Nerva performances

[Idobox]

I've been reading a bit about Nerva, and there are a few things I don't understand.

First, the ISP at sea level is really poor (well, still better than chemical). Is it a fundamental problem with using propellant that has lower molecular mass than the atmosphere, or is the result of a design optimized for vacuum operation?

Second, the NERVA 2 specs I could find say 380s ISP and 867kN, which gives us a mechanical power of 3.2GW, but the thermal power was 4.5GW. What happened to the remaining 1.3GW? That is a lot of heat, and it should cause all sorts of noticeable effects.

Third, there is power density, 4.5GW seems a lot, but the Space Shuttle main engine provided 1800kN and 366s of ISP at sea level, for a power of 6.5GW. How come a nuclear engine has a lower power density than a chemical one? Is it some limitation inherent to controlled nuclear reactions, or does it have to do with the difficulty of transferring that kind of power to a coolant?

And then, there is the thrust. Nuclear engines are heavy by design, but they have such great ISPs than single stage to orbit can be considered. The problem is that you need a decent TWR not to waste all your fuel on gravity drag. How difficult would it be to use different mixes of gases for different phases of flight, starting with heavier gases and moving to hydrogen once gravity drag stops being a big issue?

A good way to do so would be to add some oxygen to the flow, which would result in lower ISP (same temperature, higher average molecular mass), but higher thermal output (through combustion), hence a significantly higher thrust. The way I see it, oxygen would be added to the stream of hot hydrogen, because combustion instabilities are not something you want inside a nuclear reactor.

To get an idea about the numbers, the NERVA2 had a mass flow of about 230kg/s, roughly 114000 moles/s. Burning 1mole of H2 releases 240kJ, if we could burn it all, it would release about 27GW, which would make it the most powerful rocket engine ever, it would also require pumping 1.8t/s of O2. A more realistic design could burn 10% of the hydrogen, providing 2.7GW and requiring "only" 180kg/s of O2. Although the temperature would get higher than what a solid fuel nuke can survive, the larger molar mass would likely result in similar or lower ISP, especially at sea level, but a much larger thrust, making the engine a lot more versatile.

Edited November 2, 2014 by Idobox

[Kryten]

Second, the NERVA 2 specs I could find say 380s ISP and 867kN, which gives us a mechanical power of 3.2GW, but the thermal power was 4.5GW. What happened to the remaining 1.3GW? That is a lot of heat, and it should cause all sorts of noticeable effects.

Third, there is power density, 4.5GW seems a lot, but the Space Shuttle main engine provided 1800kN and 366s of ISP at sea level, for a power of 6.5GW. How come a nuclear engine has a lower power density than a chemical one? Is it some limitation inherent to controlled nuclear reactions, or does it have to do with the difficulty of transferring that kind of power to a coolant?

You just answered your own questions. In a chemical engine, the vast majority of the heat is generated within and carried away by the reaction products, you just have to worry about the small proportion radiated into the combustion chamber and nozzle. In a nuclear engine, much of the heat is generated away from the propellant and can't be easily removed by it, limiting the power levels that can be reached without the engine melting.

[K^2]

Injecting oxygen is a bad idea. It looks good until you ask yourself why they used hydrogen for propellant and not something that's much easier to store and handle. And the reason is that you get higher I_SP at the same chamber temperatures with lighter propellant. To get the same I_SP from water, which is what you'll get in exhaust if you start burning hydrogen, you need 3x higher chamber temperatures. And NERVA is already running close to the limit.

Even if you were to inject both oxygen and hydrogen from a NERVA-like system, and you happen to have some sort of an impossible to melt chamber to do the afterburn in, you'd still end up with worse performance than NERVA alone.

[Idobox]

You just answered your own questions. In a chemical engine, the vast majority of the heat is generated within and carried away by the reaction products, you just have to worry about the small proportion radiated into the combustion chamber and nozzle. In a nuclear engine, much of the heat is generated away from the propellant and can't be easily removed by it, limiting the power levels that can be reached without the engine melting.

Thanks for confirming my suspicion.

Injecting oxygen is a bad idea. It looks good until you ask yourself why they used hydrogen for propellant and not something that's much easier to store and handle. And the reason is that you get higher ISP at the same chamber temperatures with lighter propellant. To get the same ISP from water, which is what you'll get in exhaust if you start burning hydrogen, you need 3x higher chamber temperatures. And NERVA is already running close to the limit.

Even if you were to inject both oxygen and hydrogen from a NERVA-like system, and you happen to have some sort of an impossible to melt chamber to do the afterburn in, you'd still end up with worse performance than NERVA alone.

There is no question that hydrogen is optimal if you don't have to worry about gravity losses, and NERVA was planned to be used as a last stage, so it made complete sense. But if you want SSTO, you need a TWR larger than 1, and preferably around 2 or 3, otherwise you're going to waste an awful lost of deltaV to gravity drag.

My idea is to modulate the molar mass of the exhaust, a bit like VASIMR is supposed to do: high thrust low ISP at sea level to get off the ground, and once you're high enough, high ISP low thrust to get to orbital speed.

Nerva was limited by the temperature of the core, which would loose rigidity somewhere between 2500 and 3000K. RP1 Lox burns at 3700K, and that doesn't seem to be an impossible problem to solve for chemical rockets. And then again, the sea level ISP of a chemical rocket is already higher than that of Nerva.

[Kryten]

Somebody else had already had that idea, it's called LANTR.

[Guest]

But if you want SSTO, you need a TWR larger than 1, and preferably around 2 or 3, otherwise you're going to waste an awful lost of deltaV to gravity drag.

Until the atmosphere is fixed, KSP should drop the realism claim. You're wrong, 2 or 3 is way too much for any rocket that's not a missile, especially an SSTO. You want a SL TWR of 1.2, and you'll still have to throttle down at the end of the ascent.

The main idea with a LANTR is that you inject LOX into hydrogen flow after the reactor. You might heat it (and it's a good idea to do so), but no combustion takes place inside the reactor itself. Essentially, you turn your nuclear rocket into a chemical one, with performance being improved thanks to both propellants being pre-heated. Also note, you'll be running very fuel-rich anyway, like most chemical rockets do, so the exhaust would still be mostly hydrogen. Indeed, this is how they manage the 3700K combustion of LOX+RP1. The combustion chamber is never actually at that temperature, excess LOX (I think, it's got a lower molecular weight than RP1) flows through the engine, lowering the combustion temperature, but improving Isp thanks to its lower molecular weight. Hydrogen engines take this to extremes, with 3 times more LH2 being ran through the engine than the reaction equation would imply. A LANTR would give you a slightly higher thrust than a pure chemical rocket of similar characteristics, along with much better Isp. And it would work at SL just fine, though for SSTO use, it'd be good to install some altitude compensation (extendable nozzle, aerospike, you name it).

[K^2]

There is no question that hydrogen is optimal if you don't have to worry about gravity losses, and NERVA was planned to be used as a last stage, so it made complete sense. But if you want SSTO, you need a TWR larger than 1, and preferably around 2 or 3, otherwise you're going to waste an awful lost of deltaV to gravity drag.

My idea is to modulate the molar mass of the exhaust, a bit like VASIMR is supposed to do: high thrust low ISP at sea level to get off the ground, and once you're high enough, high ISP low thrust to get to orbital speed.

Hm. Might work. TWR of NERVA isn't as bad as I thought it'd be. I still think you'll get much better milage out of having a proper chemical first stage, but you might be able to make it into an SSTO with a chemical afterburner instead if you really wanted to. It'd be one big SSTO, though.

If it's just a matter of making it recyclable, I'd rather make first stage an entirely chemical space-plane. That stage wouldn't quite make it to orbit with payload, but it'd get the NERVA stage onto high enough trajectory to take over, and then return safely to the ground.

[NERVAfan]

Since we have a NERVA thread anyway...

...two questions. I love nuclear propulsion ideas (hence my username) but the problem is politics of launching nuclear material*. Could you mine thorium in the thorium-rich region of the moon to produce nuclear fuels?

And, if so, what kind of specific impulse would a NERVA using liquid oxygen get? I know it wouldn't be good, but oxygen is easily available on the moon whereas hydrogen and other light elements aren't (except in the polar craters). Would it be good enough to SSTO from the Moon (at which point you could fuel with hydrogen from Earth or asteroids or whatever).

*this is silly because we launch RTGs and plutonium is vastly more radioactive than U-235, but...

[K^2]

If isotope composition is about the same, then sure, there shouldn't be any problems with using Thorium for NTR fuel. In fact, haven't there been some discussions of Thorium-based NTR projects on this forum?

Oxygen is pretty bad for propellant, though. You'll get about 1/4 of the I_SP compared to hydrogen. All other things being the same, it drops as inverse square root of molecular mass of the propellant. Oxygen being 16 times heavier gives you that factor of 4 drop.

That's still well over 200s, though, so if it's abundant enough, it might still be a cheaper option.

[KerikBalm]

Until the atmosphere is fixed, KSP should drop the realism claim. You're wrong, 2 or 3 is way too much for any rocket that's not a missile, especially an SSTO. You want a SL TWR of 1.2, and you'll still have to throttle down at the end of the ascent.

The main idea with a LANTR is that you inject LOX into hydrogen flow after the reactor. You might heat it (and it's a good idea to do so), but no combustion takes place inside the reactor itself. Essentially, you turn your nuclear rocket into a chemical one, with performance being improved thanks to both propellants being pre-heated. Also note, you'll be running very fuel-rich anyway, like most chemical rockets do, so the exhaust would still be mostly hydrogen. Indeed, this is how they manage the 3700K combustion of LOX+RP1. The combustion chamber is never actually at that temperature, excess LOX (I think, it's got a lower molecular weight than RP1) flows through the engine, lowering the combustion temperature, but improving Isp thanks to its lower molecular weight. Hydrogen engines take this to extremes, with 3 times more LH2 being ran through the engine than the reaction equation would imply. A LANTR would give you a slightly higher thrust than a pure chemical rocket of similar characteristics, along with much better Isp. And it would work at SL just fine, though for SSTO use, it'd be good to install some altitude compensation (extendable nozzle, aerospike, you name it).

Ummm no.

The space shuttle had a TWR well over that. The engines throttle down during asccent to limit it to 4gs.

If you are doing vertical ascent, you want at least a TWR of 2.

[K^2]

Real rockets are a bad example of it, because they tend not to be very efficient away from full throttle. So it ends up being more efficient to lift up balls to the wall, rather than throttle down for optimal TWR.

TWR = 2 is optimal under two assumptions. Your ascent is perfectly vertical, and drag is quadratic. Neither of these are true. Quadratic drag model does not apply anywhere in the vicinity of transonic region, which is basically all of early ascent. Once the rocket gets into mach-independence region, you are well into gravity turn, and density starts to play a role as well.

At any rate, claiming that TWR should be exactly 2, or certainly over 2, for a real rocket is a bit reaching. But there will certainly be times when you need TWR in that vicinity during ascent. TWR of 1.2 is not going to cut it for a good ascent. Some rockets might be built to have a 1.2 from the pad, simply because they will also be running balls to the wall and will get up to higher TWR further in flight, but no rocket will ever be built to intentionally maintain 1.2 or anything like it.

[KerikBalm]

"TWR = 2 is optimal under two assumptions."

You forgot a 3rd assumption: the scale height is so high compared to your vertical velocity, that your terminal velocity does not increase significantly.

TWR of two will fight gravity drag + air drag at terminal velocity, but if terminal velocity is increasing at 500 m/s^2, a TWR of 2 is not going to cut it.

It won't cut it if terminal velocity is increasing at 50 m/s^2, and its still a bit sub optimal if TWR is increasing at 5 m/s^2 (though your lag behind terminal velocity won't be so bad in that case)

[K^2]

Yes, good catch. Thank you.

[11of10]

If isotope composition is about the same, then sure, there shouldn't be any problems with using Thorium for NTR fuel. In fact, haven't there been some discussions of Thorium-based NTR projects on this forum?

Oxygen is pretty bad for propellant, though. You'll get about 1/4 of the I_SP compared to hydrogen. All other things being the same, it drops as inverse square root of molecular mass of the propellant. Oxygen being 16 times heavier gives you that factor of 4 drop.

That's still well over 200s, though, so if it's abundant enough, it might still be a cheaper option.

You can't use O2, it's corrosive when hot. It'd wreck the engine.

[cantab]

"TWR = 2 is optimal under two assumptions."

You forgot a 3rd assumption: the scale height is so high compared to your vertical velocity, that your terminal velocity does not increase significantly.

And, I would say, there's a zeroth assumption here - what "optimal" means. In real life, "optimal" usually means cheap and reliable. If your rocket is cheap and reliable then you don't give two hoots that it uses up more delta-V and has less payload fraction. Considering that in general fuel is cheap, tankage is pretty cheap, and engines are expensive, then using as little thrust as you can get away with makes sense.

[Mazon Del]

Hmm, I could have sworn one of the proposed expansions on the original NERVA project was to combine the NERVA tech with a chemical rocket to see what sort of efficiency gain there was to be had. I could be wrong though.

[magnemoe]

Hmm, I could have sworn one of the proposed expansions on the original NERVA project was to combine the NERVA tech with a chemical rocket to see what sort of efficiency gain there was to be had. I could be wrong though.

You can add oxygen after the hydrogen leaves the reactor, this will work like an afterburner and have much of the same effect, it will increase trust but reduce ISP. it would be worth doing if you needed higher trust or if the oxygen was free anyway like if you mine ice from moon or asteroids.

[Mazon Del]

Ah! Thanks!

[rdfox]

The reason for the extremely low SL Isp. is not due to any fundamental problem with nuclear-powered reaction motors, but purely because it was (understandably!) considered politically unacceptable to be using nuclear rocket engines in any application that doesn't see them put into orbit. Consider how much the people of the Azores, for example, would like having Saturn V first stages powered by nuclear engines being dropped into the Atlantic just a few hundred miles away... and then consider what happens to Titusville and Cocoa Beach if the rocket explodes at tower clear when it's being driven by nuclear engines--the NERVA design wouldn't generate significant fallout in an explosion *before* it was lit, but once you powered up the reactor, it became, essentially, a flying Chernobyl. Indeed, even if you didn't have an abort, the pad and mobile launcher would basically be single-use items with a nuclear first stage, due to the sheer level of contamination rendering it unsafe to have personnel attempt to refurbish them or otherwise work on them. (The only reason it was considered safe for manned flights was the bulk of propellant and tankage that were between the reactor and the crew, attenuating the radiation with distance and acting as shielding. IIRC, the S-N stage designed for use included RCS tankage to keep it oriented with the engine facing directly away from the spacecraft until it was a few kilometers away.)

Nuclear reaction motors CAN be made to work well in an atmosphere; look up Project Pluto, a program to design a nuclear ramjet engine for an American cruise missile project of the 1950s. The only reason NERVA had nigh-nonexistent SL Isp. was because it was designed as a purely exoatmospheric engine, with the appropriate expansion ratio for that application.

[NERVAfan]

You can't use O2, it's corrosive when hot. It'd wreck the engine.

Oh - I thought you could make the reactor fuel elements out of oxides to prevent O2 issues (the Atomic Rockets site talks about that, I think).

Indeed, even if you didn't have an abort, the pad and mobile launcher would basically be single-use items with a nuclear first stage, due to the sheer level of contamination rendering it unsafe to have personnel attempt to refurbish them or otherwise work on them. (The only reason it was considered safe for manned flights was the bulk of propellant and tankage that were between the reactor and the crew, attenuating the radiation with distance and acting as shielding. IIRC, the S-N stage designed for use included RCS tankage to keep it oriented with the engine facing directly away from the spacecraft until it was a few kilometers away.).

A NERVA doesn't emit radioactive material unless something goes wrong, so wouldn't the pad be fine after the (terribly radioactive) active reactor departed?

[DerpenWolf]

Funny thing is I was just about to ask similar questions on NERVA before I saw this! So one thing I'm wondering is what the efficiency is for the engine if you use water instead of hydrogen (the engine is optimized for water as a propellent), does it have any advantages, whats its ISP? Also for the whole oxygen thing, why don't you just break up waste water you don't want to really recycle and use the Hydrogen for fuel and the oxygen for breathing (less strain on filters?). Would that work?

[Kulebron]

^^ You need water on board, and it's a bad idea to not recycle it. Disposing of water will be a significant waste, especially in long trips like to Mars or other planets. This will make a much bigger start mass.

The efficiency if 1 / square root( molecular mass ). MM of H is 1, MM of H2O is 18, so it's 4 times worse if same energy is applied.

[Kryten]

A NERVA doesn't emit radioactive material unless something goes wrong, so wouldn't the pad be fine after the (terribly radioactive) active reactor departed?

It would emit plenty of neutron radiation, which would produce radioactive material.

[Guest]

The space shuttle had a TWR well over that. The engines throttle down during asccent to limit it to 4gs.

If you are doing vertical ascent, you want at least a TWR of 2.

Launch TWR. On the Shuttle, it wasn't too large, either, and certainly less than 2. It did throttle down, but it was because its fuel depleted during flight and thus, its TWR increased. If you mean Engine TWR, then we're talking a different thing altogether, and 2 is too little for any practical lower stage engine. The real lower stage engines TWR range from 70 to over 150. So you're wrong either way. It's either 1.2-1.6 or 70+ depending on which TWR you mean.

Edited November 6, 2014 by Guest