[Ryan T]

So I have a combustion chamber I want to simulate where air is mixed with hydrogen gas at the inlet of the combustion chamber. The incoming Mach number is 0.235 while the outgoing Mach at the exit is 0.435. The fuel injection is sonic so Mach 1.

So aside from the fuel injection, the exit is at 0.435 and the fuel injection is Mach 1, because I have 2 conditions above Mach 0.3, does this mean I need to run the density based solver? I'm using Species Transport for fuel injection since it isn't multiphase, but I'm not sure how to adapt that for the density based solver.

When I run the pressure based solver my residuals seem to go down while my continuity residual plateaus at 10^-2

[LoGaL]

No, you don't need to use density based solver. Until Mach = 2 (take with a bit of caution the number) pressure based solver is perfectly fine

[Ryan T]

Originally Posted by LoGaL

No, you don't need to use density based solver. Until Mach = 2 (take with a bit of caution the number) pressure based solver is perfectly fine

What could be the reason for my continuity plateauing like that?

[LoGaL]

are you imposing a sonic inlet or smth?

[Ryan T]

Originally Posted by LoGaL

are you imposing a sonic inlet or smth?

Yeah, I have an inlet set as sonic which is where the hydrogen would be injected.

EDIT: Would this mean I'd have to go wit ha density based solver or is there something else causing this?

[LoGaL]

To be fair, I have no idea of what Is causing this because I can’t see your setup

My guess is that you are setting a sonic inlet for the fuel and this is bad because:
1) you are setting the inlet very near to where interesting stuff happens: your fuel flow is undeveloped, it usually requires some numerical length to transition from flat profile where your BC is imposed to a nice parabolic profile
1) you are setting a sonic inlet

My suggestion is to extend the flow inlet, first with a straight section (where you will attain Mach 1) ,then with a converging one ( to accelerate the flow) then a straight again, at the end of which you will assign a nice, clean mass flow inlet. If the flow is chocked, you will attain Mach 1 in the throat which is also where your virtual duct will dump stuff in the chamber. This time you will however inject a nice, developed flow while still being physically consistent

[Ryan T]

Originally Posted by LoGaL

To be fair, I have no idea of what Is causing this because I can’t see your setup

My guess is that you are setting a sonic inlet for the fuel and this is bad because:
1) you are setting the inlet very near to where interesting stuff happens: your fuel flow is undeveloped, it usually requires some numerical length to transition from flat profile where your BC is imposed to a nice parabolic profile
1) you are setting a sonic inlet

My suggestion is to extend the flow inlet, first with a straight section (where you will attain Mach 1) ,then with a converging one ( to accelerate the flow) then a straight again, at the end of which you will assign a nice, clean mass flow inlet. If the flow is chocked, you will attain Mach 1 in the throat which is also where your virtual duct will dump stuff in the chamber. This time you will however inject a nice, developed flow while still being physically consistent

I attached some pictures of the setup and contours I've gotten. The mesh itself is just a simple square mesh that I updated the OP with since I'm at the limit for attachments for this post.

By extending, do you mean to extend the length of the combustion chamber? This is actually part of a ramjet in Mach 2.6 freestream flow, except this current simulation is just to see if the analytical values we get at the end of the combustion chamber match with the CFD.

[LoGaL]

no, you need to extend the fuel nozzle: right now it's just a tiny edge (i assume it's 2D simulation). It does not matter that it is the real fuel nozzle, just make sure you have straight pipe --> converging section --> throat --> chamber

Same goes for the hydrogen: put those boundaries away from the chamber. Give the flow some space to develop and avoid using the sonic BC. Check the static pressure distribution: why the heck do you have 500 Pa near the injector exit? It does not make any sense to me. Speaking of, what is the operating pressure?

[Ryan T]

Originally Posted by LoGaL

no, you need to extend the fuel nozzle: right now it's just a tiny edge (i assume it's 2D simulation). It does not matter that it is the real fuel nozzle, just make sure you have straight pipe --> converging section --> throat --> chamber

Same goes for the hydrogen: put those boundaries away from the chamber. Give the flow some space to develop and avoid using the sonic BC. Check the static pressure distribution: why the heck do you have 500 Pa near the injector exit? It does not make any sense to me. Speaking of, what is the operating pressure?

Okay, so I need to increase the injector size and doing a full simulation of the ramjet may be unavoidable then, the operating pressure is 101325 Pa.

[LoGaL]

Nono, i am not saying to do a full simulation of the ramjet. You can make a fake extension with simple pipes. For the hydrogen inlet straight ones, for the fuel inlet converging duct. The do not have to resemble reality, they just need to exist for purely numerical reasons.

[Ryan T]

Originally Posted by LoGaL

Nono, i am not saying to do a full simulation of the ramjet. You can make a fake extension with simple pipes. For the hydrogen inlet straight ones, for the fuel inlet converging duct. The do not have to resemble reality, they just need to exist for purely numerical reasons.

So you're saying to try adding a fake "pipe" of sorts for the sonic fuel flow?

[LoGaL]

exactly, instead of sonic inlet, you add a fake pipe, whose section is
straight (make this one quite long) --> converging --> straight (make it enough long)--> enters the chamber. Then I want you to assign a standard mass flow inlet

[Ryan T]

Originally Posted by LoGaL

exactly, instead of sonic inlet, you add a fake pipe, whose section is
straight (make this one quite long) --> converging --> straight (make it enough long)--> enters the chamber. Then I want you to assign a standard mass flow inlet

I see, thanks for the info! I'd still have to do a species transport for the gaseous hydrogen. And the converging pipe section would be for the sonic flow? How long relative to the geometry I posted do you think it should be?

[LoGaL]

Yes, the converging section makes so that you attain sonic flow at the exit of the fake nozzle. Concerning the dimensions, I will just throw some numbers: if your fuel nozzle Diameter is D
Throat section length 3-4 Diameters
Converging section length 3-4 Diameters
Straight section length 6-7 diameters
Those are really out of blue numbers but they should work. Make sure the fake pipe is smooth. No sharp corners
Ah, also, don’t be surprised if the flame structure changes a bit, it has to. It will for sure be more realistic than before

[Ryan T]

Originally Posted by LoGaL

Yes, the converging section makes so that you attain sonic flow at the exit of the fake nozzle. Concerning the dimensions, I will just throw some numbers: if your fuel nozzle Diameter is D
Throat section length 3-4 Diameters
Converging section length 3-4 Diameters
Straight section length 6-7 diameters
Those are really out of blue numbers but they should work. Make sure the fake pipe is smooth. No sharp corners
Ah, also, don’t be surprised if the flame structure changes a bit, it has to. It will for sure be more realistic than before

Huh, issue is that it looks like the combustion is supposed to take place at the start of the combustion chamber and not the end with the combustion chamber being the middle square between the ramps in the picture I attached since making the length of the throat straight section 3/4 diameter would move it towards the end.

[LoGaL]

By throat I mean the throat of the fake fuel nozzle, not the one of the engine. I will try to make a sketch later

[Ryan T]

Originally Posted by LoGaL

By throat I mean the throat of the fake fuel nozzle, not the one of the engine. I will try to make a sketch later

Do you mean something like this?

EDIT: I get some interesting contour results, now it says it "converged" but it converges with continuity plateaued at 10^-3

[LoGaL]

Check the picture i posted xD, I think you didn't get why I meant. I of course exaggerated the nozzle size, but this is the idea.

And by the way, disable the convergence control in fluent, you stop the simulation when you decide, not when fluent decides. What you posted above (even though it was not what I meant) is definitely not converged, the simulation stopped when the residuals where about to drop further. You can also see it with the contours: symmetric domain requires symmetric solution if it converges.

[Ryan T]

Originally Posted by LoGaL

Check the picture i posted xD, I think you didn't get why I meant. I of course exaggerated the nozzle size, but this is the idea.

And by the way, disable the convergence control in fluent, you stop the simulation when you decide, not when fluent decides. What you posted above (even though it was not what I meant) is definitely not converged, the simulation stopped when the residuals where about to drop further. You can also see it with the contours: symmetric domain requires symmetric solution if it converges.

Got it. Thanks for the info. How would I simulate the air inlet? Just treat the top and bottom boundaries above the sonic nozzle as mass flow inlets for air?

[LoGaL]

Yes, Mass flow inlets also for air. They are the best. We don't like fancy boundary conditions.

Best way would be to have some fake space for the flow to develop also in that case. With your current domain however, i am not sure if it is possible.


Other thing you may look into is how to impose profiles and somehow find a way to impose a developed flow profile for the air inlet. This would avoid the fake extensions. It would require a 2D adaptation of this idea https://www.youtube.com/watch?v=opqZbOR_gNM.