[WIP] Nert's Dev Thread - Current: various updates

[Jimbodiah]

Do people use cryogenic engines for anything other than lifting and placing into orbit?

[Kerbas_ad_astra]

3 hours ago, Jimbodiah said:

Do people use cryogenic engines for anything other than lifting and placing into orbit?

I've got a probe right now that uses a 1.25m upper-stage motor for interplanetary departure and arrival.  (Not realistic, but it gets better performance.)

[Nertea]

Drag will probably not be changed to start. I will probably lower the maxTemps a bit just because I like doing that and it makes me feel powerful :P. Just boiloff vs mass ratio as a tradeoff... It turns out I will probably need to write a tiny module for this, but that will not be hard.

Rates, what do we think about these. It will probably be volume-scaled exactly, for logic's sake. I mean in theory it should be dependent on the surface area to volume ratio of the tank, but that's too much work/silly/inconsistent. In an ideal world, loss should be minimal during launch and a few orbits (~45 min), but be severe enough that a trip to Duna will put a serious dent in your tanks. 

[Captain Sierra]

1 hour ago, Nertea said:

I mean in theory it should be dependent on the surface area to volume ratio of the tank

This is why I suggested giving all tanks a flat rate, thus making larger tanks have less boiloff. It roughly simulates the effect of this without the complex math behind it (which is ultimately how most things in KSP are designed: simulate the effects to illustrate the real-world problem, while still grossly simplifying it for gameplay-sake).

[Fraz86]

19 hours ago, Nertea said:

Drag will probably not be changed to start. I will probably lower the maxTemps a bit just because I like doing that and it makes me feel powerful :P. Just boiloff vs mass ratio as a tradeoff... It turns out I will probably need to write a tiny module for this, but that will not be hard.

Excellent!

Quote

Rates, what do we think about these. It will probably be volume-scaled exactly, for logic's sake. I mean in theory it should be dependent on the surface area to volume ratio of the tank, but that's too much work/silly/inconsistent. In an ideal world, loss should be minimal during launch and a few orbits (~45 min), but be severe enough that a trip to Duna will put a serious dent in your tanks. 

If you want the trip to Duna to put a "serious dent" in the tanks, but not to completely deplete them, then I'd suggest something like 0.025% loss per hour. That translates to 45% LH2 loss after a trip to Duna, if we assume the trip takes 300 Kerbal days (1800 hours), and assuming we're talking about % loss relative to LH2 volume capacity (not relative to the remaining volume of LH2).

Edit: Alternatively, if you want the dent to be a bit less serious (potentially making it practical to bring atmospheric tanks to Duna), you could go with 0.01% loss per hour, which translates to 18% LH2 loss after a trip to Duna. Another way to look at loss rates is in terms of "time until depletion," which would be 4000 hours (667 Kerbal days) at 0.025% per hour, or 10000 hours (1667 Kerbal days) at 0.01% per hour.

Edited December 27, 2015 by Fraz86

[Captain Sierra]

@Fraz86 I think I would prefer something in the middle. I think it should be possible to execute a Duna mission on lifter tanks alone (assuming insulated orbital tanks are higher up the tech tree). Boiloff should become a serious concern that has to be designed around, but not be prohibitive. 0.015% loss per hour results in a 27% loss to Duna.

 

Another thing that came to mind is will oxidizer boil off too? Its a stock resource so depending on how the module is configured, it may not have the effect. In that case, you also have to consider the oxidizer which is now dead mass. You're still lugging it even if you cant burn it. This means less boiloff rate can still inflict the same punishment to your mission margins.

[Fraz86]

17 hours ago, Captain Sierra said:

@Fraz86 I think I would prefer something in the middle. I think it should be possible to execute a Duna mission on lifter tanks alone (assuming insulated orbital tanks are higher up the tech tree). Boiloff should become a serious concern that has to be designed around, but not be prohibitive. 0.015% loss per hour results in a 27% loss to Duna.

 

Another thing that came to mind is will oxidizer boil off too? Its a stock resource so depending on how the module is configured, it may not have the effect. In that case, you also have to consider the oxidizer which is now dead mass. You're still lugging it even if you cant burn it. This means less boiloff rate can still inflict the same punishment to your mission margins.

Oxidizer shouldn't boil off; that would be a dramatic change to the stock game, which NFT avoids.

The left over oxidizer isn't dead weight - it actually makes boiloff less punishing for LH2/OX engines, not more. The player simply needs to haul extra LH2 (presumably in LH2-only tanks) for longer duration missions, to compensate for LH2 lost to boiloff. LH2/OX engines only use 0.14kg LH2 per kg OX, so you won't need much extra LH2 to make use of the surplus OX. These considerations argue in favor of higher boiloff if we want it to have a significant impact on gameplay.

On an unrelated note:

@Nertea After further consideration, I think boiloff rate should be modeled as % loss of remaining LH2, as opposed to % loss relative to LH2 capacity. Percent loss relative to LH2 capacity incentivizes tedious, silly behavior. For example, the player could minimize LH2 loss during a trip to Duna by storing their LH2 in 100 tiny LH2-only tanks. Once 1% of the LH2 has boiled off, the player empties 1 tank (transferring its contents to the other 99 tanks), thereby effectively reducing the boiloff rate by 1%. This process is repeated each time enough LH2 has boiled off to empty another tank, always maintaining the used LH2 capacity at the minimum necessary, minimizing boiloff. Encouraging this behavior is clearly undesirable. If the boiloff rate is instead modeled as % loss of remaining LH2, there is no incentive to move LH2 between tanks, and optimal gameplay is more straightforward. I would propose 0.025% loss per hour, which (with exponential decay) corresponds to ~36% loss to Duna.

Edited December 28, 2015 by Fraz86

[Nertea]

That sounds promising. It's also more forgiving, because with a percentage of remaining, it'll take quite a while for the last, last bit of LH2 to boil off.

We want to make round trip missions to Duna using cryogenics awkward without using a ZBO tank, just as they would be. Using a LH2 departure stage with lifter tanks, as IRL these days, would be just fine with slow boiloff.

[Fraz86]

@Nertea Excellent. Now we just need to decide what loss rate would make such a mission suitably awkward. For the purpose of illustratuon, let's imagine a mission to Duna that will need 10t of LH2/OX (1.24t LH2, 8.76t OX) for its return trip, using cryo engines, 300 Kerbal days (1800 hours) after its initial launch. How much extra LH2 will be needed to compensate for bioloff, and what does this represent in terms of % increase of the original total fuel mass of the return vehicle?

LH2 loss per hour	LH2 lost after 1800 hours	Extra LH2 required to have 10t LH2/OX (1.24t LH2, 8.76t OX) after 1800 hours	Relative increase of original fuel mass
0.025%	36%	0.7t	7%
0.05%	59%	1.8t	18%
0.06%	66%	2.4t	24%
0.07%	72%	3.1t	31%
0.08%	76%	4.0t	40%
0.09%	80%	5.0t	50%

It seems that 0.025% might be inadequate to make Duna round trips truly awkward; 7% extra fuel mass on the return vehicle isn't really a substantial penalty.

Also worth noting: the impact of boiloff will be much more significant for engines that run on LH2 only, such as nuclear engines. In these cases, a 0.025% loss rate would correspond to a 56% increase in required fuel mass (for use after 1800 hours).

Edited December 29, 2015 by Fraz86
Wider range of example values.

[Jimbodiah]

If you are using cryogenic engines for interplanetary missions, you are doing it wrong. This is not a fuel system meant for long hauls.

[Streetwind]

6 minutes ago, Jimbodiah said:

If you are using cryogenic engines for interplanetary missions, you are doing it wrong. This is not a fuel system meant for long hauls.

Congratulations, you figured out why the current discussion is going on in the first place

[Jimbodiah]

Yeah, but this should be logical. I don't get why Fraz is doing all sorts of calculations on a fuel that dissappears so he can time his Duna missions. There is no worse fuel to use on long periods of transit than one that dissappears on it's own. It's not like cryos are that much more efficient that regular LOx engines, so why go through the trouble. Launch your rockter from kerbin, insert into orbit, launch towards planet, forget. Hence boil-off is not even of interest while launching as it will be over in a matter of hours and you should have enough dV with you to be negligable. Or is it just me?  I use regular LOx engines for Duna, or NERV/Lightbulb, cryo doesn't even come to mind.

Edited December 29, 2015 by Jimbodiah

[Psycho_zs]

Boiloff could be done with stock resource converter, all-in-one with cooling (in)ability and temperature curve.

BTW, parts in NFS and NFC may need maxtemp tweaking. For reference, stock crew container has maxTemp=1000, skinMaxTemp=2000. IMHO, most orbital parts should adhere to that, although it's still a bit high.

Edited December 29, 2015 by Psycho_zs

[Kerbas_ad_astra]

1 hour ago, Jimbodiah said:

It's not like cryos are that much more efficient that regular LOx engines, so why go through the trouble.

Because it can be worth the trouble, at least as the parts are now (i.e. no boiloff implemented yet) -- 350 s versus 450 s is quite substantial, and because of the log term in the rocket equation, the efficiency gain can have an even larger impact on spacecraft mass.  In my own testing, to get a Hitchhiker crew can to have 6 km/sec of delta-V (more delta-V than most space missions need in KSP, but reasonable for the real solar system -- I was doing this testing in the context of SMURFF, which is intended to rebalance fuel tanks and engines to be more useful with RSS), the total vehicle had a mass of 48 tons using LFO and an LV-909, while using the Chelyabinsk (1.25m upper-stage engine) and hydrolox, the total vehicle mass was 25.6 tons.  This is nearly a 50% reduction in vehicle mass, which allows the launch vehicle to be that much lighter as well.

(With SMURFF reducing LFO tank mass by a factor of 4 and engine mass by 62.5%, the LFO-burning and hydrolox-burning masses were 19.2 tons and 14.8 tons, respectively.  The margin is narrower because LH2 tank mass was not reduced, but it's still present.)

If I were serious about sending a hitchhiker can to Mars, I would perform further optimization that would probably lead to a multi-stage design that has a lower overall mass, but the point of the testing was to perform like comparisons of the two fuel types, not to produce viable spacecraft.  Tl;dr: As things are now, switching to hydrolox engines can offer substantial improvements in spacecraft mass.  Of course, their inferior TWR and potentially larger size (since LH2 is relatively "fluffy" compared to liquid fuel and oxidizer, even with the fudge factor increasing density) provide new challenges, so it's not a choice that's automatically worth making...your mileage delta-V may vary, so do the testing for each mission, etc.

[Elthy]

1 hour ago, Jimbodiah said:

Yeah, but this should be logical. I don't get why Fraz is doing all sorts of calculations on a fuel that dissappears so he can time his Duna missions. There is no worse fuel to use on long periods of transit than one that dissappears on it's own. It's not like cryos are that much more efficient that regular LOx engines, so why go through the trouble. Launch your rockter from kerbin, insert into orbit, launch towards planet, forget. Hence boil-off is not even of interest while launching as it will be over in a matter of hours and you should have enough dV with you to be negligable. Or is it just me?  I use regular LOx engines for Duna, or NERV/Lightbulb, cryo doesn't even come to mind.

I think NFT is often used for rescaled solar systems which require more Delta-V. There it does make sense to watch out for ISP...

[Whirligig Girl]

On 12/23/2015 at 1:40 PM, Nertea said:

Ok, testing seems to show good results. I'm now down to the description fixing and number tweaking stage, so I can hopefully declare this NFE dev round complete. As usual, I'd really like to thank everyone who helped out. They made this possible. Apply some rep!

I'd like to turn the conversation back to LH2, CryoEngines and AOYUNEM (As Of Yet Unnamed Nuclear Engines Mod), now that LH2 has been cut out of electrical engine balance this might be a little more manageable.

CryoEngines "feels" nice to me right now, but there's always room for changes.

The way it looks to me is that the mass ratio of LH2 tanks just needs to be friendlier. This will likely need some small downward revisions in CryoEngines Isp values, I'm fine with that. I'd like to aim for something slightly lower than LiquidFuel dry mass. The only annoying thing that I see is that the dry masses (and total masses) of some of the smaller LH2 tanks will be ridiculously small. It might be good to fudge the amount of LH2 going into a tank a bit more than it already does. 

This will still "nerf" the LH2 LV-N a little bit with a worse dry mass ratio, but I'm totally willing to buff the LH2 Isp up to a max of 900 and/or decrease the engine mass a bit. Or do nothing, really. The progression for AOYUNEM would indicate that the Nerv is really a bottom tier, entry level nuclear engine, and could be not the end-all be-all that it is in stock.

The other thing to solve is some sort of differentiating factor for the fancy LH2 'orbital' tanks without implementing boiloff, preferably. I'm thinking minor advantages and disadvantages, mostly to discourage atmospheric use and encourage orbital use. Options might include that drag increase (doable for sure), increasing the LH2 volume compared to the basic tanks (but preserving the same mass ratio)... fiddly things like that.

 

 

So is that nuclear engine above a liquid core NTR? Pictures I've seen of LCNTRs have spherical reactors. Of course, some solid designs I've seen as well. It would be cool to see a Liquid Core and Open-Cycle Gas Core rocket, with the latter being a huge reputation and funding drain to operate within Kerbin's atmosphere.

Edited December 29, 2015 by GregroxMun

[Jimbodiah]

If ISP (vac) is a problem, use nukes, ISP of 800-1500 and no boil-off. Or use a upscaled Ion engine with onboard reactor, ISP of 3K-11K.

I just launched a 300 ton space station (career mode, 2.5M credits just for the lifter) and will be sending up the engine section (nukes) later to get it to Duna, crew will follow later. 3500ton launch weight to get to 1000km kerbin orbit for assembly with the EDS. Lander, probes and life-support prep mission will follow soon, but also nukes to get to Duna. Same will go for the crew ship later. Ion would be an option, but burn times take too long and you need a huge/expensive reactor that will be discarded later on, so will just strap on a bigger tank with some nukes.

Edited December 29, 2015 by Jimbodiah

[billkerbinsky]

If we're getting boiloff and heavier ZBO tanks, can we get electric engines that use LH2 back?

BTW, in the real world, ULA is working on this cryogenic storage and in-space refueling - see https://en.wikipedia.org/wiki/Advanced_Cryogenic_Evolved_Stage and http://www.ulalaunch.com/uploads/docs/Published_Papers/Exploration/DepotBasedTransportationArchitecture2010.pdf .  They're planning to profit from boiloff in part by burning the boiled-off gases in an internal combustion engine to generate electric power; they're also planning to use cryogenic fuels for RCS, eliminating the need for additional tanks for hydrazine.

 

 

[Fraz86]

@Jimbodiah The Chelyabinsk is currently (without boiloff) an excellent engine for many Duna mission profiles. Compared to its LF/OX analogue (the Terrier), the Chelyabinsk has nearly the same thrust (55 vs 60), a bit more mass (0.8 vs 0.5), and higher cost (1920 vs 390), but substantially better ISP (455 vs 345). For many Duna missions, the fuel savings with the Chelyabinsk will easily justify its downsides. Nukes, in some cases, may be overkill due to their dramatically higher mass and cost. Thus, if we don't want LH2/OX engines to be ideal for these missions, we need to make sure that boiloff is suitably problematic.

Hypothetically, we could just set boiloff at a rate that makes it practically impossible to use these tanks outside the Kerbin SOI, which was my original thought (back when I suggested 0.1% loss per minute). That would keep things simple, but perhaps it's a bit boring, and it might annoy players by forcing them to make transfer burns immediately after launch (players might want a few hours or days to rendezvous with another ship section, dock at a station to pick up crew, or whatever else). Nertea asked for a rate that would "put a serious dent" in the fuel if taken to Duna, so I set about calculating what that might be.

Edited December 30, 2015 by Fraz86

[Captain Sierra]

On 12/28/2015 at 5:53 PM, Fraz86 said:

Also worth noting: the impact of boiloff will be much more significant for engines that run on LH2 only, such as nuclear engines. In these cases, a 0.025% loss rate would correspond to a 56% increase in required fuel mass (for use after 1800 hours).

One possible solution to this to equalize this may be to actually switch cryo engines to using actual LOX instead of stock oxidizer and adding the associated boiloff to that component too. Now ... this begins to stray away from stock in the interest of CryoEngines -AOYUNEM compatibility, so thats something Nert should weigh in on.

@Fraz86 That table you put together is fantastic. Great info there so take some rep for that. Given the data displayed, I think the value we should go with should be between 0.06-0.07%. Perhaps 0.065%. All three of those values look promising and are still slow enough that cryogenic stages are still fully viable for intra-kerbin operations (which I think should be a reasonable expectation).

[Fraz86]

4 hours ago, Captain Sierra said:

One possible solution to this to equalize this may be to actually switch cryo engines to using actual LOX instead of stock oxidizer and adding the associated boiloff to that component too. Now ... this begins to stray away from stock in the interest of CryoEngines -AOYUNEM compatibility, so thats something Nert should weigh in on.

I think it's perfectly reasonable to assume that stock oxidizer is LOX, and adding even more fuel types should probably be avoided unless truly warranted. I don't think it's important to "equalize" the effects of boiloff on cryo engines versus AOYUNEM. Practically, there isn't really much difference between "hugely impractical" (return from Duna with AOYUNEM and atmospheric tanks) versus "not hugely impractical, but clearly inferior to alternatives" (return from Duna with cryo engines and atmospheric tanks). There's no logical reason to use either in actual gameplay, so the distinction is academic.

Edited December 30, 2015 by Fraz86

[Jimbodiah]

Kspi LH2 tanks have an option to cool the tank at cost of Ec. Boiloff is very low when cooling is active, and higher when turned off. Maybe an option?

[Fraz86]

A small EC requirement for ZBO tanks might be an interesting mechanic, reflecting their active refrigeration, akin to the EC requirement of radiators.

Edited December 30, 2015 by Fraz86

[Jimbodiah]

I happened to notice when I was testing an engine that needs LH2 and used a KSPI tank that the rocket was dead at the launchpad for some reason, turend out it was draining current to cool the LH2. Other, non-kspi, LH2 tanks don't have this Ec effect so I didn't even notice they used that mechanic.

Without cooling active I lose about 1 unit per second.on a tank with 7K storage, don't know if it's a flat-rate or a % as I didn't test further. But it sure drains the Ec fast if you are not prepared for it.

[Starbuckminsterfullerton]

15 hours ago, Fraz86 said:

A small EC requirement for ZBO tanks might be an interesting mechanic, reflecting their active refrigeration, akin to the EC requirement of radiators.

Seconded, I actually assumed they already would consume power.

If you want to get really fancy Nert, you could make the boiloff rate a function of the tank temperature, and require radiators/insulators or power to keep them cool and avoid transferring heat generated in other parts of the ship.  For short trips like launches you wouldn't need to bother, but for longer voyages the extra mass would start to pay off.

On 12/29/2015 at 11:03 PM, Jimbodiah said:

If you are using cryogenic engines for interplanetary missions, you are doing it wrong. This is not a fuel system meant for long hauls.

True, but the tanks are used for more than cryogenic LH2/LFO engines.  Once you have a mod adding Hydrogen it only makes sense to patch NTR's to run on it, and NTR's are of course used extensively for interplanetary travel, so the properties of these fuel tanks has a significant effect on almost all non-electric interplanetary craft. 

[QUOTE:Nertea]As usual, I'd really like to thank everyone who helped out. They made this possible. Apply some rep! [/ QUOTE]

A bit off topic, but is the reputation system even still working in the new forums? I see values, but no way to add to them...

 

 

Edited December 31, 2015 by Starbuckminsterfullerton