
[Sponsors] 
December 5, 2006, 04:49 
natural convection

#1 
Guest
Posts: n/a

hi i am doing natural convection of closed domain in which wall is heated and cylinder is cooled thru natural convection of cylinder. in answer when i used gravity it is not converging but i have smooth residual criteria.in contours plot i have negative value h which seems not correct.can yu tell me the reasons mangeer verma


December 5, 2006, 05:31 
Re: natural convection

#2 
Guest
Posts: n/a

First of all, which solver you are using? Did you activiate DTRM and set density of air as "boussinesq" with 1.225 and the absorption coefficient as 0.00027?


December 6, 2006, 04:15 
Re: natural convection

#3 
Guest
Posts: n/a

hi, first thanks for giving information i am using segregatted solver i used radiation along with boussinesq it gives a big value of h .if i do not use radiation it becomes very low. i have convergence problem too nabeel


December 6, 2006, 05:03 
Re: natural convection

#4 
Guest
Posts: n/a

hi why you refered DTRm model for natural convection i want simulate shot cylinder in the air is cooled but when i modeled in gambit i cannot set boundary condition to air becuase it cannot be set to wall. kindly help me in this regard


December 6, 2006, 06:45 
Re: natural convection

#5 
Guest
Posts: n/a

For the boundary condition setting in Gambit, first create a rectangular block (acts as an outer environment, larger than the cylinder) to hold the cylinder. And then subtract the cylinder from the block. Mesh the remaining space. The walls if the block are the boundary of the air.
I have used DTRM model to simulate natural convection, the result is quite good. Anyway, did you activate the Energy Equation? 

December 6, 2006, 07:07 
Re: natural convection

#6 
Guest
Posts: n/a

Forget to mention air thermal expansion coefficient is set 0.00279.
but do your "h" stand for convection coefficient? My method is: (1) obtain heat rate (Q) from the Flux report. (2) use Q = hA DT to calculate h. DT is the difference between the temp of cylinder and the ambient temp. A is the surface area of the cylinder. Q would be negative or positive, but that only means the heat transfer direction. 

December 7, 2006, 03:16 
Re: natural convection

#7 
Guest
Posts: n/a

hi yes i cativated energy equation and i made both wall condition isothermal it covnverges but the solution is not reliable. second how can cimulate air in the cylinder considering closed domain


December 7, 2006, 03:25 
Re: natural convection

#8 
Guest
Posts: n/a

You mean the cylinder is closed and some fluid (water? liquid?) is inside the cylinder? and then the whole cylinder dissipate heat to the outer environment?


December 11, 2006, 01:45 
Re: natural convection

#9 
Guest
Posts: n/a

hi i have made a very large surronding relative to cylinder and made it adiabatic it gave me good result and second if heat transfer coeffecient comes negative what it means


December 11, 2006, 02:10 
Re: natural convection

#10 
Guest
Posts: n/a

The negative sign means heat is losting at that surface. positive & negative indicate the direction of heat transfer.
The surrounding is more than double length of the cylinder is enough. To speed up the simulations, you can use fine mesh near the cylinder and coarse mesh in the area far away the cylinder. Hope i can help you. 

December 12, 2006, 01:05 
Re: natural convection

#11 
Guest
Posts: n/a

hi thanks for coperating i am doing 3d simulation and in that i have same problem of not transferring heat from cylinder to surronding and i used wall as coupled boundary condition but it is not working yet i am trying other option


December 12, 2006, 06:08 
Flow simulation in a heat pipe

#12 
Guest
Posts: n/a

I am trying to use the FLUENT solver for a coupled problem of heat flow in the vapor region and heat conuction in the solid region of cylindrical heat pipe. We know the boundary conditions at the wall and with that boundary condition, we have to solve the energy and momentum equation for the vapor region and heat conduction in the pipe wall. The solution always diverges. Please suggest me the way to solve the problem. Thanks


December 15, 2006, 03:18 
Re: Flow simulation in a heat pipe

#13 
Guest
Posts: n/a

hi i cannot understand your problem and why are you using coupled solver. kindly tell me your problem in detail


December 20, 2006, 05:27 
Re: Flow simulation in a heat pipe

#14 
Guest
Posts: n/a

hi I'm a student on Msc of mechanical engineering and haet pipe modeling on Fluent was my Bsc thesis and work on it for one year and could model the starting section of heatpipe and want to work more than this on Msc period i will be glad if we can help each other.


December 20, 2006, 06:34 
Re: Flow simulation in a heat pipe

#15 
Guest
Posts: n/a

In a heat pipe, heat is conducted to the wall of the pipe. this heat is transferred to the liquid in the wick as a result liquid evaporates and flow towards the condenser where it is cooled down and the liquid is again moves back to the evaporator. We model heat conduction in the wall of the pipe, liquid in the wicks and vapor in the vapor low region. This makes a coupled problem.


December 21, 2006, 00:13 
Re: Flow simulation in a heat pipe

#16 
Guest
Posts: n/a

Hi all,
For problem involving heat transfer from the solid region to the fluid region and also the heat flow around the solid region, i think the mesh generated in gambit plays important role (agreed with peggy). The mesh near the wall should be very fine, and 'gradually' getting coarser when the distance from the wall is further. This is similar to modeling external flow boundary condition. In gambit, I tried to used 'boundary layer' (at mesh part) and 'size function' (at tool part). Besides, meshing using Hex cell element is always better compared to Tetra. cell element, in term of convergence speed. Below are some suggestion I obtained from this forum previously to get a converged solution in fluent 6.2: 1. In define>model>solver, selected 'nodedbased' for gradient option. 2. If you are using tetrahedral grid, u might have to input text command at bottom of the fluent window as > (rpsetvar 'temperature/secondarygradient? #f) ...including the bracket... 3. Change multigrid settings (solve>controls>multigrid) for energy. Set it to W cycle with a termination criteria of 0.01 (or smaller) Peggy, can i ask u why u use the value of 0.00027 for absorption coefficient and 0.00279 for air thermal expansion coefficient? I am searching for the appropriate value for the 2 properties...Btw, i ll try to input this 2 value into my simulation... Regards Tuw 

January 1, 2007, 22:52 
Re: Flow simulation in a heat pipe

#17 
Guest
Posts: n/a

For air properties, actually it will change slightly (you may find that it doesnt affect the overall result much) with temperature. I forget my original source for data. But the following reference is more or less the same as what I mentioned.
Thermal Expansion Coefficient: http://www.engineeringtoolbox.com/ai...iesd_156.html I really forget the source for absorption coefficient. But if you are doing buildingrelated analysis, and there's a certain location for the building. Please check the geographical data from local guidelines/code. (For example, in china, check the GB code). 

Thread Tools  
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
natural convection problem with radiation  jorien  CFX  0  October 14, 2011 09:26 
Natural Convection with heat generation  krishnachandranr  Main CFD Forum  0  July 28, 2009 04:22 
Coupled vs Seg  Natural vs. Forced Convection  Alex  Siemens  5  December 12, 2007 05:58 
natural convection at high Rayleigh  mauricio  FLUENT  2  February 23, 2005 20:43 
Mixing By Natural Convection Processes  Greg Perkins  FLUENT  0  February 12, 2003 19:40 