# Data Center Air conditioning Boundary Condition problem

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

 October 31, 2012, 13:01 Data Center Air conditioning Boundary Condition problem #1 New Member   Jay Patel Join Date: Feb 2012 Posts: 8 Rep Power: 6 Hi Foamers, I am trying to simulate the flow of air in a data center. I am ready with mesh file but now I have a problem with BC. I have 4 type of BCs. 1 = CRAC(Computer Room Air Conditioner) -> Suction side -> ? (it will be taking air from the fluid domain 2 = CRAC(Computer Room Air Conditioner) -> Discharge side -> velocity(in will be inlet to the fluid domain. 3 = Air inlet to Server Rack -> ? 4 = Air outlet from Server Rack -> ? 3 and 4 must be having same Flow rate as what ever goes in the server due to suction fan in each server comes out from the other side of server. The flow diagram is like this. Air enters the the space from discharge side of CRAC(velocity and flow rate are know) the air passes through server (3) ,get heated and exit from other side of server (4), and again this hot air is suck by the CRAC suction side (1)get cooled and supplied again(2) Thanks for reading.

 December 10, 2012, 08:24 #2 Member   Dinesh Balaji Join Date: Oct 2012 Posts: 43 Rep Power: 5 Hi Jay, I am also working on the same problem. What software are you using and have you decided upon the boundary conditions?

 December 11, 2012, 13:11 Recirc BC #3 Member   Michael Roth Join Date: Mar 2009 Location: Guelph, Ontario, Canada Posts: 46 Rep Power: 9 I guess what you need is a recirc boundary condition. Consider swak4foam, and in particular, groovyBC, and even more specifically, the example "average-t-junction": http://openfoamwiki.net/index.php/Co...age-t-junction For the air inlet to the server rack, you would specify a negative velocity, everything else outflow-like. At the air outlet from the server rack, you would code up a groovyBC for temperature that grabs the average temperature in the air inlet (Tavg), add in a suitable rise in temperature: delta T = Heat (W) / ( density (kg/m3) * Cp (J/kg/K) * volume_flow (m3/s) ) and finally apply this temperature (Tavg + deltaT). The CRAC units, something similar, but with cooling applied. Boussinesq solver assumed so that we don't have to worry about density. Hopefully enough info above to get you started.

 December 11, 2012, 20:47 #4 Member   Dinesh Balaji Join Date: Oct 2012 Posts: 43 Rep Power: 5 Roth, thanks for the info. Will try it and let you know.

 July 25, 2013, 20:02 Any update #5 New Member   Alex Lee Join Date: Sep 2012 Posts: 16 Rep Power: 5 Hi guys, interesting topic! I am wondering have you guys managed to resolve the problem faced? I am also working on the same topic and would like to team up with you all. Alex

 July 25, 2013, 20:29 #6 Member   Dinesh Balaji Join Date: Oct 2012 Posts: 43 Rep Power: 5 Hi Alex, Kind of. But there is now a problem in modeling the data center using gmsh.

 July 27, 2013, 00:26 Meshing #7 New Member   Alex Lee Join Date: Sep 2012 Posts: 16 Rep Power: 5 I can help you to do the meshing if you can provide us some information. We have developed a nice GUI for SNHM and OF. Pls contact me at alexlee@zeb-tech.com and I can provide u with a version for testing.

November 13, 2013, 01:21
Type of BC's on rack & CRAC inlet and outlet
#8
New Member

kedar manohar
Join Date: Dec 2010
Posts: 11
Blog Entries: 1
Rep Power: 7
Quote:
 Originally Posted by roth I guess what you need is a recirc boundary condition. Consider swak4foam, and in particular, groovyBC, and even more specifically, the example "average-t-junction": http://openfoamwiki.net/index.php/Co...age-t-junction For the air inlet to the server rack, you would specify a negative velocity, everything else outflow-like. At the air outlet from the server rack, you would code up a groovyBC for temperature that grabs the average temperature in the air inlet (Tavg), add in a suitable rise in temperature: delta T = Heat (W) / ( density (kg/m3) * Cp (J/kg/K) * volume_flow (m3/s) ) and finally apply this temperature (Tavg + deltaT). The CRAC units, something similar, but with cooling applied. Boussinesq solver assumed so that we don't have to worry about density. Hopefully enough info above to get you started.
Hello;

The links are fine and UDF's mentioned are also fine. Can we apply BC's without UDF?

I am working on similar project. I have modeled everything in ICEM-CFD and using Fluent for CFD analysis.

Thanks and Regards
SSM

February 20, 2014, 13:12
outflow boundary, fixed velocity, pressure boundary setting??
#9
New Member

RB
Join Date: Aug 2013
Posts: 5
Rep Power: 5
Quote:
 Originally Posted by roth For the air inlet to the server rack, you would specify a negative velocity, everything else outflow-like.
Hi,

Nice topic, thanks for the interesting tips.

I've tried something really similar to Roth's advises.
But I'm a bit struggling on the boundary condition of what you named the "rack air inlet":
U: negative velocity --> I assumed it means pointing out/leaving the domain
P: ??? outflow-like ?
Traditionally for an outlet a fixed pressure is used, but it doesn't suit there as it will become overspecify with the velocity already set at fixedValue? So I have apply zeroGradient for the pressure (for P_rgh in my case), as I would have specify in case of a fixed velocity pointing inward my domain.
The simulation runs and hits the convergence criteria but I don't think the flow is acting accordingly to nature of a suction area around my "rack inlet"... i.e weird pressure profile and velocity getting a bit crazy close to the "rack inlet".

Any hint or idea on that particular point??

Thanks all,

R

 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 jwillie2000 CFX 1 December 7, 2009 18:07 smn CFX 5 November 24, 2009 07:37 Pankaj CFX 9 November 23, 2009 05:05 sam FLUENT 2 July 20, 2003 02:19 M Rad Main CFD Forum 12 November 27, 1998 13:49

All times are GMT -4. The time now is 03:17.