Nuclear Salt Water Rockets!

[Wubslin]

Hi all,

I've been fairly certain that the unnamed "torch drive" engine that has yet to be announced is in fact Robert Zubrin's own nuclear salt water rocket. For those who aren't aware, the basic concept is that an aqueous solution of an ionic salt of fissile material (usually Uranium) is pumped into a nozzle or reaction chamber where it undergoes a supercritical reaction to create thrust from the water. In short, it's basically an Orion drive except it's a continuous nuclear explosion rather than intermittent pulses. Suffice it to say, there's serious potential for this to be a high thrust and high specific impulse torch drive.

That image is via the old game informer interview, by the way. I feel fairly confident this is what these engines are, especially considering there appear to be dozens of little apertures dotting the inside of the engines' nozzles. This "shower head" configuration would agree with what prior to KSP 2 being announced were basically the only two concept sketches in existence for such a rocket engine:

Seeing as this is a really cool rocket engine that would deservedly be the brachistochrone-between-planets Expanse drive that we all wish for in KSP 2, I thought I would give some thought as to how I personally would design such an engine.

This is my concept for a fully self-propelled nuclear salt water rocket engine. It runs on an aqueous solution of enriched Uranium Bromide, as well as a second aqueous mixture which contains a dopant that is exceedingly good at absorbing X-rays. As opposed to the NSWR engine that appears in Nertea's Far Future Technologies (I love your mod!) and what I expect will show up in KSP 2, this engine requires no external electrical input to run motors. In fact, this engine will actually generate large amounts of electrical current when operating. It 's sort of a cross between the steam Rankine cycle that's likely keeping your lights on this very moment, and the old expander cycle rocket engine.

"Now wait a minute", you might ask, "I know a thing or two about the expander cycle! Because of the square cube law, those things can only get so big. And surely a machine that makes a constant nuclear explosion isn't going to be tiny, right?" 

It is correct that under normal circumstances, at larger scales the expander cycle has insufficient surface area for even normal chemical propellants to be pumped. The key lies in the way heat transport occurs within this engine. Because the fission reaction is hot enough to produce X-rays and the interior of the nozzle is constructed from a material which is relatively transparent to these X-rays, heat transfer actually occurs volumetrically through the thickness of the nozzle channel rather than convectively through only its inner surface. Of course there is no math here but I'm assuming the heat picked up through this process will be enough to expand the shield water into a dry, intense superheated steam. One consequence of the radiative heating is that the inner wall containing the cooling pores must relatively delicate and a balance between the extreme pressures on either side of it are required to avoid collapse. Steam which went into turning the turbine is sent through a bank of radiators - because no spaceship is cool without glowing orange panels on it somewhere. The radiators allow the steam to condense down into a liquid again and be sent back into the shield water pump.

One thing I think is worth noting about this concept is that it is very miserly when it comes to dumping material overboard. The shield water which serves to both power the entire engine through turbines and also insulate the engine walls from the furious heat of the X-rays is actually the main reaction mass of the engine. Since the nuclear reaction takes up only a tiny part of the total mass flow rate through this engine, it makes sense for the vast majority of the water in the system to be of the simple inert shielding variety.

If we take a closer look at the nozzle itself, you'll see that I've taken a far different approach than what appears in the material we have so far for Kerbal Space Program 2. Rather than support the reaction inside a combustion chamber of some kind or even a more focused nozzle, I feel that the conditions generated by the fission reaction will be too extreme for anything except the slightest deviation from the classic Orion drive pusher plate. You'll also see that, outside of the main laminar flow jacket, shield water is injected normal to the inside of the nozzle rather than parallel to the flow of the nuclear fuel. The intention here is that a sort of "stagnation zone" sets up that ensures shield water does not truly exit the engine until it's hot enough to either repel the steam front developing inside the nozzle walls or else eject out the back at a significant clip.

So what do you guys think? I of course am satisfied with the design of the engine as it appears in game now, but I thought this was a fun thought exercise and really got me thinking about the possibilities.

Edited September 3, 2021 by Wubslin

[Zaffre]

Good post. The models themselves look amazing (how did I miss such a wonderful sneak peek?). The bits inside the cone do seem to suggest a jacket of water being injected. The smaller of the two is probably a lower tech and thus lower ISP, but it may also be a more atmosphere friendly version. Not that you'd be lighting one of these monsters near a populated area. I imagine the one on the right will have an ISP in the high 4, probably 5 digit range. The larger one will probably be the holy grail sitting at 500,000 ISP, which is absolutely reaching scifi torch drive levels. It's also likely vacuum only and probably expensive to run, but who cares?

I'm also curious if we'll see LSWR as a variant, though for simplicity it may just be NSW. I'm interested to see these things fully rendered.

I imagine we'll see some other torch drives such as Ultra Dense Deuterium, high-grade Orion/Medusa, axial flow Z-pinch fusion, the ever-so-coveted Inertial Confinement Fusion pulse, and perhaps a D-D-D breeder. Maybe even a CCGCNR? Thought that isn't quite int he same league as these other ones.

[Wubslin]

1 hour ago, Zaffre said:

Good post. The models themselves look amazing (how did I miss such a wonderful sneak peek?). The bits inside the cone do seem to suggest a jacket of water being injected. The smaller of the two is probably a lower tech and thus lower ISP, but it may also be a more atmosphere friendly version. Not that you'd be lighting one of these monsters near a populated area. I imagine the one on the right will have an ISP in the high 4, probably 5 digit range. The larger one will probably be the holy grail sitting at 500,000 ISP, which is absolutely reaching scifi torch drive levels. It's also likely vacuum only and probably expensive to run, but who cares?

I'm also curious if we'll see LSWR as a variant, though for simplicity it may just be NSW. I'm interested to see these things fully rendered.

I imagine we'll see some other torch drives such as Ultra Dense Deuterium, high-grade Orion/Medusa, axial flow Z-pinch fusion, the ever-so-coveted Inertial Confinement Fusion pulse, and perhaps a D-D-D breeder. Maybe even a CCGCNR? Thought that isn't quite int he same league as these other ones.

Ha, thanks. I think 500,000 seconds might be a bit extreme for these engines. After all, they are in the end thermal rockets that use water as reaction mass. I would imagine these guys have maybe 10,000 or 20,000 seconds of Isp at the very most - the real engine you'd be conducting interstellar flight with would definitely be that Daedalus drive we've been shown. Not that the NSWR engines won't be great for doing flip n' burns between planets in a given solar system, though.

Also,

Quote

it may also be a more atmosphere friendly version

AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

[SciMan]

I mean this as the highest form of praise, but have you considered talking with the guy running the Atomic Rockets website (Winchell Chung)
He'd likely be QUITE interested in your take on the NSWR as seen from a "plumbing" perspective and not just a "nuclear salt water goes in, "stuff" happens, thrust comes out" perspective like the existing data on it seems to be.

What I mean is that it actually attempts to solve problems like "how do you get the blasted thing to run?" and "how do you make it so it (hopefully) doesn't explode?", and with that 2nd sketch, the most important one "How do you get around the fact that you have to contain a continuously detonating nuclear fission explosion?"

The "Perforated Wok nozzle" as I'm going to call that arrangement in your 2nd sketch is an especially good bit of conceptual engineering.

Hopefully Winchell Chung could help you figure out the math, or help you figure out who you need to talk to to get the math right. Your concepts are great, but without the math behind them we wouldn't be able to judge if a Kerballized version of them would be reasonable or not.

EDIT: Oh yeah the back-side of that nozzle in the 2nd sketch would be a really great place to put a radiation shadow shield, which you'd need even if it was "just" an unmanned interstellar flyby probe (if you're going by the upper limits of ~500k ISP), because no electronics we know of would be able to survive that amount of radiation.

... Well, maybe you don't "need" the shadow shield, but you'd WANT it if your avionics are any more complicated than a Babbage engine (which is a nice way around the radiation problem, provided you can make it out of MEMS things instead of electronics to keep the weight down).

Edited September 2, 2021 by SciMan

[Nate Simpson]

[Davi SDF]

On 9/1/2021 at 4:22 PM, Wubslin said:

AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

What? You really expected us to not commit those Warcrimes?

56 minutes ago, Nate Simpson said:

[Wubslin]

3 hours ago, SciMan said:

I mean this as the highest form of praise, but have you considered talking with the guy running the Atomic Rockets website (Winchell Chung)
He'd likely be QUITE interested in your take on the NSWR as seen from a "plumbing" perspective and not just a "nuclear salt water goes in, "stuff" happens, thrust comes out" perspective like the existing data on it seems to be.

What I mean is that it actually attempts to solve problems like "how do you get the blasted thing to run?" and "how do you make it so it (hopefully) doesn't explode?", and with that 2nd sketch, the most important one "How do you get around the fact that you have to contain a continuously detonating nuclear fission explosion?"

The "Perforated Wok nozzle" as I'm going to call that arrangement in your 2nd sketch is an especially good bit of conceptual engineering.

Hopefully Winchell Chung could help you figure out the math, or help you figure out who you need to talk to to get the math right. Your concepts are great, but without the math behind them we wouldn't be able to judge if a Kerballized version of them would be reasonable or not.

EDIT: Oh yeah the back-side of that nozzle in the 2nd sketch would be a really great place to put a radiation shadow shield, which you'd need even if it was "just" an unmanned interstellar flyby probe (if you're going by the upper limits of ~500k ISP), because no electronics we know of would be able to survive that amount of radiation.

... Well, maybe you don't "need" the shadow shield, but you'd WANT it if your avionics are any more complicated than a Babbage engine (which is a nice way around the radiation problem, provided you can make it out of MEMS things instead of electronics to keep the weight down).

Thank you! And I don't know how I would take this wrong, considering everything I posted is completely unvalidated and you have simply pointed out this much. Yes, there are a few unsolved questions in the design. For one, it's basically a given that any prompt (aka not kept alive by the decay of daughter products like in modern power reactors) fission reaction is going to be able to propagate through the UBr4 solution far faster than any real turbopump could reasonably move it. This leads to a few things. For one, the reaction will basically climb up as close as it can to the laminar nozzle which will basically be one neutron MFP away. At this distance, the radiative heat transfer to said nozzle will be extreme. 

One way of solving this problem would be to have the nozzle be ablative and consumable, with a very large length of honeycomb constantly extruding out while the engine is in operation like a thermic lance. Another option would be to take the UBr4 nuclear fuel and mix it with an extreme amount of the dopant responsible for making the normal shield water absorb x-rays. Transport of energy out of the reaction due to x-rays will be inhibited at first, so there's no telling what kind of effect such a design change might have. The third option, which is the option I imagine to be more sensible and one which I didn't write about, is to actually have the stream of nuclear salt water be subcritical on its own and require the addition of more neutrons from an external supply like a gun pointed at the flow. That way the reactivity could be carefully controlled, and also be able to happen further away from the nozzle itself. How you can shoot a beam of neutrons through so much heat and water, I have no idea.

Another unsolved problem is the specific details regarding heat transport from the reaction to the shield water flowing through the engine bell. Excuse my language here but to quote the old plumber's adage, "excrements don't flow uphill". Simply put, the highest pressure zone in the entire coolant loop needs to be the exit of the turbopump moving the shield water.  In addition, the entirety of the layer of water flowing down through the engine bell needs to be at a higher pressure than the zone lining its inside, so that water will actually exit the pores to receive heat and turn into reaction mass while cooling the rocket. Allowing all of this to happen at the same time as the water turns into basically the stuff that came out the back of the Saturn V and have it constantly be a downward gradient in pressure is difficult, to say the least. Not to mention, the fact that the inner lining of the nozzle needs to have a low enough atomic number and be thin enough that the x-rays skip right over it and absorb into the shield water layer inside means withstanding those conditions will be even harder. The coolant heat injection process is almost like an inside-out nuclear light bulb.

2 hours ago, Nate Simpson said:

Ha, I'm glad it got your attention. By the way, I'm sorry that interview I had with you guys didn't pan out this summer. Suffice it to say, I traveled the other path and now I'm pursuing a master's degree. While doing it I'm actually conducting research for my university that's funded by NASA in support of Orion and Artemis. How cool is that? Anyway, say hi to Shana for me!

Edit: My plumber's adage appears to have been autocensored. Shame

Edited September 3, 2021 by Wubslin

[Hyperspace Industries]

Nuclear salt water rockets, or as I've nicknamed them "Chernobyl drives" are probably the most kerbal way to make a torch drive and they are more realistic than antimatter drives, so, I wouldn't be surprised if this theory about the torch drives being Chernobyl drives (or C-drives for short) in ksp2 is correct.

By the way, your drawings and concepts for a Chernobyl drive are awesome!!! 

[Wubslin]

On 9/5/2021 at 7:50 AM, Hyperspace Industries said:

Nuclear salt water rockets, or as I've nicknamed them "Chernobyl drives" are probably the most kerbal way to make a torch drive and they are more realistic than antimatter drives, so, I wouldn't be surprised if this theory about the torch drives being Chernobyl drives (or C-drives for short) in ksp2 is correct.

By the way, your drawings and concepts for a Chernobyl drive are awesome!!! 

I dunno, "Chernobyl drive" feels a tiny bit in poor taste to me. Besides, I'd argue it doesn't even do the system justice in terms of power output. If you wanted to do historically inspired names though, U.S. nuclear test cycles always had a nice ring to them. You could call it the 'Ivy King' for the largest pure fission test ever conducted or maybe the 'Buster Jangle' just because that's a super fun name.  In keeping with the KSP 1 convention there'll probably be some kind of creature-themed title. What do you call this vacuum ripper of an engine?  The only thing that comes to my mind is atomic breath, so my money's on it being called the 'Godzilla'.

[AtomicTech]

I'm down for a sip of "Salt Water".

46 minutes ago, Wubslin said:

I dunno, "Chernobyl drive" feels a tiny bit in poor taste to me. Besides, I'd argue it doesn't even do the system justice in terms of power output. If you wanted to do historically inspired names though, U.S. nuclear test cycles always had a nice ring to them. You could call it the 'Ivy King' for the largest pure fission test ever conducted or maybe the 'Buster Jangle' just because that's a super fun name.  In keeping with the KSP 1 convention there'll probably be some kind of creature-themed title. What do you call this vacuum ripper of an engine?  The only thing that comes to my mind is atomic breath, so my money's on it being called the 'Godzilla'.

I like "Buster Jangle", it's really fun to say!

[Wubslin]

7 hours ago, AtomicTech said:

I'm down for a sip of "Salt Water".

I like "Buster Jangle", it's really fun to say!

Actually, on second thought that one's already taken. Back in high school I had come up with an idea for a nuclear powered sniper rifle that was guaranteed not to miss its target when you pulled the trigger. I named it the 'Buster Jangle'. 

 

[AtomicTech]

11 hours ago, Wubslin said:

Actually, on second thought that one's already taken. Back in high school I had come up with an idea for a nuclear powered sniper rifle that was guaranteed not to miss its target when you pulled the trigger. I named it the 'Buster Jangle'. 

 

Hmmm, that's a good flag idea

[Nertea]

On 9/1/2021 at 10:56 AM, Wubslin said:

If we take a closer look at the nozzle itself, you'll see that I've taken a far different approach than what appears in the material we have so far for Kerbal Space Program 2. Rather than support the reaction inside a combustion chamber of some kind or even a more focused nozzle, I feel that the conditions generated by the fission reaction will be too extreme for anything except the slightest deviation from the classic Orion drive pusher plate. You'll also see that, outside of the main laminar flow jacket, shield water is injected normal to the inside of the nozzle rather than parallel to the flow of the nuclear fuel. The intention here is that a sort of "stagnation zone" sets up that ensures shield water does not truly exit the engine until it's hot enough to either repel the steam front developing inside the nozzle walls or else eject out the back at a significant clip.

 

This is sweet and I kinda wish I'd seen this when I developed my larger NSWR. 

[Wubslin]

3 hours ago, Nertea said:

This is sweet and I kinda wish I'd seen this when I developed my larger NSWR. 

Hey, thanks! I'm a HUGE fan of your work, by the way. Seriously, No install of mine ever feels complete to me without the complete near and far future packs. Sometimes I'll spend 10 straight minutes single staging a nuclear salt boat into Low Earth Orbit just so I can admire the plume work and the soft glow of those radiators. Not to mention that without the Heinlein engine, drunkenly tooling around the outer Solar System would have been way harder. The Intercept team is sorely missing out not having you on board.

[Nertea]

17 hours ago, Wubslin said:

Hey, thanks! I'm a HUGE fan of your work, by the way. Seriously, No install of mine ever feels complete to me without the complete near and far future packs. Sometimes I'll spend 10 straight minutes single staging a nuclear salt boat into Low Earth Orbit just so I can admire the plume work and the soft glow of those radiators. 

Thanks, but I think the new FX in the latest FFT version is better!

I mean conceptually taking the direction of pusher plate vs. magnetic nozzle is the interesting trade. Nobody has studied the silly engine enough to say but I had taken the general idea that the plasma-fied exhaust would be sufficiently ionized for exhaust direction and the skeletal nature of the nozzle would reduce direct heating of the nozzle coils. You could then run the added water through each coil's reaction-area oriented local shield, which would help with protecting the coils from neutron damage further. You could also immediately spray the shield water directly along the pipe to avoid the challenges of recirculating it into the area directly at the reaction point (less stuff to flow 'uphill' so to speak, and well, you have impulse to spare). This may only be suitable for an up to 50-60% enriched engine as I have modeled though.

 

17 hours ago, Wubslin said:

The Intercept team is sorely missing out not having you on board.

Edited September 10, 2021 by Nertea

[sturmhauke]

On 9/2/2021 at 4:20 PM, SciMan said:

... Well, maybe you don't "need" the shadow shield, but you'd WANT it if your avionics are any more complicated than a Babbage engine (which is a nice way around the radiation problem, provided you can make it out of MEMS things instead of electronics to keep the weight down).

Is there anything more Kerbal than pairing 23nd century rocket technology with 19th century computing technology?

54 minutes ago, Nertea said:

It was Nertea all along!

[Wubslin]

On 9/10/2021 at 12:18 PM, Nertea said:

Thanks, but I think the new FX in the latest FFT version is better!

Oh my god, that's amazing! And yeah, if you can direct a completely ionized exhaust via magnetic fields then that would obviously be the best shot at preventing your engines from picking up heat and getting destroyed.  I'm not exactly sure where you'd get the current to run those nozzles without an external power supply though. Maybe absorptive surfaces could still be placed near or just downstream of the nozzle such that a traditional Rankine power plant could be run? You would probably want to swap NaK instead of water for miniaturization's sake. There's certainly no making the nozzle from superconductors, considering how hot it gets. Maybe electrical current could somehow be tapped off of moving charged exhaust products? I have no idea, especially considering the exhaust has no net electrical charge. My knowledge of plasma dynamics stops about at the right hand rule. I do agree though that running shield water through the inside-facing walls of the magnetic coil nozzles would be a good idea to prevent overheating and neutron embrittlement like you say. Maybe if you have some extreme pressure of water you can tap off from somewhere then perforations could be added to create a hybrid of the two designs. Who's to say?

By the way, have you thought about adding some more expanded, less luminous "cones" of exhaust coming out the back in vacuum to represent where the confinement ends? For such an extreme exhaust I imagine the plasma/gas/reaction mass would still be under an immense amount of pressure as it leaves the nozzle. It would definitely be cool to see that. Maybe even that central exhaust cone could be seen to peel away with its angle of divergence increasing slightly as the gas expands? I don't know how modding works so don't listen to me if I'm not making any sense.

 

On 9/10/2021 at 12:18 PM, Nertea said:

Okay, all is well in the world.

On 9/10/2021 at 1:12 PM, sturmhauke said:

Is there anything more Kerbal than pairing 23nd century rocket technology with 19th century computing technology?

Don't freak out or anything, but I've had this idea kicking around in my head of a spacecraft that didn't have a single electrical circuit anywhere. Not one blinky light or battery or wire or instance of V=IR or anything. I'm talking life support from Einstein refrigerators that are powered by waste heat from a fission reactor and regulated via an analog computer loop made out of refrigerant in lines. Cockpits with actual accelerator pedals which are mechanical linkages coupled directly to the rotating control rods and LH2 valves in the solid core NTR reactor. Gyroscopic navigation powered by H2 boiloff going into high speed turbines like you might find in a dentist's office. And of course everything needs a brass, steampunk 20,000 leagues under the sea aesthetic. The very first PID controllers to be invented used compressed air to close their control loop. Who's to say that couldn't be done again in three axes to give us SAS authority?

[sturmhauke]

I'd play a steampunk mod if it was done well.

[Wubslin]

Okay wow. That's a lot of Isp, Nert. Just yesterday I flew out to Enceladus using your new engine without one maneuver node and I've expended, what, like 3 percent of my fuel? This thing has like half a million meters of delta-v.  Holy smokes, and it looks awesome too.

Thing's ridonculous. Also, I hope the KSP 2 devs make reactors that shut down automatically when you're idle for a really long time. Imagine getting all the way out here, and then you do a three year coast only to find you have no power because you forgot to turn the reactor off and it slowly petered out of fuel after all that time. After smashing F9 and having nothing happen, I could almost feel the clown wig descending down onto my head! If only I'd known how dumb 120,000 seconds is I could've dispensed with 95% of the fuel and gotten the mission done via 10 minutes of 5 G burns rather than 100 minutes of 0.5 G burns. 

[Nertea]

On 9/13/2021 at 10:34 PM, Wubslin said:

Thing's ridonculous. Also, I hope the KSP 2 devs make reactors that shut down automatically when you're idle for a really long time. Imagine getting all the way out here, and then you do a three year coast only to find you have no power because you forgot to turn the reactor off and it slowly petered out of fuel after all that time. After smashing F9 and having nothing happen, I could almost feel the clown wig descending down onto my head! If only I'd known how dumb 120,000 seconds is I could've dispensed with 95% of the fuel and gotten the mission done via 10 minutes of 5 G burns rather than 100 minutes of 0.5 G burns. 

There's actually a 'disable in timewarp' option that you can turn on with the reactors at the moment.

On 9/11/2021 at 4:44 PM, Wubslin said:

By the way, have you thought about adding some more expanded, less luminous "cones" of exhaust coming out the back in vacuum to represent where the confinement ends? For such an extreme exhaust I imagine the plasma/gas/reaction mass would still be under an immense amount of pressure as it leaves the nozzle. It would definitely be cool to see that. Maybe even that central exhaust cone could be seen to peel away with its angle of divergence increasing slightly as the gas expands? I don't know how modding works so don't listen to me if I'm not making any sense.

 

This is functionally not very hard to do with WF right now (you could even do it!) The answer as to why is probably that I didn't think about it. 

[Wubslin]

12 hours ago, Nertea said:

There's actually a 'disable in timewarp' option that you can turn on with the reactors at the moment.

Okay, so this game has been one giant parade of me being made a fool of myself through my own unbelievable idiocy. Did you know that only last night, at the ripe old age of 1,164 hours of playtime on record in steam, did I realize that you can force a timewarp with Alt+. during a burn? I think I need to go to drop by one of those year-round halloween stores and pick up a nice clown wig for myself.

12 hours ago, Nertea said:

This is functionally not very hard to do with WF right now (you could even do it!) The answer as to why is probably that I didn't think about it. 

Really? Huh. Waterfall has always been pretty intimidating to me in the past, so I'll have to check it out. I know it can't be done in the first game but I hope KSP 2 supports dynamic enough lighting that in certain situations your plume can be seen to be dozens, even hundreds of times longer than your ship. I mean, look what this guy posted on reddit last night after the Inspiration 4 launch:

That's a big structure. Vacuum exhaust is nothing short of awe inspiring to me.