combustion ..............flame not achieving the desired flame velocity ?

  • 153 Views
  • Last Post 31 July 2018
sooraj546 posted this 30 July 2018

I am trying to replicate the simulated combustion of methane through fine tubes(16 mm dia) into a combustion chamber but in my case my flame are not attaining the desired velocity contour compared to the developed model, even though the temperature developed seems ok

i dnt know where i have gone wrong.........

models used 
1. species transport-volumetric- diffusion energy source- -eddy dissipation model
2. air velocity 141.75 m/s temp 566 K
3. fuel velocity 14.19 temp 295 K
4. standard k-epsilon with std wall function
6. Radiation- discrete ordinate model.

the velocity which i am supposed to get are as shown in 1st figure and what i am getting is shown in last figure both are on same scale for easy comparison ?

Order By: Standard | Newest | Votes
abenhadj posted this 30 July 2018

Hi,

It looks like the Fluent simulation is either non well converged or the combustion has newly started and you still run for more iterations. Is in the model you are using to compare is the Eddy Dissipation model used there?

Best regards, Amine

  • Liked by
  • sooraj546
sooraj546 posted this 30 July 2018

A study has already been developed using Eddy dissipation model i am just trying to replicate it ..........but i am facing the above mentioned problem. I have already run the problem for 17000 iteration, moreover i have selected 3 points in the combustion chamber at different heights that too has converged with less than 1 % variation.

abenhadj posted this 30 July 2018

If in the other code the Eddy Dissipation Model is used then please for the coefficients accompanying that model ( A and B ). Moreover, are you seeing the convection rolls as in the model you are using for comparison? Are you replicating from Fluent to Fluent or from another code. If from another code you need to list all physical models used there. Please check for heat flux imbalance and mass flow imbalance in the Flux report.

Best regards, Amine

  • Liked by
  • sooraj546
sooraj546 posted this 30 July 2018

I am replicating from fluent to fluent...................A=4 and B= 0.5 as specified in paper.......https://link.springer.com/article/10.1007/s11663-013-0021-8  this is the paper sir. They have almost given everything.......all the inputs needed etc

i am able to replicate temperature rest all some what similar..................except velocity

vganore posted this 30 July 2018

Hi Sooraj, we are getting an internal flag on your post due to the username. You have added an email address in place of username. Could you please change that by going through edit profile? The username could be anything other than email address.

Thanks in Advance!

Vishal Ganore, ansys.com/student

  • Liked by
  • sooraj546
abenhadj posted this 30 July 2018

I cannot access the paper. You need to identify why you are not getting similar results and perhaps contact the authors of the paper. Instead of replicating it you can do it better by applying more sensitive wall treatment, adding some finite rate kinetics, relaxing to chemical equilibrium, using a turbulence model which can deal with swirl effects and streamline curvature, like realizable k-epsilon or omega based models and at last but not least high order numerics and grid sensitivity analysis.

Best regards, Amine

  • Liked by
  • sooraj546
sooraj546 posted this 30 July 2018

RWOOLHOU: Images removed for copyright reasons. 

 

 

Sir u r absolutely rt i can do many thing ...................but before all that first of all i should validate this model. I have attached the images of the paper. Thanks a lot for replying

Y

sooraj546 posted this 30 July 2018

I have changed it sir

abenhadj posted this 30 July 2018

You need to be aware that the paper is IP protected. Moreover, the pictures are compressed so that it is hard to identify anything.

I have read in the abstract that they are using UDF's to model pellet's layer under the hearth by porous zone with source terms for energy to mimic the depletion of the pellets.

Best regards, Amine

  • Liked by
  • sooraj546
sooraj546 posted this 31 July 2018

UDF are nothing sir just a heat source and a CO source .It doesn't have that much effect on combustion flame velocity profile which i am getting. Thanks a lot for ur reply sir!

My air velocity 141.75 m/s temp 566 K and  fuel velocity 14.19 temp 295 K, still my combustion flame velocity profile is less than 60 m/s..............I simply cant understand why this is happening sir. 

 

 

abenhadj posted this 31 July 2018

Hi,

As I do not have the paper the only two things which I recommend:

 

1/Check your setup ensure for good convergence and run it for a longer while

2/Contact the authors of the paper: perhaps they added something not published into their model.

 

 

Best regards, Amine

sooraj546 posted this 31 July 2018

I will sir thanks a lot....... i really appreciate taking this much time to read my post and reply to each of my comments.I have uploaded my residual and another temperature plot (The 2 nd figure below shows the variation of temperature with iterations at 3 different heights inside the combustion chamber. Ie. 0.020 from bottom, 0.1 from bottom and 0.95 from bottom etc)

 

rwoolhou posted this 31 July 2018

Hi,

I'll delete the paper images shortly to avoid any copyright issues: please don't include images from behind password barriers or paywalls on here. 

Looking at your residuals I think you've got a transient in the solution, and this means it's not well converged. Given the swirl & combustion I'd also expect you to need a good quality & well refined hex or poly mesh. Have you checked the documentation for recommended settings for the combustion models, and that all of the boundaries tie up with the paper (including inlet pressures)? I'd also advise increasing the DO discretisation to 3 by 3 as a minimum (DO panel). Have you set the gas absorption coefficients & other properties to be the same as in the paper? 

As ANSYS staff we're unable to take the case and check it over. Once you've checked the above and run on for much longer I'd recommend talking to your supervisor or department CFD expert for their input if the wider community aren't able to assist further. 

Best wishes,
Rob

 

 

 

 

 

 

  • Liked by
  • sooraj546
sooraj546 posted this 31 July 2018

Thanks a lot for replying,

I will go for a refined hex or ploy mesh and compare my results and will let you know

Gas absorption coefficient are domain based as given in paper

Close