hi,
Unfortunately i cant give you the grid the reason being its confidentiality.
I think getfac command is the cause of the crash.
I tried using usdpst to have extra mesh of my wall boundary and for the plane. when i run code saturne, I see ensight folders named plane.ensight and wall.ensight. but when 1 open it in paraview and extract the block (wall or plane) but i dont see anything. Its all blank. Please find my usdpst subroutine attached with this comment
Sorry, i trouble you a lot :(
calculating Drag over an airfoil
Forum rules
Please read the forum usage recommendations before posting.
Please read the forum usage recommendations before posting.
-
Mubashir Ali
Re: calculating Drag over an airfoil
- Attachments
-
- usdpst.F
- (15.01 KiB) Downloaded 178 times
-
Yvan Fournier
Re: calculating Drag over an airfoil
Hello,
Be careful, use GETFACwith NLFAC and LSTFAC, and GETCEL with NLCEL and LSTCEL (in usdpst.F, the three lists are different, and sized according to the maximum number of cell or faces; in usproj.F, LSTELT is sized after the maximum of NCEL, NFAC ,and NFABOR).
You crash remains surprising, as NFAC is usually greater than NCEL, so using an array of size NFAC for a maiximum of NCEL values should be OK, but when building the postprocessing mesh, you still have the wrong list.
So I would recommend using NLCEL and LSTCEL for part 1 (selected with GETCEL) and NLFBR and LSTFBR for part 2 (selected with GETFBR) and see if it solves your problem.
Also, again, be careful of using a mesh in millimeters instead of meters. I recomment rescaling the mesh rather than having to adjust all physical parameters.
Regards,
Yvan
Be careful, use GETFACwith NLFAC and LSTFAC, and GETCEL with NLCEL and LSTCEL (in usdpst.F, the three lists are different, and sized according to the maximum number of cell or faces; in usproj.F, LSTELT is sized after the maximum of NCEL, NFAC ,and NFABOR).
You crash remains surprising, as NFAC is usually greater than NCEL, so using an array of size NFAC for a maiximum of NCEL values should be OK, but when building the postprocessing mesh, you still have the wrong list.
So I would recommend using NLCEL and LSTCEL for part 1 (selected with GETCEL) and NLFBR and LSTFBR for part 2 (selected with GETFBR) and see if it solves your problem.
Also, again, be careful of using a mesh in millimeters instead of meters. I recomment rescaling the mesh rather than having to adjust all physical parameters.
Regards,
Yvan
-
Mubashir Ali
Re: calculating Drag over an airfoil
hi,
the above suggestions have solved my problem. I am able to see the meshes made by usdpst.F. the wall mesh is perfect but plane is still a problem. I see a plane but its at the place where specified, may be this a problem of units as rightly said by you.
Regarding units, i checked Gambit..... (where i made geometry and mesh) that it has dimensionless units. so, if i would like to specify units for example SI units in code saturne how is it possible ???
for example when i transfer mesh made in gambit to FLUENT.... there in FLUENT we have a option to dimensionalize everything for example SI units.
the above suggestions have solved my problem. I am able to see the meshes made by usdpst.F. the wall mesh is perfect but plane is still a problem. I see a plane but its at the place where specified, may be this a problem of units as rightly said by you.
Regarding units, i checked Gambit..... (where i made geometry and mesh) that it has dimensionless units. so, if i would like to specify units for example SI units in code saturne how is it possible ???
for example when i transfer mesh made in gambit to FLUENT.... there in FLUENT we have a option to dimensionalize everything for example SI units.
-
Yvan Fournier
Re: calculating Drag over an airfoil
Hello,
Code_Saturne assumes the input mesh is already in SI units. if you can export in in SI in Gambit, that would be the simplest solution.
Otherwise, you can scale the mesh with Code_Saturne either by using a meta-mesh file for the preprocessor (see user documentation), or use the USMPST user subroutine to rescale the mesh to SI units (only if you do not have periodicity, as using USMPST updates the mesh but not the periodicity transformation).
Regards,
Yvan
Code_Saturne assumes the input mesh is already in SI units. if you can export in in SI in Gambit, that would be the simplest solution.
Otherwise, you can scale the mesh with Code_Saturne either by using a meta-mesh file for the preprocessor (see user documentation), or use the USMPST user subroutine to rescale the mesh to SI units (only if you do not have periodicity, as using USMPST updates the mesh but not the periodicity transformation).
Regards,
Yvan