Dedicated Radiator Panel parts

[Angelo Kerman]

I have my radiators! The next update of DSEV, which will be released tonight, contains the foundation for my active radiator system. In a nutshell, the radiators take on the heat of the vessel's parts, keeping them as close to ambient temperature as possible while still accounting for heat conductivity of the part. Right now, that means kerbals are going to get quite cold, but it's step one in my plan.

Thank you Red Iron Crown and NathanKell for your valuable information!

[NecroBones]

Yep, I would love some radiators to be added, stock. We can always mod some in, but now heat is a stock engineering problem.

[Hunting.Targ]

Are radiators active or passive? To avoid the "killer solution" you could make radiators require electricity. While unpowered it would act like a normal part radiating heat at some (probably good) rate, but when you powered it up it would accumulate and thus radiate heat more quickly. When you turned it off it would continue to radiate out but would also share any leftover heat back into the craft. In my head that sounds like a good balance - it'd be powerful and useful but with the costs of high power requirement and danger when you turn it off that it doesn't overheat other parts.

That's pretty much what your average refrigerator does; it's a heat pump that expends electricity to reduce temperature. The main problem is that heat pumps use Convection cooling, where radiative panels use radiative cooling, which is what Angel-125 and Red Iron Crown are discussing, and is ultimately the only principle to remove waste heat from a spacecraft in vacuum.

The thing about a heat pump is it can imbalance the heat distribution of a craft, at the cost of electricity as you suggested. Run the heat pump and you can force the waste energy out of your spaceframe and into the radiators - this of course raises their temperature. The rub is that if this is taken too far you can wreck the heat pump if it is overworked or the radiators if they get too hot. Sounds like a neat idea though.

- - - Updated - - -

Another idea that might be more complicated but also interesting would be liquid coolant that can be vented... perhaps an engine that naturally runs hot, like a Mainsail, could benefit from this, and a coolant tank would automatically consume its coolant to keep the engine below overheat level. This would add more weigh to a rocket, but it wouldn't induce drag the way that radiator panels would, so it'd be a better choice for liftoff stages and boosters. It still might be too complicated to bother with, though... unless it consumes monopropellant, which stretches believability unless monopropellant is pressurized gas instead of hypergolic chemicals. It's a thought, anyway.

I do really want heat radiator panels, not only for functionality but aesthetics too. I've used structural panels to represent them in the past, especially on space station trusses orthogonal to the solar panel arrays, but those are going to be really hard to get into orbit now with the new aerodynamics.

Venting coolant seems almost a pyrric proposition, since there is only a finite amount of coolant that can be carried. Passive systems work indefinitely; active systems, especially ones that use up a fluid or other resource, can only operate as long as the resource lasts. We've all seen what happened in stock .24-.90 when you turned on SAS for attitude control; it burned up monopropellant and did nothing you couldn't do yourself with a little careful attention; something you might do in a pinch, but not all the time.

Same here. Venting coolant could be something that is done either only manually or under specific conditions (like with Smart Parts!); if you've ever played with a thermostat or digital climate control panel, it should work pretty much the same way. Your spaceframe gets dangerously hot, so you vent some reserve coolant to get out of the emergency and then close the valve.

- - - Updated - - -

I would have to point out that a combo radiator/solar panel part would be interesting; a part that was a solar panel on one side and a radiative heat sink on the other. This is mainly because the solar panel would absorb a large swath of solar radiation, creating a 'blue shadow' zone where a radiator could function reasonably well. It could potentially be equally or even more efficient than an equivalent radiator edge-on to solar radiation, because it would be radiating into a space with lower energy density than the ambient environment. It would just have to flip 180° each time the craft moved from shadow to Kerbolight (two-sided panels don't have this problem). So if a craft ran out of electric charge during a night cycle, it would be out of luck until about late morning-midday. Small risk to take for a 2-in-1 part, I think.

Edited May 4, 2015 by Hunting.Targ
accidentally gave Angel's handle a -2. Sorry! You're now 125 again

[Sonny_Jim]

I started another thread about this, but just wanted to add my voice to the others and say it's a little perverse to add heat effects and not give us a dedicated part to deal with the heat.

[Laie]

Coming late to the party... Yesterday, I spent an hour watching how heat flows around my vessel and... well.

Within a three minutes after igniting the Nerva, the whole vessel was insanely hot, even the most remote parts were over 1000K. IIRC, at that temperature the paint should come off, the metal should be glowing red, and the fuel in the tanks should long be gone. Shouldn't the whole thing radiate like mad? I'm expecting hundreds if not thousands of Watt/squaremeter. Likewise, I didn't calculate how much heat per thrust a LV-N gives you, but I expect it to be preposterous.

In a nutshell: I don't think we need a heatsink, we need much reduced heat generation or much better radiation or something. Or maybe even a whole new heat system. The current one seems credible as long as we're talking about re-entry heat (which probably is superficial); but a hot engine boiling the vessel through and through until batteries pop at the far end, well, that's not even wrong. ©

I don't know how 60kN of thrust translate into Joule or Watt, but I wonder if there isn't more energy going into the vessel than out of it.

[Kuu Lightwing]

Well, if you talk about NERVA, being hot is the main feature of the engine! And since it heats by nuclear fission, it's probably doesn't matter how efficient is it in terms of waste heat. But I agree that spacecraft shouldn't heat up to 1000K, it's just ridiculous.

Does anybody knows what temperature is Merlin 1D Vacuum nozzle when it runs? You can see it being red hot, but that's just the nozzle, the rest of spacecraft doesn't heat like that. Granted, it's a chemical engine, it doesn't need to be as hot as the exhaust, unlike NERVA.

[Brotoro]

Coming late to the party... Yesterday, I spent an hour watching how heat flows around my vessel and... well.

Within a three minutes after igniting the Nerva, the whole vessel was insanely hot, even the most remote parts were over 1000K. IIRC, at that temperature the paint should come off, the metal should be glowing red, and the fuel in the tanks should long be gone. Shouldn't the whole thing radiate like mad? I'm expecting hundreds if not thousands of Watt/squaremeter. Likewise, I didn't calculate how much heat per thrust a LV-N gives you, but I expect it to be preposterous.

In a nutshell: I don't think we need a heatsink, we need much reduced heat generation or much better radiation or something. Or maybe even a whole new heat system. The current one seems credible as long as we're talking about re-entry heat (which probably is superficial); but a hot engine boiling the vessel through and through until batteries pop at the far end, well, that's not even wrong. ©

I don't know how 60kN of thrust translate into Joule or Watt, but I wonder if there isn't more energy going into the vessel than out of it.

Yes, it's goofy. But you should have seen how goofy it was in 1.0.

- - - Updated - - -

Well, if you talk about NERVA, being hot is the main feature of the engine! And since it heats by nuclear fission, it's probably doesn't matter how efficient is it in terms of waste heat. But I agree that spacecraft shouldn't heat up to 1000K, it's just ridiculous.

Does anybody knows what temperature is Merlin 1D Vacuum nozzle when it runs? You can see it being red hot, but that's just the nozzle, the rest of spacecraft doesn't heat like that. Granted, it's a chemical engine, it doesn't need to be as hot as the exhaust, unlike NERVA.

No, the exhaust of a NERVA engine is COOLER than the exhaust of a chemical engine like the Space Shuttle Main Engines. The core of the NERVA is solid material, which limits how hot you can let it get. The combustion gasses in a chemical engine can be hotter (you just need to keep the walls of the combustion chamber cool enough to not fail).

[Chris_2]

Coming late to the party... Yesterday, I spent an hour watching how heat flows around my vessel and... well.

Within a three minutes after igniting the Nerva, the whole vessel was insanely hot, even the most remote parts were over 1000K. IIRC, at that temperature the paint should come off, the metal should be glowing red, and the fuel in the tanks should long be gone. Shouldn't the whole thing radiate like mad? I'm expecting hundreds if not thousands of Watt/squaremeter. Likewise, I didn't calculate how much heat per thrust a LV-N gives you, but I expect it to be preposterous.

In a nutshell: I don't think we need a heatsink, we need much reduced heat generation or much better radiation or something. Or maybe even a whole new heat system. The current one seems credible as long as we're talking about re-entry heat (which probably is superficial); but a hot engine boiling the vessel through and through until batteries pop at the far end, well, that's not even wrong. ©

I don't know how 60kN of thrust translate into Joule or Watt, but I wonder if there isn't more energy going into the vessel than out of it.

I totally agree. The solution could be to decrease the thermal conduction between parts, so the heat becomes much more localized to where it is generated. That would be much more realistic.

[Kuu Lightwing]

Exhaust of NERVA is cooler, but the problem is that NERVA core HAS to be hotter than its exhaust, while chemical engine can (and usually will) be way cooler than its exhaust.

[Brotoro]

Exhaust of NERVA is cooler, but the problem is that NERVA core HAS to be hotter than its exhaust, while chemical engine can (and usually will) be way cooler than its exhaust.

No. Look up the numbers. The hottest end of the core of a NERVA is at 3,000 Kelvin. The combustion going on in the SSME gets up to 3588 Kelvin.

The exhaust is always cooler than the temperature in the combustion chamber for any rocket engine because part of what the nozzle does is speed up the exhaust gasses, and it does this at the expense of heat energy in the exhaust gasses.

[Kuu Lightwing]

Exhaust of the NERVA core can't be hotter than NERVA core, because NERVA core heats the gases. If the core not hotter, it would cool the gases. For chemical engines, you can cool the parts using regenerative cooling or whatever you like way below the temperature of gases. The temperature you listed for SSME chamber is the temperature of the gases, not engine parts. Inconel 718 (the combustion chamber material) melts at ~1700K, but it works because of regenerative cooling of the chamber.

[Brotoro]

Exhaust of the NERVA core can't be hotter than NERVA core, because NERVA core heats the gases. If the core not hotter, it would cool the gases. For chemical engines, you can cool the parts using regenerative cooling or whatever you like way below the temperature of gases. The temperature you listed for SSME chamber is the temperature of the gases, not engine parts. Inconel 718 (the combustion chamber material) melts at ~1700K, but it works because of regenerative cooling of the chamber.

Nobody said the exhaust of the NERVA is hotter than its core. In fact, I said just the opposite in my last post:

"The exhaust is always cooler than the temperature in the combustion chamber for any rocket engine because part of what the nozzle does is speed up the exhaust gasses, and it does this at the expense of heat energy in the exhaust gasses."

The walls of the NERVA engine's 'combustion' chamber, just like the walls of the SSME chemical engine's combustion chamber, use regenerative cooling.

[Kuu Lightwing]

My point was that some parts of the NERVA (core) are getting hotter than any part of the chemical engine. Walls could be cooler, okay, but you still need some crazy materials for the core. What did they use, tungsten?

However the thing I didn't think about is that the rest of the engine could be cooled down using the same methods, so technically NERVA can be build without overheating everything...

[Laie]

Yes, it's goofy. But you should have seen how goofy it was in 1.0.

I'm not complaining about the heat-generating engine (though I'd like to, but that's a different discussion to which I have nothing new to add).

This is about the high temperatures and how they spread across the vessel... to repeat myself: it looks good and seems credible in the context of re-entry heat, where one naturally assumes that it's surface temperature. 1200...1500...1700K on re-entry? No problem with that. I'm sometimes annoyed at how the heat quickly spreads throughout the vessel, reaching every nook and cranny even on the leeward side, but still, it's not bad.

However, in the context of parts running hot and dumping their heat into the vessel, this breaks down. Completely. If it says "this tank has 1200K" I assume that this is not only the surface, but the whole tank and it's contents. I don't have words for this, "silly" doesn't quite cut it.

Maybe it would help if parts had a surface and a core temperature. Maybe it would help if conductivity was lower, allowing you to have sections of your vessel that won't run hot (though others would, of course, grow even hotter). I don't know, and perhaps this is not the thread for it. But I do know that the current system is fundamentally flawed, and papering it over with a radiator part would only make it worse.

[Brotoro]

My point was that some parts of the NERVA (core) are getting hotter than any part of the chemical engine. Walls could be cooler, okay, but you still need some crazy materials for the core. What did they use, tungsten?

However the thing I didn't think about is that the rest of the engine could be cooled down using the same methods, so technically NERVA can be build without overheating everything...

The fuel elements of the NERVA (the parts that get to 3000K when the engine is running) were uranium dioxide embedded in graphite coated with niobium carbide. Later tests were done with zirconium carbide coatings.

To shutdown the NERVA, first the fission in the reactor core is stopped by rotating the control drums into position so that the neutron-absorbing sides face the core. Liquid hydrogen continues to run through the core for a while as the reactor core cools down (and the engine produces a decreasing amount of thrust during this phase). When shutdown (after cooldown), the temperature of the core is below 505 Celsius (according to the design documents), and generates less and less heat as the short-half-lived nuclides in the core decay away. All the heat spreads out to the rest of the engine's surface, its temperature will be even lower (not even red hot). During the operation of the engine, the outer parts of the engine will not be very hot because of the regenerative cooling by the liquid hydrogen, just like with a chemical engine (except for the nozzle, which is radiatively cooled).

[Dooz]

My NERVA Powered mun lander glows in the dark Luckily the small fins prevent it from overheating

But either radiators should be added or nerf the heat because the exhaust removes heat from the reactor core.

[Kuu Lightwing]

Thanks, for the additional information! So, NERVA doesn't need to be much hotter than the chemical engines externally and shouldn't heat the spacecraft like it does in KSP. About shutdown part, that's also an interesting piece of information, which I was thinking about - indeed with chemical engines, you just stop fuel/oxidizer and it will cool itself, while NERVA can't do that. Do you have any data on how long the shutdown period would be? It can affect the performance quite significantly... And makes the engine incapable of precision maneuvers like we do in KSP. Actually, even though overheating turns out to not be a realistic mechanic for NERVA, I would prefer it over not being able to do burns precisely...

[Laie]

My point was that some parts of the NERVA (core) are getting hotter than any part of the chemical engine.

This discussion has been had so often, and everytime the most difficult part is to figure out what the other side already knows, or believes. Please forgive me if I'm talking past you, but I don't know where you're coming from.

Basic functional principle of a NTR: cold propellant -> nuclear pile -> hot exhaust

The nuclear pile is nothing fancy: tubes filled with uranium pellets. You have to keep temperatures in a range where the tubes still hold. I can't tell you what material they used. Some data sheet I found online mentions temperatures of ~2400K. Wikipedia says over 2000K.

However, please note that a NTR is much simpler than the reactors that power the grid or a submarine. You don't move hot water from the pile to a turbine or something, but just heat the propellant and that's that. During operation, all heat should be carried away. There's a substantial amount of cold propellant coming in, I can't imagine that much heat could spread against this onrush. By the way, the last NTRs that were built were good for hours of continuous operation; the limit wasn't heat, but how much propellant you could store.

However, there is a heat problem: Uranium is radiaoctive, and the fission products even more so. Decay generates heat. So the time when you have to worry about heat is not when it's running, but rather when it's not. Especially for the first minutes after shutdown.

- - - Updated - - -

About shutdown part, that's also an interesting piece of information, which I was thinking about [...] Do you have any data on how long the shutdown period would be?

http://en.wikipedia.org/wiki/Decay_heat

This graph is applicable to reactors that have run for weeks or months at a time. In case of a NTR, it will probably drop faster. Can't tell you by how much.

Hmm. Maybe we should take this to the Science Subforum?

[Brotoro]

Thanks, for the additional information! So, NERVA doesn't need to be much hotter than the chemical engines externally and shouldn't heat the spacecraft like it does in KSP. About shutdown part, that's also an interesting piece of information, which I was thinking about - indeed with chemical engines, you just stop fuel/oxidizer and it will cool itself, while NERVA can't do that. Do you have any data on how long the shutdown period would be? It can affect the performance quite significantly... And makes the engine incapable of precision maneuvers like we do in KSP. Actually, even though overheating turns out to not be a realistic mechanic for NERVA, I would prefer it over not being able to do burns precisely...

In real-life, the NERVA could be used for precision burns requiring a certain impulse...you just need to factor in the impulse you get during the shutdown. What we have in KSP is a simplification (just like real chemical engines can't be instantly throttle from 0% to 100% with unlimited restarts). I always thought it would be fun if the nukes in KSP had a spool-up/spool-down lag like the jet engines do.

The shutdown phase of a NERVA engine took 45 seconds. This is where most of the tail off thrust occurs. This was followed by the cool down phase (with much lower thrust) that lasts anywhere from 9 seconds (if the main burn was 49 seconds long) to 158 seconds (if the main burn was 1497 seconds long)...or so the table says.

Edited May 7, 2015 by Brotoro

[Aerindel]

This has turned into a fantastic post.

Its also depressing to remember that the spaceprogram was over before I was born.

[Kuu Lightwing]

Laie, I understand that stuff. The thing I was wrong about is that I assumed that the entire engine would be very hot, but there's no need in that - only core which actually heats the hydrogen needs to withstand those temeratures - the rest of the engine would be just cooled the regular way. And thanks for the graph, but the main question is how long you need to pump hydrogen through it, because that means running it cooler and thus with reduced ISP.

[Brotoro]

Yes...the propellant used during the shutdown phase gets expended at a lower Isp. So long burns are more efficient for a NERVA.

[Kuu Lightwing]

In real-life, the NERVA could be used for precision burns requiring a certain impulse...you just need to factor in the impulse you get during the shutdown. What we have in KSP is a simplification (just like real chemical engines can't be instantly throttle from 0% to 100% with unlimited restarts). I always thought it would be fun if the nukes in KSP had a spool-up/spool-down lag like the jet engines do.

The shutdown phase of a NERVA engine took 45 seconds. This is where most of the tail off thrust occurs. This was followed by the cool down phase (with much lower thrust) that lasts anywhere from 9 seconds (if the main burn was 49 seconds long) to 158 seconds (if the main burn was 1497 seconds long)...or so the table says.

Thanks! Yes, I understand that it's simplification. Merlin only throttles down to 70% and requires TEA-TEB to restart it. But once again, since our missions are not programmed, but manually-controlled, that would not be very fun to play with...

Spool-up/down mechanic might be interesting, but would be quite hard to predict the exact burn time because of that. Maybe we would even need to use RCS to correct issues caused by that, not sure if many players would like it...

[Laie]

Well, looking at that graph I'd say a couple of minutes at worst. Possibly 30 seconds. We're definitely not talking about hours.

Someone once suggested spooling as the determining feature of in-game Nervas. You know, much like Jets do. This would make a lot more sense... and have you tried landing a jet-powered VTOL at any time?

EDIT to add: I'm just too damn slow.

Brotoro, where do you get that stuff? Are there links you could share?

Edited May 7, 2015 by Laie

[Brotoro]

Thanks! Yes, I understand that it's simplification. Merlin only throttles down to 70% and requires TEA-TEB to restart it. But once again, since our missions are not programmed, but manually-controlled, that would not be very fun to play with...

Spool-up/down mechanic might be interesting, but would be quite hard to predict the exact burn time because of that. Maybe we would even need to use RCS to correct issues caused by that, not sure if many players would like it...

Oh, yes. I'm not recommending it be included in stock KSP because it WOULD make the gameplay more difficult, and people wouldn't like it (just like I have trouble with the jet engine spool-down lag when I'm trying to land a ship vertically on jet engines). Besides...if I want to play with a spool-up/spool-down lag, I could edit that into the config file for the NERV and have fun...I don't want to rain on anybody else's parade.