[endron]

Hi everyone,

I am trying to simulate flow around a train in the OpenFoam. I have set up my case in such a way that a rectangular-shaped domain has an inlet, outlet, moving ground, and slip-wall sides and top.

Pressure distribution on these slip walls is non-uniform, there is a difference of several Pascals across each boundary face.

Is this something that could influence the result (f. ex. drag force coefficient)? Can the solution independence with respect to the domain size be assessed without resizing the domain?

[mAlletto]

The only way to check if the solution is independent of the domain size is to successively enlarge the domain and look the what size of the domain the solution does not change anymore substantially

[Tobermory]

The only other things to add to Michael's suggestion is that:
1. You could consider the "rules of thumb" for wind tunnel modelling; typically you would want to achieve a blockage ratio of 1% or less. What is the blockage ratio in your domain (blocked area over cross sectional area)?
2. You could run a potential flow simulation and look at the velocity perturbation created by the train at the distance of the boundary - is it negligible? If not, then the boundary is probably too close.

But ultimately, Michael's suggestion is the best one - the only way you can be sure is to increase it and run again.

[endron]

Originally Posted by Tobermory

The only other things to add to Michael's suggestion is that:
1. You could consider the "rules of thumb" for wind tunnel modelling; typically you would want to achieve a blockage ratio of 1% or less. What is the blockage ratio in your domain (blocked area over cross sectional area)?
2. You could run a potential flow simulation and look at the velocity perturbation created by the train at the distance of the boundary - is it negligible? If not, then the boundary is probably too close.

But ultimately, Michael's suggestion is the best one - the only way you can be sure is to increase it and run again.

Thank you for the answer!

1. The blockage ratio in my domain is equal to 0.15% (see attached picture).
2. I have already run a steady RANS case with the kw-SST turbulence model. Velocity at the inlet was set to uniform (-80 0 0) and pressure to 0 at the outlet. I have attached pictures of pressure and velocity distribution at the boundaries.

blockage.jpg

boundary-pressure-1.jpg

boundary-pressure-2.jpg

boundary-velocity-half-1.PNG