Fuel flow rate for jet engines incorrect

[OhioBob]

Please be advised that the fuel flow rates for jet engines are incorrectly computed. Below is an example of how the fuel flow rates are currently calculated. This example is for the Turbojet engine.

Thrust = 225,000 N

Specific impulse = 800 s (sl)

Air/fuel mixture ratio = 15

Mass flow rate = 225000 / (800 * 9.80665) = 28.6795 kg/s

Volumetric flow rate = 28.6795 / 5 = 5.7359 l/s

Air flow rate = 5.7359 x 15/16 = 5.3774 l/s

Fuel flow rate = 5.7359 x 1/16 = 0.3585 l/s

Both the game and KSP Wiki article shows the fuel flow rate as 0.358 l/s, agreeing with the above. Note that the above computation assumes that the specific impulse is based on the total mass flow (air + fuel).

The specific impulse of jet engines when computed using total mass flow rate are lower than rocket engines. They must be lower because they operate at lower pressures, lower temperatures, and expel gases with a higher average molecular weight. The specific impulse for a jet engine might be about 100-200 seconds (depending on operating conditions) versus about 250-350 seconds for a rocket engine.

The specific impulses for jet engines are so high because they are calculated based on the fuel mass only (i.e. the part that's carried aboard the aircraft). The air that comes from the atmosphere is not included when computing jet engine specific impulse. With this in mind, the fuel flow rate in the preceding example should be the entire 5.7359 l/s. The air flow rate would be 15 times this amount, or 86.0385 l/s.

Using these corrected flow rates, the specific impulse of the Turbojet engine works out to be 800 s when based on fuel flow only, but only 50 s when based on total mass flow. This 50 s number seems unrealistically low.

I recommend that the specific impulses for the jet engines be increased to somewhere in the 1600 to 3200 s range, which would result in total mass flow specific impulses of 100 to 200 s. Of course the fuel flow rates would have to be correctly computed. If, for example, the Turbojet specific impulse were changed to 2000 seconds, then fuel flow rate would be,

Fuel mass flow rate = 225000 / (2000 * 9.80665) = 11.4718 kg/s

Fuel volumetric flow rate = 11.4718 / 5 = 2.2944 l/s

I hope you take this recommendation under advisement and modify KSP accordingly.

[OhioBob]

Bumping this thread in the hope that Squad addresses this issue before the release of version 1.0.

[FleshJeb]

Excellent job finding it, but it's a long-term known issue. I couldn't find any reference on the bugtracker, but I've seen it in plenty of spaceplane threads..

[Claw]

Yes, this is a known issue. Keep your fingers crossed!

Cheers,

~Claw

[Renegrade]

Mass flow rate = 225000 / (800 * 9.80665) = 28.6795 kg/s

What? The correct value for g₀?? Isn't that against the rules??

If I weren't so lazy, I'd rewrite the isp curves for the jet engines with an MM config using these values to compensate for the air mass....

[FleshJeb]

What? The correct value for g₀?? Isn't that against the rules??

That's not the correct value in this universe ;-)

http://wiki.kerbalspaceprogram.com/wiki/Specific_impulse#Conversion_factor

[Renegrade]

That's not the correct value in this universe ;-)

Yeah, yeah, I know. I'm rather regex-y in my opinion of the 9.82 fiasco, FYI~

(that article also uses a rather approximate standard gravity too, OhioBob's value is the exact and correct one. While I don't mind it if people type '9.81' into their calculators for rough calculations, any semi- or fully-permanent scientific documentation or code should use the full, real value for g₀. It's not the 70s anymore, there is such a thing as (nearly)universal copy and paste now)

[NathanKell]

RF best F.

(Amongst other things, it force-corrects g0 to 9.80665f )

I also am keeping my fingers crossed on this one. And, Renegade, the MM code is not that bad, you can use regular expressions to get the Isps involved (shorn of their x values), divide, then re-add the x-value (0 or 1).

You can even use the sum of the LF and Air ratios as the divisor.

[Laie]

Using these corrected flow rates, the specific impulse of the Turbojet engine works out to be 800 s when based on fuel flow only, but only 50 s when based on total mass flow. This 50 s number seems unrealistically low.

(emphasis mine)

Yes, this is a known issue. Keep your fingers crossed!

Sure this is the same old well-known issue? OhioBob brought the interesting twist that the resulting ISP is too low rather than too high. Getting a somewhat realistic-like ISP without making it totally OP will be fun.

Adding my own to cents: clean combustion requires about 18 weight of air per 1 weight of fuel. I suggest that basic jets could require and use much more intake_air per unit of fuel (say 30:1), thus raising the basic jet's fuel-relative ISP, while Turbos could use air at something like a 8:1 ratio (less incentive for intake spam, more consumption across the board).

I also suggest that Turbos be re-designated to something that has "Ram" in it's name.

Edited March 1, 2015 by Laie

[Renegrade]

RF best F.

(Amongst other things, it force-corrects g0 to 9.80665f )

RF? RealFuels? It certainly gains high marks from me (tech advancement and thrust correction are big plusses for me, plus I tend to fiddle with the files so I can add batteries to fuel tanks for part reduction lulz ), I'm just a little nervous about resource propagation (I generally only use kero-lox, nto/mmh, and hydro-lox), and my last install (very recently actually)kinda hit some sort of weirdness with capsules and TAC_LS (doubling of tanks and vanishing of buttons)... er, I think it was TAC_LS which wasn't agreeing with it. It may have been a different mod.. Procedural parts perhaps?

(TAC_LS also represents resource proliferation and I'd be terrified to put it, RF, MKS/OKS and Interstellar all on the same install for fear of the resource panel becoming sentient and destroying the real universe )

And, Renegade, the MM code is not that bad, you can use regular expressions to get the Isps involved (shorn of their x values), divide, then re-add the x-value (0 or 1).

You can even use the sum of the LF and Air ratios as the divisor.

Hmm, good point. Maybe I will take a crack at that (and yeah, MM has really advanced quite a bit).

[OhioBob]

Excellent job finding it, but it's a long-term known issue. I couldn't find any reference on the bugtracker, but I've seen it in plenty of spaceplane threads..

Thanks for letting me know. I searched for any mention of this issue but couldn't find it. Sorry for being repetitive. Hopefully Squad will fix it.

[NathanKell]

Sorry, I brushed over the OP's calculations. Let me try to address them now.

First, the 9.82 thing was mentioned, so let's fix that:

225000 / (800 * 9.82) = 28.64053

Volumetric flow rate is undefined however, since we don't know the volume of a KSP fuel unit. All we know is the mass ratio, which since air and fuel have the same "density" (aka tons per KSP unit) will be the same as the volumetric ratio. So we know there's 1.79kg/sec mass flow of fuel.

(We know in real units volumetric flow is 1.79 / 0.81 = 2.21 liters/sec flow rate for fuel, and if we assume based on various factors discussed elsewhere that 1 KSP unit is five liters, that means we should expect 1.79/.81/5 = 0.442 units/sec of Kerosene, or 0.358 units/sec of 1kg/liter LiquidFuel.)

That said, OP's conclusion--that like in real life jets should only count fuel flow towards specific impulse, and that specific impulse should then be raised to realistic levels--I agree with entirely. So Laie, I don't think there's anything different from the original way this was phrased, just pointing out that merely correcting the "air calculated towards Isp" bug would make jets not efficient enough, and thus once that is done jet Isp should be raised. Because (800 * 16 = ) 12,800s is not "too little" by any stretch of the imagination.

OhioBob: as above, while the "air counted for Isp" bug has been discussed before, few have gone into the calculations as much as you, nor emphasized the need to increase Isp once that issue is fixed. I salute you!

Edited March 2, 2015 by NathanKell

[NoMrBond]

Parhaps forward to Ted for consideration during the current Q&A pass they're undertaking in conjunction with the aero overhaul?

[FleshJeb]

Thanks for letting me know. I searched for any mention of this issue but couldn't find it. Sorry for being repetitive. Hopefully Squad will fix it.

Sure thing. Sorry if I appeared terse...I was typing in a hurry. I wish I could have found a reference for you.

Just because it's not a new issue doesn't mean a lot of people aren't aware of it, or that it isn't worth discussing further. Thank you for rekindling it so thoroughly.

- - - Updated - - -

Yeah, yeah, I know. I'm rather regex-y [LOL] in my opinion of the 9.82 fiasco

Is there a story behind the 9.82 thing? I was only aware it was an issue, but not the origin of it.