The caterpillar drive was a silent submarine propulsion system in
The Hunt for Red October by Tom Clancy. I have never read the book, but I saw the
movie when it came out in theaters and have always really liked it, because I was into
submarines. The idea was that the engine was so quiet because it somehow
moved water with no moving mechanical parts. It's the sounds from the moving parts of the
engine that are what normally allow a submarine to be detected with passive sonar.
I believe that in the movie they refer to this as a
magnetohydrodynamic drive. At the time I didn't know what the hell
that meant, but I did see something else that made me think that maybe you could do this
with electromagnetism (which I now know would, by definition, be the case with a
magnetohydrodynamic drive). So, as a final project in my high school AP
physics class, I built a working model of a caterpillar drive. Now let me describe the
basic idea of how it works, then how mine worked specifically, and some of the
technical problems involved.
The basic idea of a magnetohydrodynamic drive
It's really simple, based on the Lorentz force law. This is the physical law that
says that if you have electric current flowing in one direction and a magnetic field
perpendicular to that, a force will be exerted on the moving charges in a third direction,
perpendicular to both the electric and magnetic fields. You can find the
direction of the force as follows by using the right hand rule. This force means that
if you take a current carrying wire and put a magnetic field at a right angle to it,
the wire will get pulled in one direction. So, the basic idea of the drive is this:
If water has the proper things dissolved in it (like salt) it makes a pretty
good conductor (because it has enough free ions in it). So, if you used water
inside a tube as your "wire" by passing current through it one direction, then you put a
magnetic field perpendicular to that, the water would get pushed out one of the ends of the
tube. If you make that tube your engine, and orient things so that it gets
pushed out the back, you've got an engine. Here's a sketch of the basic scheme:
/^\ B
|
/ | /
/ | /
/ /
--------------------
| |
| |
| |
| | ----------> I
| / |
| / |
| / | /
| / | /
|_____/______________|/
/
|---/----------------|
/ L
/
/
|/ F
Where I is the current (also the direction of electric field and decreasing
electrical potential), B is magnetic field (or magnetic flux density if
you like), F is the force exerted on the water, and L is just the
distance between the electrodes (the width). Assuming you had a uniform
current and magnetic field in the region (which you almost certainly wouldn't), you force
produced by the engine would be F=B*I*L. If
you had current of 1 ampere, magnetic field of one tesla, and a distance of ten
centimeters you'd get 0.1 netwons of force. That's about 1/50 of one
pound. Not a lot of force, and a tesla is a LOT of magnetic field.
What we made
Now, let me preface this by saying that I was in high school. I was only beginning to
understand physics and I had no practical knowledge about materials, electronics, or
engineering in general. Well..ok, so the stuff about practical knowledge hasn't changed a
lot. So, I wanted to build, together with my partner from my physics class (a
very bright girl), a working version of this engine and have it propel a little
boat. We
quickly were forced to abandon the idea of the boat. First, we spent
way too much
time worrying about the
design of the boat before we realized that we didn't have the
ability to build an engine small enough and strong enough to be carried, and propel, a
small
model boat. Thus, we just settled for building the engine itself, as a
demonstration of
principle. It consisted of a
plexiglass tube about six inches
long, with a
square cross section that was probably about 1.5 by 1.5 inches (I'm
trying to remember from about 5 years ago). On the inside, in about the middle of the
tube, were mounted two electrodes, which were two rectangular
copper plates about as tall
as the side each was
mounted on and 1-2 inches long. We used copper because
that's what we had available. It was one of our
biggest mistakes. The electrodes had wires
soldered to the
back of them that led to the
power supply. We also spent a lot of time trying to figure
out how to generate the magnetic field: whether to use an
electromagnet or a
permanent
magnet, and if we used an electromagnet what sort of core to use. In the end
we used a permanent magnet, because we couldn't get our hands on a suitable core for the
right size electromagnet for the engine we'd constructed. The magnet we used was a huge,
heavy
industrial strength horseshoe magnet that weighed a lot and still only had a
strength of about 0.1 tesla (1000 gauss) at the center of its field.
The problems
There's several basic problems. You want to maximize the quantity
F=B*I*L subject to a couple of
constraints. I'll go term by term to describe the problems:
- B
The main problem is that B just isn't a very large quantity, and, worse yet,
magnetic field falls off quickly as a function of distance. Basically, you can use an
electromagnet or a permanent magnet. For an electromagnet, you need wire that will carry
enough current, a way to cool the wire if it gets too hot, and then a good core to
amplify the field. For a permanent magnet, you just want something that
has a strong field for its weight, and hopefully one that doesn't disperse too fast with
distance so that you can have a good field strength across the entire cross section of the
engine. We should have used a permanent niobium magnet, because it's easy and you get a
LOT of strength for little weight. If I were going to build a real one to power a boat,
probably the only way to get enough thrust would be to use
superconducting electromagnets (that encompasses the cooling problem
too).
- I
Again, one constraint here is wire. You need to have wire connecting your power source
and electrodes that is capable of carrying the current you want and cooled properly. This
is not the main problem, though. The main problems relate to the water. First, if you
apply too high a voltage to the water you will split it into hydrogen and oxygen
(hydrolysis). When this happens you will loose energy to this
reaction (the same amount you'd get per molecule by burning
hydrogen), which won't go into thrust. Also, then you'll have a bunch of hydrogen gas
forming, which will cause turbulence, noise, and God knows what else (it
is flammable after all). A way to get around this problem, as well as the
resistance of the water, is to make the region through which the current passes (and
thus the electrodes and array of magnets) comprise as much of the length of the engine tube
as possible. That way you can have a high total current, but the
current per unit area will be lower, which is what matters for the
problems of resistivity and hydrolysis I just talked about. The other main problem, the
one that thwarted us, was chemical reactions between the metal in the
electrodes and the water. We had this problem because we used copper plates (that was all
we had). Ideally you'd like something very non-reactive like gold or platinum
at least coating the surface of your electrodes. Something like nickel or high
grade stainless steel might work ok, I don't know. But don't use copper!
- L
Well, now this will depend more on your particular design. Basically, if you had a
completely uniform magnetic field you'd want as large an L as possible. In
practice, though, magnetic fields aren't uniform and usually weaken very quickly
with distance. For dipoles (basically like bar magnets) the field strength goes
down as one over the cube of the distance. If the field from your magnet was like that,
you'd want to make L as small as possible, because the larger you made
L, the smaller B would be, and it would get small faster than
L would get big. So, the decision of what to make L is complicated
because it depends on the geometry of your magnetic field. A good rule of thumb in
most cases would probably be to keep L pretty small, approximately the same
size as the width of the magnet, or smaller. The only thing is that you can't make the
tube of the engine too narrow or flow inside will become turbulent, which will make the
engine louder and less efficient.
As far as I can tell, the main overall problem of the design if the making a strong and
uniform enough magnetic field. I say this really because most of the other terms
governing the thrust are either easily maximized or governed by stuff you can't do much
about, like the chemical properties of water. The magnetic field is the main area where
there seems to be room for significant improvement, since it seems there are many ways you
could try to make bigger and stronger magnets with more favorable geometry. Not that it
would be easy. The only other way to make the thing better would be if you could find a
better overall design geometry, like maybe the best way isn't some plates and stuff inside
a tube.
The result
In the end, when we tested it, we found that it did move water!
However, it also built up all kinds of corrosion on the electrodes very quickly. This
made the resistance jump, the current drop, and, thus, the thrust drop, then eventually
this precipitate would form a thick enough layer that it would flake off, and the engine
would spew black flakes out. It was like a pollution engine, like something that would be
invented by a villain on the cartoon Captain Planet. In any case, it didn't have much
thrust, but it did move water, and the corrosion problem could be fixed with the proper
materials. Refer to the discussion of "problems" above.
If I had to do it again
For a model, I'd definitely use a much smaller tube for my engine, maybe a centimeter
across. I'd use permanent niobium magnets to generate the magnetic field
because you can get pretty small ones, and they have an insane ratio of field strength to
weight. I might make the electrodes go pretty far along the length of the engine tube and
get a bunch of niobium magnets to go along the length with the electrodes. I'd definitely
invest in getting electrodes that were at least plated with something fairly non-reactive.
If it were light enough to be carried by a model boat, then I'd probably power it with a
battery. For an actual vessel I have very much less idea. Definitely, you'd use
superconductors as the wiring for your electromagnets.
Has/Can/Should the military build such a thing
Well, one of the main points of my write-up is that they certainly can and
probably have built working models and/or prototypes, because I did it in
high school. There's a good chance, though, that you can't really make a
caterpillar drive feasible for military use, mainly because you probably can't make the
thrust high enough for it to be useful. There might also be some question about its
tactical value, especially if it were slow. Attack subs need to be
able to move pretty fast, and they can already move pretty silently. You could try to use
it for a ballistic missile submarine like in The Hunt for Red October, but really those
don't need to get too close to their target (so silence might not be that much of an
advantage). Also, in the story, the captain Ramius defects because he believes this
technology would only be used for a first strike (surprise attack), which he believes
is wrong. I must agree. I think that starting global thermonuclear war is always a
strategic mistake. Another reason you might not want such a sub is that it might still be
detectable by other means. For example, the US navy is currently in the process of testing a
low frequency active sonar system (LFA sonar), which detects all ships and submarines for
tens or even hundreds of miles or more using active sonar. In this case, no matter how
silent you are, it won't help you. Also, given the massive magnetic fields that would be
used in the engine, together with the metal hull of the ship, you might be able to detect
the submarine magnetically using a sophisticated instrument like a SQUID
(Superconducting QUantum Interference Device). So, to put it succinctly, it might be
unfeasible, or it might be possible but not worth the effort.
Odds and Ends
- Ok, like I said, I'm not an engineer, and I haven't thought much about
this since high school. There may be other technical aspects I have missed, and there may
be easier and more practical ways of doing things I didn't see. I'm not putting
this up as a definitive guide of how to do this, but just how I did it and it can be
done. I very much encourage people to post write-ups if they have enough knowledge to
suggest better how a real, working version might be constructed and operate.
- I will respond to one other write-up that said you might want to use
AC, rather than DC. I'd say probably not. You
certainly COULD use AC (then you MUST use electromagnets), but as I stated in the
"problems" section above, there's effectively a fixed upper limit to the voltage you can
apply, so in either AC or DC this would set your maximum voltage. The problem is that for
the same maximum voltage, you get lower average (RMS) power from AC. Normally this
isn't an issue, but in this case it would be. On the up side, using AC might help a bit
to lessen the corrosion problems.
- Just by way of information, both my lab partner and I got As for this project
and the class, and we both went on to study physics in college. I went to the
University of Maryland; she went to Princeton. And, we are both continuing to pursue a
Ph.D. now.