Tweet
Effect of mesh size on Fluid and FSI models
Hello!
I am using FEBio version 2.9 on Windows 8.1 Pro to study the flow of two Newtonian liquids flowing inside two concentric cylindrical tubes in a two-dimensional plane. The inner tube has a finite thickness. I used two models-
1. Fluid Mechanics- the material of inner tube is rigid (no rigid body has been used). This is achieved simply by fixing X, Y velocities=0 at those nodes where the fluids interact with walls of inner tube. Fluid_TwoLiquids_CoarseMesh.feb
2. FSI- the material of inner tube is solid neo-Hookean. FSI interface traction is applied along the walls of tube. FSI_TwoLiquids_CoarseMesh.feb
Other common boundary conditions are-No slip at the outer walls, zero dilation at the outlets and at those nodes which are common to the no-slip boundaries and inlet.
Common surface loads-normal velocity at the inlet (Magnitude 1, smooth from 0 to 1), Backflow and Tangential Flow stabilizations at inlets and outlets.
I expected a parabolic velocity profile in both cases. I am puzzled by the effect of mesh refinement on my solutions. I observed that a finer mesh gives a plug velocity profile while a coarse mesh gives the expected parabolic profile.
I tried with one liquid flowing in a channel, the results were opposite; a finer mesh gave the exact parabolic profile while a coarse mesh gave a plug profile.
Kindly help me in understanding this ambiguity or if there is any mistake in my model. Thank you for your patient read.
Fluid_TwoLiquids_CoarseMesh.pngFluid_TwoLiquids_FineMesh.png
Abha
I am using FEBio version 2.9 on Windows 8.1 Pro to study the flow of two Newtonian liquids flowing inside two concentric cylindrical tubes in a two-dimensional plane. The inner tube has a finite thickness. I used two models-
1. Fluid Mechanics- the material of inner tube is rigid (no rigid body has been used). This is achieved simply by fixing X, Y velocities=0 at those nodes where the fluids interact with walls of inner tube. Fluid_TwoLiquids_CoarseMesh.feb
2. FSI- the material of inner tube is solid neo-Hookean. FSI interface traction is applied along the walls of tube. FSI_TwoLiquids_CoarseMesh.feb
Other common boundary conditions are-No slip at the outer walls, zero dilation at the outlets and at those nodes which are common to the no-slip boundaries and inlet.
Common surface loads-normal velocity at the inlet (Magnitude 1, smooth from 0 to 1), Backflow and Tangential Flow stabilizations at inlets and outlets.
I expected a parabolic velocity profile in both cases. I am puzzled by the effect of mesh refinement on my solutions. I observed that a finer mesh gives a plug velocity profile while a coarse mesh gives the expected parabolic profile.
I tried with one liquid flowing in a channel, the results were opposite; a finer mesh gave the exact parabolic profile while a coarse mesh gave a plug profile.
Kindly help me in understanding this ambiguity or if there is any mistake in my model. Thank you for your patient read.
Fluid_TwoLiquids_CoarseMesh.pngFluid_TwoLiquids_FineMesh.png
Abha