Sunday, March 18, 2007

Mr. Fusion

Instapundit says that we will need coal, oil, and natural gas for some time even if some one invents Mr. Fusion tomorrow. Which is true.

However, Mr. Fusion was invented yesterday (several years ago actually).

It was invented by Dr Robert Bussard formerly of the AEC fusion office. He did the work under a Naval contract.

The Bussard Fusion Reactor will lower electrical costs at the busbar by at least 10X over coal or fission nuke power plants. Capital cost for electical plants using the Bussard Fusion design will decline by at least 5X mainly because no turbines, condensers, steam generators or electrical generators are required. With such a lowering of costs and simplicity of required equipment, roll out will be very fast.

The reactor is just a big sphere surrounded by electro-magnets. The main cost of the plant is converting the 2 million volts DC output to AC for local use. The direct 2 million volt output would be great for long distance transmission. Although the plants could be sited in just about any reasonably sized electrical yard since any required cooling would not requre a water supply. Air cooling would work fine.

The power generator is about 10 to 12 ft across for an output between 100 MW and 1,000 MW. Power output scales as the 7th power of size. Double the size and you get 128X as much power.

No thermal plant is required. Thermal plants - steam generators, turbines etc. - are long lead time items. They can take from 3 to 5 years from start of production to delivery. The Bussard Fusion Reactor output is direct 2 million volts DC. (a very large battery).

Unlike fission plants there is no fuel stored in the reactor core = no Three Mile Island kind of problems. Turn off the electricity or turn off the fuel and the reaction stops.

It would make a good rocket engine for fast interplanetary travel.

Easy Low Cost No Radiation Fusion - video plus technical details.

Dr Bussard needs $2 million in start up funds to verify reaction constants. He will need $200 million for a test reactor.

The fuel is Boron 11 which is very abundant. We have 200,000+ years of reserves on the planet if it is used exclusively for power. Most borax is used now for borosilicate glass.

Let me take this time to specifically thank the crew at Classical Values, Justin and Eric, for giving me a heads up on this.

Update: 11 May 007 2002z

There is a petition started on 20 Nov 2006 calling for Congress to support the Bussard Fusion Reactor.

The short version: Dr. Bussard has a Navy contract that is unfunded.

Also please write your Government to urge them to fund the Navy contract:

House of Representatives
The Senate
The President

In addition Dr. Bussard is taking donations to help fund his work at EMC2 Fusion.

Update: 30 Aug 007 0032z

The US Navy has funded the next phase of Polywell research. This is no reason to let up. The Navy plans a five year program to construct a 100 MW test reactor. With more money they could speed up development. With enough cash a three year time line ought not be difficult. Two years is an outside possibility if we really pour it on.

Cross Posted at Classical Values

30 comments:

Anonymous said...

Not sure it's a good idea to trivialize the equipment required to harness 2MVDC. The high-voltage DC transmission lines usually don't top 250KVDC, AFAIK. You're looking at some serious R&D time to develop your output inverters. It might even be cheaper and faster to dump the energy to heat through resistive banks and go with steam turbines, or some other thermal engine.

M. Simon said...

The highest DC voltage used for transmission so far is 600 KV.

Scaling that up to 2000 KV (about 3X) should be do able.

Going thermal wastes about 60% of the output and raises costs and lead times (turbines, steam generators, condensers, etc) Plus you then require a lot of water or lots and lots of water. Siting then becomes more difficult.

However, connecting such a unit to an already built steam plant cout work out well as an interim measure.

M. Simon said...

Here is a nice bit on High voltage - high power - static inverters

Anonymous said...

I agree it can probably be done. But it's some tricky engineering, might require some invention, and you probably ought to work the inverter into your math when you are looking at scaling up. The hair on the back of my neck stands up thinking about a 2MVDC capacitor bank (no pun intended). A 2MVDC synchronous condenser may or may not be possible--think about the issues involved just in insulating the windings. In general, inverter losses scale linearly with power, and the silicon required scales by the acre.

Oh, and I agree about the losses with thermal output, but if the inverter proves tricky, it wouldn't do to let perfect be the enemy of good...

M. Simon said...

I have designed a lot of inverters (low and very low power). So I know a little.

The 600 KVDC line was built in 1987; we have come a ways since then.

Really high voltage stuff (>2MV) is done often in particle research. The power is low but the voltage problems are well understood.

The advent of Silicon Carbide FETs, IGBTs, and SCRs should simplify things since they can support higher voltages than ordinairy silicon.

I read about a 500 KVDC high power (300 MW) inverter that used a series string of ~280 SCRs in series per phase. That probably means using 6KV rated SCRs at 2KV for high reliability.

BTW phase correction and harmonic supression can be done at the transformed AC voltage side at 100 to 300 KV - well developed technology.

On the high voltage side what you need is just copper, iron and controlled switches. Smoothing can be done on the AC side.

Depending on space available the eqpt can be oil or air insulated.

Anonymous said...

I've designed several inverters, but never one at high voltage, either. Sounds like you've done a lot of good thinking about this, but I still think the lead time on the inverter is not going to be negligible. I also think that the inverter is going to be an important part of your calculations of scale.

LarryD said...

Even if the full scale prototype were switched on today, and worked without a hitch, how long would it take to actually start building power plants based on this technology? And how long would it take to complete one?

And we have to factor in the inevitable opposition from the greens, who are really opposed to modern technology, they're just more subtle about it than Luddites. Except for the "watermelons" (green on the outside, red on the inside), who are after power of a different kind.

I sure hope that the Bussard Fusion Reactor works as hoped, at full scale. But roll-out is still going to take years, maybe a decade or two. And petroleum will still be needed as feedstock for plastics, et al

It would make it practical to convert to hydrogen as a transportation fuel, though. Don't expect any kudos from the greens though, their support for any power source ends when it becomes deployable.

M. Simon said...

Larry D,

If it worked today as a 100 MW experimental reactor I'd estimate about 3 to 5 more years to get an energy production design. Depending on funding.

Once that was operational and tested long enough to get the reliablity up to snuff then roll out could begin. A few reactors in the first years 10s in the second. 100s in the third. And by the 5th year 1,000s.

A lot will depend on the cost differential from a traditional steam plant. Capital and operating.

What makes a faster roll out possible is direct conversion of particle energy to electrical energy. Thus no turbines.Which are a big expense and take a long time to produce.

The basic design is just a big hollow sphere, some power supplies, and vacuum pumps. With some cooling for waste heat. The waste heat might amount to as much as 20% of plant output vs around 65% of plant output for a nuke plant or 60% of output for a coal fired plant.

The power conversion equipment is just a bunch of semiconductors and transformers.

The whole thing is amenable to mass production. Any kind of steam plant is harder.

Did you know that for a nuke plant the reactor only amounts to 20% of plant costs? The rest is steam generators, turbines, condensers, and pumps, plus auxiliaries.

For the Bussard Reactor no steam plant. Only conversion of 1.2 million volts DC to AC.

The highest DC voltage used in commercial service is 600 KV (positive and negative) so 1,200 KV is not a big stretch.

Plus the DC is ideal for long distance transmission and a grid intertie system that does not have all the problems of AC (which I will not go into at this time except to say that an AC plant could find itself working against its own output due to AC phase differences over different transmission paths - DC does not have that problem).

So yeah. It may take 20 years or longer to convert the whole system to this kind of plant.

Less time if production ramps up to fill the demand in energy scarce regions.

Lots of ifs. However, you are in the right ball park.

NobobyExpects said...

Some info about UHVDC transport systems. Currently 800kV is under study.

https://www.energy-portal.siemens.com/irj/portal/ptd/public/en/target_12f09d586e5690881ff98721ce155ce3?NavigationTarget=ROLES://portal_content/grp/ptd/cont/p/com.siemens.pct.ptd.sip_gen.admin.global.p.redirect&buildTree=false&int_param=https://concert.siemens.com/conductor/servlet/sbs.concert.CallConcert?prod_name=KN03011202%2526service=997%2526language=en%2526country=ZZ%2526cmid=1

Note the large size of those facilities.

About the reduction of piping, etc. I think we should consider the vacuum system. A power of 100MW require 50A @ 2MV.

Some calculations:

1A = 1Q/s
1Q = 6.242×10^18 e-
Every alfa particle needs 2 e- to neutralize, and becoming a He4 atom.
1 mol (He4) = 6.022×10^23 atoms of He4 = 4g of He4

we're talking about evacuating:

4 × 50 × 6.242×10^18 / (2 × 6.022×10^23) = 1.037mg of He4 per second

I know zilch about high vacuum pumps, and so I don't know if this is a big quantity of He gas at ultra low pressure, but I suppose the lion's share of pumping efforts will be removing neutrals.

NobobyExpects said...

Sorry, the link to the Siemens site has been cropped.

One can find useful info also at:

http://www.siemens.com/HVDC

M. Simon said...

Jose,

Here is the Siemens link:

Siemens

Anonymous said...

This thing smells like snake oil, but as far as the high voltage inverters are concerned, it's a lot easier than you guys are thinking. They simply have a stack of fiber-optic driven photoIGBTs, that are all insulated at the gate. With that, there's no theoretical limit on voltage. Just don't open circuit on one of them before the others (kaboom!). They have to be perfectly coordinated.

A 2MV transformer would be a beast, but I don't think it would require any new technology.

NobobyExpects said...

Curious. I knew of the use of Light Triggered Thyristors (LTT) in HVDC converters, but this is the first notice I have about the existence of photo-IGBTs.

Could you elaborate?

OTH, the issue of very high DC voltages could be addressed using a relatively simple Buck converter between the collecting grid and somewhat standard HVDC switchgear at 500-600kV connecting the power station to a standard HVDC transport line.

But those lines carry high amounts of power, from 500MW to 3GW, so possibly the optimum economical size of a Bussard reactor will be around 1000MW, i.e. like a medium-big nuke fission power plant.

M. Simon said...

At this time there is no "standard" DC line.

I believe there are one or two DC lines in America.

The Ghost said...

Since there has never been a sustained contained fusion reation that lasted more than seconds I'm not sure they are as close to "building" one of these things as you seem to imply ... They have been studing/testing fusion for 20+ years with tons of money both here and in the old Soviet U. with nothing to show for it in the way of a commercial application.

M. Simon said...

Ghost,

The Fusion experiments that have been getting a ton of money are based on heating plasma. Tokamaks.

This approach is much better because it is based on ion acceleration. All the ions are at the same relative speed. It has been getting hardly any money.

Ion acceleration fusion was first proved in 1959. It has since languished.

NobobyExpects said...

For "standard HVDC power line" I meant something in the 500kV range. Now I'm thinking about the HV transport grid in Europe. Here in Spain we use HVAC lines at 220 and 400kV. This reactor will produce power in the 1.2MV range, as we only need a 1.2MV grid to collect the energy of a 2.4MeV alfa particle (2 positive charges). A step-down converter from 1200kV to 566kV (peak value of a 400kVrms voltage) with the relatively low currents involved looks feasible with current technologies-it only needs a static switch, and there are a number of 400kV inverters commissioned as of today. Those are bulky yes, but they work. As in USA you use 500kV AC, your case is slighty easier.

The next step will be a direct converter from 1.2MV DC to 400kV AC or to 500kV AC for y'all USians ;-)

M. Simon said...

Jose Luis,

Yes. I got the number wrong in the article. There will actually be two voltages. Ine around 1.23 MV DC and the other around 1.84 MV.

The reason is you have two different ion energies from the reaction.Yoy will get 1/3 the direct conversion energy at 1.84MV and 2/3 at 1.23 MV.

Actual voltages will be somewhat less due to various losses.

NobobyExpects said...

m. simon

Don't apologize. We both made the same mistake. See my calculations for the vacuum system, on which I said something about 2MV.

And it's an important aspect because cutting the voltage in half increases greatly the feasibility of a power station.

And yes, according to the transcript of the Google talk, there is the fact that the three alphas produced in p B11 reaction are not equal: the first one has 3.76 MeV, and the other two a median energy of 2.46 MeV. If we want to harness the full energy of the first one, we'll need a grid charged to 1.88 MV.

Bussard has stated there'll be a number of grids, charged to different voltages in order to capture all those alphas. Make that two *systems* of grids to account for some scattering due to the variability of the velocities of the reactives. If we don't want to use two step-down DC-DC converters, the electric design could be tricky: maybe it can be solved connecting the grids series-like with intermediate capacitors in-between but I have my doubts.

M. Simon said...

Jose Luis,

I would be willing to design for higher losses to reduce the design effort and construction complicatins for the first model.

In fact I might just use one collector voltage to start and have that collector be an outer wall which would be easy to cool.

May I suggest joining us at

IEC Fusion Newsgroup

and at blog:

IEC Fusion Technology

we are a group of amateurs (at reactor design) working out some of the details in order to pre-engineer as much as possible.

NobobyExpects said...

m. simon

Thanks for the kind offer, but since I discovered the loooong thread at nasaspaceflight.com, I've been almost "following" you from site to site.

In fact, I'm jl_domingo_f in the yahoo group, but I find this blog entry the more EE oriented. Also I post as joseluis at fusor.net.

NobobyExpects said...

Using the wall of the vacuum chamber as collecting grid for an initial proof-of-concept device seems logical, especially if one can use that as a first step, expanding later the collection system by doubling the DC voltage of the wall, and inserting a grid. After all, high vacuum is an excellent dielectric media, so the isolation distances are quite low, but only after providing means of preventing Paschen discharges in transients, for the sake of the integrity of the device.

Anonymous said...

Particle beam gun for fusion

http://www.p2pnet.net/story/12661

I'll bet the device can supply the
necessary electrons to Bussard's
fusion reactor.

Anonymous said...

Bussard's polywell concept is flawed. He's a nice guy and has come up with interesting ideas but they are full of holes, which is why the fusion researchers of the world haven't jumped on it. IEC is a very well-investigated science and there are plenty of people willing to fund it.

You folks should spend less time talking about DC converters -- those are a tiny part of the design. The bigger issue is: would the polywell design actually create net energy. So far, the consensus is no.

M. Simon said...

Doesn't matter what the consensus is.

The consensus was in 1895 that heavier than air craft could never fly significant distances under their own power.

In any case we will know what the Navy thinks in about 9 months more or less.

Bussard Reactor Funded

You might want to look up Rostoker's response to Rider.

If that last sentence means nothing to you you are not deep enough into this to discuss it.

James said...

My favorite line to challenge is:
"So far, the consensus is no."

Who gives a damn about what the consensus says, if No One Has Actually Studied The Damn Thing?

How about anonymous someone's provide a shred of evidence of their own, who assert tripe against third parties who are publishing results, regardless of the political implications of what they're saying?

Anonymous said...

Just a thought, instead of going for large devices producing lots of power, how small could one of these become, household / vehicle usage? If it went small I could see it becoming used in a widespread way very rapidly, without reduced infrastructure needs.

M. Simon said...

The problem with really small devices is that the vacuum system alone comes in at $20K to $100K.

Even the "no neutron" reaction produces some neutrons. So you need shielding.

I'd say that the home or neighborhood reactor may some day be a possibility. Right now the technology is too raw.

gstaw said...

Two observations: The boron - proton fusion reaction requires a much higher energy than D-D or D-T. If I remember correctly, it is approximately 150 KeV. This presents both the problem of providing 150 KeV electrons and the magnetic field to contain them. Either the WB7 ro WB8 machine will have to have superconducting magnets to provide the fields necessary. In my opinion, this will be the major engineering challenge: 4K helium to cool the magnets while producing 2.8 MeV electrons only a foot or two away. Doable, but not easy.

Second observation: Based on the scaling factor and type of technology, I would expect the deployment to be by distributed generation. Put a 10 - 50 MW unit at each substation in your electrical grid with voltage stepped down to whatever the local distribution voltage is (usually 8 - 15 kV) and the only power flow over the transmission lines would be to compensate for local outages as individual lines or generating units come off line for maintenance or repair.

M. Simon said...

gstaw,

You can find more technical info and links to discussion groups at:

IEC Fusion Technology blog

Check the sidebar.