[Jegan]

Hello Everyone,
Kerosene-air combustion is my present work, just I wanna capture radical emission intensity from the flame . mixture consist of kerosene-0.03 and air-0.97 lean combustion. this was data of an experimental work im trying to simulate this and that title was partially premixed ...so now im trying with 21 step reaction with 14 species in partially premixed combustion but I couldn't get the appropriate results ,then I tried with species transport that results was considerable and results seems somewhat good with experimental results....now is there any problem on getting results with species transport than partially premixed combustion model...please explain what is the exact difference between these two..?

[LuckyTran]

You need to provide more details.

The species transport model is exactly that, a species transport model. By itself, it does not have any reactions or combustion.

The partially premixed combustion model is also a species transport model with a combustion model (the partially premixed combustion model).

In the partially premixed combustion model you are transporting a mixture fraction. In the species transport model you (normally) solve a transport equation for each species and give some detailed reaction model. But again, you need to provide these details. Because you can also transport a single species that represents the mixture and this would be like one of the combustion models.

[Jegan]

Thank you so much for your kind reply.

In experimental case they investigated the combustion phenomenon of droplet spray obtained by atomized kerosene, a representative hydrocarbon liquid fuel, into the atmosphere through a slit nozzle.
Now, I want to simulate this so I created a simple domain with slit area, that slit area as velocity inlet and gave mixture as C12H23(vapour)-0.03 and O2-0.21and bulk species N2. and outlet as pressure outlet.
And also I patched small area near the slit nozzle inlet with high temperature.
Now the problem is …when I use species transport included volumetric(26 step reaction) and eddy dissipation concept ,the results seems somewhat considerable. But when I was trying in partially premixed ,steady diffusion flamelet the flame structure and other radical capture completely different from experimental data...
Please help me with some ideas..

[LuckyTran]

So you imported the same chemkin file for both cases right? Well the answer is there are a ton of differences without



With species transport, you are solving 13 (14-1) species transport equations, one for each species in the mechanism. At the same time, you are brute forcing the system of ode's in the chemkin mechanism (the 24 reactions). This is the highest fidelity approach with minimum modeling except there is a turbulence-chemistry interaction, which is still quite significant. But in a laminar flow this would be equivalent to a DNS if you used this procedure. But this way is really expensive to calculate. But if you can brute force it, why not?



With the partially premixed combustion model, you lump everything together into a fuzzy mixture and treat it with a mixture fraction. The mixture faction is the volume fraction of "fuel" in each cell. Only this mixture fraction gets transported. You do not solve for the individual concentrations of anything, just this funky mixture. Here there is a combustion model which describes the flame separating the burnt/unburnt regions and this shows up as a progress variable. And then there is yet another model that decides how you calculate the properties of each cell based on how much fuel and how burnt it is. This is a very crude way, but that's the model.



The steady diffusion flamelet model is very close to being a chemical equilibrium model. If your flow is not at chemical equilibrium, then it will depart drastically from what you get when you brute force the chemistry. Are you sure the steady diffusion flamelet and a partially premixed combustion mode is the right choice for your problem? Is your case non-remixed or pre-mixed? The steady diffusion flamelet is more applicable to non-premixed combustion and FGM better for premixed cases. The partially premixed model combines both premixed and non-premixed models and gives you a lot of flexibility, but this does not mean that it is forgiving of poor decisions.

[Jegan]

Thank you so much for a clear explanation.

In experiment initially some volume of kerosene at bottom of chamber was stimulated with certain frequency using ultrasonic transducer, then it forms droplet cloud,meanwhile carrier gas (air) coming at certain velocity from side wall (perpendicular direction)gets mixed in near molecular level ,when carrier gas velocity reaches 2.2m/s blow out occur from slit jet nozzle into atmosphere, at that time that mixture contains3% weight kerosene and remaining is air and they ignited near the slit.

Now I created a domain with slit as velocity inlet and imported chemkin mechanism and patched a small area with 1000k temperature. I tried partially premixed approach with steady diffusion flamelet relation .but I couldn't get good results.

[ruchit@15847]

@Jegan Where did u find CHEMKIN mechanism file for Kerosene-air mixture reaction?? Can u please tell me. I want it. Thanks in advance.

[Jegan]

Cerfacs - U can find detailed mechanism

https://pubs.acs.org/doi/10.1021/acs...yfuels.6b01664 - and try this one.

[ruchit@15847]

Thank you so much. I will try.