How can I model the regression of a combusting solid fuel? Any advice gladly recieved.

In Fluent this is not easy out of the box.

Assumming you're happy with a steady state solution, the easiest way is probably to use Fluent 5.x and write a udf to calculate the consumption of a gas species at a wall due to the hetrogeneous reaction there. You would need to write several source udfs - one for each gas phase particpating in the reaction and one for the energy release/consumption. You could apply these udfs at a wall - where your solid is, or maybe even at a porous boundary so that you might also have flow across the solid interface - given that solid is really a porous medium during reaction.

From the reaction rates you can determine the solid consumtipn rate and hence steady-state interface recession.

If you want to model transient behaviour, the evolution of the solid over time, things get more complicated. Its possible to do this in both 4.5 and 5.x if you assume a generalised reacting porous medium and are prepared to write a fair amount of code. I have developed codes for this as part of my research and its been a fairly large headache. If you go this route expect to spend several or more months writing code (in fact I've spent about 12months looking at various models in 4.5 and 5.x!) and integrating it into fluent. Its probably easiest to do in Fluent 4.5, since the multiphase capability already exits, but its not straightforward and of course then you are stuck with a structured grid and poor parallel performance. On the other hand, in Fluent 5.x which is what I now use, you need to solve lots of additional scalar transport eqns to model the solid phase - and while the udf documentation is quite good, it still takes a long time to get things working . . .

So there are the options - I think this is a very challenging area at the moment. And very little work has been done in 2d and 3d with these types of models.

Interestingly, in 2d/3d models you can capture the 'fingering' phenomena at the reaction front - where the chemical reaction has a self-enforcing role - something akin to 'fingering' in immisible viscous fluids. This is a phenomena which you can't model in 1D - and for which very few numerical models/results have been published.

Regards

Greg

Greg,

Thanks for the info. Yes, I am happy with a steady state solution and have read up a bit on using UDF's. I saw your UDF in the forum (udf_scalar_dos.c) and wondered how I could modify it for the combustion of PMMA with oxygen? You state in the code that it can be used for 2d/3d. It may be a silly question but does that include 2d axisymmetric-swirl?

Peter

Without looking at the code, probably the only thing you need to modify is any reference to C_VOLUME macros - they should be C_VOLUME_2D for axi-symmetric problems - otherwise should be ok.

That's a really early version of some macros I've developed though - and its for a two phase Eulerian-Eulerian type model. It might be suitable - in this case I think you'll need to add some reaction rate calculations and species sources.

Greg

Could you tell me how your combustiom happen?

Hi Greg,

Thank for your information. I totally agree with you regarding this aspect of modelling - this has always been a challenging one to me as well...

Peter Moss mentioned in this forum about some of the macros you developed for this type of problem. I have been looking for it all over the forum and try to see if this can help me solve some of my present problem in modelling my underfeed stoker.

I hope you wont mind if you can tell me what topic in this forum where you posted that UDF.

Regards, Stanley

[kookguy]

Just wondering since this thread had been so long, are there any development for 3D solid fuel regression rate recently? ( especially in Fluent)

I just started my thesis on the effect of swirl flow (in hybrid rocket engines), and I am very new in this research area.

Thank you very much.

[svk16]

Hello Greg
I'm trying to model solid fuel combustion in CFX as part my of BE Project & there is very less information about this out there.
Can you tell me more about writing UDF or any trusted book/website where I can find the Relative material ?
I'm using ANSYS 16.2
Thanks in Advance

Originally Posted by Greg Perkins
;96061

In Fluent this is not easy out of the box.

Assumming you're happy with a steady state solution, the easiest way is probably to use Fluent 5.x and write a udf to calculate the consumption of a gas species at a wall due to the hetrogeneous reaction there. You would need to write several source udfs - one for each gas phase particpating in the reaction and one for the energy release/consumption. You could apply these udfs at a wall - where your solid is, or maybe even at a porous boundary so that you might also have flow across the solid interface - given that solid is really a porous medium during reaction.

From the reaction rates you can determine the solid consumtipn rate and hence steady-state interface recession.

If you want to model transient behaviour, the evolution of the solid over time, things get more complicated. Its possible to do this in both 4.5 and 5.x if you assume a generalised reacting porous medium and are prepared to write a fair amount of code. I have developed codes for this as part of my research and its been a fairly large headache. If you go this route expect to spend several or more months writing code (in fact I've spent about 12months looking at various models in 4.5 and 5.x!) and integrating it into fluent. Its probably easiest to do in Fluent 4.5, since the multiphase capability already exits, but its not straightforward and of course then you are stuck with a structured grid and poor parallel performance. On the other hand, in Fluent 5.x which is what I now use, you need to solve lots of additional scalar transport eqns to model the solid phase - and while the udf documentation is quite good, it still takes a long time to get things working . . .

So there are the options - I think this is a very challenging area at the moment. And very little work has been done in 2d and 3d with these types of models.

Interestingly, in 2d/3d models you can capture the 'fingering' phenomena at the reaction front - where the chemical reaction has a self-enforcing role - something akin to 'fingering' in immisible viscous fluids. This is a phenomena which you can't model in 1D - and for which very few numerical models/results have been published.

Regards

Greg

[samiye]

How can I model non-regression assumption of a combusting solid fuel? Any advice gladly received

[Hrtz]

Hi,
I am trying to model solid fuel combustion during my research.
Can someone advise how it can be done in Fluent? Is there any new models for that in new versions of Fluent?
I thought to consider the fuel surface as a wall BC and add to it sources of species & momentum, and sink for the heat of pyrolysis, but i am not sure how to define it within Fluent.
Thanks.

[472631447@qq.com]

Where can I see the code.Is your problem solved now?

[472631447@qq.com]

Originally Posted by Peter Moss
;96065

Greg,

Thanks for the info. Yes, I am happy with a steady state solution and have read up a bit on using UDF's. I saw your UDF in the forum (udf_scalar_dos.c) and wondered how I could modify it for the combustion of PMMA with oxygen? You state in the code that it can be used for 2d/3d. It may be a silly question but does that include 2d axisymmetric-swirl?

Peter

Where can I see the code.Is your problem solved now?