[kuleke]

Hello,

I am trying to run a combustion simulation for kerosene in the combustion can of a small turbojet engine. The main problem I've found is that there doesn't seem to be any combustion taking place. I have used a few tutorials and read a manual or two but it seems to me like the fuel isn't actually burning. In fact, it seems like nothing is happening (temperature field not changing). I have attached a screenshot of the can mesh and any help would be appreciated.

Thank you

[kuleke]

Bump. Still no joy :/

[diamondx]

What model are you using for combustion ? transport species ? DPM ?

[sescobar]

Is your simulation steady state or transient? Did you make sure you are solving for species transport? Did you enable volumetric reactions? If so what turbulence-chemistry interaction are you using?
Additionally, in order to initiate the reactions you need to.precribe an initiation patch with high temperature and appropiate species concentration.

[ansys_matt]

I am running transient combustion models with a spark ignition. How are you igniting the mixture?

[sescobar]

I first mark a region of the domain where I expect the flame to be. Then I patch that region with a high temperature value and species mass fraction corresponding to the combustion products.
The value for temperature and mass fractions can be based on adiabatic flame calculations or chemical equilibrium calculations.

[kuleke]

Hi everyone.

Thank you for all your replies and sorry I have taken so long to respond. I came across this issue previously but I think FLUENT will burn the mixture of air and fuel as soon as they come into contact in my model. I'm not very good at the combustion side of it in general but when I am setting up non-premixed combustion, I am asked to generate a pdf table - I think this constitutes part of the solution system for combustion because it asks for a new table every time I make a change (adiabatic setting, species mass fractions, etc).

The reason nothing was burning previously was something in the way the boundary conditions were set up, I've now changed that around and I ran it a few times, playing around with URFs and solution methods until it wasn't reporting any errors but not converging either. I looked at the solutions and the temperature fields look like combustion is in fact taking place.

I understand that there might be transient effects at work here - I am running a steady state simulation as far as I know. I'm gonna attach a shot of one of the residual plots to give you an idea of what I'm saying.

[kuleke]

Some more information on the model

[macfly]

According to Fluent Users Guide 15.0, section 16.2.5: "An RFL (rich flamability limit) value of approximately twice the stoichiometric mixture fraction is appropriate." Did you verify your value?

I use low URFs in order to obtain convergence for combustion:
pressure, density, momentum => 0.2
k, epsilon, turb. viscosity => 0.5
temperature, mixture fraction/variance => 0.9
Start with low URFs then increase them.

[kuleke]

Originally Posted by macfly

According to Fluent Users Guide 15.0, section 16.2.5: "An RFL value of approximately twice the stoichiometric mixture fraction is appropriate."

Did you verify that?

I did some calculations around that somewhere, but I will double check thanks.

Been playing around with solution schemes and URFs today and its getting closer to the convergence criteria. Any other tips from anyone?

[sescobar]

Do not limit yourself to check the residual in order to determine convergence. Make sure that the total mass and energy balance is appropriate. Additionally I would place monitor probes in the flame region to verify that the velocity, temperature and main specie do not change significantly with the iterations.

[kuleke]

I have looked at the results from previous simulations where it didn't converge and it looks 'correct' (as expected). The temperature and flow fields look correct, I think my issue is now with trying to get more accurate results.