Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Modeling Fsi physics with viscoelastic material and low-Re-K-epsilon turbulence model
Posted 2015年1月4日 GMT-5 15:04 3 Replies
Please login with a confirmed email address before reporting spam
I am trying to model fsi physics. For fluid part, "low-Re K-epsilon" turbulence model is proper in this case and for solid mechanic part linear viscoelastic model must be chosen.
unfortunately, I encounter an error message "duplicate parameter/variable name"! it refers to auxiliary wall distance and instantaneous shear modulus which is "fsi.G0".
is there any way to solve this problem?
Thank You,
--
SY
unfortunately, I encounter an error message "duplicate parameter/variable name"! it refers to auxiliary wall distance and instantaneous shear modulus which is "fsi.G0".
is there any way to solve this problem?
Thank You,
--
SY
3 Replies Last Post 2015年1月8日 GMT-5 05:36
Please login with a confirmed email address before reporting spam
Posted:
10 years ago
Try k-e first to get used to how the FSI module works. The low-Rey k-e is much more difficult.
Then try the low-Rey k-e solution on your geometry without the structural feedback on the fluid just to see if you can get the fluid problem to solve with your geometry. The low-Rey k-e requires a separate computation for the wall distance to solve for G. This will need to be updated with each time step/iteration of the FSI problem since G changes with the structure.
Also, I have found the SST model to be more stable than the low-Re k-e model.
Then try the low-Rey k-e solution on your geometry without the structural feedback on the fluid just to see if you can get the fluid problem to solve with your geometry. The low-Rey k-e requires a separate computation for the wall distance to solve for G. This will need to be updated with each time step/iteration of the FSI problem since G changes with the structure.
Also, I have found the SST model to be more stable than the low-Re k-e model.
Please login with a confirmed email address before reporting spam
Posted:
10 years ago
Mr. Freels, Thank you so much for your reply
In fact, I dont' have any problem using k-epsilon and viscoelastic structure and also using low-Re-k-epsilon and elastic structure.
as you know both sst and low re k epsilon need to solve wall distance initialization but I guess my problem refers to a bug in COMSOL. Because of the same parameter name for wall distance initialization and instantaneous shear modulus I can't select both simultaneously. (You can test it)
--
SY
In fact, I dont' have any problem using k-epsilon and viscoelastic structure and also using low-Re-k-epsilon and elastic structure.
as you know both sst and low re k epsilon need to solve wall distance initialization but I guess my problem refers to a bug in COMSOL. Because of the same parameter name for wall distance initialization and instantaneous shear modulus I can't select both simultaneously. (You can test it)
--
SY
Please login with a confirmed email address before reporting spam
Posted:
10 years ago
Hi Soheyl,
There are two possible workarounds for this problem.
The first (and simpler) assumes that
a) You are using version 5.0
b) The deformations are so small that that the geometry of the fluid domain does not change.
In this case you can switch to the new 'Fluid-Structure Interaction, Fixed Geometry' instead of the more general 'Fluid-Structure Interaction'.
The second (and more general) workaround:
1. Add an extra Solid Mechanics interface to your model.
2. Change the name of the dependent variable in the Solid Mechanics interface so that it matches the name in the FSI interface. Typically: Replace 'u' by 'u_solid'. This is done in the 'Dependent Variables' section of the Solid Mechanics interface.
3. In the Solid Mechanics interface, select only the viscoelastic domains. Assign a viscoelastic material to all domains, and copy the viscoelastic data from the FSI interface.
4. In the FSI interface, replace the material for the viscoelastic domains by a Linear Elastic material with very low stiffness.
In effect, this approach will give you an overlay of two materials (one elastic from FSI and one viscoelastic from Solid Mechanics) for each viscoelastic domain. As long as the elastic material has a low enough stiffness, it will however not affect the results. Since there is no longer a viscoelastic material in the FSI interface, the variable name conflict is removed.
Regards,
Henrik
There are two possible workarounds for this problem.
The first (and simpler) assumes that
a) You are using version 5.0
b) The deformations are so small that that the geometry of the fluid domain does not change.
In this case you can switch to the new 'Fluid-Structure Interaction, Fixed Geometry' instead of the more general 'Fluid-Structure Interaction'.
The second (and more general) workaround:
1. Add an extra Solid Mechanics interface to your model.
2. Change the name of the dependent variable in the Solid Mechanics interface so that it matches the name in the FSI interface. Typically: Replace 'u' by 'u_solid'. This is done in the 'Dependent Variables' section of the Solid Mechanics interface.
3. In the Solid Mechanics interface, select only the viscoelastic domains. Assign a viscoelastic material to all domains, and copy the viscoelastic data from the FSI interface.
4. In the FSI interface, replace the material for the viscoelastic domains by a Linear Elastic material with very low stiffness.
In effect, this approach will give you an overlay of two materials (one elastic from FSI and one viscoelastic from Solid Mechanics) for each viscoelastic domain. As long as the elastic material has a low enough stiffness, it will however not affect the results. Since there is no longer a viscoelastic material in the FSI interface, the variable name conflict is removed.
Regards,
Henrik