Preserving Composite Layer Properties in Pre-Stressed 2-Way FSI

[leviverreck]

Hi everyone,

For my master’s thesis, I’m running a 2-way FSI simulation of a composite airplane wing in Ansys. The wing structure consists of ~25 parts, modeled in Workbench using ACP-Pre to assign composite material layers. The ACP-Pre setup can be linked to a Static or Transient Structural module for structural analysis, while a Fluent module handles the CFD. System Coupling connects both for the 2-way FSI.

In System Coupling, I quickly get the following error from the Fluent module: “Update-Dynamic-Mesh failed. Negative cell volume detected”. From what I gathered online, this is caused by too large mesh deformations during Fluent’s remeshing in between iterations. The usual solutions are either coarsening the mesh (which isn’t really an option for me) or reducing the time step. However, I would need an extremely small time step to have it work, which would cause the simulation to take an extremely long time.

A potential workaround I found is doing a pre-stressed FSI:

1. Run a 1-way FSI (Fluent → Static Structural) to obtain an initial deformation.
2. Use the deformed geometry to start the 2-way FSI analysis.

This approach is described in these forum posts:

• How to include a pre-stressed initial state in 2-way FSI
• How to apply a pre-stressed geometry in FSI

The problem: When importing the deformed geometry from Static Structural into Transient Structural, it seems to lose the composite ply definitions from ACP-Pre. This results in incorrect structural properties, making the solution meaningless.

My question:

Is there a way to export the deformed composite plies from Static Structural into Transient Structural? Or could I use MAPDL to extract stiffness properties from each part and reapply them in Transient Structural?

Or are there any other solutions which allow me to use a larger time step, without getting the “Negative cell volume” error?

 

Any suggestions would be greatly appreciated!

[Ashish Khemka]

Hello,

I am not an expert in this topic but what if you replace the Static Structural Analysis with a Transient Structural one:

Click on the arrow indicated below and then try replacing the system:

Please let me know how it goes.

Regards,

Ashish Khemka