CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Fluidized Bed With Gas-Air Combustion (http://www.cfd-online.com/Forums/fluent/29737-fluidized-bed-gas-air-combustion.html)

Fred Kang May 9, 2002 11:31

Fluidized Bed With Gas-Air Combustion
 
Dear all,

I am going to model a fluidized bed incinerator, which is used to burn off the trace amount of solvent on solid particle surface. The case description are as below: (1) There is a gaseous combustion: natural gas is introduced to the fluidization zone, reacts with the fluidization air supplied through an air distributor. (2) The mass fraction of the solvent on the particle surface is negligible, so that we can forget its impact on any aspects of modeling. (3) The volume fraction of the discrete phase is extremely low (<1.0%). The round particles are 0.5~1.0 mm in diameter. (4) The vessel for fluidization is 1.0 m in diameter and 4.0 m in height. The air distributor is installed at the lower part of the vessel.

I read the FLUENT manual yesterday, and found that both the Eulerian model and the DPM model are of option. For general fluidization beds, Eulerian model is preferred. But, it seems there is no way to count heat transfer and combustion in Eulerian model, which is fatal to this application. For the DPM, I barely know the expense to run it for tracking so many particles.

I expect your suggestion/experience/thought to direct the modeling of such a fluidization bed with gas-air combustion attached. I appreciate your time and expertise.

Fred

Jin-Wook LEE May 10, 2002 01:02

Re: Fluidized Bed With Gas-Air Combustion
 
1. I am doubt that 'he volume fraction of the discrete phase is extremely low(<1.0%)' for fludized bed incinerator.

2. Anyway, my suggestion is that : You don't need to track each particle. Actually, as you guess, it is impossible to track such many particles. Instead, in general, you may define 100~1000 injection points. Assume that you define 100 injection points, then each each injection represents 1% of total mass flow rate.

3. Mathematically, 1% volume fraction of particle phase is no problem. But ...... Even though volume fraction of dispersed phase is low, the mass fraction might be fairly high. Assume volume fraction as 1%, gas density 1kg/m^3, particle density 1000kg/m^3. Then mass flow rate of dispersed phase is about 10 times that of gas phase. Then you may have serious problem for convergence. As far as my experience is concerned, when mass flow rate of particle exceeds that of gas phase(the ratio exceeds 1), the convergence was very very very difficult. I am not sure that you can obtain converged solution for such a high particle mass flow rate by using DPM.

Sincerely, Jinwook


Fred Kang May 10, 2002 12:45

Re: Fluidized Bed With Gas-Air Combustion
 
Thank you, Jinwook. My case is even worse--the mass flow rate of discrete pahse 20 time more than gas'. I doubt that DPM is a right option for this case, then will Eulerian model be the right one?

Jin-Wook LEE May 12, 2002 21:26

Re: Fluidized Bed With Gas-Air Combustion
 
As my personal point of view, Eulerian model would be the better approach. But, unfortunately, Fluent can not accept chemical reaction in Eulerian-Eulerian model. My acquaintance is doing fluidized bed simulation with chemical reaction by E-E approach by his own developed code.

DPM is difficulty to obtain the converged solution with high mass flow rate of dispersed phase. But it does not mean that 'it is impossible'.

So, I can not give you conclusion but I hope that the above information would be helpful.

Sincerely, Jinwook


Fred Kang May 12, 2002 23:10

Re: Fluidized Bed With Gas-Air Combustion
 
Jinwook, thanks for having given me a hand. I will start to try and will be back on this topic later. Fred

stevenshi11 August 21, 2011 15:14

Quote:

Originally Posted by Fred Kang
;100862
Jinwook, thanks for having given me a hand. I will start to try and will be back on this topic later. Fred

What about now? If it can be simulated in Fluent 13.0? Thanks for ur help!


All times are GMT -4. The time now is 03:44.