Flow is not converging in natural convection problem of boussinesq model for air

rahulsharma · January 25, 2010, 02:16

hey guys
this is rahul working on cfd analysis of natural convection cross ventilation of a room inside a computational domain. the heat source is provided on a wall.
1. is it fine to define a heat source at wall in boundary condition panel as heat source W/m2
2. Solution is not converging than what should be the problem.
can anyone help me.
thanks,.

rlahlou · December 15, 2015, 10:19

Hey rahulsharma,

did you finally solve your issue?
I am facing convergence issue as well using Fluent for a free convection problem in a side-heated vertical channel. When I use Boussinesq assumption with pressure inlet and pressure outlet BC, I get the AMG solver divergence ( Although I reduced the under-relaxation factors, checked the mesh quality and a bunch of other changes recommended on forums for AMG solver divergence ...).
Using the ideal gas law for density instead of Boussinesq, I obtain a solution but it is unphysical.

Thanks a lot for your (or anyone else's) feedback!

DB99 · December 2, 2020, 19:19

Fluent does not converge well at low mass flow values.

JuPa · December 2, 2020, 20:05

If it's a closed domain (i.e. no inlets and no outlets) then it will struggle to converge.

DB99 · December 2, 2020, 22:07

Yes, I have seen the Fluent solver claim to meet convergence criteria but if a conservation of mass flow is done, in vs out, there can be a large difference. Factor of 7 is not unusual. These were in the low CFM range (4.5 e-4 kg/s if I converted that right from memory) in somewhat confined spaces with arrays of electronic enclosure vents as inlet and outlets.
Ansys tech support said sorry, the flow is too low. FloTHERM XT and FloEFD do not have such problems.

January 25, 2010, 02:16	Flow is not converging in natural convection problem of boussinesq model for air	#1
rahulsharma New Member rahul Join Date: Oct 2009 Posts: 10 Rep Power: 17	hey guys this is rahul working on cfd analysis of natural convection cross ventilation of a room inside a computational domain. the heat source is provided on a wall. 1. is it fine to define a heat source at wall in boundary condition panel as heat source W/m2 2. Solution is not converging than what should be the problem. can anyone help me. thanks,.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Natural Convection in a storage tank - boussinesq?	Jervds	CFX	6	October 2, 2016 03:53
Natural Convection problem in Fluent - urgent	NSV	FLUENT	10	May 6, 2014 05:25
lenght and number of nodes for an orifice meter flow problem - k-eps model	fofo	FLUENT	0	November 15, 2009 04:49
Adiabatic and Rotating wall (Convection problem)	ParodDav	CFX	5	April 29, 2007 20:13
Boussinesq - natural convection	Jeremie	FLUENT	2	March 8, 2007 09:12

December 15, 2015, 10:19		#2
rlahlou New Member Radia Join Date: Feb 2015 Posts: 2 Rep Power: 0	Hey rahulsharma, did you finally solve your issue? I am facing convergence issue as well using Fluent for a free convection problem in a side-heated vertical channel. When I use Boussinesq assumption with pressure inlet and pressure outlet BC, I get the AMG solver divergence ( Although I reduced the under-relaxation factors, checked the mesh quality and a bunch of other changes recommended on forums for AMG solver divergence ...). Using the ideal gas law for density instead of Boussinesq, I obtain a solution but it is unphysical. Thanks a lot for your (or anyone else's) feedback!

December 2, 2020, 19:19		#3
DB99 New Member DB Join Date: Jan 2019 Posts: 28 Rep Power: 7	Fluent does not converge well at low mass flow values.

December 2, 2020, 20:05		#4
JuPa Senior Member Mr CFD Join Date: Jun 2012 Location: Britain Posts: 361 Rep Power: 15	If it's a closed domain (i.e. no inlets and no outlets) then it will struggle to converge.

December 2, 2020, 22:07		#5
DB99 New Member DB Join Date: Jan 2019 Posts: 28 Rep Power: 7	Yes, I have seen the Fluent solver claim to meet convergence criteria but if a conservation of mass flow is done, in vs out, there can be a large difference. Factor of 7 is not unusual. These were in the low CFM range (4.5 e-4 kg/s if I converted that right from memory) in somewhat confined spaces with arrays of electronic enclosure vents as inlet and outlets. Ansys tech support said sorry, the flow is too low. FloTHERM XT and FloEFD do not have such problems.