# CFD-problem in ANSYS

 Register Blogs Members List Search Today's Posts Mark Forums Read

January 17, 2011, 11:41
CFD-problem in ANSYS
#1
New Member

Join Date: Jan 2011
Location: Germany
Posts: 12
Rep Power: 7
Dear all,

I'm new here (and to Ansys), so let me say "hi" to you all.

I just started using Ansys to model a gas flow under reduced pressure conditions for a simulation chapter in my PhD-thesis. Let me explain the procedure:

I have a gas flow that is running through a tube (from left to rigth) with a diameter of 0.046 and a length of 0.06. Gasvelocity varies between 0.001 and 1000.0, density and viscosity are known.

Here's my question: how do I put in the pressure (60 Pa)? Up to now, I put the pressure (DOF) on the outlet line (in the 2 D modeled tube, on the right hand). But nothing changes, when I change the pressure. Thus I think that my way of putting in the pressure is not correct.

Please have a look at the pic of my model which I uploaded. You will see that I have put the velocity (red arrows, vx= 1, vy, vz=0) on the left side of the model, velocities on the horizontal borders of the tube are all 0! Pressure (green arrows) are on the right side of my tube.

Can anyone give me a hint on how to put in the pressure effect, so it affects the flow?

Thank you all

Slavius
Attached Images
 cfx.jpg (44.6 KB, 27 views)

 January 17, 2011, 17:47 #2 Super Moderator   Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 12,638 Rep Power: 98 Usually you set the reference pressure to be a typical pressure in the flow, maybe the inlet or outlet pressures. Then the pressure at that inlet or outlet is zero - as pressure is relative to the reference pressure. All pressures in the simulation are measured relative to the reference pressure, use the "absolute pressure" variable if you want the absolute pressure.

 January 18, 2011, 02:26 #3 New Member   Join Date: Jan 2011 Location: Germany Posts: 12 Rep Power: 7 Good morning from Germany, thanks for your reply. The pressure is not just at the in- our the outlet. The whole tube is pumped down to and held at 60 Pa during the gas flow. Do I have to set the ref. pressure to 60 Pa? Regards

 January 18, 2011, 19:13 #4 Super Moderator   Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 12,638 Rep Power: 98 Hello from Australia! In that case set the reference pressure to 60Pa, and all pressures are set relative to that.

 January 19, 2011, 13:19 #5 New Member   Join Date: Jan 2011 Location: Germany Posts: 12 Rep Power: 7 Hi there, thanks for the answer. I made two simulations w. 2 different ref. pressures: 1st simulation) 101350 Pa, 2nd simulation) 60 Pa ref. pressure. I applied the pressure to all lines of my geometrical model (via pick all lines), because the atmosphere in the pipe has a constant pressure (independently of the gas flow velocity). I entered a 0 (zero) as the pressure value in the up poping input field. I've expected that flow characteristics change w. alternating pressure, but they don't. Flow profiles look perfectly the same at both ref. pressures with equal flow velocities. I think that these pressure invariant flow profiles originate from a wrong pressure definition in my model. Do I have to enter a 0 (zero) in the input field, or do I leave it blank? Is the pressure definition on all the lines correct after all, when the pressure inside the pipe is constantly 10150 Pa or 60 Pa everywhere in the pipe? Best regards, and thank you so much for your support .

 January 19, 2011, 13:46 #6 Senior Member   Edmund Singer P.E. Join Date: Aug 2010 Location: Minneapolis, MN Posts: 512 Rep Power: 12 Pressure invarient flow profile might indicate you have constant properties gas. When you state that density and viscosity are known, is that at a point and constant, or tabular, or ideal gas? Also, what do you mean your entire section of pipe is constant pressure? What is driving the flow?

 January 19, 2011, 14:03 #7 New Member   Join Date: Jan 2011 Location: Germany Posts: 12 Rep Power: 7 Hi there, it is a vacuum chamber with a butterly control valve and a pumping system. The gas enters the chamber by a magnetic valve. The control valve can compansate the pressure increase by reducing the suction cross-section at the pump. I'm only interested on the flow conditions after reaching the 60 Pa chamber pressure. I'm not interested what happens before. The gas is ArH2, the density and viscosity at 60 Pa are calculated externally by a fluid property program. The question is, whether Ansys Flotran can compute flows at 60 Pa at all?

 January 19, 2011, 15:12 #8 Senior Member   Edmund Singer P.E. Join Date: Aug 2010 Location: Minneapolis, MN Posts: 512 Rep Power: 12 You are using Flotran!? This is a CFX forum. I can't comment on the applicability of Flotran in near vaccuum conditions.

 January 19, 2011, 15:31 #9 New Member   Join Date: Jan 2011 Location: Germany Posts: 12 Rep Power: 7 Hi there, yes FLOTRAN. Is CFX capable of near vacuum conds? Greets

 January 19, 2011, 15:49 #10 Senior Member   Edmund Singer P.E. Join Date: Aug 2010 Location: Minneapolis, MN Posts: 512 Rep Power: 12 I would doubt most commercial codes (CFX and FLOTRAN included) would not handle non-continuum regimes. What is the Knudsen number of your flow? If it is near .001 or greater, then you probably need something that deals with a more statistical flow and the contiuum assumptions that these codes are built around break down.

 January 20, 2011, 07:59 #11 New Member   Join Date: Jan 2011 Location: Germany Posts: 12 Rep Power: 7 Hi, Kn lies between 0.0037 and 0.0008 (gas flows through a pipe with 2 diameters). Literature states that with these Kn flow character is viscouos, so a distinction between laminar and turbulent flow should easily be possible by the application of the (navier-)stokes eqns. and the Re-number. Maybe that is the reason why the flow character is invariant of the pressure! Regards

 January 20, 2011, 21:15 #12 Super Moderator   Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 12,638 Rep Power: 98 I think Fluent has a modified wall boundary condition which is designed to allow running at low Knudsen numbers. Not sure it is any use for you, but you might want to check it out.

January 24, 2011, 05:12
#13
New Member

Join Date: Jan 2011
Location: Germany
Posts: 12
Rep Power: 7
Hi there,

well now, thanks for the quick replies and the guidance, so far I solved the problem.

But only on laminar basis.

My Re-number calculations tell that my flow becomes turbulent, beyond a certain flow velocity. So I switched the calculations from laminar to turbulent (Pic1).

The solution runs well for very small gas velocities (0.001 - 1 m/s).

An error message pops up at flow velocities beyond 1 m/s, telling me that the solution diverges (ENKE=1), c.p. Pic2.

The boundary conditions were selected like in Pic 3. Execution controls were set like in Pic 4.

Can anyone of you tell me what went wrong?

Many thanks and greetings to you all

Attached Images
 Pic1.jpg (35.5 KB, 10 views) Pic2.jpg (49.4 KB, 8 views) Pic3.jpg (58.3 KB, 10 views) Pic4.jpg (40.2 KB, 7 views)

 January 24, 2011, 06:44 #14 Super Moderator   Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 12,638 Rep Power: 98 As previously stated, nobody here on this forum has seen Flotran since the dinosaurs walked the earth. Time to upgrade to a modern solver I think.

 January 24, 2011, 12:02 #15 New Member   Join Date: Jan 2011 Location: Germany Posts: 12 Rep Power: 7 Hi, I can't tell how much I'd love to switch to CFD. But CFD seems to be accessible through ANSYS Workbench. And I cannot access CFD through Workbench, although I can model and mesh the geometry. But when it comes to solving: ah ah! No CFD! Greets

 January 24, 2011, 17:36 #16 Super Moderator   Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 12,638 Rep Power: 98 CFX runs just fine stand-alone without workbench. I never run it inside workbench.

 January 25, 2011, 11:38 #17 New Member   Join Date: Jan 2011 Location: Germany Posts: 12 Rep Power: 7 No, it doesn't. There is not even a shortcut in the menu. But anyway. Thanks for the support.

 January 25, 2011, 16:57 #18 Super Moderator   Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 12,638 Rep Power: 98 I think I know what is happening - you must have a really old version of ANSYS from before the CFX merger, that is why you have Flotran. Just goes to show how old your version is.

 Thread Tools Display Modes Linear Mode

 Posting Rules You may not post new threads You may not post replies You may not post attachments You may not edit your posts BB code is On Smilies are On [IMG] code is On HTML code is OffTrackbacks are On Pingbacks are On Refbacks are On Forum Rules

 Similar Threads Thread Thread Starter Forum Replies Last Post adeban CFX 17 June 13, 2016 07:51 compizer CFX 0 May 5, 2010 03:02 Kordou CFX 4 July 17, 2008 12:49 Stephen Lau CFX 1 April 18, 2007 03:02 Ben Main CFD Forum 17 September 20, 2006 03:36

All times are GMT -4. The time now is 17:33.