Fluids

Fluids

System Coupling (SC) – Transient Structural (TS) – Fluent (FL) – Overset Meshing

    • max_n
      Subscriber

      I am currently running a 2-way FSI simulation with pressure inlet and outlet boundary conditions. These flaps come into contact which is taken care of with contact detection in FL and contact offset in TS. The problem described below happen with contact detection off too.

      I am currently using an overset mesh in FL. Before setting up System Coupling, I made sure the mesh worked in standalone FL. I was running that simulation with a compressible formulation for the fluid, which is essential for the work. The overset mesh was not dynamically moved in the standalone FL test setups.

      When testing was finished for TS and FL subsystems. The SC assembly was created. I realized that after the first mesh motion in the coupling iterations FL would diverge to floating point exception. To be clear, the simulation did not go to another time-step, at the second coupling iteration in the first time-step it failed. I would assume that the incremental displacement would be minimal at this point.

      When the overset interface and mesh associated with it (region of finer resolution) is set to stationary. The simulation would run a few time-steps until the SC interface wall would cause negative cell volume because it would intersect the stationary mesh. This leads me to believe that the dynamic meshing movement of the overset meshes causes this divergence.


      Oddly enough, using an incompressible fluid worked and the simulation ran for the desired time, but a compressible formulation is needed for the work.

      Since SC requires 3D geometries, but a 2D system would have been adequate, FL and TS are thin and 1 element thick. With symmetric BC on the "z" plane. There are no problems in TS at the moment and since SC runs with incompressible formulation, I think the problem lies in FL and not in TS or SC.

      Things that have been tried to fix the simulation

      1) Decrease time-step

      2) Ramp and Decrease pressure BC (1/1000th of what the original value was)

      3) Decrease relaxation factors.

      Thank for any help, it is appreciated.

    • Stephen Orlando
      Ansys Employee

      Can you provide some more pictures showing the whole mesh and label the boundaries, ie inlet, outlet, FSI interface? I'm surprised the incompressible case worked but the compressible case didn't.
      FSI simulations with very soft materials or membranes are prone to numerical instabilities. In 2020R1 we have introduced a stabilization method in System Coupling called the Quasi-Newton Stabilization Algorithm. Note that this has to be used with the new System Coupling GUI or Command Line Interface that is run outside of Workbench. More information here: https://ansyshelp.ansys.com/account/secured?returnurl=/Views/Secured/corp/v211/en/sysc_ug/sysc_gen_scservice_dt_supplemental_iqnils.html
      System Coupling User's Guide \\ System Coupling Data Transfers \\ Supplemental Processing Algorithms \\ Quasi-Newton Stabilization Algorithm
      You can also look at the following for a similar tutorial but run with the System Coupling GUI outside of Workbench. The new System Coupling GUI (run outside of Workbench) is available by searching for "System Coupling 2019R3 (or newer)" in the Windows Start menu. https://ansyshelp.ansys.com/account/secured?returnurl=/Views/Secured/corp/v211/en/sysc_tut/sysc_tut_reedvalve_fluent.html
      Steve


    • max_n
      Subscriber

      The image above is the whole fluid domain. I am not showing the actual mesh because it would be harder to see. The yellow lines outline the FSI interface (there is a FSI interface on the gap above it also) and the BCs are on the left-most and right-most walls. There is a no slip condition on the bottom and top walls. I am using an overset model now so the top is not exactly how the top model looks, but with the overlapping meshes without hole cutting it would be hard to show visualize the boundary after the hole cutting. I am not sure how the Quasi-Newton method would help. With the compressible flow, the system crashes during the second coupling iteration.

Viewing 2 reply threads
  • You must be logged in to reply to this topic.