CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   SU2 (http://www.cfd-online.com/Forums/su2/)
-   -   Non-uniform inlet conditions? (http://www.cfd-online.com/Forums/su2/123574-non-uniform-inlet-conditions.html)

winter September 16, 2013 11:38

Non-uniform inlet conditions?
 
Hi everyone!

I was interested in validating SU2's incompressible solver to the Pitz Daily (backward facing step) experiment. To conduct the simulations it would be beneficial to use a profile for the velocity and the turbulence quantities (k and omega), as the entrance length is quite long.

I haven't, so far, found that SU2 is capable to use profiled inlet conditions. Has anybody found a way to use non-uniform inlet conditions in SU2?

fpalacios October 3, 2013 14:18

Quote:

Originally Posted by winter (Post 452050)
Hi everyone!

I was interested in validating SU2's incompressible solver to the Pitz Daily (backward facing step) experiment. To conduct the simulations it would be beneficial to use a profile for the velocity and the turbulence quantities (k and omega), as the entrance length is quite long.

I haven't, so far, found that SU2 is capable to use profiled inlet conditions. Has anybody found a way to use non-uniform inlet conditions in SU2?

Dear Magnus,
Thanks a lot for your interest in SU2. Most of the work in SU2 has been done for external flows, and so far we don't have a method to introduce an inlet profile.

However, it is not complicate to add this capability, and we are looking forward for contributions. A good starting point would be to modify.

void CEulerSolver::BC_Inlet(CGeometry *geometry, CSolver **solver_container, CNumerics *conv_numerics, CNumerics *visc_numerics, CConfig *config, unsigned short val_marker)

in solver_direct_mean.cpp and instead of using the velocity at the Infinity (GetVelocity_Inf) create a function that reads the velocity profile from a file and interpolate the solution given a (X,YZ) coordinate geometry->node[iPoint]->GetCoord().

As you can see, SU2 use the weak formulation, so a numerical flux will be computed between the value that you provide and the actual value at the node, and then the result will be added to the total residual at the point.

Our recommendation is to fork the SU2 repository in https://github.com/su2code and contact us if you need further support.

Thanks again for your interest,
Francisco


All times are GMT -4. The time now is 00:51.