CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   SU2 (http://www.cfd-online.com/Forums/su2/)
-   -   The initialization of flow field for time accurate, unsteady simulation and BC (http://www.cfd-online.com/Forums/su2/115324-initialization-flow-field-time-accurate-unsteady-simulation-bc.html)

zhengjg March 28, 2013 06:35

The initialization of flow field for time accurate, unsteady simulation and BC
 
Dear SU2 developers and users,
I wish to carry out the time accurate, unsteady simulation of an flow. I need to set up the simulation in SU2 and my questions are:
1. How can I implement the user-specified initial solution throughout the computational domain? The initial solution is the function of coordinates of mesh element. One possible way is to write the initial condition into the SU2 restart file using external program and following the SU2 restart file format. Then we carry out the calculation as a restart simulation. Is this correct?
2. I need to apply zero-gradient boundary condition. Is this type of boundary condition implemented in SU2?
3. Your AIAA paper indicates that the explicit R-K time integration scheme is included in SU2. Is this scheme developed by Jamson? Is the TVD R-K algorithm by Shu Chiwang used in your software?
4. For unsteady calculation, which limiter is better, Barth's limiter or Venk's algorithm?

Thank you

Amrita Lonkar April 11, 2013 13:38

Hi Jianguo,

Yes, the way to implement a set of grid dependent initial condition is through the restart file.


There are zero-gradient boundary conditions for some solvers such as the electric potential solver, there are adiabatic boundary conditions for zero temperature gradient at the walls. Can you please give us more detail about your equations and boundary conditions?

The explicit R-K time integration scheme in SU2 is the Jameson implementation with division between convective and viscous terms.

Venkatakrishnan limiter is generally more popular because it have a free parameter that allows convergence to machine precision. It is worth mentioning here that the limiter is only applied to the spatial discretization of the convective fluxes and NOT applied to the time discretization in SU2.


Hope this helps,

Amrita

fpalacios April 13, 2013 14:15

Quote:

Originally Posted by zhengjg (Post 416930)
Dear SU2 developers and users,
I wish to carry out the time accurate, unsteady simulation of an flow. I need to set up the simulation in SU2 and my questions are:
1. How can I implement the user-specified initial solution throughout the computational domain? The initial solution is the function of coordinates of mesh element. One possible way is to write the initial condition into the SU2 restart file using external program and following the SU2 restart file format. Then we carry out the calculation as a restart simulation. Is this correct?
2. I need to apply zero-gradient boundary condition. Is this type of boundary condition implemented in SU2?
3. Your AIAA paper indicates that the explicit R-K time integration scheme is included in SU2. Is this scheme developed by Jamson? Is the TVD R-K algorithm by Shu Chiwang used in your software?
4. For unsteady calculation, which limiter is better, Barth's limiter or Venk's algorithm?

Thank you

An initial condition can be also defined using the function:

solution_container[iZone][MESH_0][FLOW_SOL]->SetInitialCondition(geometry_container[iZone], solution_container[iZone], config_container[iZone], ExtIter);

that you will find in the next developer version of the code (it will be available during the next week).

Cheers,
Francisco


All times are GMT -4. The time now is 19:09.