Why do NERVA engines require heat radiators during a burn?

[SomeGuy123]

So you have a nuclear reactor core.  The core is very hot, and special ceramics and coatings separate the core elements from gas channels.  You send hydrogen gas through the channels.  The core is just on the verge of melting.  The hydrogen gas cools the core to just below melting, which superheats the gas and it exits the rocket at high velocity.

I actually know that for entropy reasons you need a heat radiator, I just can't quite grasp why.  If the rate of heat extraction = heat production, the reactor doesn't overheat as long as the hydrogen gas is flowing.

Sure, after the burn is finishes, the decay heat means you cannot throttle the reactor below about 30% power.  You need to vent the heat from 30% of the reactor's power output through heat radiators until it cools.  I just don't see why you need heat radiators during the burn.  And if you don't need heat radiators, a logical thing to do would be to throw away NERVA engines after a burn is complete, letting them melt down as they do so.

[magnemoe]

No expert however it might be to remove heat, not from the core but from the surrounding structure who also heat up and don't like the high temperatures.
Second is that its an good idea to be able to run the engine dry without an meltdown. 

[Streetwind]

45 minutes ago, SomeGuy123 said:

Sure, after the burn is finishes, the decay heat means you cannot throttle the reactor below about 30% power.  You need to vent the heat from 30% of the reactor's power output through heat radiators until it cools.  I just don't see why you need heat radiators during the burn.  And if you don't need heat radiators, a logical thing to do would be to throw away NERVA engines after a burn is complete, letting them melt down as they do so.

That is not really a logical thing, no. Not when you specifically chose to add an engine more than ten times as expensive as a chemical alternative to your spacecraft.

You might as well suggest to simply switch off a nuclear power plant on Earth after it burned through one batch of fuel rods and build a new one instead of refueling the old one, just because removing and storing spent fuel rods is such a major hassle

Using a nuclear thermal rocket for a single burn is, I am sorry to say, quite far away from getting a proper return for your investment. If you can't refuel it, developing and deploying a NTR is (in my opinion) pointless in all but one single use case: the "money is no object for this mission, do whatever it takes" use case. The aerospace world hasn't done that since the Apollo program.

Also, even non-refuelling missions with NTRs in mind, such as those explored in NASA's Mars Design Reference Architectures (which I don't think make economic sense), use the NTRs for the entire mission - single-stage Earth-Mars-Earth, with a drop tank. The point is, for it to even begin to be a valid option, you need to milk this engine for all it is worth, using it for every single burn in your entire mission profile. Even if it's projected to take multiple years.

Edited August 22, 2016 by Streetwind

[DDE]

16 minutes ago, Streetwind said:

You might as well suggest to simply switch off a nuclear power plant on Earth after it burned through one batch of fuel rods and build a new one instead of refueling the old one, just because removing and storing spent fuel rods is such a major hassle

Well, the problem is, many NERVA models burn through their fuel rods in a single burn. Hence all the multi-stage designs from the 1960s.

@SomeGuy123, while nominally the propellant is supposed to be the coolant, the radiators help with any escaping heat, and then there are the start-up/cool-down cycles, which require either radiators or active cooling at the cost of propellant.

Classic NERVA designs indeed have no radiators but a pair of opposed nozzles that I think are used for expendable coolant.

Edited August 22, 2016 by DDE

[shynung]

2 hours ago, DDE said:

Well, the problem is, many NERVA models burn through their fuel rods in a single burn. Hence all the multi-stage designs from the 1960s.

I thought the NERVA designs were rated for 4-5 hours of core life at maximum output? Surely a Terra-Mars burn wouldn't last even an hour?

[TheSaint]

2 hours ago, DDE said:

Well, the problem is, many NERVA models burn through their fuel rods in a single burn. Hence all the multi-stage designs from the 1960s.

I don't believe that was true, I think that they were originally designed for multiple burns. But even if they weren't, that's just a design issue. 

Actually, IIRC, the plan with the original NERVA was to have non-directional nozzles (i.e. two opposing nozzles on the sides of the reactor vessel) that would allow them to vent reaction mass through the reactor after shutdown to control decay heat. You can also shut down the reactor near the end of the burn to use the beginning of the decay heat cycle for the end of your burn. 

Yeah, the whole "Nuclear engines need radiators" isn't a reality thing, it's a play balance thing. I find it annoying too. Nuclear power plants don't use radiators to cool the reactor, they use it to cool the steam that is exhausted from the steam turbines. The reactors are completely cooled by the coolant that is flowing through the core. There is no reason to believe that a nuclear rocket engine would operate any differently. 

[Laie]

2 hours ago, TheSaint said:

the whole "Nuclear engines need radiators" isn't a reality thing, it's a play balance thing.

Exactly.

While running, the heat goes into the propellant, that's the whole point. Last time I checked, even the amount of cooling required after shutdown was surprisingly small. The inventory after a short high-power run (or several) is quite different from what you get in a power plant that's effectively simmering for weeks or months.

[DDE]

@Laie, but, err, since when the NERVs in KSP need radiators? Seriously, do people use huge clusters of them? Am I the only one who doesn't have and never had overheating issues?

[Bill Phil]

45 minutes ago, DDE said:

@Laie, but, err, since when the NERVs in KSP need radiators? Seriously, do people use huge clusters of them? Am I the only one who doesn't have and never had overheating issues?

It was kind of insane for a short time, since they had a tendency to blow up.

[Laie]

@DDE Browse the forum a bit, I think it was when 1.0 came out. Lots of heat from nukes and no means to get rid of it -- they'd blow up within two minutes or so. People quickly learned how to use wings as radiators, but that meant that every nuke necessitated, like, 10 parts? Much outcry. Eventually the devs gave in, the same update that gave us radiators also reduced LV-N heat so much that you could often do without.

I don't play stock much, but last time I did their heat couldn't be completely ignored, at least not with burn times of several minutes.

[magnemoe]

46 minutes ago, Laie said:

@DDE Browse the forum a bit, I think it was when 1.0 came out. Lots of heat from nukes and no means to get rid of it -- they'd blow up within two minutes or so. People quickly learned how to use wings as radiators, but that meant that every nuke necessitated, like, 10 parts? Much outcry. Eventually the devs gave in, the same update that gave us radiators also reduced LV-N heat so much that you could often do without.

I don't play stock much, but last time I did their heat couldn't be completely ignored, at least not with burn times of several minutes.

Yes, nowday its not an major issue, if you have an low mass LV-N craft at Moho its smart to cool it as it run hot, same if you plan for an 10 minutes burn, else its no issue. 
Back at 1.0 the LV-N required lots of cooling, the twin boar also run very hot, but you mostly dropped it before going into orbit. 

And I thought someguy talked about real world engines there I see that radiators has some mission for more efficient and safer operations. 

[tater]

The later LANL designs solved many of the ablation of nuclear fuel issues, actually. I think they are fundamentally limited by propellant issues (not the least of which would be boil off).

[jwenting]

However good your thermal efficiency is in transferring heat to the propellant gas, it's not perfect so excess heat builds up and you need to get rid of it. Hence the requirement for additional cooling systems.

[wumpus]

51 minutes ago, jwenting said:

However good your thermal efficiency is in transferring heat to the propellant gas, it's not perfect so excess heat builds up and you need to get rid of it. Hence the requirement for additional cooling systems.

While this is certainly true, a better question is if radiators can begin to overcome it.  It looks more and more like single use nerva stages are much more likely (regardless of the cost).  The obvious other questions are how you cool the reactor *after* shutdown.  You need to separate the rods sufficiently to shut down the reaction, and then somehow cool the rods as well.  As well as [presumably slowly] cooling the entire structure.  Eject before meltdown keeps looking better and better.

As far as black body cooling while in operation, you might need it but don't expect significant cooling (such cooling might just stretch you to your goal, but little more).

[Laie]

2 hours ago, jwenting said:

heat builds up

[citation needed]

Yup, it's hot in there. It's also hot in any other rocket motor, or at least the chemical ones. Actually, the burners have even higher temperatures in their combustion chamber than the Nerva would reach on any part.

Does any of that heat carry over to other parts? You bet.

It it so much that it has to be dealt with in the shape of forced cooling? Last I heard, no. At least when I watch a rocket launch, I fail to notice any radiators.

There's no reason why more heat should seep out of a NTR compared to a Kerolox burner. At least while it's running. Decay heat after shutdown is another matter, of course.

[Brotoro]

The decay heat of the NERVA engine was to be handled by a "cool down" phase where they would send pressure-fed hydrogen through the reactor core in a number of pulses over many hours (or as long as a few days) to carry off the heat. The number of pulses and length of the cool down phase would depend on the length of the previous burn. When used as a nuclear tug to carry payloads to the Moon, for example, the NERVA might do over a hundred cool down pulses during the three-day cruise.  Once the NERVA reactor has cooled to the point where the decay heat can be handled by the radiant losses from the engine's outer casing, the pulses are no longer needed.

The cool down pulses would produce thrust (the hydrogen is expelled through the main nozzle...those two small canted nozzles you see on the NERVA are where the hydrogen from the two turbines of the redundant turbo pumps gets exhausted), perhaps totaling as much as a few percent of the main burn's impulse, and this would be calculated into the length of the main burn needed. They would not vent the hydrogen in opposing nozzles (as some have suggested here), because that would just be wasting good impluse. 

[Bill Phil]

On ‎8‎/‎23‎/‎2016 at 3:03 PM, Laie said:

[citation needed]

Yup, it's hot in there. It's also hot in any other rocket motor, or at least the chemical ones. Actually, the burners have even higher temperatures in their combustion chamber than the Nerva would reach on any part.

Does any of that heat carry over to other parts? You bet.

It it so much that it has to be dealt with in the shape of forced cooling? Last I heard, no. At least when I watch a rocket launch, I fail to notice any radiators.

There's no reason why more heat should seep out of a NTR compared to a Kerolox burner. At least while it's running. Decay heat after shutdown is another matter, of course.

Thing is, hydrogen is a bad coolant, and you're not running much of it through the reactor anyways. So, while the temperature may not be much higher, the heat flux likely is.

[Laie]

On 8/24/2016 at 11:10 PM, Bill Phil said:

Thing is, hydrogen is a bad coolant, and you're not running much of it through the reactor anyways. So, while the temperature may not be much higher, the heat flux likely is.

What the...?

You're aware that you have control over both the heat flux from the reactor, and the amount of hydrogen you pipe through it, right? There's no reason why they shouldn't match each other; actually, it would be a bad design if they didn't.

[Bill Phil]

2 hours ago, Laie said:

What the...?

You're aware that you have control over both the heat flux from the reactor, and the amount of hydrogen you pipe through it, right? There's no reason why they shouldn't match each other; actually, it would be a bad design if they didn't.

There's a maximum amount of hydrogen you can push through it.

[KSK]

2 hours ago, Laie said:

What the...?

You're aware that you have control over both the heat flux from the reactor, and the amount of hydrogen you pipe through it, right? There's no reason why they shouldn't match each other; actually, it would be a bad design if they didn't.

 I've been reading up on this recently and found a very helpful thesis on simulated control system designs for NERVA startup. Can't find it right now but I'll have another look when I get home. Basically, you have a set of control drums around the reactor core, which are mostly neutron absorbing but include a wedge of neutron reflecting material. Rotating the control drums lets you control the amount of neutron flux reflected back into the reactor and thus the intensity (for want of a better word) of the fission reaction and the amount of heat created by that fission reaction.

Because hydrogen is also a neutron moderator, engine start involves a delicate balance between propellant flow, heat absorption by that propellant and reactor criticality. So yes heat flux and hydrogen flow are matched and have to be matched unless you want a nicely molten NERVA.

Edited August 26, 2016 by KSK