Transcript The current - AB-BDI-BL

Electrical principles, magnet
components and schematics, risks to
and from magnets, protection
MOPS Training Session 1
21.8.2008
KHM
The nice ideas and pictures are stolen from M. Wilson , A.
Siemko., R. Denz and P. Schmueser. The mistakes and the
rest of it are mine.
Apologies for the quality of pictures and talk. It had to be prepared in
a hurry, parallel to HC.
Electrical principles, magnet
components and schematics, risks to
and from magnets, protection
MOPS Training Session 1
21.8.2008
KHM
Outline
Components in a typical circuit
Energies
Risks
Energy Management (Protection)
Quench Detection
Reminder
The basic components:
Consider a superconductor, already immersed in LHe:
The basic components:
Consider a superconductor, already immersed in LHe:
As such pretty useless, but the picture is incomplete, anyhow:
The basic components:
Consider a superconductor, already immersed in LHe:
We need: Current leads and all the warm parts
We will have in addition: Inductance, resistance and capacitance
A single wire in details
90
80
70
60
50
40
30
C
R
C
L
R
East
West
North
C
20
10
R
0
1st Qtr
2nd Qtr
3rd Qtr
4th Qtr
A single wire in detail
Frequency dependence
Stored magnetic
energy
C
R
C
L
R
C
R
Stored electrical
energy
Stored Magnetic Energy
LHC dipole magnet (twin apertures)
E = ½ LI2
L = 0.12H
I = 11.5kA
E = 7.8 x 106 Joules
the magnet weighs 26 tonnes
so the magnetic stored energy is equivalent to the kinetic energy of:
26 tonnes travelling at 88km/hr
Stored Magnetic Energy
LHC dipole magnet (twin apertures)
E = ½ LI2
L = 0.12H
I = 11.5kA
E = 7.8 x 106 Joules
the magnet weighs 26 tonnes
so the magnetic stored energy is equivalent to the kinetic energy of:
26 tonnes travelling at 88km/hr
Stored Magnetic Energy
In a sector we have 154 magnets…in LHC we have 154*8 magnets
with a total stored energy of
E=9.6 GJ
Stored Magnetic Energy
In a sector we have 154 magnets…in LHC we have 154*8 magnets
with a total stored energy of
E=9.6 GJ
This corresponds a 100 000 to ship running at 27 knots.
Stored Magnetic Energy
In a sector we have 154 magnets…in LHC we have 154*8 magnets
with a total stored energy of
E=9.6 GJ
This corresponds a 100 000 to ship running at 27 knots.
Stored Magnetic Energy
Magnetic energy can be converted
to electrical energy by a fast change of
the current
(break of busbar, opening of a switch….).
U=L dI/dt
3.6.03
K H Mess, LHC days 2003
15
In 2003:
About 15…20% of all cold tested magnets
have isolation problems. They can (with
some exceptions) not be used in the tunnel.
Why are these faults not detected earlier in the
manufacturing?
Reason 1: The faults are produced during cool
(heater, omega)
down.
Reason 2: It is difficult, because we use
Helium or
measure lousy transmission
lines.
In 2008:
Not all were found during the tests!!!
3.6.03
K H Mess, LHC days 2003
16
Back to the basics
Consider a superconductor, already immersed in LHe:
Kamerlingh Onnes liquifies for the first
time (1908) Helium and studies the
temperature dependence of the electrical
resistance of metals. (1911)
Below a critical temperature
the resistance (voltage drop)
seems to disappear. He calls
the phenomenon
“Superconductivity”.
Nobel Price in 1913
18
Critical Temperature, Meissner Ochsenfeld
Low temperature superconductivity is due
Critical Field Bc:
to a phase transition. Phase transitions
Type 1 superconductors show the
happen to keep the relevant
Meissner effect. Field is expelled
thermodynamic energy (Gibbs energy) low.
when sample is cooled down to
Here pairs of electrons of opposite
become superconducting.
momenta and spin form a macroscopic (nm)
boson, the Cooper Pair.
The binding energy determines the critical
temperature.
Critical Temperature qc
3.5k Bq c  2(0)
where kB = 1.38
J/K is the
Boltzmann's constant
and (0) is the energy gap
(binding energy of Cooper pairs)
of at q = 0
10-23
19
The thermodynamic energy due to
superconductivity Gsup increases with
the magnetic energy, which is expelled
i.e. with B2
Gsup reaches Gnormal at the maximal field
Bc, which is small. (~0.2 T)
Type 1 superconductors are useless for magnets!
London Penetration depth, Coherence Length
•Very thin (<) slabs do not expel the field completely. Hence less
energy needed.
•Thick slabs should subdivide to lower the energy.
•But we pay in Cooper pair condensation energy to build sc boundaries of
thickness energy .
•We gain due to the not expelled magnetic energy in the penetration depth
Material
In
Pb
Nb
 Sn
.
is a net30
gain
24 nm •There
32 nm
nm if  >32.nm


20
360 nm
510 nm
170 nm
39 nm
Ginzburg Landau refine the
argument::
If the ratio between the distance the
magnetic field penetrates ( )
London penetration depth
and the characteristic distance 
Coherence length
over which the electronic state can
change from superconducting to
normal is larger than 1/2, the
magnetic field can penetrate in the
form of discrete fluxoids - Type 2
Ginzburg Landau refine the
argument::
If the ratio between the distance the
magnetic field penetrates ( )
London penetration depth
and the characteristic distance 
Coherence length
over which the electronic state can
change from
superconducting
The coherence
length  istoproportional to the mean free path
normal is larger than 1/2,
of thethe
conduction electrons.
2
is penetrate
the area ofinathe
fluxoid. The flux in a fluxoid is
magnetic fieldcan
form of discrete fluxoids - Type quantised.
2
The upper critical field is reached, when all fluxoid touch.
Bc2=0/(22).
Hence, good superconductors are always bad conductors (short
free path).
Type 2 Superconductors are mostly alloys.
Transport current creates a gradient in the fluxoid pattern.
Fluxoids must be movable to do that. However not too much,
otherwise the field decays …..
Here starts the black magic.
Current Density
The current (density) depends on
the field and on the temperature
and is a property of the sample.
(here shown for NbTi)
7
6
5
4
3
temperature K
Current density kAmm-2
2
1
10
8
23
6
4
2
2
Field T
4
6
8
10
12
14
16
Working Point and Temperature Margin
Blue plane: constant temperature, green plane: constant field
Red arrow: “load line”= constant ratio field/current
If the “working point” leaves the tent (is outside the phase
transition) => “Quench”
•Too far on the load line:
•Magnet Limit
2
•Energy deposition increases
temperature
•Temperature margin
1
2
4
6
8
2
Deposited Energy: 2 mJ ~106 p/m
~1 A4 sheet falling 4 cm
4
6
8
10
10
24
•Movement
•Eddy current warming
•Radiation (all sorts)
Material Constants, Copper
Low ρ
Copper Resistivity
High λ
Copper Thermal Conductivity
Material Constants, specific heat
0.1
Cu
10
He
4
Scales differ, Specific heat of He is by far
bigger than of Cu
Compares with Water 4.2 J/g K
Quench Development
  2T  2T 
T x, y, z 
 2T
C
 r  2  2   z 2  g ( J , T , z )  Q(T , z )
t
y 
z
 x
•Heat Capacity <= small
•Heat Conductivity, radial<= small
•Heat Conductivity, longitudinal<= good
•Cooling<= depends
•The Quench expands (if the current is above the recovery limit)
•The Temperature at the origin (Thot-spot) continues to rise
dT  (T ) 2

J (t ) 

dt C (T )
27
Thotspot

T0

C (T )
dT   J 2 dt
 (T )
0
Only material
constants, can
be calculated.
Measurement
of the max
temperature
(MIITS)
Material Constants, specific heat
Highest at the  point and
around the boiling point
Water
Magnet Quench – Quench Signal
Introduction to testing the
LHC magnets - Info
Sessions 2002, A. Siemko
Threshold
Slide 29
Introduction to testing
the LHC magnets - Info
10ms
validation
window
P
R
O
T
E
C
T
I
O
N
How to keep the temperature down?
High temperature results in:
Movement, friction
Insulation damage
Magnet destruction
•Keep the MIITS down by Heatcapacity and Resistivity (too late now)
•Keep the MIITS down by shortening the current flow
•Increase the bulk resistivity (Heating, spread the energy)
•Fast, complicated, energy into He
•Bypass the energy of the rest of the sector (if applicable)
using Diodes or Resistors
•Using Resistors <= Attention, introduces a time delay L/R
and Quench back
30
•Extract the energy (External Resistors and Switches)
•Slow, energy into air/water, needed to protect the diodes
Voltage
High resistance means
and
high I*R
high L*dI/dt
High voltage is dangerous for the insulation
Local damage => ground short or winding short
Global damage => Diodes reverse voltage
Voltage taps
Overvoltage can be/ can develop to be a global phenomenon.
Can cause considerable damage.
31
Voltage breakdown
Current I
-
3.6.03
K H Mess, LHC days 2003
U
+
32
Voltage breakdown
3.6.03
K H Mess, LHC days 2003
33
U.V. light
Electron avalanche
Ne(x)=Ne(0)* eax
Ion Bombardment
Per electron (ead-1)
ions hit the Cathode
In total
ead/(1-g(ead-1))
Breakdown for (1-g(ead-1)) = 0 ,
3.6.03
ead>> 1
K H Mess, LHC days 2003
=> g
e ad ~ 1
34
U.V. light
Electron avalanche
Ne(x)=Ne(0)* eax
Ion Bombardment
Per electron (ead-1)
ions hit the Cathode
In total
ead/(1-g(ead-1))
Breakdown for (1-g(ead-1)) = 0 ,
ais proportional to
ead>> 1 the
=> gdensity
e ad ~ n.
1 It
varies with the
a field E (geometry!)
and depends on the
gas
3.6.03
K H Mess, LHC days 2003
35
 1
g e  1  1  ad  ln 1  
 g

ad

a  n Ae

Bn
E
Combine it to obtain:
 1 
Bn
ln ln    1  ln n d A  
E
 g  
In uniform gaps
E=V/d
B nd
VBreakDown 




1
ln ndA  ln ln 
 1
g

 
3.6.03
K H Mess, LHC days 2003
Paschens
law
36
a  n Ae

Bn
E
Combine it
 1 
Bn
ln ln    1  ln n d A  
E
 g  
VBreakDown 
B nd
ln ndA  ln ln  1  1
  g 
Paschens
law
V  1.2 * (  * d )0.98Approx. in LHC-PM-ES-1, in kg/l and mm
3.6.03
K H Mess, LHC days 2003
37
In air at this
density
Vb=6.6kV !!!
3.6.03
K H Mess, LHC days 2003
38
Values differ,
because of different
Cathodes and geometries
3.6.03
K H Mess, LHC days 2003
39
A Data Compilation
Paschen Curve Helium and N2
100000
Breakdown Voltage [V]
10000
1000
100
1E+20
1E+21
1E+22
1E+23
Number Density * Distance [m^-2]
Bortnik et al
Gerhold et al
ES
N2
1E+24
Minimal detectable distance for various scenarios in He
Breakdown Distance for various conditions
10
1 bar
2
bar
1
Distance [mm]
6 bar
4.2 K gas
0.1
Liquid He
0.01
0
200
400
600
800
1000
1200
1400
1600
1800
Voltage [V]
Distance @ 1 bar
Distance @ 2 bar
Distance @ 6bar
Distance @ cold
Distance in Lhe
2000
• The break down voltage of air is 6 * bigger than that
of He.
• Tests at elevated voltages run into problems at other
spots.
• Magnets that have seen Helium, may not be tested
again at “air voltages”.
• Voltages during operation (quench) may be locally
higher than can be applied globally. Interturn shorts
are particularly difficult.
• We have observed problems with the heater strips.
3.6.03
K H Mess, LHC days 2003
42
Evidence of the insulation deficiency
3.6.03
43
K H Mess, LHC days
2003
• The break down voltage of air is 6 * bigger than that
of He.
• Tests at elevated voltages run into problems at other
spots.
• Magnets that have seen Helium, may not be tested
again at “air voltages”.
• Voltages during operation (quench) may be locally
higher than can be applied globally. Interturn shorts
are particularly difficult.
• We have observed problems with the heater strips.
3.6.03
K H Mess, LHC days 2003
44
Energy Management
•
•
•
•
Divide et impera!
Treat sectors separately!
Detect resistive the transistion asap
Divide the energy in a magnet over many windings, using
heaters (if necessary).
• Guide the energy of all other 153 (or so) magnets around
using a diode or resistor.
• Protect the diode by a fast extraction of the energy.
Voltage over one aperture
Introduction to testing the
LHC magnets - Info
Sessions 2002, A. Siemko
Spike
Slide 46
Introduction to testing
the LHC magnets - Info
Irreversibl
e quench
Example of the mechanical activity in dipoles
Circa 1 spike per 1ms
Slide 47
Introduction to testing the LHC magnets Info Sessions 2002
Quench - What Went Wrong?
•
Abnormal voltage signals recorded during the provoked quench
Courtesy: A. Siemko
How does it look at LHC?
Symbolic Circuit
Inventory
• Current Leads
–
–
–
–
–
–
13 kA
6 kA
600 A
120 A in DFB
120 A in magnet
60 A in magnet
• Busbars
– Big busbars
– Small busbars
Difficult, because CL need a
working cooling environment to run
current. To establish this the load
parameters have to varied, which
in turn requires various currents
through a working magnet circuit.
To be discussed.
Form part of the circuit, but
tested only globally.
Inventory
• Magnets
– 13 kA circuits
– 6 kA circuits
– 600 A circuits
– 120 A circuits
– 60 A circuits
Inventory
• Magnets
– 13 kA circuits
– 6 kA circuits
– 600 A circuits
– 120 A circuits
– 60 A circuits
“Easy”, Freddy takes
care.
The 60 A circuits and
most 120 A circuits (
including the current leads
and bus bars) are
protected by the
overvoltage detection of
the powerconverter.
Its AB-PO.
Inventory
• Magnets
– 13 kA circuits
– 6 kA circuits
– 600 A circuits
– 120 A circuits
– 60 A circuits
The 120 A MO and the
600 A circuits have a
“global quench
protection”, that means
the current is
measured and the first
and second derivative
are calculated to
predict the inductive
voltage. Note that the
inductance depends on
the current.
Difficult
Global Quench Protection
ΔV
ΔV
L dI/dt
DSP
Interlock
24 bit ADC
Fieldbus
Inventory
• Magnets
– 13 kA circuits
– 6 kA circuits
– 600 A circuits
– 120 A circuits
– 60 A circuits
6 kA quadrupoles
ΔU
ΔU
Long voltage tap,
Problems to be expected
Inventory
• Magnets
– 13 kA circuits
– 6 kA circuits
– 600 A circuits
– 120 A circuits
– 60 A circuits
13 kA busbar protection
Note that the reference magnets
Courtesy R. Denz
have to represent an average magnet!
Problem after a quench!
Local quench detector for main magnets
Note that only one of the two channels is
Visible in the CCC.
Courtesy R. Denz
The “hidden” card may have “seen” things, invisible for you
Summary
What is special with superconducting circuits?
Large inductance, large stored energy, low resistance,
long time constants, extremely high current density
What are the specifically dangerous issues?
Shorts, opening connections, high voltage, high energy
density, hydraulic problems
Keep on telling the operation crew:
We are pulling a tigers tail!.
62
References
H. Brechna, Superconducting Magnet Systems, Springer, Berlin 1973
P. Schmueser, Superconducting magnets for particle accelerators, Rep. Prog.
Phys. 54 (191) 683
M. N. Wilson, Superconducting Magnets, Clarendon Press, Oxford, 1983
See also his lectures here and at CAS
A. Siemko, Introduction to testing the LHC magnets - Info Sessions 2002
http://nobelprize.org/nobel_prizes/physics/laureates/1913/onnes-lecture.pdf
http://www.bnl.gov/magnets/Staff/Gupta/cryogenic-data-handbook
KHM et al, Superconducting Accelerator Magnets, World Scientific, Singapore,
1996