February 11, 2008

Magnetic Catapult a feasible advanced earth launcher

A 9000 meter long magnetic catapult was proposed in 2003 by Warren D Smith, mathematician at Temple University. It was designed to launch 5 meter long, 1 meter diameter projectiles at 2250 gees of constant acceleration with a launch velocity of 20km/s (twice earth escape velocity, Mach 58). It would cost about $2-20 billion to build and operating costs would be $10-100 million/year. This system is not a railgun and is not a coilgun. It is similar to a superconducting coilgun but is better than the quench gun proposal.

2250 gees would allow electronics to be launched

It is an engineering project on the same scale as the largest particle accelerators. So it is difficult but achievable. The system should be cheaper and safer than chemical rocket launches.

The magnetic catapult would be different from an electromagnetic railgun or linear electric motors.

The magnetic catapult has the following advantages:
1. No electrical or mechanical contact between the projectile and anything else
2. No capacitors or other external energy storage devices. The superconducting magnets of the launcher are the energy storage device and the energy stored is at essentially uniform density in the form of magnetic field.
3. the accumulation of electric power may be accomplished gradually, losslessly and purely mechanically
4. Essentially 100% of the stored energy is converted to projectile kinetic energy. Guns, railguns and rotary pellet launchers suffer damage from friction and other wasted energy.
5. Magnetic catapult has inherently stable operation. Instabilities of other designs at these velocities could be catastrophic.
6. No switching at large current and voltages (like Linear electric motors). Switching only occurs when current and voltage are zero. This minimizes stress on the switch and maximizes energy efficiency.

Railguns use two sliding contacts that permit a large electric current to pass through the projectile. This current interacts with the strong magnetic fields generated by the rails and this accelerates the projectile.

The Magnetic catapult is similar to a superconducting coilguns (quench gun), which are contactless and which use a magnetic field generated by external coils arranged along the barrel to accelerate a magnetic projectile. However, the quench gun releases a lot of heat during coil quenching.

The picture to the left is a standard copper coil coilgun. The magnetic catapult would need cryogenics (to cool the superconductors) and a lot of superconducting material. The magnetic catapult design will need to address the issue of magnetic quenching. During projectile flyby of a superconducting ring, the magnetic environment near that ring can change as much as plus or minus 12 Tesla. to protect against quenching a finely intermixed composite material made of both YCBO and an electrically and magnetically inert companion material is needed. TiO2 might be a suitable material. With a ratio of twice as much TiO2 to YCBO then the inner rings would be 3 times bigger. The total energy losses related to the quenching issue are about 8%.

Another disadvantage is the system does not scale down well, so some test sections could be built but only a nearly full scale system would really indicate if the system would work. However, it seems promising and worth detailed testing and modeling.

The superconducting magnets that are needed for the project are on the leading edge of developments in that area There are record superconducting magnets with 50 cm bore sizes, 7 Tesla and 20 cm bore size, 8.1 Tesla. The full scale system needs 17 Telsa magnets with 100 cm bore size. There is an Atlas hybrid magnet being built to hold 1.2 Gj with 22 meters (2200cm) diameter. The highest power superconducting magnet is 26.8 Tesla and they believe they can soon reach 50 Tesla

Magnet Lab researchers tested a small coil (9.5-millimeter clear bore) in the lab’s unique, 19-tesla, 20-centimeter, wide-bore, 20-megawatt Bitter magnet. However, I think the main issue is one of cost that larger bore sizes mean more wire and more expense. More information on superconducting magnets

A Cern superconducting magnet has an inside diameter of 6.3 meters, a length of 12.5 meters and generates a magnetic field of 4 T (about 80,000 times stronger than the Earth’s). Once completed, the Cms superconducting magnet will boast a notable record: with its 2.6 Gigajoule of energy it will hold the world record of energy ever stored in a magnet.

The magnetic catapult is described in a 42 page postscript file.

The Magnetic Catapult Design
The projectile will be a cylindrical shaped permanent magnet. The accelerator will consist of a sequence of coaxial stationary rings, each of which is also a supercurrent loop magnet. The projectile and the rings all generate the same amount of magnetic flux, which will be assured by initially magnetizing all the rings.

[Thread a ring shaped piece of superconductor with another magnet whose North end is on one side and whose South end is on the other of the ring. cool the ring to ists superconducting temperature, then remove the magnet. Its flux no longer traverses the ring but the total magenetic flux through ring must remain unchanged. More discussion on page 11 of the paper.]

During launch, the projectile passes through the superconducting rings. The South end of the first ring attracts the North end of the projectile. Once the projectile has reached a central position inside that ring (the supercurrent is now zero) and we switch off the superconductivity in that ring converting it into an insulator. The projectile continues on without deaccelerating. This repeats along each of the magnetic rings.

The entire accelerator is enclused in a pipe with a superconducive inner coating. The outer pipe has the following purposes:
1. Is a vacuum vessel
2. It is an EM shield
3. It is a thermal insulator
4. It would help to levitate the projectile
5. The sequence of rings are a long solenoid, the field must come back the otehr way on the outside of the solenoid.

The projectile bursts a membrane at the end of the tube or it passes a double door airlock.

The system should be made in a mountain like Annapurna or Dhaulagiri in Nepal. The projectile would exist above half to two thirds of the atmosphere. The projectile would need to have material that would burn off (ablate) to take away the heat. Only a few percent of the total projectile would need to be sacrificed.

Cost estimate
The Superconducting supercollider (SSC) was to cost $8.25 billion but ran into cost overruns and was cancelled. The SSC was to be in an 87 kilometer long tunnel with 10,000 7m long 6.6 Tesla magnets that were Helium cooled. The Magnetic Launcher would be almost ten times shorter, the magnets need not be as precise, the outer vacuum shells are smaller and the vacuum need not be as high. The Magnetic launcher would be built very robustly and be located on a high mountain. The Brookhaven Relativistic Heavy Ion Collider (RHIC) was 4 kilometers long with 1740 superconducting magnets. The RHIC cost $600 million for $155,000 per linear meter. Those prices would enable the 9 km launcher to be made for $1.4 billion. Another cost estimate looks at costs of components (tunneling, Dewaring, structural support and superconductors) for a maximum cost of $800,000 per meter or $7.2 billion for a 9km long launcher.

The previous article that I had on scaling up railguns still had the issue of wearing out the launch tube and replacing the worn launch tube was the primary cost driver for the advanced railgun proposal.

Overview of electromagnetic guns


al fin said...

Articles and proposals for electromagnetic launch go back to the 70s and perhaps the 60s.

Gerard O'Neill, when alive and at Princeton, was an advocate of space based EM launch for getting lunar resources to L4, L5. K. Eric Drexler did some work with the O'Neill crew back in the 70s, although he was at MIT.

That was before Drexler got interested in Nano.

Snake Oil Baron said...

It is an idea that continues to make a lot of sense in theory but there is always the near certainty that if you wait a decade the materials which it is being made of and launching will be more advanced, the manufacturing techniques will be easier and cheaper etc. How does one know when the right time to build it has arrived?

It may be cheaper and safer but the chemical rockets are familiar and have a supporting industry developed around them. That is a lot of economic economic inertia to overcome. I hope the idea makes it.

bw said...

Al, The 42 page paper addresses the flaws of the O'Neill designs and proposals. Of the big gun launching approaches this seems to be superior. I believe big guns can be made to work and of the EM approaches this appears to be the best.

Snake Oil Baron, I think it probably does make sense to wait for the full build until we get superconducting magnets down an order of magnitude in cost. But I think this idea is worth spending a few million on per year refining it. My current best mid-term (6-10 years) hope is that this year we get news of Bussard IEC success, then that would be the best path forward. Hypersonic rockets will take longer than 10 years. Hypersonic missiles in the 6-10 year timeframe. In the near term (1-5 years) the best bet is SpaceX and Dnepr and using cheap rockets with a fuel depots and Bigelow inflatable space stations. Although it seems that the near term is clogged with bad budgets and political pork.

I believe that improving nanomaterial control will let us make room temperature superconductors. (Phonon mediated designs). The recent work where they found small islands of continuing superconductance at far higher temperatures suggest that nanostructuring could expand those islands.

The other thing about this idea is that a big system on the moon looks like it would not be limited to this speed. The design looks like it scales up to even higher speeds when it is vacuum.

Anonymous said...

I have thought about this problems, and specifically about an ideal launch site. It seems to me that two launch sites might be better than one in the Himalayas.

For launches that eventually be directed to geostationary orbit, the Andes might be a better bet, specifically Mt Cayambe in Ecuador, whose peak is over 18,000 feet high.

For launches that are polar, one might consider at or near Mt. McKinley. While the climate is not ideal in this case, it has the advantage of a thinner atmosphere at the polar regions

Joseph said...

Brian, what is the projected maximum launch capacity? (and contstant power draw to achieve it?)
And what does the projected cost minimum go down to at maximum throughput?

bw said...

If you were using about a net (any losses or inefficiencies of 100MW to charge the superconductors then it would take about 8000 seconds to get to 800 GJ. So about 3 hours to charge. The paper speculated about two shots per day. Each shot is 4 tons. So 8 tons per day. About 2800 tons per year. Some of that ablates away and some is projectile and not cargo. 1500-2000tons of electronics or supply cargo. So if it cost 15 billion to build then over ten years the cost to get stuff into space would be about $1100/kg. The main gotcha is that you have to have the constant stream of launches. The purpose and use for all of those things.

If you charged it with 1GW then the charge time is 800 seconds and with the airlocks for faster cycling 10 shots per day is possible. Then the ten year costs could go down to $220/kg.

also, the costs are halved if the lower end build estimate worked out say $7.5 billion.

Anon: in the paper about 20 mountains are considered. The postscript can be viewed when one uses ghostviewer and ghostscript.