Fuel Tank Mass: Kerbal Space Program vs Science

[TChapman500]

I just did a few calculations on the Jumbo 64 tank and found that the tank (without internal supporting structures) should have a mass of only 371 kilograms. Yet the tank with the internal supporting structures and the separate fuel and oxidizer cells and their supporting structure and the feed lines going to either end of the tank, somehow gives the tank a total mass of 4,000 kilograms.

Tank Radius: 1.25 meters

Tank Height: 7.50 meters

Tank Material: Aluminum (Assumed)

Tank Thickness: 0.002 meters (Assumed)

Tank Volume: ~36,815 Liters

Volume For Fuel: 6,400 Liters (17.4 %) [Note: Using realistic densities without reducing the mass would make this figure about 80%.]

This fuel tank has only 1.4% of the volume of the external tank on the space shuttle (if that tank was cylindrical instead of tapered at the ends). 371 kilograms would put the mass of the Jumbo 64 tank at 1.4% of the mass of the space shuttle's external tank. But come to think of it, perhaps 2 millimeters is too thick. If we go with 1 millimeter, that might give us about 371 kilograms of mass with all of the internal components in place (minus the fuel). But my research suggests that the mass using solid aluminum would still be much higher than 371 kilograms (perhaps in the order of 1,000 kilograms?). There are also several repercussions to using a non-alloy such as the metal atoms sliding past each-other under stress. In which case you can forget about adding struts! And you can definitely forget about putting large things (such as landers and Scott Manley's super-heavy satellites) on top of the launchers. The rocket would fall apart in the VAB!

The solution for the Kerbals seems to be "add more aluminum." Kerbal engineers can't seem to make the alloys they need to make light, but strong, rocket components, so they put so much aluminum into the internal supports that you get a 4 tonne tank that should weigh less than 1/10 that much. It also explains why the Kerbals need unrealistically dense fuels. Their vehicles would fall apart in the VAB if they tried using real fuels because they wouldn't be able to put enough aluminum in to make the tanks ridged.

Note: Most of the references here can be found at Wikipedia.

Manufacturer's Note: Warranty of the Jumbo 64 tank is void if anything is attached to the sides of the tank or if anything but decouples, other tanks of similar size or engines are attached to the ends of the tanks. Warranty is also void if an engine is pointed at the tank. Rockomax is not responsible for any rapid, unplanned disassemblies. Warranty void if used by Scott Manley.

[Russoft]

Nevermind... I was going to question the radius of the tank, then realized you got it right.

Edited June 10, 2014 by Russoft
bad at math

[malkuth]

Sorta assuming that a unit of measurement tank wise is = liter.

Pretty sure this is why NathanKell changed these values in realFuels, because had no idea what 1 unit actual represents in real world examples.

[Red Iron Crown]

Despite the Wiki saying otherwise, the fuel unit is a mass (5kg). Otherwise the densities and tank sizes are silly.

[regex]

If you want proper, real-world tank sizes and volumes, get Real Fuels and an engine pack to go along with it. Stock KSP is balanced against a tiny, completely unrealistic solar system and thus will not, in any way, match the real world as far as equipment stats are concerned.

To give you an example, I have an RSS config that scales the Kerbin solar system up to 6.4 times its stock size. Under this solar system, which is still unrealistic and kind of ridiculous (only less so), it takes a good 7km/s to get to orbit using FAR (roughly double stock + FAR) and the first Dres transfer (I'm doing a mission right now, so it's in my mind) takes some 15km/s, including braking (optimal is somewhere around 10km/s), whereas the first stock Dres transfer requires some 4km/s, including braking. In our actual solar system, a transfer to Ceres (or "Dres") takes some 9.5km/s optimal (if my light reading this morning is correct).

Basically its kind of silly trying to rationalize things in KSP scientifically when the solar system isn't really accurate to begin with. That's okay, KSP is a game, not a simulator, and it should be balanced as such.

[Sandworm]

Reality, lol.

KSP is the training ground for RSS. Anyone willing to break out the calculator to criticizes KPS's lack of physics is more than ready for the challenges of RSS.

[weezl]

I just did a few calculations on the Jumbo 64 tank and found that the tank (without internal supporting structures) should have a mass of only 371 kilograms. Yet the tank with the internal supporting structures and the separate fuel and oxidizer cells and their supporting structure and the feed lines going to either end of the tank, somehow gives the tank a total mass of 4,000 kilograms.

Tank Radius: 1.25 meters

Tank Height: 7.50 meters

Tank Material: Aluminum (Assumed)

Tank Thickness: 0.002 meters (Assumed)

Tank Volume: ~36,815 Liters

Volume For Fuel: 6,400 Liters (17.4 %) [Note: Using realistic densities without reducing the mass would make this figure about 80%.]

This fuel tank has only 1.4% of the volume of the external tank on the space shuttle (if that tank was cylindrical instead of tapered at the ends). 371 kilograms would put the mass of the Jumbo 64 tank at 1.4% of the mass of the space shuttle's external tank. But come to think of it, perhaps 2 millimeters is too thick. If we go with 1 millimeter, that might give us about 371 kilograms of mass with all of the internal components in place (minus the fuel). But my research suggests that the mass using solid aluminum would still be much higher than 371 kilograms (perhaps in the order of 1,000 kilograms?). There are also several repercussions to using a non-alloy such as the metal atoms sliding past each-other under stress. In which case you can forget about adding struts! And you can definitely forget about putting large things (such as landers and Scott Manley's super-heavy satellites) on top of the launchers. The rocket would fall apart in the VAB!

The solution for the Kerbals seems to be "add more aluminum." Kerbal engineers can't seem to make the alloys they need to make light, but strong, rocket components, so they put so much aluminum into the internal supports that you get a 4 tonne tank that should weigh less than 1/10 that much. It also explains why the Kerbals need unrealistically dense fuels. Their vehicles would fall apart in the VAB if they tried using real fuels because they wouldn't be able to put enough aluminum in to make the tanks ridged.

Note: Most of the references here can be found at Wikipedia.

Manufacturer's Note: Warranty of the Jumbo 64 tank is void if anything is attached to the sides of the tank or if anything but decouples, other tanks of similar size or engines are attached to the ends of the tanks. Warranty is also void if an engine is pointed at the tank. Rockomax is not responsible for any rapid, unplanned disassemblies. Warranty void if used by Scott Manley.

Maybe I'm stating the obvious here.. but its a game, yes? Games approximate reality in some form.. until they don't. And when they don't they really don't.

Squad is on to something beautiful with KSP.. it approximates reality where it is important.. and departs from reality where it is important.. that is called good game design.. or simply 'designing for fun'. I understand folks that want reality.. I really do.. but I feel ( like many ) that fun would be compromised if we get too caught up in making things mimic reality. If you are just commenting.. fine. But if this is a suggestion for change, I most heartily disagree with it. Right now KSP seems very well balanced between reality and fun.. There are things I'd suggest.. but they are more career oriented features and things focused on the game aspect.. what they've done to approximate physics but still keep things simple is just rocking from my perspective.

For what its worth.

Finally:

Rockomax is not responsible for any rapid, unplanned disassemblies.

LOL!

Edited June 10, 2014 by weezl

[TChapman500]

Yes, I know it's a game. No, I'm not suggesting a change. All this is doing is comparing KSP to real life. I did another thread a little while back where I compared fuel densities. I referenced that thread but didn't provide a link.

[problemecium]

I think most of the disparity between KSP and your data probably results from your assumption regarding tank wall thickness. I've looked at the ends of the Kerbodyne tanks, and while I know those aren't the Jumbo-64, I can presume they're not going to be terribly different. The walls on these appear to be a lot thicker than 2mm - more like 20mm. That'll easily make the tank several times heavier than in your calculation.

Also, given that Kerbin is made out of some super-dense material, it's not unlikely that the tank is made out of something denser than Aluminum.

You may or may not have covered this, but I mention it just to make sure.

[Dkmdlb]

They have to be heavy like that in order to be able to be stacked on top of each other in a modular fashion. They need extra piping and structural elements in order to be able to move fuel between tanks. They need to be over-engineered so that any of a number of engines can just be slapped on the bottom in a way that won't cause the whole thing to splode.

Yes, they are heavier than earthling fuel tanks, but that's because they can do a lot more. Considering the delta-v requirements to get around, I'd say they made the right decision to go that route.

[J4CKAL2014]

If you want proper, real-world tank sizes and volumes, get Real Fuels and an engine pack to go along with it. Stock KSP is balanced against a tiny, completely unrealistic solar system and thus will not, in any way, match the real world as far as equipment stats are concerned.

To give you an example, I have an RSS config that scales the Kerbin solar system up to 6.4 times its stock size. Under this solar system, which is still unrealistic and kind of ridiculous (only less so), it takes a good 7km/s to get to orbit using FAR (roughly double stock + FAR) and the first Dres transfer (I'm doing a mission right now, so it's in my mind) takes some 15km/s, including braking (optimal is somewhere around 10km/s), whereas the first stock Dres transfer requires some 4km/s, including braking. In our actual solar system, a transfer to Ceres (or "Dres") takes some 9.5km/s optimal (if my light reading this morning is correct).

Basically its kind of silly trying to rationalize things in KSP scientifically when the solar system isn't really accurate to begin with. That's okay, KSP is a game, not a simulator, and it should be balanced as such.

I think that, if the Kerbal system was scaled to the same size as our solar system, the stock rockets and fuel tank sizes wouldn't be sufficient for anything other than the simplest of orbital tasks. Even rocket engines like the Griffon XX - the biggest and most powerful rocket engine in the KW Rocketry mod has a thrust of 4,900 kN - has a thrust which is over 6 times less than that of the Saturn V's first stage, so you'd have to make gigantic rockets just to get stuff into orbit. Kerbin's universe seems to be well balanced when compared to the tech you have available. If the rockets available were more powerful (i.e. as powerful as stuff we have IRL) then sure, the Kerbin System would be ridiculously small.

[Sandworm]

It isn't small to accommodate week engines. The week engines were created to accommodate the small planet. The reason for it being small have to do with the limitations of ancient versions of KSP.. Those limitations are long forgotten but, as with many things in KSP, we are stuck. Squad is unwilling to revisit working systems no matter what the community wants. (Hence the rule about discussing planet size changes).

You don't need giant engines to get into realworld-sized orbits. You do need giant engines to lift the massively-overpowered and overweight orbiters seen in KSP. Realworld spacecraft have on-orbit deltaV of perhaps 300m/s and most of that is needed for de-orbiting.

[regex]

I think that, if the Kerbal system was scaled to the same size as our solar system, the stock rockets and fuel tank sizes wouldn't be sufficient for anything other than the simplest of orbital tasks. Even rocket engines like the Griffon XX - the biggest and most powerful rocket engine in the KW Rocketry mod has a thrust of 4,900 kN - has a thrust which is over 6 times less than that of the Saturn V's first stage, so you'd have to make gigantic rockets just to get stuff into orbit. Kerbin's universe seems to be well balanced when compared to the tech you have available. If the rockets available were more powerful (i.e. as powerful as stuff we have IRL) then sure, the Kerbin System would be ridiculously small.

Yes, that's exactly what I said. The parts in KSP are balanced to the smaller solar system scale.

E: Side note - The Saturn V first stage had five F-1 engines and was some ten meters in diameter, so the Griffon XX engines aren't exactly puny by comparison.

[J4CKAL2014]

Yes, that's exactly what I said. The parts in KSP are balanced to the smaller solar system scale.

E: Side note - The Saturn V first stage had five F-1 engines and was some ten meters in diameter, so the Griffon XX engines aren't exactly puny by comparison.

Yeah that is true. That said, the XX is comprised of 4 engines, so the S-I has a lot more power per engine AND more of them

[regex]

Yeah that is true. That said, the XX is comprised of 4 engines, so the S-I has a lot more power per engine AND more of them

The XX looks more like a single engine with four combustion chambers, much like many Russian/Soviet engines. Perhaps the Griffon XX is modeled after the RD-170...

[J4CKAL2014]

You don't need giant engines to get into realworld-sized orbits. You do need giant engines to lift the massively-overpowered and overweight orbiters seen in KSP. Realworld spacecraft have on-orbit deltaV of perhaps 300m/s and most of that is needed for de-orbiting.

Well that's true, I normally try and take more fuel than I'm gonna need. And as for the overweighred orbiters, fair point. I was kinda frustrated that I haven't been able to make a serial-stage-only Mun/Minmus vehicle, but considering the amount of weight they have to get up there I suppose it isn't all that surprising. Guaranteed that people have managed it though, I probably just lack the engineering skill

[J4CKAL2014]

The XX looks more like a single engine with four combustion chambers, much like many Russian/Soviet engines. Perhaps the Griffon XX is modeled after the RD-170...

If it was, I would demand the same fuel efficiency

[weezl]

Yes, I know it's a game. No, I'm not suggesting a change. All this is doing is comparing KSP to real life. I did another thread a little while back where I compared fuel densities. I referenced that thread but didn't provide a link.

Very good.. and as I alluded in the previous post.. good humor!.

[Jesrad]

Mass fraction, people ! It rules over your maximum deltaV per stage.

The external tank for the SLS is a marvel of engineering:

Whereas in KSP, tanks have all the same fixed mass fraction of 8/9, or 81.1%.

[Red Iron Crown]

Whereas in KSP, tanks have all the same fixed mass fraction of 8/9, or 81.1%.

Except for the 0.625m and 3.75m tanks.

[Pecan]

Well that's true, I normally try and take more fuel than I'm gonna need. And as for the overweighred orbiters, fair point. I was kinda frustrated that I haven't been able to make a serial-stage-only Mun/Minmus vehicle, but considering the amount of weight they have to get up there I suppose it isn't all that surprising. Guaranteed that people have managed it though, I probably just lack the engineering skill

Yep - see chapter 5 of the tutorial in my signature - Long Tom is a small serial-staged Minmus/Mun vehicle launched on the, also serial-staged, LV-6-S.

It isn't at all optimised but is used as part of the discussion on staging and lunar-missions.

[ArcFurnace]

Whereas in KSP, tanks have all the same fixed mass fraction of 8/9, or 81.1%.

I like the KW Rocketry mod tanks, with a mass ratio of 10 (propellant fraction 90%).

[Starwaster]

Sorta assuming that a unit of measurement tank wise is = liter.

Pretty sure this is why NathanKell changed these values in realFuels, because had no idea what 1 unit actual represents in real world examples.

Correct, that's an assumption and it's an erroneous one. The OP has available fuel volume as ~1/5.75 of total available volume. I have no idea what Squad's real basis for their volume unit was (or indeed whether it's anything but arbitrary) but it's not liters.

Despite the Wiki saying otherwise, the fuel unit is a mass (5kg). Otherwise the densities and tank sizes are silly.

You derive fuel mass from volume x density, not from mass x density. The capacity of the fuel tank is measured as volume, not mass. People don't like to think that we're dealing with some arbitrary unexplained unit of measurement (it makes them uncomfortable) but we are and always have been.

[Red Iron Crown]

You derive fuel mass from volume x density, not from mass x density. The capacity of the fuel tank is measured as volume, not mass. People don't like to think that we're dealing with some arbitrary unexplained unit of measurement (it makes them uncomfortable) but we are and always have been.

Fuel mass per unit doesn't need to be derived, it can be directly observed in game. Remove one unit of fuel or oxidizer from a tank and the ship's mass decreases by exactly 5kg. There's no reason to think that the units of fuel and oxidizer are volumes, especially since the fuel and oxidizer would then have exactly the same density (an unlikely coincidence).