Buffer Zone Convergence problem

akhan · August 23, 2012, 16:17

Hi,

Friends I need your help with a problem I am facing for a while now.
I have made Buffer zone and wanted to see how the temperatures and the flow acts around the buffer zone, completely passive.

I am getting really high temperatures not sure as what error I have made
another problem is I have about 1.7 million nodes and it takes over a month to converge or reach 0% imbalance in the Buffer zone. It is currently at 17.7 % imbalance.

Here is the all the info from geometry to set up and I have added a pic of the results I obtained.

my computer is quite fast i7 16 gb of memory.

Geometry:
The geometry used in the 12-domain method involved modeling the TBZ along with all the physical walls and the south side and north side ambient regions

Mesh Settings:
Default Body Spacing 0.002 m
Default Face Spacing Angular Resolution, 30 degrees, 0.0005 m minimum edge, 0.002 m maximum edge
Mesh Options Delaunay Surface Meshing with Advancing Front Meshing Strategy and Volume Meshing
The final mesh resulted in 1482494 nodes and 5829675 elements.

Setup
The setup for this simulation involved 14 Solid-Fluid boundaries with the use of default "Glass Plate" and "Wood Board" materials. All other boundaries were set to symmetry except the top and bottom (roof and ceiling) boundaries which, for our purposes, were set to be adiabatic. Energy inputs into the system were done by use of subdomains with "Specific Source" Energy Sources (380 W/m2).

Solver Controls
Solver controls used a timescale factor of 0.1 with the target residual RMS of 10-5. Solver run used the Double Precision setting as to aid the convergence of a buoyancy driven system.

Thanks

ghorrocks · August 23, 2012, 19:53

I have no idea what a Buffer Zone is. Can you explain what it is?

But from your test I suspect it is a CHT simulation, and steady state. Is this correct? In that case, are you using a solid time scale factor?

If your simulation takes a month to run then I trust you have done a careful sensitivity analysis of everything (convergence, mesh, physics) before starting otherwise you are going to only find out in a month's time that your results are rubbish.

akhan · August 23, 2012, 20:46

Basically what I am trying to simulate is
A room within a room, a buffer zone around the room where the air heats up on the south side of the building and therefore rises and the opposite happens on the south side (cools and drops) therefore creating a convention loop (buoyancy).
so I have a heat source on the south side which heats up the air, glass on all four walls (2 south and 2 north) with different heat transfer coefficients, same goes for the roof and floor. The Top and bottom are adiabatic.
It is a stead state analysis.
Buffer Zone area:
turbulence model SST is used
heat transfer: Thermal Energy

akhan · August 23, 2012, 23:17

Is there anyone I can contact on the phone .. that way it will be better explaining the problem I am facing.

Thanks

ghorrocks · August 23, 2012, 23:52

Did you see the other comments in my post - sensitivity analysis, solid time scale factor?

If you are getting very high temperatures and the factors I mention above are under control then have a look at this FAQ:
http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

In your case I would be suspicious that:
1) Your radiation is not correct. This is commonly forgotten in these sort of simulations.
2) The simulation is not steady state. Natural convection simulations are often inherently transient.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[ICEM] volume mesh leaking out of the surface, in a multi zone problem	premchandpendota	ANSYS Meshing & Geometry	6	July 31, 2012 02:36
Porosity profile, dividing a zone, or getting zone location from zone	khoopes	FLUENT	0	June 2, 2012 20:39
[ICEM] issue occur after extrude 2D airfoil mesh and convergence problem in CFX	shiyun	ANSYS Meshing & Geometry	4	May 9, 2012 20:55
Error to re-open fluent case file	J.Gimbun	FLUENT	0	April 27, 2006 09:42
convergence problem	Trushar	Phoenics	5	August 28, 2002 00:40

August 23, 2012, 19:53		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	I have no idea what a Buffer Zone is. Can you explain what it is? But from your test I suspect it is a CHT simulation, and steady state. Is this correct? In that case, are you using a solid time scale factor? If your simulation takes a month to run then I trust you have done a careful sensitivity analysis of everything (convergence, mesh, physics) before starting otherwise you are going to only find out in a month's time that your results are rubbish.

August 23, 2012, 20:46		#3
akhan New Member A Khan Join Date: Aug 2012 Posts: 3 Rep Power: 14	Basically what I am trying to simulate is A room within a room, a buffer zone around the room where the air heats up on the south side of the building and therefore rises and the opposite happens on the south side (cools and drops) therefore creating a convention loop (buoyancy). so I have a heat source on the south side which heats up the air, glass on all four walls (2 south and 2 north) with different heat transfer coefficients, same goes for the roof and floor. The Top and bottom are adiabatic. It is a stead state analysis. Buffer Zone area: turbulence model SST is used heat transfer: Thermal Energy

August 23, 2012, 23:17		#4
akhan New Member A Khan Join Date: Aug 2012 Posts: 3 Rep Power: 14	Is there anyone I can contact on the phone .. that way it will be better explaining the problem I am facing. Thanks

August 23, 2012, 23:52		#5
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Did you see the other comments in my post - sensitivity analysis, solid time scale factor? If you are getting very high temperatures and the factors I mention above are under control then have a look at this FAQ: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F In your case I would be suspicious that: 1) Your radiation is not correct. This is commonly forgotten in these sort of simulations. 2) The simulation is not steady state. Natural convection simulations are often inherently transient.