[1.12.x] Cryogenic Engines: Liquid Hydrogen and Methane Rockets! (Jan 22, 2022)

[Fraz86]

47 minutes ago, Psycho_zs said:

If LH2 fudge factor were 1.5, it would result in much more handy (matching-wise and a bit closer to RL) 2:1 tankage ratio with OX. Just a thought occurred while trying to mash up a ship with cool looks, separate  LH2/OX tanks and sane part count.

The LH2/OX tankage ratio is currently 3:2, which I think is plenty handy, and much easier than the stock LF/OX ratio of 9:11. It's true that real-world LH2/OX tankage ratio is closer to 2:1, but Nertea's cryogenic engines run a bit more fuel-rich than their RL counterparts. Regardless, these ratios are reasonably close to RL, and precise realism has never been a high priority in Nertea's mods - balance take precedence, and I believe the current volume ratio offers the best balance.

[Bluebottle]

10 hours ago, Fraz86 said:

The fuel:dry mass ratio for oxidizer (8:1) is derived from stock KSP, and as such is not something we would alter. Nor would we add boiloff to a stock resource. Adding boiloff for LH2 but not LOX makes a certain amount of sense anyway, as boiloff is a much less severe problem for LOX compared to LH2 in the real world. Moreover, the mass "penalty" for extra insulation/refrigeration for LH2 is actually relatively minor - that is, the interval mass difference between LH2 lifting tanks and ZBOs is equivalent to only 10% of the dry mass of a volume-matched LFO tank.

The true dry mass "penalty" for LH2 tanks is primarily related to fuel density. LH2 is far less dense than other fuels, and therefore requires much greater (~700%) tank volume for an equivalent fuel mass. Greater tank volume means greater tank mass, though the mass per volume is actually only half as much (or less), as mentioned above, which partially offsets this penalty. The penalty is further mitigated by using 2x the real-world density of LH2. What we're left with should be a bit inferior to LFO as a fuel type, but reasonably easy to balance by giving NTRs attractive properties.

Thanks for your considered answer here. LOX can hardly be considered a "storable" - usually it would be gone in a couple of weeks - but I understand that altering a stock resource would be wrong. This isn't Realism Overhaul. Then again, LH2 is now the only propellant with RO-like properties, i.e. boiloff. Those NTR "attractive properties"; would that mean lowering the mass of all NTR engines, increasing thrust, etc? Stock ones, or the Nova from Ven's revamp, are, as things stand, completely non-viable.

Hydrolox engines are also non-viable though. I cannot currently design any craft, or any stage, where hydrolox makes more sense than LFO. Nertea talked about un-nerfing the cryo engines and I hope that happens. LFO stages are lighter, higher TWR, lower part count, and higher dV right now, even without ZBO EC penalties. More on that below.

10 hours ago, Fraz86 said:

What kind of craft are you building that requires 2000 EC/s for ZBOs? That should correspond to 18 of the ridiculous 10 meter tanks.

This petite darling, my "Olympus" station...

Spoiler

Yes, 18 spherical tanks wrapped around Speedy's hex truss in the LH2 section for 20,760,000 LH2, and another 4 million in the hydrolox section with the tweakscaled Mondo tanks. Each section is only 1 welded part (I love that Ubio plugin) which works well after some part file editing. Everything except LFO will be dry for the Minmus-Kerbin transfer, and then I'll have to make some similarly large tankers to service it. That's where I'm running into problems with the feasibility of LH2/NTR and hydrolox.

 

7 hours ago, Fraz86 said:

8 hours ago, blowfish said:

I think it's more meaningful to look at the tank mass per unit of fuel mass rather than per volume.  Since just about everything in rocketry is about mass (volume affects drag a bit but it's usually not a major concern).

Fuel:dry mass ratios are certainly much more meaningful for analysis of performance. However, when arguing that LH2 tanks ought to be lighter, based on comparisons to specific real-world rocket stages or lack of need for internal bracing (as Bluebottle was doing), I believe it's useful to step back and remember that empty LH2 tanks are actually the lightest objects in the game for their size.

Yes, they are. And the mass and volume ratios you've implemented are remarkably realistic (very close to the Delta IV CBC). A Centi-2 + Centi-3 tweakscaled (I know, I know) to 5m give me something very close to a D-IV CBC tank (http://spaceflight101.com/spacerockets/delta-iv-heavy/), so the 2x fudge isn't actually unrealistic in that case. The engine TWRs, though, have relegated cryo engines to being niche parts - suitable only for a 2nd/3rd stage lifter - which is somewhat realistic, but therefore unbalanced versus the now-magical-by-comparison LFO. What I'm saying is: nobody's going to use them, at least nobody who uses KER or Mechjeb for dV calculations, or anybody cares about part count and doesn't weld (so, so many tanks on interplanetary vessels). That makes me sad because I love the idea/concept of the cryo engines.

 

5 hours ago, NathanKell said:

Besides, OX isn't liquid oxygen anyway, it's nitrogen tetroxide

Oh dear, that makes me feel bad for venting it right into Bill's face!

Edited February 4, 2016 by Bluebottle
missed a negation

[Streetwind]

4 hours ago, Bluebottle said:

This petite darling, my "Olympus" station...

  Reveal hidden contents

Yes, 18 spherical tanks wrapped around Speedy's hex truss in the LH2 section for 20,760,000 LH2, and another 4 million in the hydrolox section with the tweakscaled Mondo tanks. Each section is only 1 welded part (I love that Ubio plugin) which works well after some part file editing. Everything except LFO will be dry for the Minmus-Kerbin transfer, and then I'll have to make some similarly large tankers to service it. That's where I'm running into problems with the feasibility of LH2/NTR and hydrolox.

Holy Saint in a camper van! I certainly have not seen much in my years of KSP'ing that would exclipse this thing in sheer physical size...

However, you do realize that your power draw here is perfectly "working as intended", right? Supermassive stations will require supermassive upkeep. Powering this with a 2.5m reactor is about comparable to a normal spacecraft being powered by a 0.625m reactor, both in relative power envelopes and in terms of physical size of the reactors. So I would say that everything is scaling properly and just fine.

I recommend investing in a Whirlijig. Not only does it recycle 50% of your expensive uranium, but it can also produce new uranium out of locally sourced ore.

[Bluebottle]

2 hours ago, Streetwind said:

Holy Saint in a camper van! I certainly have not seen much in my years of KSP'ing that would exclipse this thing in sheer physical size...

However, you do realize that your power draw here is perfectly "working as intended", right? Supermassive stations will require supermassive upkeep. Powering this with a 2.5m reactor is about comparable to a normal spacecraft being powered by a 0.625m reactor, both in relative power envelopes and in terms of physical size of the reactors. So I would say that everything is scaling properly and just fine.

I recommend investing in a Whirlijig. Not only does it recycle 50% of your expensive uranium, but it can also produce new uranium out of locally sourced ore.

Thanks. My ultimate aim is to recreate something like the Pegasus from the BBC's "Space Odyssey" mini series in 2004. That was about 1.2km long, fueled with LH2, powered by NTR engines.

There are two spots on that station for reactors to be installed later, once it reaches Kerbin orbit. I will probably add a Whirlijig.

As for scaling, and I know this is likely 'too much', but could it be possible to scale the cooling EC demand down, as fuel mass and/or volume rises? Surface area increases at a much lower rate than volume (http://www.tiem.utk.edu/~gross/bioed/bealsmodules/area_volume.html), so the relative losses at larger volumes should be far below the current, linear relationship of 0.1 EC per 1000 LH2. There would have to be two metrics, I suppose: an absolute cooling EC demand that tapers down as the absolute volume increases, and a relative, per-vessel EC rate that compares the current LH2 remaining to the MaxAmount. A full tank means the volume:surface-area ratio is at its most favorable, and the bigger that tank is (in absolute terms), the better. This is because, per unit of LH2, a full Centi-2 should require much more cooling than a full Mondo, simply due to the volume:surface-area ratio. Just pie-in-the-sky stuff that would require large and unpleasant changes, I guess, but it could be a way to help NTR viability on large vessels.

Edited February 4, 2016 by Bluebottle

[Streetwind]

I wasn't much involved, but I believe the flat rate was an intentional choice intentional to keep the system straightforward and simple. You can check the dev thread for details.

I don't think anyone thought someone would build a vessel quite as large as yours   But your NTR is not any less viable than it would be on a smaller craft... it just doesn't scale to be more viable through cooling efficiency gains.

[Bluebottle]

1 hour ago, Streetwind said:

I wasn't much involved, but I believe the flat rate was an intentional choice intentional to keep the system straightforward and simple. You can check the dev thread for details.

I don't think anyone thought someone would build a vessel quite as large as yours   But your NTR is not any less viable than it would be on a smaller craft... it just doesn't scale to be more viable through cooling efficiency gains.

Completely understandable. I'll lurk on the dev thread instead of clogging up the release one.

Hmmm, I know. I once built an 860m-long Minmus base extension down to some flats which were better as landing pads, but anything that was undocked from it (or the extension itself) instantly teleported to the other side of Minmus, lat/long/altitude 0. Probably bad to undock when the new vessel's CoM is outside the 200m physics force simulation range - kind of a hard limit. As for the NTR viability, I was hoping to save on the tonnage and parts associated with the reactors, thereby increasing the TWR. This will require some... more inventive solutions now.

[Ixenzo]

I'm currently trying to internalize the niche of cryo engines. They have less TWR, are more expensive, require relatively large tankage and EC to combat boil-off and return higher Isp and less wet mass overall. Conventional LFO setups give more thrust and dV assuming the same tankage volume with no drawbacks. I haven't played with cryo that much and can't quite wrap my head around in what situations would a cryo propulsion setup be better than an LFO one.

Lifters? Good ASL Isp, don't need fancy tanks but the stage is quite large and has less TWR. I tried using the first cryo engine you get in the career but ended up going LFO with a Swivel simply because it was cheaper, dV-richer and smaller in size. In an old save I had, when there wasn't boiling off, I "engineered" an SSTO to Duna which worked quite well and it was fun to try out but obviously in real missions a design like that would be kinda silly.

Deep space stages? Need to pack up additional EC which is mass, which is less TWR and dV. Nukes, LFO and electric engines do that job quite well too.

It feels as if cryos are a sidegrade to conventional engines but the stats are balanced in a way it's just easier to use the existing designs. A TF2 analogy comes to me: there is a default rocket launcher. Well, it's been there forever, a good, reliable weapon. Then later a sidegrade is introduced. The projectile travels 80% faster, does 25% more damage but the splash radius is basically non existent. In serious gameplay (competitive) practically nobody uses the last one because it's very situational and on top of that is highly dependent on one's ability to aim good very fast and score direct hits which is instantly thrown off by the 80% faster projectile. I hope that made sense.

Am I missing an obvious application of cryos or I just haven't had a mission where cryos would be advantageous to use?

[blowfish]

@Ixenzo Maybe look at the roles cryogenic engines fill in real life - usually not primary lifters or long-term interplanetary stages, but long-burning sustainers paired with boosters and upper/transfer stages to be used within a few hours of launch.  That's generally where the I_sp benefit will give you largest advantage while the boiloff won't matter much.

And the comparison with nukes sort of highlights how OP they are in stock.  If you're using CryoEngines already, I'd recommend Nertea's Kerbal Atomics mod which changes them to run on LH2, so you actually have to balance their high I_sp against large fuel volume and boiloff/refrigeration requirements.

All that being said, Nertea has said that CryoEngines may be due for a rebalance.  Not sure if/when that will happen though, and exactly what might change.

[Fraz86]

13 hours ago, Bluebottle said:

Thanks for your considered answer here. LOX can hardly be considered a "storable" - usually it would be gone in a couple of weeks - but I understand that altering a stock resource would be wrong. This isn't Realism Overhaul. Then again, LH2 is now the only propellant with RO-like properties, i.e. boiloff.

LH2 boiloff isn't intended to feel like an "RO-like property." It's intentionally simplified and easily prevented, so hopefully it feels more like a small factor that sets LH2 apart from other fuels in an interesting way, as opposed to a rigorous simulation of real-world challenges.

Quote

This petite darling, my "Olympus" station...

  Reveal hidden contents

Yes, 18 spherical tanks wrapped around Speedy's hex truss in the LH2 section for 20,760,000 LH2, and another 4 million in the hydrolox section with the tweakscaled Mondo tanks. Each section is only 1 welded part (I love that Ubio plugin) which works well after some part file editing. Everything except LFO will be dry for the Minmus-Kerbin transfer, and then I'll have to make some similarly large tankers to service it. That's where I'm running into problems with the feasibility of LH2/NTR and hydrolox.

Haha, well, as I'm sure you're aware, that thing is very far from a typical craft.

Quote

Yes, they are. And the mass and volume ratios you've implemented are remarkably realistic (very close to the Delta IV CBC). A Centi-2 + Centi-3 tweakscaled (I know, I know) to 5m give me something very close to a D-IV CBC tank (http://spaceflight101.com/spacerockets/delta-iv-heavy/), so the 2x fudge isn't actually unrealistic in that case. The engine TWRs, though, have relegated cryo engines to being niche parts - suitable only for a 2nd/3rd stage lifter - which is somewhat realistic, but therefore unbalanced versus the now-magical-by-comparison LFO. What I'm saying is: nobody's going to use them, at least nobody who uses KER or Mechjeb for dV calculations, or anybody cares about part count and doesn't weld (so, so many tanks on interplanetary vessels). That makes me sad because I love the idea/concept of the cryo engines.

It may be the case that hydrolox engines and NTRs need some buffing to make them more attractive. I think Nertea has wisely approached these balancing issues conservatively, because we absolutely don't want to create uber-engines that outright obsolete LFO.

[Nertea]

I was hoping to get to this sometime this week, but have failed to find the time. Been very busy. I will probably update it on the weekend because I need to do so to update bundled dependencies (CRP).

Planned changes are ~ +10s Isp for most cryogenics, a revisit to TWR for many engines, a *slight* decrease in cryo cooling costs (~10%). 

[mikegarrison]

In Real Life, LH2+LOX stages are lighter for the same total stage delta-v. You take advantage of this by reducing your requirements for the stages that come before. So they aren't used as first stages, they are used as upper stages.

Obvious examples are the Saturn V with the second and third stages LH2+LOX and also the Space Shuttle. (While the Space Shuttle did burn LH2 all the way from the ground, the "first stage" was really mainly the SRBs.)

I think a well-balanced Cryo engines mod would encourage the use of SRBs or Kero stages for launch and cryo stages for upper stages. You know the balance is right if that gives you a better overall cost for the payload to orbit than either all-cryo or all-kero.

It's hard to do in stock KSP, though, because the tiny size of Kerbin (and resulting low orbital velocity) means that it's often easier and not much less efficient to just build a SSTO kero rocket.

Edited February 5, 2016 by mikegarrison

[blowfish]

8 minutes ago, mikegarrison said:

So they aren't used as first stages, they are used as upper stages.

Well, they're sometimes used as sustainers paired with long-burning boosters.  Examples: Space Shuttle/SLS, Ariane V.  Occasionally used as a first stage only: Delta IV.  None of those launch vehicles are known for their cost effectiveness though.

[StickyScissors]

So i read the changelog AFTER updating, like a scrub, and now i know why my sub-assemblied interplanetary transfer stages can't push as much as they used to.

• Changes to the base masses, mix ratios and capacities of LiquidHydrogen tanks

Back to the drawing board, i guess :3

[Sudragon]

I use cryo rockets because they're better for the environment.  

Kerbals are, after all, 

<sunglasses>

Green.

[Captain Sierra]

19 hours ago, blowfish said:

Well, they're sometimes used as sustainers paired with long-burning boosters.  Examples: Space Shuttle/SLS, Ariane V.  Occasionally used as a first stage only: Delta IV.  None of those launch vehicles are known for their cost effectiveness though.

I've never assumed these engines were supposed to be cost effective.  I care more about raw dV than I do about my space program's budget.

[sashan]

The engines are incompatible with FAR, in some sense. The gimbal range is so ridiculously small that rockets flip even after FAR increases the range 1.5times. Real engines have t on the order of 7-10 degrees, and most of these ones have it on 1.5 degrees!

[Ixenzo]

I personally have never had any problems with cryo engines while using FAR. In fact, it is hard to flip an average rocket in FAR atmo while following a standard gravity turn: higher gimbal range here would actually contribute to the flippening (is that a word?).

[Drew Kerman]

On 2/5/2016 at 4:18 PM, sashan said:

The engines are incompatible with FAR, in some sense. The gimbal range is so ridiculously small that rockets flip even after FAR increases the range 1.5times. Real engines have t on the order of 7-10 degrees, and most of these ones have it on 1.5 degrees!

 

13 hours ago, Ixenzo said:

I personally have never had any problems with cryo engines while using FAR. In fact, it is hard to flip an average rocket in FAR atmo while following a standard gravity turn: higher gimbal range here would actually contribute to the flippening (is that a word?).

No one is mentioning if they are using fins or not. I'm guessing sashan isn't and lxenzo maybe is. But if gimbal isn't enough, add fins. Or dampen your control inputs.

[sashan]

14 hours ago, Gaiiden said:

I'm guessing sashan isn't and lxenzo maybe is

No, fin's don't help unless they are sized like half the rocket. 1.5 degrees of gimbal are ridiculous, even for an upper stage engine.
I have 2 years of experience with FAR, all rocket designs imaginable. I almost never use fins, proper gimbal ranges and low angles of attack are a way to go. Do you know that all real rockets are aerodynamically unstable?

[Ixenzo]

18 hours ago, Gaiiden said:

fins

Actually, no. They're useless mass and drag and add no useful value to the craft. I consider fins on a conventional rocket more of a crutch for a bad design.

[Raptor9]

On ‎2‎/‎5‎/‎2016 at 3:05 PM, Captain Sierra said:

I've never assumed these engines were supposed to be cost effective.  I care more about raw dV than I do about my space program's budget.

Spoken like a true engineer, lol

[Nertea]

0.2.1: Balance tweaks and other minor fixes. This tested pretty well last night, should make cryo a lot more attractive given the boiloff parameters.

• Updated bundled CRP to 0.4.9.0
• Moved ISRU patch into CryoTanks folder, fixed ratios to new 15:1 mode from 0.2.0
• Fixed ISRU patch not affecting the mini ISRU
• Reduced mass of Volcano to 1.6 from 1.8, decreased thrust to 230 from 275, increased SL/Vac Isp to 385/422 from 360/410
• Reduced mass of Chelyabinsk to 0.7 from 0.8, increased Vac Isp to 465 from 455
• Reduced mass of Odin to 5.75 from 7.25, increased SL/Vac Isp to 385/420 from 365/415
• Reduced mass of Tunguska to 2.75 from 3.45, increased Vac Isp to 455 from 450
• Reduced mass of Mars to 13.25 from 12.5, increased SL/Vac Isp to 365/415 from 355/405
• Reduced mass of Yucutan to 8.65 from 8.95, increased SL/Vac Isp to 245/445 from 235/435
• Reduced boiloff cooling cost to 0.08 Ec/1000u from 0.1 Ec/1000u

[Kirondoll]

Maybe there is a problem with the amount of oxidizer inside the cryo tanks.

This is a test on a LR cryo probe.

Spoiler

 

[Fraz86]

12 minutes ago, Kirondoll said:

Maybe there is a problem with the amount of oxidizer inside the cryo tanks.

This is a test on a LR cryo probe.

  Hide contents

 

That tank should hold 4800 LH2 and 320 OX. Either something went wrong with the latest release, or you have some other mod that is overwriting the tank capacities.

[Kirondoll]

8 minutes ago, Fraz86 said:

That tank should hold 4800 LH2 and 320 OX. Either something went wrong with the latest release, or you have some other mod that is overwriting the tank capacities.

Found it, was an old Interstellar fuel switch configuration file.