Learning Forum

[tortju]

Hi,

 

I am running a simulation of a steel pipe casted in concrete. In order to reduce the computional cost of the simulation, I have taken advantage of symmetry in geometry, loading, material and the expected response.

 

The symmetrical section of the pipe is shown here:

 

What is the best way to simulate this situation?

 

I have tried both to include the concrete in the model and to use fixed supports directly on the ribs on the pipe (leaving the concrete out of the simulation). Either way, the von-Mises stresses in my model does not seem to converge. I have tried to round all sharp corners in order to fix this, but the problem still persists. I have also tried other support conditions in order to avoid stress singularities from arising.

 

I have attached a .wbpz-file in case this might help in identifying the problem: https://drive.google.com/drive/folders/1gQrU1Cj09HNPmTgwvYGFs-C1EuOFRRLc?usp=share_link

 

Thanks!

[Ashish Khemka]

Hi,

Please see if you can share a snapshot of the stress results and mesh. 

 

Thanks,

Ashish Khemka

[tortju]

The stress results (von-Mises):

The mesh:

As can be seen, the mesh is very fine (a body sizing control of 10 mm is used). However, the solution do not seem to converge:

[Ashish Khemka]

Hi,

At which locations ar eyou expecting converged results. If you can clarify the regions of interest then the mesh can be refined at those locations. For example if fillet is region of interest then we can refine the mesh there.

Regards,

Ashish Khemka

[tortju]

 

I am most interested in the stresses near and at the ribs. This also includes the fillets. In order to obtain the convergence plot above, the mesh near high values of equivalent stress are automatically refined:

The mesh is becoming very fine and I have replaced sharp corners with fillets to avoid any stress singularities. Still, the solution do not seem to converge.

 

Is there anything that can be done to aid in convergence here?