Cryogenic technology for LHC - Cern

actorrattleΠολεοδομικά Έργα

15 Νοε 2013 (πριν από 3 χρόνια και 6 μήνες)

205 εμφανίσεις

Basic Cryogenic Document

Date: 27

Prepared By


INDIA Collaborator, Department of Atomic Energy,
VECC, Kolkata, India.





Page No.



Helium and its Superfluidity


Liquefaction of helium and Superfluid helium………………………………04

Dipole Cold Mass and Cryostat


Dipole Cooling at 1.9 K


AntiCryostat for Dipole


LHC Cryogenic Architecture


Superconductor and its Critical parameters


Superconductor for LHC Dipole


Quench in magnet









Cryogenic technology to produce and maintain low temperature plays a very
crucial role in modern large
scale accelerators like LHC, Tevatron, HERA, etc
magnets are used to bend and mainta
in the path of accelerating particle beams. The
higher the field strength of the magnets result tighter the arc of the beam. With stronger
dipole magnets, an accelerator can push particles to much higher relativistic energies
around the same
sized circular

beam path. The use of high
field strength superconducting
electromagnets has always been a considerable technical challenge, because
superconductivity has a tendency to weaken and disappear in the presence of a strong
magnetic field. Nonetheless, the inhe
rent limitations of conventional electromagnets that
they can not attain a dipole field strength above 2 Tesla

has prompted a continuing
development for new and better superconducting alloys.

The large hadron collider (LHC) at CERN is the largest proje
ct to use
superconducting magnets at 1.8 K superfluid helium temperature. Total inventory of
liquid helium during the operation of LHC would be around 700,000 litres[1]. To supply
the cooling power to strings of magnets, eight helium liquefier stations of
power 18 kW (at 4.5 K) each will be located at several points along the accelerator.

1. Helium and its Superfluidity:

Helium is the only gas that makes a good superfluid because it has very weak
intermolecular forces. Superfluidity was first

demonstrated in Helium
4 in 1962 by
Landau. Helium condenses to a liquid at 4.2 K, and turns into a superfluid at 2.17 K. The
point at which it becomes superfluid is called the lambda point, because its specific heat
graph looks like a lambda (

). The poi
nt is characterized by a jump in specific heat, and a
discontinuity in its density graph. Liquid helium remains in the liquid phase under its
own vapor pressure and would apparently do so right down to absolute zero temperature
(Figure 1). Due to the small

mass and extremely weak forces between the helium atoms,
pressure is required to produce solid helium (25 atmospheres or more)




The He II phase is referred to as superfluid helium because of its remarkable properties,

i)Very high t
hermal conductivity… is able to conduct away heat a thousand times better
than a metallic conductor like copper.

ii) Very low coefficient of viscosity… can penetrate tiny cracks, deep inside the magnet
coils to absorb any generated heat.

iii)Very high heat

capacity…prevents small transient temperature fluctuations.


Phase diagram of

2.Liquefaction of helium and superfluid helium:

The basic principle of the helium cooling and liquefaction process are illustrated in Fig.2.
Highly purified (g
rade A

99.995 % purity) high
pressure (around 15 bar) helium gas
from the screw compressor flows along path 2, 3, and 16 through a series of plate
fin heat
exchangers and turbo expanders to an expansion valve (JT). After isenthalpic expansion
at JT, fract
ion of gas is liquefied and stored in the 4.5 K liquid helium dewar. The
uncondensed portion of the helium, which is very cold, is returned to the compressor
through the heat exchangers (as cold fluid).

The result is that the warm high
helium st
ream is cooled by the cold low
pressure helium stream which itself warms up.
This process will only function if the gas temperature at point 16 is low enough for Joule


Thomson cooling to take place (Should be around 5 K, 3
3.5 bar(a), but depending on the
design it may vary). The pre
cooling is carried out using the Brayton process. In the
diagram it may be seen that a portion of the high pressure gas stream is diverted at point
4, 8, and 10 into the expansion turbine to undergo isentropic expansion. Both t
he pressure
and the temperature of the gas fall after it has done work in the turbine, and the cooled
gas is returned to the low
pressure part of the circulation system. Helium flowing back
from the low temperature region, and the combined flows cool the h
igh pressure stream
via heat exchangers. This ensures that the high pressure gas is at a low enough
temperature to obtain cooling when it is expanded through the valve.


Conceptual helium liquefaction process

The combination of the Brayton pro
cess and the liquefaction by Joule
expansion is called the
Claude process
. In the presence of liquid nitrogen pre


liquid helium production rate increases more than 1.5 times. To obtain one litre of liquid
helium, however, around 700 liters

of the gas (at atmospheric pressure) is needed.

Liquid helium at 4.5 K at 1.2 bar(a) is expanded through JT valve again to a pressure of
20 mbar(a) or less (maintained by pumping in a controlled way) to get superfluid helium
at 1.8 K.

Dipole Cold M
ass and Cryostat:

Cryostat designs for superconducting magnets are largely driven by thermal and
structural considerations. A cryostat consists of a 15 m long, 1 m diameter mild
vacuum vessel, aluminium thermal shields wrapped with multi
layer insul
surrounding the cold mass and three support posts carrying the magnet. The dipole cold
mass consists of two dipole coils, a common non
magnetic, force
retaining laminated
structure (collar) made up of stainless steel, laminated iron yoke, all surroun
ded by
shrinking cylinder (HeII vessel), made up of two stainless steel half cylinders welded
together, thermally insulated support system, the cryogenic piping, and corrugated copper
heat exchanger tube. The cold mass of dipole MBA contains corrector sext
octupole and decapole, whereas MBB cold mass contains sextupole corrector only. In
order to reduce the heat generated by proton beams from their synchrotron radiation,
beam screen has been provided inside both the dipole coil apertures. When inserte
d in its
cryostat, the dipole cold mass is supported by three posts which also provides the thermal
insulation. Once installed in their cryostats, the cold masses must be equipped with
instrumentation feedthroughs, and various ancillary equipments as shown

Fig.3.1 and

The thermal radiation shield surrounding the shrinking cylinder is kept at around 15
20 K
(By gaseous He flow) to minimize the radiation heat load to 1.9 K superfluid helium. An
intermediate aluminium thermal screen surrounded in
between outer vacuum chamber
(OVC) and 15
20 K radiation shielding is maintained at around 55
77 K (By gaseous He
at 50 K, 20 bar) to reduce the heat load to the radiation shield around shrinking cylinder.
In addition, multilayer super insulation (MLI) aro
und the shrinking cylinder and radiation
shield are provided to reduce the conduction and radiation heat load to the system. The


vacuum vessel contains insulation vacuum at a pressure of around 10

mbar. It is
equipped with safety relief valve (SRV) oper
ating at about 0.5 bar overpressure.

Fig. 3.1

Three dimensional view of cryodipole


Fig.3.2 Dipole cross
sectional view



4. Dipole Cooling at 1.9 K:


superconducting devices operating in superfluid helium can be cooled with
saturated or pressurized superfluid helium. Devices cooled with superfluid
helium have to operate at a pressure below 20 mbar. At this pressure, the gaseous helium
phase h
as a poor dielectric strength. Consequently, to avoid electrical breakdown, this
cooling method has to be applied only for devices, which can operate in any
circumstances with a voltage difference of about 100 V [2]. Moreover, working in sub
atmospheric co
nditions increases the risk of air inleaks through the different electrical
and instrument feedthroughs, which are unavoidable in superconducting apparatus.
Devices cooled with pressurized superfluid helium are prevented from air inleaks and can
be operate
d with voltage differences in the kV range. Liquid
liquid heat exchangers are
required, introducing an additional temperature difference for the heat extraction.
However, the volume of saturated liquid in these heat exchangers can be small in
comparison wi
th the liquid needed at the device side.

The simplified cryogenic flow scheme of an LHC dipole is shown in Figure 4.1, where as
the actual cooling circuit followed in SM18 power test is shown in Figure 4.2. The
pressurized superfluid helium bath at 1.9 K,

in which the superconducting magnets are
immersed is cooled by saturated two
phase liquid helium flowing in heat exchanger tubes
extending along the string of magnets and supplied by line A through the expansion valve
TCV1. The low saturation pressure is
maintained by pumping the vapour through line B.
Cooldown and warm
up is achieved by forced circulation of high
pressure gaseous
helium supplied at variable temperature by line C, tapped through valve CFV and
returned to the refrigerator by valve SRV and l
ine D. In case of magnet’s resistive
transition, the resulting pressure rise is contained below the 2 MPa design pressure by
discharging the liquid helium inventory of the magnet into line D through the SRV

The temperature levels are:

thermal sh
ielding between 50 K and 75 K as a first major heat intercept, lines E and F.

distribution of supercritical helium at 4.5 K by line C for initial filling of the magnet.



isothermal superfluid helium cooling the superconducting magnets at a maximum

temperature of 1.9 K and transporting the heat loads across the length of the magnet at
1.8 K.


LHC dipole cooling scheme for power test

Fig.4.2 Actual magnet cooling circuit for power test

(Courtesy: Cryo
operation group, SM18)



gaseous helium cooling the resistive upper sections of Bi

HTS ceramic current leads between 20 K and ambient.

5. Anticryostat for Dipole:

All the

LHC dipoles are tested at cryogenic temperature, while the test equipment
for m
agnetic field measurement works at ambient temperature. The necessary
measurement conditions are provided by warm bore anticryostats (one for each aperture).
An anticryostat is built as a coaxial tube system, consisting of seamless 316L stainless
steel (wi
th an inner diameter of 40 mm and a wall thickness of 0.7 mm) maintained at
ambient temperature with the help of four mineral
insulated coaxial heater cables with an
outside stainless steel jacket of 0.5 mm diameter and inside conductor of 0.09 mm
. Each of them forms over its total length a loop, which is soft soldered in the
form of a helix around the outside surface of the warm bore. The heater wires are excited
with maximum current of 1 A and a maximum voltage of 30 V. During normal operation,
the average dissipation is about 0.8 W/m for all the four heaters in one aperture, though
the maximum power dissipation for single heater is up to 2 W/m [6]. A 10 µm thick
silver layer deposited on the outer and an inner surface of the outer tube (screen t
reduces the total heat losses of anticryostats by about 20% [6] at operation temperature
relative to non
treated tubes. An aluminum ribbon is wrapped around the warm bore and
the heaters in order to reduce the emissivity of the surface and hence radi
ation heat loads
to the cold mass. In addition, on top of the aluminum ribbon two blankets of multi layer
insulation (MLI) with four layers separated by mylar net have been provided. Finally for
mechanical protection during handling, the MLI is covered by
a thin walled stainless steel
tube. In order to facilitate the assembly, this tube is cut in 700 mm long segments and
mounted step by step together with the MLI. For the mechanical stability of the assembly,
polyimide insulation spacers are provided radial
ly in between inner and outer tube
through out the length. The cross
sectional view of anticryostat is shown in Figure 5.

The electrical heating system is equipped with a DC power supply, which feeds three of
the four heaters with a current up to 1A. T
he fourth is used as a temperature gauge.
Current and voltage of the gauge are controlled and regulated by PLC w.r.t. to a reference


temperature. Temperatures in three heaters are transferred to a supervision system in the
control room of the test station


sectional view of antocryostat.

6.LHC Cryogenic Architecture:

6.1 Cooling for Half Cell:

Eight large cryogenic refrigerators (18 kW each at 4.5 K) will produce
refrigeration for the LHC ring (Fig.6.1). Each of these refrigerators will h
ave to cool
down and warm up about 4500 tons of stainless steel and aluminium and to liquefy about
eight tons of helium [4].

Each plant normally supplies a whole LHC machine sector of
about 3.3 km length via a cryogenic distribution line, with interconnec
tions at every basic
machine cell length of 107 m. The elementary block of cryogenic refrigeration system
corresponding to a half
cell (about 53.5 m) of the machine lattice is shown in Fig. 6.2.
The magnets are immersed in static pressurized superfluid hel
ium, cooled by heat
exchange with saturated superfluid helium flowing inside the tubes running all along the
length of the half
cell string. Sub
cooled liquid helium arrives at each 53.5 m long loop
through line A. It first expands in the JT valve (LCV). I
f all magnets are to be adequately
cooled irrespective of their position in the string, the liquid flow through this valve must
exceed the boil
off rate in the cooling loop. The excess liquid is then sent in a phase
separator and maintained at saturation p
ressure by cold pumping line B. The heat
exchange loop itself consists of two tubes, one inside the other. Saturated phase II helium
goes out through the inner tube and returns through the outer tube. The tubes thread


through the upper median hole in the m
agnet yoke and outer one forms the surface of
heat exchanger.

6.2 Beam Screen Cooling:

The screen which surround the beam inside each of the two magnet apertures are
cooled by forced flow of supercritical helium (4.5 K, 3 bar) tapped from distribution li
C and controlled as a function of outlet temperature by valve TCV. The gaseous helium
flow is returned to the low pressure side of the octant refrigerator through line D.

6.3 Thermal shield:

The thermal shield of the magnets is cooled to a temperature

in the range from 50
75 K by forced flow of cold gaseous helium, tapped at a suitable level from the octant
refrigerator. No liquid nitrogen will be used in the tunnel in order to simplify the flow
scheme of cryogen and to avoid asphyxiation (oxygen defic
iency in case of a leakage in
the line).

6.4 Cooldown:

Cooling of magnets from ambient is carried out by forced circulation of gaseous
helium at progressively decreasing temperature, tapped by Cooldown and fill valves
(CFV) at each half cell from supply l
ine C, and returned through valve SRV to the low
pressure side of the octant refrigerator via line D. Once they have reached sufficiently
low temperature, the magnets will be filled with helium at 4.5 K through CFV, and
exhaust valve SRV will then be close

Final cooling is achieved by initiating the flow in the cooling loop and then lowering its
saturated vapour pressure by pumping through line B. The initial cooling
down of the
machine from ambient to 1.8 K will take around 25 days [5] and the total inv
entory of
liquid helium is around 700,000 litres. This amount of helium can easily accommodate in
the gap between the laminations of magnet yokes.



Fig. 6.2

Cryogenic distribution system for the

half cell [4]


Helium refrigerators around LHC



7. Superconductor and its critical parameters:

, materials that have no resistance to the flow of electricity, are
one of the last great frontiers of scientific discovery. In 1911 superconductivity was first
observed in me
rcury by Dutch physicist Kamerlingh Onnes.

For a superconductor, it is important to remember the three critical parameters, T
, B

and J
. They define the boundaries of the environment within which a superconductor
can operate, i.e. in temperatures up to


(Critical temperature), magnetic fields up to B

(Critical magnetic field) and with currents up to J

(Critical current density). It must also
be noted that B

varies as a function of temperature and J

varies as a function of field
and temperature.
The superconducting state is an ordered state of the conduction
electrons of the metal. The order is in the formation of loosely associated

pairs of
electrons. Below the transition temperature, the electrons are ordered and form cooper
pairs while disorder

occurs above the transition

A sufficiently strong
magnetic field can destroy the superconductivity. For some superconductors, the values



too low to have any useful technical application in superconducting magnets.
These are called
type I superconductors
. In contrast,

type II superconductors

exhibit a
magnetization curve, which makes them far more suitable for magnets. They tend to be
alloys with high resistivity (short mean free path) in a normal state. The superconductor
used for t
he LHC dipole is made of NbTi (type II).

Superconductors have two outstanding features:


Zero electrical resistivity

This means that an electrical current in a
superconducting ring continues indefinitely until a force is applied to oppose the
current (Fig


The magnetic field inside a bulk sample is zero


When a
magnetic field is applied current flows in the outer skin of the material leading to
an induced magneti
c field that exactly opposes the applied field. The material
becomes strongly diamagnetic as a result (as shown in fig. 7.2).




Temperature response of resistance

Fig. 7.2

Magnetic flux exclusion in superconductor

8. Superconductor for LHC

The LHC dipole uses superconducting cable, composed of strands where NbTi
(the superconductive alloy) filaments are embedded in a copper matrix. The specification
of the cable is shown below.

Cable Characteristics:

type NbTi cabl

28 strands

edge thickness: 2.064 + 0.006 mm

edge thickness: 1.736 + 0.006 mm

Width: 15.1 + 0/
0.02 mm


Strand Characteristics:

Diameter: 1.065 + 0.0025 mm

Filament diameter : ~7 µm

Number of filaments: ~8900

Stabrite coating (bet
ween 0.4 and 0.6 µm)

Cable Characteristics:

type NbTi cable

36 strands

edge thickness: 1.598 + 0.006 mm

edge thickness: 1.362 + 0.006 mm

Width: 15.1 + 0/

0.02 mm


Strand Characteristics:

Diameter : 0.8250 + 0.0025

Filament diameter: ~6 µm

Number of filament: ~6400

Stabrite coating (between 0.4 and 0.6 µm)


9. Quench in magnet:

The process when a superconductor becomes resistive i
s known as quench. The
superconducting cables are exposed to a variety of disturbances which may heat the coil
locally beyond the critical temperature, i.e. wire motion during excitation of the magnet,
cracking of the epoxy insulation or beam losses. In or
der to avoid movements of a
winding that can cause a quench, the magnet coils are pre
stressed with collars at ambient
temperature such that a stress remains at cold and the windings cannot move. Otherwise
the Lorentz force (electromagnetic) that is built
up while increasing the current (current
ramp) causes a few turns to move towards a more stable position, which can provoke a


quench. If the prestress is too high, the critical current is reduced because of the impact of
the critical parameters of the supe
rconductor on the applied strain. In some places the
cable can still move depending on the quality of the coil winding. One of the basic
reasons of quench is that the heat capacity (C) of composite conductor at 1.9 K is around
2000 times less than at ambie
nt value. Thus a small evolution of heat energy (~

J) inside
the strand or cable is sufficient to make it resistive. Therefore, temperature locally in the
resistive region increases, which in turn reduces the critical current capacity of the
magnet and the
refore quench in the whole magnet occurs.

During the quench, all the helium inside cryostat evaporates within a second and the
pressure inside the cryostat builds up instantaneously to about 15
20 bar depending on
the current in the coil. Thus immediately
on quench detection (10
15 ms), quench valve
should be opened so that the evaporated gas can be collected in the gas bag (balloon). As
the helium pressure inside is maximum during a quench, personnel are advised not to
stand close to the longitudinal shaft

position (mechanically weakest) during quench.

9.1 Different Quenches:

9.1.1 Training Quench:

When the current is increased in the magnet a movement appears at a position
with little mechanical stability. The mechanism is known as a
training quench
. Du
the next magnet excitation a quench can start at another location. After several
excitations, the dipole is capable to sustain ultimate field without any quench. The
magnet is said to be trained. The number of quenches needed to reach this ultimate fi
and the closeness with which it reaches the critical current at that temperature is a
measure for the quality of the magnet.

9.1.2 Heater provoked quenches:

Quenches can also be provoked by the protection system when quench heaters are
installed. Thi
s quench is generally performed at low current in the magnet to analyze the
effectiveness of the protection heaters (All LF and All HF provoked quench). Heaters are
also utilized to determine the Minimum Quench Energy (MQE) at a given current in the
to study the quench propagation velocity as well as the effectiveness of the heater. In


this case, energy to the magnet coil is provided by LF heaters and is protected by HF

9.1.3 Cryogenic failure:

If the cryogenic system fails, the temperature
increases which can provoke
quenches. The temperature margin in case of transient heat deposition is about 1.4 K at
nominal field. The temperature margin for continuous heat deposition is however 0.26 K
at nominal field and at 1.9 K temperature [3].


Beam induced quenches:

In an accelerator, quenches in superconducting magnets can occur because of
beam losses, which would lead to an energy deposition in the superconductor high
enough to exceed the critical temperature.

9.1.5 Internal joule dissipatio

Quench may occur during ramping due to cable hysteresis loss, inter
coupling, inter
filament coupling, resistive joints (splices), etc. For main dipole, magnet
total joule dissipation (energy) is 340 J/m. Out of this, 240 J/m is due to hysteresis


9.1.6 Electric faults:

Quench may also happen due to false quench detection, false triggering of magnet
protection, or powering the coil above allowable current.

9.2. Quench Parameters (MIITs, Hotspot) of a magnet:

The source of conductor hea
ting in a quenching magnet is power dissipation by
the Joule effect. Hence, to eliminate the heat source and limit temperature rise, it is
necessary to ramp down the magnet current. To prevent burnout, it is desirable to
maximize the volume in which the en
ergy is dissipated by ensuring that the normal
resistive zone spreads rapidly throughout the quenching coil. This can be done by means
of heaters, installed near the magnet coils and fired as soon as a quench is detected. These


heaters are referred to as
uench protection heaters
. The volume of conductor that heats
up most significantly during a quench is the hot spot where the quench first originated.

The maximum hot spot temperature for LHC dipole or most of the other accelerator
superconducting magnets s
hould not exceed beyond 350
400 K, and whenever possible,
the limit is set to 100 K. The maximum
hot spot

temperature (T
) of the magnet can be
calculated from the adiabatic approximation as following


ere, QI is known as quench integral (expressed in unit of
), i

is the initial current
before decay, τ

is the detection time delay (along with heater delay), i is the current at
any instant of time in the magnet during discharging after a quench, A is the cross
sectional area of the composite conductor, C is the heat

capacity per unit volume of the
composite conductor,

is the resistivity of composite conductor, λ is copper to
superconductor ratio in the composite, T

and T

are the initial bath temperature and final
hotspot temperature respectively.

The MIIT number

of a magnet is dependent on the protection system.

As seen from the
basic equation, lower the value of MIIT results reduction of hot spot temperature. The
maximum value of MIIT can be estimated with the designed allowable hot spot
temperature for ultimat
e current in the magnet and accordingly, the protection system is
designed. If the MIIT value estimated from time integral of current exceeds the designed
estimated MIIT value, then it is called “
MIIT exceeds upper limit

9.3 Quench Heaters

Quench heat
ers are installed to protect magnet coils from overheating and
excessive voltages (By increasing quench propagation velocity). They consist of stainless
steel strips that are positioned along the magnet between the coil and the collars and are
heated by a
capacitor bank discharge after quench detection. The heat is transferred
through the insulation layer into the coil, which provokes a quench. Main magnets
connected in series with a string of magnets are additionally protected with diodes that


are mounted
parallel to the magnet coil. The current bypasses the quenching magnet
when the turn
on voltage of the parallel diode (6V at 1.9K) is reached. There are two
types of quench heaters used in LHC dipole viz. High Field (HF) heaters (placed in the
higher magne
tic field region) and Low Field (LF) heaters (placed in the lower magnetic
field region). As the resistivity of a metal increases with magnetic field, HF heaters are
more effective than the LF heaters. The efficiency of a heater is basically characterized
by heater delays (It’s the time between the discharge of heater and real quench occurred
in the coil) and MIIT value (which in turn gives the hot spot temperature). Lower the
value of heater delay and MIIT, the heaters are termed as more effective. Typical
ly for
1500A provoked quench (All LF and All HF), the maximum value of heater delay is
around 130 ms, whereas for higher current (>10000 A), the delays comes down to 30 ms
or less. The position and dimension of the heater strip is schematically shown in fi

Fig. 8

Position of quench heater strip and its design respectively (Courtesy of F. Sonneman)


1. Ph. Lebrun et al., “Cooling string of superconducting devices below 2 K: the helium II
bayonet heat exchanger”, Adv. Cryo. Eng. 43A (1
998) 419



2. L. Tavian, “Large Cryogenic System at 1.8 K”, Proceedings of EPAC 2000, Vienna,
Austria, 212

3. A. Seimko, “Magnet Quench Process”, Chamonix XI, 245

The Large Hadron Collider


Conceptual Design, CERN/AC/95
05(LHC) October

5. Design Study of Large Hadron Collider (LHC), CERN 91
03, 2 May 1991.

6. O.Dunkel et al., “A warm bore anticryostat for series magnetic measurements of LHC
superconducting dipole and short straight section magnets”, Paper C2
2003, Septem
ber 23
26, 2003.


I wish to extend my gratitude to Dr. Vinod Chohan for his encouragement in
preparing this document. I would also like to thank my colleagues at CERN for their
advice and criticism in preparing this document.