are linked together.
I want to use the same idea of multi-step simulation, in the first one I apply a pressure to open the valve without interpenetration, then in the second run I activate the contact and apply a pressure to get the close configuration.
In order to do this I moved the *INCLUDE, INPUT=add_contact.inp after the first step. Unfortunately Calculix doesn’t wait the end of the first run to activate them and send me back the error on negative Jacobian due to iinterpenetration…
Do you now how to avoid this problem?
I appreciate any hints.
Can you uppload all the input files? In the first pictures looks lije the two parts are two separated entities.
Hello, indeed on the first picture the two parts are two separated entities so I can apply your method, but in reality, I’m not working on this case, I’m trying to solve a more complex geometry problem where I can’t move all the parts because they are all clamped to a structure. (prosthesis of aortic valve)
As you understood at the initial state the two tips are interpenetrated but the goal is to get the real configuration where the two tips are in contacts, creating like a waterproof valve.
In order to do that, the mulltistep seems the best way, in a first step I apply a pressure to open it and in a second step i activate the contact and cancel the load.
But apparently we can’t add the contact only in one step…
I found the card *CHANGE SURFACE BEHAVIOR which could help but with no result so far.
I link here the inp file with just one step and the contact deactivated.
Thank for your help and I appreciate any hints.
I’m sorry, without the complete model I can’t give you more option to solve your problem. Regards