Hello, I have a rectangular open channel and I want this channel to be infinitely wide and long. Can I assign both side walls as symmetric boundary condition? OR I can assign only one side wall as symmetric boundary condition? I don't want any influence of side walls on flow. Please help me out. I am using Ansys. Thanks.
I think that what you want is to prescribe the periodicity over a sufficient extension in spanwise direction,
In general, the symmetric boundary condition is applied on the center line NOT on the wall. However, if the viscous force is taken into account, the symmetric about the center line does not exist because the bifurcation phenomenon.
practically, yes. Your description is equivalent to ship simulations in open ocean.
Filippo, I don't have any pressure gradient in span wise direction, so theoretically mass flow rate is zero in span wise direction. In stream wise direction, I can apply periodic BC but I have porous bed under the free surface flow and I want to see interaction with porous bed. As per my understanding, in such case we cannot apply periodic BC condition in stream wise direction as well (bottom of free flow and top of porous bed make an interface). I am left with symmetric BC in span wise direction to avoid side wall effect. So can I apply symmetric BC on both side walls to make channel infinitely wide? (probably YES)
Narendra Patel
no, symmetric BC is somehow like a fictious wall having slip. It mimics a mirror behavior https://www.afs.enea.it/project/neptunius/docs/fluent/html/ug/node251.htm
Filippo Maria Denaro Thank you for help. Ideally my problem is 2D, flow in stream wise and vertical direction. In span wise direction, flux is zero. But I'm solving as 3D problem and don't want influence of side walls.
Narendra Patel
If you have a laminar steady flow you can solve directly the 2D case (that assumes you have an infinite extension in spanwise direction), thus why are you using a 3D extension?
On the other hand, if the flow is transitional or turbulent, you have a 2D flow only for the statistically mean fields and that must treated in a RANS formulation.
But if you want to solve the 3D case to study transition, the unsteady flow develops 3D features even if you do not have a forcing in spanwise direction.
In conclusion, the definition of the BCs depends on the goal of your study.
Filippo Maria Denaro I am solving 3D problem as I had difficulty in convergence in 2D case. I have turbulent flow. Assume a river flowing over porous sand bed and at the mid of river where side effect is null, I cut a block 1m long, 1m wide and depth=depth of river + 25cm depth of top part of porous bed. I want to simulate this to generate velocity profile (both inside the bed and above the bed) for steady state. What BCs you would suggest for all sides? Thanks.
Narendra Patel
If you have a turbulent flow in statistical energy equilibrium, your configuration is congruent to a 2D steady RANS formulation. If that does not converge turning to a 3D problem does not solve the issue.
Hovewer, if you insist to use the 3D configuration, you have to consider your 3D block as one of an infinite series of blocks on both sides. That means you have to set periodical conditions in spanwise direction. But this configuration makes not really sense for a RANS formulation since you final statistical solution must be 2D
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Condensed Matter Physics
- Soft Matter
- Liquids