LHC about to restart - some frequently asked questions

[rootone]

I believe it's liquid helium, not water.

Edit:
Found an interesting details page from the CERN website.
Liquid Nitrogen is also involved, not directly in cooling the magnets, but in the process of keeping the helium cooled.
http://home.web.cern.ch/about/engineering/cryogenics-low-temperatures-high-performance

 

[mfb]

The main dipoles are superconducting, water-cooling would not help. About 100 tons of liquid helium at 2 K are used (and liquid nitrogen for pre-cooling), the LHC cryosystem is the largest in the world.
The required electric power is spread out over several components along the ring, they don't need large cooling towers.

 

[rollingstein]

The main dipoles are superconducting, water-cooling would not help. About 100 tons of liquid helium at 2 K are used (and liquid nitrogen for pre-cooling), the LHC cryosystem is the largest in the world.

I see. So, they keep the He in a closed circuit & reliquify whatever evaporates via expansion turbines etc.? He might be too expensive to lose.

OTOH the liq. N2 they just let evaporate away & order tanker trucks from Linde / Praxair etc.? Or is there a cryogenic air separation unit on site? At those tonnages might make sense.

 

[rollingstein]

The more I read about the LHC the more I'm fascinated by the Engineering rather than the actual Physics it is trying to do. :)

 

[mfb]

Helium is too expensive to lose, that is in a closed cycle.
CERN makes the liquid nitrogen on site - I don't know how much gets recycled and how much is extracted from air, but 10,000 tons of liquid nitrogen would be impractical to deliver with trucks.

See the LHC cooling page for more details.

 

[Vanadium 50]

I meant what is the max peak load that LHC pulls from the grid.

But the LHC itself, being superconducting, draws almost no power at flattop. It's all in the other systems - cooling, injectors, etc.

 

[dlgoff]

120 MW in regular operation is a number given by several sources ...
It is taken from the grid.

So in bringing the system up to this, I assume there's a time schedule/procedure to allow grid connected generators to respond?

 

[BilboBombadillo]

How did you come to work there?
What/where did you study?
What first got you interested in all this, and at what age?
(trying to decipher my chances of ever doing anything science related professionally)
Thanks for your work, and for this great thread!

 

[rollingstein]

So in bringing the system up to this, I assume there's a time schedule/procedure to allow grid connected generators to respond?

I suppose the ramp up is inherently slow anyways. The cryo plants seem modular & they will get brought online successively. So the 120 MW won't be a step loading,

 

[mfb]

What first got you interested in all this, and at what age?

I don't remember, earlier than 2008.

What/where did you study?

Physics at a German university.

How did you come to work there?

I applied for a PhD position and got it. Nothing special.

I suppose the ramp up is inherently slow anyways.

It took about 20 minutes for the old energy, probably longer for the higher energy now. Ramping up the magnets does not increase the power load so much.

 

[docK]

360 megajoules is nominal.
Not this year, and likely never. The magnets made by "Firm 3" seem to be struggling.

You know the answer to that should always initially be 1.21 Gigawatts

 

[rollingstein]

It took about 20 minutes for the old energy, probably longer for the higher energy now. Ramping up the magnets does not increase the power load so much.

I was assuming the biggest power ramp up came during initial cool down?

 

[mfb]

Cooling down the whole machine takes weeks.

 

[rollingstein]

What happens to accelerators when they grow old? Do they rip them apart for parts? Scrap them? Keep upgrading them? Convert to museums?

How long is the LHC projected to stay useful enough to fund?

 

[mfb]

What happens to accelerators when they grow old? Do they rip them apart for parts? Scrap them? Keep upgrading them? Convert to museums?

Depends on the accelerators.
Often a larger accelerator is built and the old one is used as preaccelerator for the newer one. PS at CERN for example is over 50 years old, it got upgraded multiple times and it is used as preaccelerator for the larger SPS now (which itself is a preaccelerator for the LHC, but it is still used for other experiments).
Sometimes they get disassembled (LEP for example, as the tunnel is used for the LHC now)
Sometimes they are just left in place (Tevatron at Fermilab, HERA at DESY), parts might get used for exhibitions and/or the original site can be visited.
Sometimes parts are re-used elsewhere (a ring moved from Brookhaven to Fermilab, a HERA magnet was used for ALPS).

In general, accelerator parts close to the beam get radioactive over time. Unless you really need the tunnel or the components, it is easier to leave the parts where they are.

How long is the LHC projected to stay useful enough to fund?

The current plan ends ~2035. New discoveries might lead to extensions, but radiation damage to the detectors will get a serious issue then.

 

[mheslep]

Helium is too expensive to lose, that is in a closed cycle.
CERN makes the liquid nitrogen on site ...

Apparently low temperature superconductors (versus HTS) are still the only way to play in an actual application. I suspect this 2005 article stating that "HTS are nearing their commercial viability" is still accurate in 2015: "nearing", but no actual commercial viability yet.

 

[rollingstein]

Apparently low temperature superconductors (versus HTS) are still the only way to play in an actual application. I suspect this 2005 article stating that "HTS are nearing their commercial viability" is still accurate in 2015: "nearing", but no actual commercial viability yet.

There's an astounding number of technologies that seem to be sold as being on the brink of commercial viability and stolidly stay on that brink for decades.

 

[mheslep]

Helium is too expensive to lose, that is in a closed cycle.
CERN makes the liquid nitrogen on site - I don't know how much gets recycled and how much is extracted from air, but 10,000 tons of liquid nitrogen would be impractical to deliver with trucks.

See the LHC cooling page for more details.

As that link indicates, nitrogen is an intermediate step for cooling the LHC superconductors by helium.

...During the first stage, some 10,000 tonnes of liquid nitrogen are used in heat exchangers in the refrigerating equipment to bring the temperature of the helium down to 80 K.

The helium is then cooled to 4.5 K (-268.7°C) using turbines. Once the magnets have been filled, the 1.8 K refrigeration units bring the temperature down yet further to 1.9 K (-271.3°C).

The system is apparently closed loop once up and charged:

The LHC's cryogenic system requires 40,000 leak-tight pipe seals, 40 MW of electricity – 10 times more than is needed to power a locomotive – and 120 tonnes of helium to keep the magnets at 1.9 K.

 

[mfb]

Apparently low temperature superconductors (versus HTS) are still the only way to play in an actual application. I suspect this 2005 article stating that "HTS are nearing their commercial viability" is still accurate in 2015: "nearing", but no actual commercial viability yet.

I think they were considered for the possible high-energy LHC to get ~15+ Tesla, but that idea probably died together with the high-energy LHC (didn't see that for a while now).

There is a 1km HTS cable in Essen, Germany, as part of the local power grid (http://www.suptech.com/Cables_Oct_10.pdf). A pilot project, sure, but at least something.

The system is apparently closed loop once up and charged:

For helium, sure, for nitrogen we don't know (but I would expect it).

 

[TheDemx27]

What is the main obstacle/bottleneck in terms of upping the eV's?

 

[mfb]

What is the main obstacle/bottleneck in terms of upping the eV's?

The field strength of the dipole magnets and the curvature of the tunnel.
Protons with more energy would hit the outer wall of the beam pipe as the magnets are not strong enough to keep them on track.

Tunnel:
For LEP, the limiting factor was acceleration versus synchrotron radiation, so the tunnel was built with 8 curves and 8 straight sections for acceleration. That design is not optimal for the LHC, but making the ring "more circular" to reach a higher energy would have needed a lot of time and money, so they decided to take the old geometry.

Field strength:
The dipole magnets are supposed to be strong enough to handle 7 TeV per proton. Unfortunately, those magnets need "training" - if you send too much current through them, they lose their superconductivity, this is called quench. A quench comes with mechanical changes of the precise geometry of the coils, and it is known that those quenches (usually) increase the current the magnets tolerate. That is repeated until the magnets are trained for the current you plan to run through them. For some magnets, training for 6.5 TeV took a long time, it is unclear if the design value can be reached at all (but they will certainly try).

 

[rootone]

That's interesting, does it mean that micro impreciseness in the construction of each dipole, it needs to be carefully adjusted for either in a hard or soft way, and that can be and is done, in order to get best quality results from the experiments?

 

[Vanadium 50]

"nearing", but no actual commercial viability yet.

You can buy commercial HTS leads today. They are a common way of powering superconducting magnets.

 

[rollingstein]

The training bit is very interesting? Is there a theory why the geometry of a magnet should adjust itself in this way?

 

[Vanadium 50]

Is there a theory why the geometry of a magnet should adjust itself in this way?

No. It probably has to do with mechanical stresses, but the phenomenon is not well understood.

 

[rollingstein]

There is a 1km HTS cable in Essen, Germany, as part of the local power grid (http://www.suptech.com/Cables_Oct_10.pdf). A pilot project, sure, but at least something.

Interesting indeed. HTS seems more realistic than I had thought.

Do these transmission wires need a liq. N2 pumping source continuously?

 

[mfb]

The training bit is very interesting? Is there a theory why the geometry of a magnet should adjust itself in this way?

You can find some articles with keywords like "magnet training superconductors", but as Vanadium said - it is not well understood. Training works, so it is done.

Do these transmission wires need a liq. N2 pumping source continuously?

They need constant cooling, I don't know details of it. While this requires constant power, using normal conducting cables needs power as well.

 

[QuasiParticle]

HTS magnet technology is not quite yet ready for high-field accelerator magnets, but development at CERN and elsewhere is ongoing.

Like mfb mentioned, there are no plans to increase the LHC energy. Instead they are studying the feasibility of a new 80-100 km accelerator, the FCC. The LHC will undergo a few updates in the future, for example to increase the luminosity, but the main dipoles will not be replaced.

The mentioned 15 tesla magnets would likely be possible to achieve with Nb3Sn technology (low-temperature superconductor). With HTS it's possible to go up to 20 T, but the cost per tesla increases rapidly at high fields (also true for LTS-only magnets).

That's interesting, does it mean that micro impreciseness in the construction of each dipole, it needs to be carefully adjusted for either in a hard or soft way, and that can be and is done, in order to get best quality results from the experiments?

The field quality requirements for the magnets are quite stringent and they are carefully designed and manufactured. There are also orbit corrector magnets in the LHC, which can adjust the beam positions if deviations are observed.

 

[mheslep]

You can buy commercial HTS leads today. They are a common way of powering superconducting magnets.

HTS has some current application as a way of powering the *leads* as you say, not yet the bulk magnet, as the additional HTS cost is worth the heat loss savings at the lead (per Sheahen). Still no nitrogen cooling in the case of HTS leads on GE MRIs.

 

[rollingstein]

Still no nitrogen cooling in the case of HTS leads on GE MRIs.

Interesting. How high a temperature can these leads remain superconducting at? At their typical currents.