Learning Forum

[donToto]

Hello,

i am currently working on a script which allows me to model a construction sequence of a modular assembly in 2D (PLANE182 elements). I am working with PyMAPDL. Basically, some concrete parts are connected by slip-resistant shear joints.

The tricky part about this is that each part is generated at a discrete time-step with varying geometry and material properties. This information is not known prior of a specific time step. The basic assembly of geometry already works out. But now my questions: How can i simulate stresses, strains and deformations properly at each time step? How can i use the results from a previous time step in the current one? How can i just add elements to the mesh without changing the already existinmg mesh?

Correct me if i am wrong but in my opinion Birth and death technique should not work out since i don't have a final model that is known in advance. I know of the command INISTATE but here I am lacking on some theoretical knowledge how to superimpose stresses and strains correctly.

I would be very grateful for any advice or literature. 

Thank you in advance!

[dlooman]

You are correct that the "Birth and Death" feature is the only way to introduce new elements during solution and preserve the solution history.  Is it possible that you could create a mesh that would envelope the potential domain and then select elements from that domain as needed.  It seems odd though that you don't know the future geometry and material properties in advance.   Why is that?

[donToto]

 

Hi dlooman,

thanks for your reply and suggestion 🙂 

All parts vary randomly and slightly around some target dimensions and are created before each construction stage. Furthermore, all parts that are already placed deform due to dead load, creep and shrinkage. I want to feedback the resulting coordinates of the current assembly to python where the placement of a new part is optimzed and then, again, input to Ansys. So it is kind of a sequential placement strategy considering transient deformations and fuzzy geometries.

Your suggestion using Birth and Death with an enveloping domain sounds interesting. But would I be able to update material constants like youngs modulus at each stage?

 

[dlooman]

It is possible to respecify young's modulus during solution, but you have to keep in mind that the new young's modulus will be applied to the total strain (to compute stress), not just the increment of strain at the new value.