how to get a upwind node

lzw2003 · March 24, 2011, 03:11

hi,
recently, i want to program a project using the high resolution scheme

, which is used in CFX.
some work has been done but i have no idea about how to get the upwind node to calculate phi_up,especially in hexahedron mesh.

maybe i will check the openFoam source code .

anyone have some suggestion ?

lzw2003 · May 4, 2011, 23:43

CFX uses CVFEM( control volume FEM)method, and this method is more closer to FEM, i think the common method for finding an upwind node in fem is eough.

could anyone give me some clue about finding an upwind node in FEM or CVFEM method? is it a pure mathematic problem?

lzw2003 · May 5, 2011, 04:51

i am sorry. after read more paper,now i think cvfem use the same methods for finding upwind nodes as used in fvm.
so i need to read openfoam code carefully to find the strategy used for 3D upwind node finding.

CFDtoy · May 6, 2011, 14:04

Hi there,
are you coding on your own? if so, did you create a list of face and cell connectivity..then do the following

face loop - on a given face..check flux from cell1 to cell2..if its positive cell1 is upwind cell - else cell2 is the downwind cell and take the properties from cell2.

Evne if you are not coding on your own (i mean basic solver structure is something like openFoams or others), accessing face / cell structure should be straightforward.

My 2c !

Thanks,

CFDtoy

Quote:

Originally Posted by lzw2003

i am sorry. after read more paper,now i think cvfem use the same methods for finding upwind nodes as used in fvm.
so i need to read openfoam code carefully to find the strategy used for 3D upwind node finding.

lzw2003 · May 8, 2011, 22:08

Hi CFDtoy,

i create my list of face and cell connectivity. if the direction of velocity on face is not parellel to the face normal or there is a big angle between them, ie. the face normal between cell1 and cell2 is parellel to x direction , the velocity is nearly vertical to x directoin and it is positive. shall we also choose the cell1's centre as the upwind node?

how does it handle in 3D mesh, is it the same?

thanks!

lalupp · May 10, 2011, 02:17

In 3d case For each hex cell which has got 6 faces ,there are 6 upstream cells possible. In most of the the codes collocated schemes are used in which pressure and velocity are evaluated at the cell center and face velocity always normal to the common face which has two option either inward or outward (-ve or +ve)

Thus to find upstream cell find face velocity (uf)
if uf > 0 upstream cell is the parent cell otherwise neighbor cell

Try openfvm documentation clearly described

March 24, 2011, 03:11	how to get a upwind node	#1
lzw2003 New Member zhiwei liu Join Date: Mar 2009 Posts: 22 Rep Power: 17	hi, recently, i want to program a project using the high resolution scheme , which is used in CFX. some work has been done but i have no idea about how to get the upwind node to calculate phi_up,especially in hexahedron mesh. maybe i will check the openFoam source code . anyone have some suggestion ?

May 4, 2011, 23:43	Again,how to find upwind node in FEM?	#2
lzw2003 New Member zhiwei liu Join Date: Mar 2009 Posts: 22 Rep Power: 17	CFX uses CVFEM( control volume FEM)method, and this method is more closer to FEM, i think the common method for finding an upwind node in fem is eough. could anyone give me some clue about finding an upwind node in FEM or CVFEM method? is it a pure mathematic problem?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Why results differs between upwind and vanLeerV	kjetil	OpenFOAM Running, Solving & CFD	1	November 10, 2012 04:55
solution diverges when linear upwind interpolation scheme is used	subash	OpenFOAM	0	May 29, 2010 01:23
License server not visible from master node	Charles	FLUENT	0	October 30, 2007 17:48
how can i list node temperatur by node location	Ali Doustdar	FLUENT	2	July 12, 2007 23:48
Roe upwind fluxes!!	Magesh	Main CFD Forum	2	July 17, 2003 11:35

May 5, 2011, 04:51		#3
lzw2003 New Member zhiwei liu Join Date: Mar 2009 Posts: 22 Rep Power: 17	i am sorry. after read more paper,now i think cvfem use the same methods for finding upwind nodes as used in fvm. so i need to read openfoam code carefully to find the strategy used for 3D upwind node finding.

May 8, 2011, 22:08		#5
lzw2003 New Member zhiwei liu Join Date: Mar 2009 Posts: 22 Rep Power: 17	Hi CFDtoy, i create my list of face and cell connectivity. if the direction of velocity on face is not parellel to the face normal or there is a big angle between them, ie. the face normal between cell1 and cell2 is parellel to x direction , the velocity is nearly vertical to x directoin and it is positive. shall we also choose the cell1's centre as the upwind node? how does it handle in 3D mesh, is it the same? thanks!

May 10, 2011, 02:17		#6
lalupp Member lalupp Join Date: Jul 2010 Location: India Posts: 44 Rep Power: 15	In 3d case For each hex cell which has got 6 faces ,there are 6 upstream cells possible. In most of the the codes collocated schemes are used in which pressure and velocity are evaluated at the cell center and face velocity always normal to the common face which has two option either inward or outward (-ve or +ve) Thus to find upstream cell find face velocity (uf) if uf > 0 upstream cell is the parent cell otherwise neighbor cell Try openfvm documentation clearly described