Plasma physicist discusses the Wendelstein 7-X stellarator | Hacker News Reader

Plasma physicist discusses the Wendelstein 7-X stellarator

phys.org

99 pointsAnchor19 comments

"We had to search every corner of Europe to find a company that could mill the segments of the central ring with such accuracy. In the north of Italy we found CLP, a family business in a small village"

Such amusing imagery. When you think of small villages in Italy you tend not to think of them having companies with the ability to do such cutting edge engineering.

This is the company: http://www.officineclp.it/en/

Northern Italy has its own character, so maybe not so surprising to find precision engineering.

Trust me, Italy is FULL of highly talented engineers, mechanics, etc, and small companies like CLP.

Unfortunately for them, they're not the best at selling their services globally.

Source: I am Italian.

Kind of funny that it's not even listed on their news section, or anywhere about their skills. :)

Italians were fine machinists through their history. Da Vince, etc - long line of mechanics and precision millwork.

A lot of aerospace companies making satellites, rockets, parts for ESA are in Northern Italy. Three of the ISS modules[1] were made near Turin.

Also, Ferraris and Lambos are made in Northern Italy (in relatively small towns, ex. Maranello is 17,000 people). High performance and precision engineering.

[1] https://en.wikipedia.org/wiki/Columbus_%28ISS_module%29 https://en.wikipedia.org/wiki/Harmony_%28ISS_module%29 https://en.wikipedia.org/wiki/Tranquility_%28ISS_module%29

And lets not forget the musical domain: Stradivarius [1]

[1]https://en.wikipedia.org/wiki/Stradivarius

The best Dutch machinists that I'm aware of (when it comes to precision) are located in a tiny little village in the south of Groningen, about 2 km from the German border. It's as out-of-the-way as it gets.

The problem why they searched all over Europe is that it wasn't just the accuracy, it was the part size combined with the accuracy. Making small parts with a lot of precision is relatively easy if you're willing to discard a whole pile of them after QA. Making big parts with any precision at all is tremendously hard (because the piece warms up as it is being worked and large pieces will expand quite a bit as they warm up) and discarding a single piece is going to be ridiculously expensive.

No, no, northern Italy is legendary for its small-to-medium businesses in fields like machine tools http://www.industryweek.com/companies-amp-executives/italian... and, you know, the automotive industry. This is illustrated in epic style by the true story of how Ferruccio Lamborghini decided to get into making cars: http://www.400gt.com/articles/compare/interview.htm .

Very interesting interview about very interesting topic.

One minor complaint: answer to the second question seems to have missed/ignored the question? I assume the true answer lies in the fact that the technical realization of the stellarator are even more complex than that of a tokamak and there is more global experience with tokamaks.

I can give this a shot. There is a sort of simpler explanation of what happens in a stellarator, or at least, why the shape is so weird:

When a plasma is confined in a torus, it is spinning (naturally). However, this creates an effective centrifuge, which means that the hottest atoms -- which are also the "lightest" -- will move towards the outside of the ring, and ultimately will escape confinement. This cools the plasma.

The stellarator solution is basically to turn the torus into a Möbius strip, so that ions which have drifted outside are naturally shuttled to the inside (for the same reason that the "inside" and "outside" of a Möbius strip are the same). However, in order to retain rotational symmetry, the strip has five twists, rather than just one (the topology remains the same).

Unfortunately, calculating the dynamics of a plasma shaped like this is hard -- really hard. It requires good computers, so for many years stellarators were mostly an object of theoretical interest. Additionally in order to deal with other instabilities the Wendelstein has an even more refined shape than the five-twisted strip.

ITER was conceived in 1987, and at the time it was necessary to go with a technology that was well-tested and understood, and which people felt confident would perform as expected. At that time, using a stellarator made as much sense as writing Firefox OS in Rust or something.

Thanks for a layman-lucid explanation. Does the Polywell design also fit into the overall fusion research picture as yet another design that might one day contribute parts of itself to the final working design as the Max Planck director suggested would happen with the tokamak and stellarator designs?

The polywell is an interesting concept, but it is extremely different from every other confinement mechanism, so it's hard to see how the ideas could be incorporated. To be precise, there are two "shapes" of "electromagnetic fields": curl-free (electric (unless magnetic monopoles exist)) and divergence-free (magnetic or electric), and each kind of field lends itself to a different kind of degree of symmetry in a confinement mechanism. Specifically, a divergence-free field cannot admit a spherical confinement because of Green's theorem, whereas a curl-free field requires a very unrealistic charge configuration to generate a toroidal confinement.

Tokamaks use a combination of divergence-free magnetic fields and divergence-free electric fields to make a donut. Stellarators use only magnetic fields to make a twisted donut. The polywell uses curl-free electric fields generated by an electron gun to make a sphere. There are other variants of inertial electrostatic confinement fusion, including that currently under development by Lockheed Martin, but it's generally received less attention.

tokamak

+ can be heated using the plasma current (although you have to add additional heating to reach the required temperature and control the plasma)

+ simpler construction

+ older and therefore people have more experience with it

- pulsed operation

- collapse of plasma current (disruption) can cause massive mechanical stress (hundreds of tons)

stellerator:

+ continuous operation

+ no risk of disruptions

- complicated construction

- external heating of plasma

- has to be larger than a tokamak to retain the same volume of plasma

Nitpick: IIRC the stellarator is in fact the slightly older design. It's just that it was more or less given up on in favor of the tokamak, due to the design and construction issues.

You are right: stellerator 1950, tokamak 1952 according to wikipedia

I wonder what's up with Lockheed's reactor prototype at this point. This is some truly badass engineering, but the complexities involved inevitably make you wonder if there's an easier way. "Only Germans can build it" does not bode well for mass production worldwide.

Since this is a research reactor, they probably had to make more structural compromises to accommodate more sensors, injection ports, etc than what a power plant version would need. I bet once they played with it for a few years they would have s better idea of what they actually need for a more practical version.

This design is fundamentally hard, I don't think the removal of ports will make it much simpler. Same with Tokamak. Lockheed's design (as disclosed) is much simpler. OTOH Lockheed's design is also vaporware until they publish their experimental data. They've supposedly done hundreds of firings by now, but haven't disclosed any data.

>"Only Germans can build it" does not bode well for mass production worldwide.

That is in the context of first try though. i.e. It sounds like they need to over-engineering it in the face of uncertainty. Once someone pulls it off successfully & publishes the details the 2nd attempt should be a lot easier.

Reply on news.ycombinator.com