« CNN Assault Weapon Poll | Main | A Counter Argument »

May 18, 2004

Warship Power Systems

Steven Den Beste wrote a bit more on spaceships when someone suggested using optics for power transmission. However, if both ends run on electricity and aren't hundreds of miles apart then optics can't compete for transmitting large amounts of power. This is one of the reasons we use transformers instead of having a high-voltage, high-intensity lamp shining on a low-voltage, high-current photo-cell. As SDB points out, the major power consumption isn't in the wiring. But the whole debate gets back to the question of where the spaceship's power goes, and there's actually only a few places a warship would use major amounts of power.

First off, let's eliminate the crew as a major component of the power requirements. Take a look at your crew size and note that you'll need maybe two kilowatts or so per person. They'll each have computers and other personal electronics at 100W or far less, lighting that might amount to a couple hundred watts, and they'll occasionally cook something, so you can figure a few intermittent surges in personal use. The Space Shuttle runs on about 14 kW, or roughly 2 kW per person, including all the navigation systems and controls, and a radiator running at roughly room temperature would only need to be about 5 square meters per person. Of course near the earth the energy coming from sunlight would affect this, as would the hundred or so watts per square meter reflected up from the earth.

If you cut the ships power and wait, everyone will of course turn into human popsicles as the laws of thermodynamics equalize the temperature of the ship with that of the cold background of space, so although the power they use becomes waste heat, it's waste heat you want to keep around, and least to a degree, or even 70 degrees. You can't let the heat build up forever or they'll bake, and this is why you have to strike a balance between heat generation, absorption from the sun, and radiation back into space. Fortunately the skin of the ship facing away from the sun tries to approach 5 degrees Kelvin, but of course the sunward side gets hot enough to fry an egg, so keeping things in balance is a trick, especially when cruising to Mercury before swinging out to Jupiter.

I'll also assume they're not going to be growing crops on a warship, or at least not dragging them into direct combat, but rather abandoning their crops before a fleet action commences, to be tended by three little robots and maybe Bruce Dern. So we have no other major sources of power required by life support. That leaves us with other major power hogs on the ship, which is really going to come down to just a few items, namely propulsion, active sensors, and weapons.

If propulsion is based on conventional chemical propellants your fleet combat will likely resemble two rafts with guys on board taking pot shots at each other, except of course the rafts will be stacked with drums of fuel and oxidizer. However, with chemical rockets you also don't get into serious overheating issues where you need giant glowing radiators. Just because you generate tremendous amounts of energy doesn't mean it has to turn into waste heat. A liquid fueled rocket, for example an S-IVB upper stage from a Saturn V, generates massive amounts of heat, but it blows it out of the engine as high-velocity exhaust, and thus thrust. The waste heat from the combustion heats the engine, which is cooled by fresh fuel, right before the fuel is fed into the engine and likewise blown out the back, and all without raising the remaining liquid hydrogen to its boiling point. It's really kind of clever.

You could use a nuclear thermal rocket like the NERVA, which will give you about twice the maximum delta-V as conventional rocket propellants, and frees you from having giant tanks of both a fuel and an oxidizer. Given their increased performance, I'd think a competent navy would choose the nuclear engine over the chemical, at least for capital ships. Even so, they still wouldn't give good enough performance to give you a good long range space fleet, which is the whole point of the exercise.

So let's get really fancy and use the new VASIMR plasma engine. The VASIMR concept is essentially to use RF energy to superheat hydrogen propellant to millions of degrees while magnetically constricting it before it accelerates out the back. It has no wear and no moving parts, can give a delta-V that's over 10 times better than even the nuclear rocket, and it looks like we'll be using them anyway. More on that here and here, with other advantages listed here, and an excellent diagram here. Indeed. Of course we'll probably not use any one of these exactly, because by the time the first fleet action rolls around we'll have thought up something new, but I'll take it as a starting point.

Happily a VASIMR is a high-voltage low-current engine, so resistive losses are smaller. To be capable of high-accelerations the power dumped into the engine will be huge, yet since it uses superconductors the engine itself isn't allowed to get very hot. It might require all sorts of cooling if enclosed, but that cooling would probably be liquid nitrogen or some other cryogenic liquid. Even though the exhaust temperature of a VASIMR can be in the millions of degrees, and the engine itself may consume many megawatts of thermal energy, you're not ever going to let it get very hot. But to keep your cryogenic cooling fluids cold you'll either have to use very large and cold radiators, or else use a conventional cryogenic cooling system, just like we do on earth, which cools the fluid by making another fluid very hot. We're basically talking a heat-pump, and of course the conversion isn't that efficient, but it would let you use a smaller and hotter radiator to dump the waste heat, and at these cryogenic temperatures any heat you need to dump is definitely waste heat. So the engine probably isn't our major heat problem.

Next up is weapons, which could be a variety of types. Missiles won't dump their heat into the ship that launches them, so no worries there, and any point defense guns will likely radiate their own heat into space, since that's where the barrels will be. Lasers might represent a major thermal load, since current diode lasers are only about 50% efficient, but hopefully moving up to the 80% range. I blogged some of the laser issues here. Now considering that the military is currently looking at having a 100 KW laser on a Humvee by 2007, any laser on a capital ship is certainly going to be in the high megawatt class. Even at 80% efficiency it's still generating some tremendous thermal problems.

Finally I come to sensors. Passive sensors are little things like radios, microwave receivers, IR sensors, radiation detectors, and optical in IR telescopes. The biggest power requirement you might have is the little motors to pan and slew the antennas and scopes. Since these are obviously mounted on your hull, what little heat they generate will likely just radiate away, which is a good thing since IR detectors have to keep cool, lest they get swamped by looking at their own thermal radiation. So you've got no real power or heat problems in your passive sensors, except possibly a tiny heater to keep the electronics from getting way too cold.

Active sensors, however, could be another major power hog. Some of the NOAA over-the-horizon radars are a million watts, and even your local TV station's Doppler weather radar is likely running 250 KW or so. An Aegis warship's AN/SPY-1 radar is at least 4 million watts, and I'd think a future warship would be able to radiate at least to the levels of a 1980's Navy destroyer. The Aegis radar can already track objects up into orbit, so it might be a good starting point. We're starting with at least 4 MW radiated power, and will likely up it from there to reach out further and pick up very high velocity inbound projectiles.

One point I'd make about all three systems is that they are fairly efficient at radiating power into space. The ship doesn't have to merely rely on blackbody radiation to shed energy, but can instead beam it into space. The laser is dumping out potential heat as a coherent beam, and the radar is sending it out as a microwave signal. The lasers might get to 80% efficiency or higher, while a class-C amplifier or Klystron might run 75% efficient. So if you had 100 watts to spend, you could directly shed 75% of it into space and then deal with the rest as heat.

Anyway, it looks like all our major heat problems would come from just three systems, active sensors, beam weapons, and possibly propulsion. But one thing is still missing on my warship, a power source. Something has to generate all the electricity to power the ship. I agree with SDB that it will be a while before fusion is both available and an attractive alternative to advanced fission systems. You could of course use a variety of solar power systems, but these can't ramp up their output during a battle, perform well in deep space where sunlight is minimal, and are very large, fragile, and vulnerable.

So let's put a fission plant in the ship, and see what problems can be anticipated. The first one I see is hippies, but thank goodness our ship has laser weapons to deal with them. The second is that the ship doesn't have a right-side up, may pull rather strange maneuvers, and could have some of its active cooling drop offline while weightless. Most of the passive cooling systems for the safer modern reactors rely on provisions for natural convection, and natural convection needs gravity or acceleration to move the fluid as it heats up and changes density.

My quick recommendation around this problem with current technology is to go with a new concept called a thorium based accelerator-driven systems. These use thorium instead of uranium or plutonium, and thorium is just short of being able to sustain a chain reaction, because it doesn't toss out quite enough extra neutrons when it fissions. But you can use a particle accelerator to hit the reactor core with high energy alpha particles. These generate neutrons by spallation, and these neutrons then hit other thorium nuclei which fission, releasing almost enough neutrons to sustain the reaction indefinitely, but not quite. Yet the reaction produces energy in the usual way, with the only big difference being that turning off the particle accelerator immediately stops the reaction. That could be handy, on top of the fact that thorium is a heck of a lot more abundant than uranium. There are already several thorium reactors in operation, but so far these rely on a uranium or plutonium core to provide the extra neutrons.

In any event, we've now got a nuclear reactor of some sort on our warship, and it's a big honking heat source. For the power levels we need we'll probably go with one of several existing alternatives to convert the nuclear energy into power, all of which involve turning turbines, whether with steam, helium, or other material. Currently we're limited to around 30% to 50% efficiency, and one of the newer and more efficient systems is shown here. But at least this gives us a start on a design.

Let's pick a warship size and see how the thermal problems might work out. Let's just use a ship somewhat analogous to a Nimitz class aircraft carrier, say a cylinder 1100 feet long and 150 feet in diameter. It has a surface area of 50,000 square meters and a crew of 4,000. You might wonder what 4,000 people would possibly be doing on a warship, but I have an idea for that later on. We need about 8 MW for the crew's lighting and life support and such. This energy ends up as waste heat, which could be radiated into space from a surface at normal room temperature by a radiator that's only about half the size of the ship. But if our reactor is only 50% efficient then we've got 16 MW worth of thermal energy to deal with, so our ship can just barely radiate the heat while keeping the outer hull the same temperature as the quarters.

And as we add power the radiator area has to increase. With an emissivity of 0.85 (on a scale of 0 to 1) this hull would lose 17 to 20 MW of waste heat if held at room temperature. I'll guess we want something like 20 MW for the radar, maybe 20 MW more for lasers, and probably the same for propulsion, just as a wild guess. But given the intermittent operation of the lasers, let's just ignore them, assuming we'll have to cut our engines to use them continuously, while just dealing with the increased thermal load if using them in bursts or on short notice.

So I've added 40 MW of required power to the ship, of which maybe 75% is transmitted away as radiation, leaving 10 MW to contend with as extra head. However, our inefficient reactor plant, running at 50% thermal efficiency, means we've another 40 MW to deal with as waste heat. Our thermal efficiency problem is becoming pretty obvious. We've got 18 MW of heat we can't avoid, which we might need to keep the ship warm, but 48 MW to get rid of as a result of our inefficient power source. If we don't increase the size of our radiators the hull will be running at over 340 degrees F, and we aren't really using serious power yet. A Nimitz class carrier can produce 190 MW, yet these levels would have our radiators at over 600 degrees F, so we're getting in trouble and we're not even up to current navy capabilities.

To dump this heat we have to increase either the radiator area or the temperature of the radiator. However, for a heat engine the maximum efficiency depends on its two temperature extremes, and if you keep upping your outlet temperature, using a hot radiator instead of a cool one, the efficiency drops. Our thermal reactor will run even more inefficiently if we let the radiator temperature, which limits the low-end output temperature, climb higher and higher. So instead of using a small radiator that's running white hot, which implies an even higher temperature elsewhere in the system, you could use that waste heat to run a conventional steam plant using a Rankin cycle or more importantly a Kalina cycle®, which uses a water-ammonia mix to achieve far higher efficiencies than we're used to getting. This cycle is rather new and we're only now bringing many plants on line with it, but in some cases it creates tremendous efficiency gains, and is even a good system for scavenging waste heat, such as from fluids that are only around the boiling point of water, all the way up to a thousand degrees or so. These Kalina cycle® engines are indeed interesting, in a steam-plant kind of way (you know what that means), and more can be found here, here, and here.

So if our turbine reactor was only 50% efficient, it's got to be generating a lot of waste heat, so we might be able to surround it with a jacket that heats water-ammonia for a second Kalina plant, to try and recover maybe half of that 50% we've been radiating away. I'm not at all sure this would work, but many of the Kalina plants have been installed to run off the waste heat left over from other power systems, including diesel exhaust. If our radiators can run cool then we even have a far lower ambient temperature we can try to take advantage of, possibly with a condensing cycle similar to the Kalina but using a mix of atmospheric gases or even the ships onboard air supply. However, I think it would be hitting a point of diminishing returns, and you have to weigh the slight increases in efficiency, and thus potential thrust for a given amount of heat, versus the increased mass and area of the required radiators.

You could also run such waste heat scavenging systems through your lasers or other concentrated heat sources, and possibly generate a bit more savings. But the more important point is giving those 4,000 crewmen a bunch of valves to turn and gages to watch, plus running lots of pipes everywhere. After all, we want that British Fleet feel to the ships, and what better way to do that than build a warship that always has a steam rupture every time we need some dramatic effects. Even better, the Kalina cycle® uses ammonia and water, and ammonia leaks would have the crewmen running around with respirators making muffled grunts at each other while madly turning valves, and what could be better than that? Well, maybe a cyanide gas cycle or something, but you get the idea.

The point is that anytime you have a radiator that's running very hot, you could use it to power a heat engine running at least at 50% efficiency, the output of which you can radiate away at 75% or 80% efficiency as a laser weapon or radar signal. So you can possibly radiate 40% of the radiator's original energy as non-black body beam emissions, leaving you with just around 60% of the original heat to get rid of, which lets your radiators run cooler and your main engines run more efficiently, assuming they're heat engines. But the design pressure seems to be on the efficiency of power generation and the power usage in a couple of critical systems, not the overall spread of power throughout the ship.

In the coming days I'll try to post some more on the fundamental question of whether the ship should be maneuverable and small or just well armed and tough, and how you figure these things out.

May 18, 2004 in Science | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Warship Power Systems:

» Putting the sciene in science fiction from Physics Geek
The title of the post is how Arthur C. Clarke has been described. Not long ago, den Beste posted an article about space ships and weapons that could potentially be used; a related post can be found here. Bastard Sword... [Read More]

Tracked on May 19, 2004 8:46:55 AM


I'm no physicist or electrical engineer (or even electrician), so while I was intrigued by this whole SDB spaceship discussion, I was a little lost until I read this quote: "This is one of the reasons we use transformers instead of having a high-voltage, high-intensity lamp shining on a low-voltage, high-current photo-cell". The light bulb in my head flicked on when I read that...thank you for providing me with "layman's term" type understanding!

Posted by: Tartan69 at May 18, 2004 10:01:10 AM

With a smaller ship, the surface-to-volume ratio will increase and the power-to-mass can increase; this may set a ceiling on ship size. A destroyer perhaps 1/8 the size of the dreadnought you propose would have 1/4 the radiating area...

Assuming a constant crewman-to-volume ratio there would be a particular size that would give the maximum power available for non-life-support systems. Too small, and it can't even stay warm; too large, and it overheats.

Posted by: Mike Earl at May 18, 2004 11:40:33 AM

Wow, 1100 feet long is a big ship! I'd like to see some ROOM DeltaV calculations to find out how much energy would be required to move that thing. I have a feeling that acheiving useful accelerations will require power on the order of terawatts, but I don't have any orbital mechanics textbooks handy to check.

As far as personnel requiremnets, I think that by the time we are building warships like this the vast majority of the vehicle will be automated. Forget life support except for a very small percentage of the vehicle, maximize firepower and "shielding", and maximize maneuverability.

The kalina cycle link was very interesting. I've heard of water-ammonia cycle power generation but I have never heard it by that name.

Posted by: sleepy at May 18, 2004 2:26:09 PM