CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   SU2 News & Announcements (http://www.cfd-online.com/Forums/su2-news-announcements/)
-   -   SU2 Version 2.0.4 Released Today (Developer Release) (http://www.cfd-online.com/Forums/su2-news-announcements/118100-su2-version-2-0-4-released-today-developer-release.html)

economon May 21, 2013 04:04

SU2 Version 2.0.4 Released Today (Developer Release)
 
Dear All,

Another note to say that we have posted the latest developer release of the code, version 2.0.4, today on our website, su2.stanford.edu. Apart from general bug fixes and improvements, this version has several notable new features, including:

- much improved stability (give a Krylov linear solver a try with a larger CFL number than usual for your implicit calculations)
- further improvements to the python scripts
- nitrogen gas simulations in the plasma solver
- RANS continuous adjoint improvements
- a new grid deformation technique based on linear elasticity (choose GRID_DEFORM_METHOD= FEA, 2-D only at the moment)
- geometrical constraints in shape optimization problems
- new capability to perform low-fidelity simulations by taking advantage of the multigrid structure (roughly one order of magnitude faster than a normal simulation, trade-off between speed and accuracy)

Lastly, there is one important update to the viscous wall boundary conditions. SU2 now uses two different marker definitions for no-slip walls that have replaced the original MARKER_NS option:

% Navier-Stokes (no-slip), constant heat flux wall marker(s) (NONE = no marker)
% Format: ( marker name, constant heat flux (J/m^2), ... )
MARKER_HEATFLUX= ( NONE )
%
% Navier-Stokes (no-slip), isothermal wall marker(s) (NONE = no marker)
% Format: ( marker name, constant wall temperature (K), ... )
MARKER_ISOTHERMAL= ( NONE )

These two BC options allow for the user to specify a heat flux or temperature directly for no-slip walls, respectively. To simply recover the old behavior for an adiabatic wall, choose MARKER_HEATFLUX= ( your_marker_name, 0.0 ). Please update your config files accordingly.

Again, thanks to everyone that is working with SU2 and posting on the forum. The code continues to improve with your help!

Cheers,
Tom

Tommy Chen May 21, 2013 12:18

Cool !!!!!!!!!!

curky May 21, 2013 15:57

Could somebody compile these sources for windows (serial and parallel)? I have tried to do this by myself but got only some errors from compiler.

shirazbj May 21, 2013 23:11

Quote:

Originally Posted by curky (Post 428986)
Could somebody compile these sources for windows (serial and parallel)? I have tried to do this by myself but got only some errors from compiler.

Hi Curky,

You could try my one. I only made one serial version of su2_cfd.exe under win7@32. You can found it here:

https://sites.google.com/site/gmsh2su2/download

Enjoy.

Cean

Tommy Chen May 22, 2013 11:36

Quote:

Originally Posted by economon (Post 428818)
Dear All,

Another note to say that we have posted the latest developer release of the code, version 2.0.4, today on our website, su2.stanford.edu. Apart from general bug fixes and improvements, this version has several notable new features, including:

- much improved stability (give a Krylov linear solver a try with a larger CFL number than usual for your implicit calculations)
- further improvements to the python scripts
- nitrogen gas simulations in the plasma solver
- RANS continuous adjoint improvements
- a new grid deformation technique based on linear elasticity (choose GRID_DEFORM_METHOD= FEA, 2-D only at the moment)
- geometrical constraints in shape optimization problems
- new capability to perform low-fidelity simulations by taking advantage of the multigrid structure (roughly one order of magnitude faster than a normal simulation, trade-off between speed and accuracy)

Lastly, there is one important update to the viscous wall boundary conditions. SU2 now uses two different marker definitions for no-slip walls that have replaced the original MARKER_NS option:

% Navier-Stokes (no-slip), constant heat flux wall marker(s) (NONE = no marker)
% Format: ( marker name, constant heat flux (J/m^2), ... )
MARKER_HEATFLUX= ( NONE )
%
% Navier-Stokes (no-slip), isothermal wall marker(s) (NONE = no marker)
% Format: ( marker name, constant wall temperature (K), ... )
MARKER_ISOTHERMAL= ( NONE )

These two BC options allow for the user to specify a heat flux or temperature directly for no-slip walls, respectively. To simply recover the old behavior for an adiabatic wall, choose MARKER_HEATFLUX= ( your_marker_name, 0.0 ). Please update your config files accordingly.

Again, thanks to everyone that is working with SU2 and posting on the forum. The code continues to improve with your help!

Cheers,
Tom


Dear economon:

As you mentioned above, several new features are involved in the new version SU2 (2.0.4) are cool, however , How could the users use these new features?

For instance, how can I set the geometrical constraints in shape optimization problems, how to set the corresponding configuration file format? I have not found the corresponding setting in the configuration template file in the new version SU2 package that I just downloaded.

I suggest SU2 developer team release a note with a simple tutorial of how to use the new features in a new version, at least with the format to activate the new feature.

What do you think, economon?

Tommy

martinl May 22, 2013 14:49

economon

I've just compiled version 2.0.4 on windows using cygwin. When I run the euler/naca0012 test case the flow.plt and flow.vtk files are not created. The surface_flow files are.

The same problem seems to be present in 2.0.3 but works OK in 2.0.0 using the same compilation environment.

Any thoughts on what is causing this?

Thanks
Martin

taxalian May 22, 2013 16:20

Quote:

Originally Posted by martinl (Post 429337)
economon

I've just compiled version 2.0.4 on windows using cygwin. When I run the euler/naca0012 test case the flow.plt and flow.vtk files are not created. The surface_flow files are.

The same problem seems to be present in 2.0.3 but works OK in 2.0.0 using the same compilation environment.

Any thoughts on what is causing this?

Thanks
Martin

If you need to use tecplot output with .plt extension, i guess for the new release you should compile SU2 code as :

--with-Tecio-lib=ARG Tecplot i/o library directory, ARG = path to tecio.a

For further details refer to:

http://adl.stanford.edu/docs/display...on+from+Source

Alternatively, you can still read the "flow" file using tecplot.

Hope this helps.

diwakaranant May 24, 2013 01:47

Issue plottting flow.dat file
 
Hi

I have been using SU2 v 2.0.4 developer version, but there is some issue
regarding plotting flow.dat file.

Firstly, some of the variables are listed as Conservative_1, Residual_1 etc.
which is a bit difficult to interpret.

Secondly, the range of the plotting variables always starts from 0 to max value
(instead it should be from min to max value). This is problematic
sometimes (like for plotting pressure) when we intend to know the minimum value of some variable.
This also affects the quality of contour plots.

Please look into these things

Thanks
Anant

CapSizer May 30, 2013 08:00

To follow on from this topic, I have also just installed and compiled SU2 v 2.0.4, and I am having similar difficulties. It does not seem to be possible to get a Paraview volume output file, nor does it seem to be possible to read the Tecplot files into Paraview. Any suggestions? ---OK, edit on this --- Paraview will read the Tecplot files, but you need to rename them to .tec first. That's a useable workaround.

fpalacios June 23, 2013 17:19

Quote:

Originally Posted by CapSizer (Post 430959)
To follow on from this topic, I have also just installed and compiled SU2 v 2.0.4, and I am having similar difficulties. It does not seem to be possible to get a Paraview volume output file, nor does it seem to be possible to read the Tecplot files into Paraview. Any suggestions? ---OK, edit on this --- Paraview will read the Tecplot files, but you need to rename them to .tec first. That's a useable workaround.

Hi, we have posted in our web page the developers release 2.0.5 which includes relevant changes in the I/O. If you have a while, please take a look at the new release.

Best,
Francisco

CapSizer June 24, 2013 00:20

Thanks Francisco, I will give that a try!

CapSizer June 25, 2013 14:32

Francisco, I have installed and compiled v 2.0.5 ... now I'm still not getting any vtk volume output files, but also (and this is new) not even any surface file. Am I missing something here?

mcell July 11, 2013 05:40

Quote:

Originally Posted by economon (Post 428818)
- a new grid deformation technique based on linear elasticity (choose GRID_DEFORM_METHOD= FEA, 2-D only at the moment)

I would be very interested in getting a reference for this technique as I am using SU2 in my master thesis and want to use this grid deformation :)

Thanks!
Manuel

economon August 10, 2013 13:30

Hi Manuel,

There is a small section in our recent paper here that describes a bit about this method of grid deformation: http://adl.stanford.edu/papers/AIAA-2013-2580.pdf. The 2-D version of the linear elasticity-based grid deformation was working well in V2.0.5 of the code, and with V2.0.6, the technique is working more robustly in both 2- and 3-D.

Thanks for working with SU2 and cheers,
Tom


All times are GMT -4. The time now is 11:45.