Approximating 3D models to 2D models

m_daniyal_alam · December 13, 2014, 22:33

can someone guide me that when can we simulate 3D geometries into 2D geometries. I have a project for college and i want to know that whether i can model it in 2d

i have attached the geometry,i need to heat the middle section as it is a resistojet

t.teschner · December 14, 2014, 09:18

what are you interested in? (mean) velocity profiles? pressure drop? ... is it a RANS or LES (DNS?) simulation?
is the wall only heated from the top and bottom or also from the sidewalls?
from the geometry it looks as if you have a some kind of a diffuser in the third direction therefore it is a bit tricky / risky to approximate it with a 2d model

m_daniyal_alam · December 14, 2014, 22:01

I need pressure drop and velocity profiles at nozzle exit,the walls are heated from the sides only,can u please guide me in general when can we approximate a given geometry to 2d models

t.teschner · December 15, 2014, 04:54

in that case i don't think you can approximate it with a 2d model.

i would say in general you can approximate two dimensionality when you can make your mesh two dimensional and then extrude it in the 3rd direction in order to get a 3d mesh. now for that you would also need at least symmetry or probably even better, periodic boundary conditions (always depending on the flow you have).
in your case you said that the side walls need to be heated so you can't reduce this direction to obtain a 2d grid and you can't reduce it in any other direction as you have geometrical features in all three coordinate directions.

another point (general speaking), you can't actually approximate a flow by a 2d model / mesh. for high reynolds number flows you encounter turbulences that are inherently 3d, when approximating it by a 2d mesh, the vortex stretching term will disappear and you no longer have a cascade of large energy eddies to smaller ones.

thankfully that is not a primary concern for engineers as we are most interested in "mean" flow profiles and hence using the RANS equations allows 2d meshes as they are just modelling the reynolds stresses but not calculating them exactly.

so again, in your case you should proceed with a 3d model, i can't see any easier 2d way.
cheers

December 13, 2014, 22:33	Approximating 3D models to 2D models	#1
m_daniyal_alam New Member Sindh Join Date: Dec 2014 Posts: 2 Rep Power: 0	can someone guide me that when can we simulate 3D geometries into 2D geometries. I have a project for college and i want to know that whether i can model it in 2d i have attached the geometry,i need to heat the middle section as it is a resistojet

December 14, 2014, 09:18		#2
t.teschner Senior Member Tom-Robin Teschner Join Date: Dec 2011 Location: Cranfield, UK Posts: 204 Rep Power: 16	what are you interested in? (mean) velocity profiles? pressure drop? ... is it a RANS or LES (DNS?) simulation? is the wall only heated from the top and bottom or also from the sidewalls? from the geometry it looks as if you have a some kind of a diffuser in the third direction therefore it is a bit tricky / risky to approximate it with a 2d model m_daniyal_alam likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Mesh requirements for k-w models	Jeremie84	FLUENT	5	June 13, 2014 09:49
Eddy Viscosity Models and Reynolds Stress Models	JuPa	CFX	1	August 20, 2013 18:56
Eddy viscosity hypothesis versus Reynolds stress models	JuPa	ANSYS	0	August 12, 2013 06:20
Incompressible Turbulence models	achinta	OpenFOAM	4	May 27, 2010 10:35
Two-fluid models vs mixture models for bubbly flows	Hansong Tang	Main CFD Forum	6	December 8, 2009 03:21

December 14, 2014, 22:01		#3
m_daniyal_alam New Member Sindh Join Date: Dec 2014 Posts: 2 Rep Power: 0	I need pressure drop and velocity profiles at nozzle exit,the walls are heated from the sides only,can u please guide me in general when can we approximate a given geometry to 2d models

December 15, 2014, 04:54		#4
t.teschner Senior Member Tom-Robin Teschner Join Date: Dec 2011 Location: Cranfield, UK Posts: 204 Rep Power: 16	in that case i don't think you can approximate it with a 2d model. i would say in general you can approximate two dimensionality when you can make your mesh two dimensional and then extrude it in the 3rd direction in order to get a 3d mesh. now for that you would also need at least symmetry or probably even better, periodic boundary conditions (always depending on the flow you have). in your case you said that the side walls need to be heated so you can't reduce this direction to obtain a 2d grid and you can't reduce it in any other direction as you have geometrical features in all three coordinate directions. another point (general speaking), you can't actually approximate a flow by a 2d model / mesh. for high reynolds number flows you encounter turbulences that are inherently 3d, when approximating it by a 2d mesh, the vortex stretching term will disappear and you no longer have a cascade of large energy eddies to smaller ones. thankfully that is not a primary concern for engineers as we are most interested in "mean" flow profiles and hence using the RANS equations allows 2d meshes as they are just modelling the reynolds stresses but not calculating them exactly. so again, in your case you should proceed with a 3d model, i can't see any easier 2d way. cheers