Title: Accidental beam losses and protection in the LHC
1Accidental beam losses and protection in the LHC
- R.Schmidt and J.Wenninger
- for the Working Group on Machine Protection
- HB 2004 Bensheim
LHC parameters and associated risks Overview
accidental beam losses Aperture and accidental
beam losses Protection and redundancy Conclusions
2Beam dump tunnel
LHC tunnel
3Some numbers for 7 TeV
- Momentum at collision 7 TeV/c
- Beam intensity 2808 ? 1.1 ? 1011 protons per
beam - Luminosity 1034 cm-2s-1
- Dipole field at 7 TeV 8.33 Tesla
- Typical beam size 200-300 µm
- Energy stored in the magnet system 10 GJoule
- Energy stored in one (of 8) dipole circuit 1.1
GJoule - Energy stored in one beam 350 MJoule
- Average beam power to compare with
- high power accelerators, both beams
some 10 kWatt - Instantaneous beam power for one beam 3.9
TWatt - .during 89 µs
- .corresponds to the power of 1700 nuclear power
plants - Energy to heat and melt one kg of copper 700 kJ
4Bunch intensities, quench and damage level
- Intensity one pilot bunch 5?109
- Nominal bunch intensity 1.1?1011
- Batch from SPS (216/288 bunches at 450
GeV) 3?1013 - Nominal beam intensity with 2808 bunches 3?1014
- Damage level for fast losses at 450 GeV 1-2?1012
- Damage level for fast losses at 7 TeV 1-2?1010
- Quench level for fast losses at 450 GeV 2-3?109
- Quench level for fast losses at 7 TeV 1-2?106
- Damage and quench assessment approximative,
supported by experience in SPS and calculations - Further calculations and material tests at SPS in
two weeks planned
5Livingston type plot Energy stored in the beam
6Failure scenarios and accidental beam losses
- A large number of different mechanisms can cause
accidental particle losses Classification of
accidental beam losses according to time constant
for the loss
- Ultra fast beam losses (single turn or less)
- to be avoided, beam dump block is the only
element that can safely absorb the 7 TeV LHC beam - passive protection with collimators and beam
absorber
- Very fast beam losses (some turns to some
milliseconds) - Fast beam losses (5 ms several seconds)
- Slow beam losses (several seconds 0.2 hours)
- active protection, by detecting failure and
extracting the beams into beam dump block
7Single turn accidental beam losses
- Failure mechanisms
- Failure of beam dump kicker (prefiring,
asynchronous beam dump) - Failure of kickers for tune measurements and
aperture exploration - During transfer and injection
- failure of injection kicker
- wrong trajectory or mismatch of beam energy
- obstruction of beam passage
- Recent studies on protection during transfer and
injection of the beams from SPS at 450 GeV to the
LHC (see H.Burkhardt) - Strategy for protection
- Avoid such failures (systems with high
reliability) - Block beam transfer from SPS to LHC if parameters
are not correct (i.e. magnet current) - Beam trajectory after such failure is reasonably
well defined - Passive protection rely on collimators and beam
absorbers
8Consequence of a failure scenario Full 7 TeV LHC
beam deflected into copper target
2808 bunches 7 TeV 350 MJoule
Copper target
2 m
Energy density GeV/cm3 on target axis
vaporisation
melting
Target length cm
collaboration with N.Tahir (GSI) et al.
9Density change in target after impact of 100
bunches
Copper target
copper solid state
radial
100 bunches target density reduced to 10
collaboration with N.Tahir (GSI) et al.
Target radial coordinate cm
- Energy deposition calculations using FLUKA
- Numerical simulations of the hydrodynamic and
thermodynamic response of the - target with two-dimensional hydrodynamic
computer code - From this calculations one can estimate the
longitudinal range of full beam in copper between
10m and 40m
10LHC Layout eight arcs (sectors with a length of
about 2300 m) eight long straight sections (about
700 m long)
IR5CMS
IR6 Beam dumping system
IR4 RF Beam instrumentation
IR3 Momentum Cleaning (normal conducting magnets)
IR7 Betatron Cleaning (normal conducting magnets)
IR8 LHC-B
IR2ALICE
IR1 ATLAS
Injection
Injection
Transfer Line
Transfer Line
11- Collimators for cleaning the beam halo
- close to the beam between 5-10 ?
- must be accurately adjusted (within a fraction
of one ?) - position depends on optics and possibly on energy
IR5CMS
IR6 Beam dumping system
IR4 RF Beam instrumentation
IR3 Momentum Cleaning (normal conducting magnets)
IR7 Betatron Cleaning (normal conducting magnets)
IR8 LHC-B
IR2ALICE
IR1 ATLAS
Injection
Injection
Transfer Line
Transfer Line
12- Collimators for protection of equipment against
single turn beam losses - shadow equipment downstream
- must be adjusted (better than one s)
- position depends on LHC operational mode
(injection, energy ramp, ) and on optics
IR5CMS
IR6 Beam dumping system
IR4 RF Beam instrumentation
IR3 Momentum Cleaning (normal conducting magnets)
IR7 Betatron Cleaning (normal conducting magnets)
IR8 LHC-B
IR2ALICE
IR1 ATLAS
Injection
Injection
Transfer Line
Transfer Line
13- For protection of equipment against multiturn
beam losses - all collimators limiting the aperture contribute
to this function
IR5CMS
IR6 Beam dumping system
IR4 RF Beam instrumentation
IR3 Momentum Cleaning (normal conducting magnets)
IR7 Betatron Cleaning (normal conducting magnets)
IR8 LHC-B
IR2ALICE
IR1 ATLAS
Injection
Injection
Transfer Line
Transfer Line
14- Collimators for protection and cleaning of the
low-beta insertions, mainly in IR1 and IR5 - close to the beam about 10 ?
- must be accurately adjusted (within about one ?)
- mainly required during squeeze and for squeezed
beams
IR5CMS
IR6 Beam dumping system
IR4 RF Beam instrumentation
IR3 Momentum Cleaning (normal conducting magnets)
IR7 Betatron Cleaning (normal conducting magnets)
IR8 LHC-B
IR2ALICE
IR1 ATLAS
Injection
Injection
Transfer Line
Transfer Line
15Lifetime of the beam - for nominal intensity at
7 TeV
Accidental beam losses
16Accidental multiturn beam losses
- Closed orbit grows around the ring
- Fast emittance growth beam size explodes
- Both
- Can happen very fast
- Can be detected around the entire accelerator
- Local orbit bump
- cannot happen very fast
- might be detected only locally
- Protection Detect failure and dump beam
- Detection, transmission to beam dump, and beam
dump takes at least 3 turns 270 ?s
17Multiple turn failures Magnet powering failures
- Quench of superconducting magnets
- Discharge of superconducting magnets switching a
resistance into the circuit (after quench, or by
accident) - Failure of magnet powering
- For some magnets very fast beam loss (several
turns) D1 normal conducting magnet - Electric short in the coil of a normal conducting
magnet
18Multiple turn failures Other failures
- Aperture limitation in beam pipe (circulating
beam) - Vacuum valve moves into beam
- Collimator moves into beam
- Other element moves into beam
- Loss of beam vacuum
- Failure in the RF system
- Debunching of beam and number of protons in the
abort gap could lead to single turn failure when
beam is dumped - Operational failures
- Combined failures, for example after Mains
Disturbances (thunderstorm, )
19Beam losses and aperture
- The aperture of the LHC at 450 GeV is limited
(about 7.5s, assuming closed orbit excursions4
mm, beta-beating, .) - Critical operation at 7 TeV with squeezed optics
- ?-function up to 4850 m in insertions IR1 and IR5
- very strong low-? quadrupole magnets with orbit
offset - normal conducting dipole magnets
- superconducting dipole magnets
- In general, particle losses first at collimators
- Fast orbit changes are the most critical failures
- collimators at about 6-9 ? from the beam
- 1 of the beam would damage the collimators for
fast beam loss
20Critical apertures around the LHC (illustration
drawing)in units of beam size ? at 450 TeV
collimators (betatron cleaning)
collimators (momentum cleaning)
arc aperture down to about ? 7.5 ?
? 6-9 ?
aperture in cleaning insertions about ? 6-9 ?
aperture in cleaning insertions about ? 6-9 ?
IR1
IR2
IR3
IR4
IR5
IR6
IR7
IR8
21Critical apertures around the LHC (illustration
drawing)in units of beam size ?
7 TeV and ? 0.55 m in IR1 and IR5
Triplet
Triplet
arc aperture about ? 50 ?
triplet aperture about ?14 ?
IR1
IR2
IR3
IR4
IR5
IR6
IR7
IR8
22Most likely failures for fast losses
quenchesFailures leading to the fastest
multiturn losses D1 magnet
orbit mm
MB quench fast loss
Quench of - MQX - D2 - MB Powering Failure of
D1 normal conducting
D1 normalconducting very fast loss
D2 quench fast loss
MQX 2 quads quench fast loss
time seconds
V.Kain Diploma thesis 2001 / O.Brüning
Squeezed optics with max beta of 4.8 km All 4
quadrupole magnets (inner triplet MQX) quench ,
approximately Gaussian current decay with time
constant 0.2 s Powering failure for D1,
exponential current decay, time constant 2.5
s Quench of one MB, approximately Gaussian
current decay with time constant 0.2 s
23Particles that touch collimator after failure of
normal conducting D1 magnets After about 13
turns 3109 protons touch collimator, about 6
turns later 1011 protons touch collimator
1011 protons at collimator
Dump beam level
V.Kain
HERA experience confirmes worries very fast beam
losses
24Protection and redundancy what triggers a beam
dump?
- Hardware diagnostics
- Quench signal from Quench Protection System
- Beam loss monitors at the collimators and other
aperture limitations - Beam loss monitors in the arcs
- Magnet current change monitors
- Beam position change monitors
- Fast beam current decay (lifetime) monitors
25Hardware failure diagnostics
- Vacuum valve leaving the OUT position (away
from end switch) - Other movable devices leaving the OUT position
- Powering failures detected by the power
converter, requesting a beam dump (typical times
in the order of 10 ms) - Failures of cooling for normal conducting magnets
- Failure in the RF system
- Anticipated failure in the beam dumping system
(before it is too late), e.g. when 1 out of 15
kicker is lost - Failure in critical beam absorbers and
collimators
26Hardware failure diagnostics
- PLUS
- Does not require collimators to have correct
settings - If early enough, can dump the beam before
particle losses - MINUS
- For many type of failures the beam dump comes too
late - Complexity of hardware not all failures are
detected - Too many channels too many False Beam Dumps
- Risk of including failures that would not lead to
particle losses
27Quench detection
- Magnet starts to quench
- Resistive Voltage across magnet gt 0.1 V
- 10 ms quench detection
- fire quench heater
- requests energy extraction
- requests a beam dump
The quench heaters become effective
3 ms the interlock system transmits the
request to the beam dump
300 ?s the beam dump kicker extracts beam
Magnet current starts to debypass magnet by diode
5 ms current starts to decay exponentially
28Quench detection
- PLUS
- Does not require collimators to have correct
settings - If early enough, beam gone before losses
- Dumps beam for failures of the quench protection
system - Does not reduce the availability of LHC Quench
protection is always required. After a quench the
beam must be dumped - MINUS
- Only covers beam losses due to magnet quenches
- Might be too late (being further analysed,
efficiency depends on quench process, magnet
field, beam loss pattern, etc) - Large complexity (several 1000 channels) good
post mortem analysis required
29Beam loss monitoring at aperture limitations
- In general, collimators are limiting the aperture
- Always true for beam blow up
- Mostly true for orbit changes
- Beam loss monitors at aperture restrictions
continuously measuring beam losses - Losses are detected within less than a turn
- After detection it takes 2-3 turns to extract all
particles into beam dump block
30Beam loss monitoring at aperture limitations
- PLUS
- Should capture (nearly) all types of accidental
beam losses - Dumps the beam if there are really particle
losses - Very fast (lt 100 ?s)
- Limited complexity (some 100 channels)
- Expected to be very reliable
- MINUS
- Does require collimators to have correct settings
and defining the aperture - Does not catch beam losses in the arcs (for
example, closed orbit bumps) - Random spikes might trigger beam dump
- Setting of thresholds not obvious - if too low,
False Beam Dumps if too high - risk of damage
31Beam loss monitoring around the accelerator
- Beam loss monitors continuously measuring beam
losses - Together with the BLMs at aperture limitations,
covers most of the LHC (all arcs) - Losses can be detected within less than a turn
- After detection it takes 2-3 turns to extract all
particles into beam dump block
32Beam loss monitoring around the accelerator
- PLUS
- Dumps the beam if there are really particle
losses - For failures leading to orbit changes and
emittance growth - Detection can be made very fast (lt 100 ?s)
- Does not require collimators to have correct
settings - Catches failures that appear only in the arcs
(for example, closed bump) - MINUS
- Large complexity (some 1000 channels)
- Could increase number of False Beam Dumps
- Setting thresholds delicate balance between
avoiding magnet quenches and avoiding False Beam
Dumps
33Magnet current decay monitoring for critical
magnets
- Very fast detection of power converter / magnet
failures - Monitors change of magnet current (Hall probes,
voltage, ) - Prototype quick and dirty gave promising
results (M.Zerlauth) - Similar technique recently successfully
implemented at HERA (M.Werner) - Should be possible to detect powering failures in
less than one millisecond
34Magnet current decay monitoring for critical
magnets
- PLUS
- Independent method to monitor failures in the
powering system power converter fault /
thunderstorm / short circuit in magnet / other
problems - Does not require collimators to have correct
settings - Can be made fast (lt 1ms)
- Mainly for normal conducting magnets
- MINUS
- Needs to be demonstrated if practical (EMC, )
wait for HERA experience - Setting of thresholds required could be
delicate - Should be limited to a few electrical circuits
with normal conducting magnets otherwise too
complex
35Beam position change monitors
- If the orbit start to moves very fast, dump the
beam - Fast orbit changes can be observed anywhere
around the LHC - Observation for each beam, each plane, two
monitors with 90 degrees phase advance in total
8 BPMs - system with limited complexity
- BPMs at location of high beta function, using the
same monitors that are already required for
machine protection (to ensure x lt 4 mm in
the insertion IR6 for the beam dumping system)
36Beam position change monitors thresholds
450 GeV fastest orbit movement during normal
operation by an orbit corrector
magnet Superconducting orbit correctors 2
mm/s 15 mm/s Normal conducting orbit
correctors 0.6 mm/s 1.7 mm/s 450 GeV if
the change of the orbit exceeds, say, some 10
mm/s corresponding to 0.01 mm/msec, there is
something wrong Detection of very fast orbit
drifts (IF dx/dt gt 0.1 mm/ms) OR (IF dx/dt
gt 1 mm/100ms) THEN beam dump 7 TeV if the
change of the orbit exceeds, say, some 1 mm/s
corresponding to 0.001 mm/msec, there is
something wrong (IF dx/dt gt 0.05 mm/ms) OR
(IF dx/dt gt 0.3 mm/100ms) THEN beam dump
numbers preliminary
37Beam position change monitors
- PLUS
- Independent method to measure fast orbit drifts
due to failures - Does not require collimators to have correct
settings - Can be made fast (lt 1ms)
- Beam dump before particles are lost
- Limited complexity
- MINUS
- Is it practical ? False Beam Dumps ?
- Setting of thresholds delicate
- What to do during injection, during kicks for
Q-measurements, to be studied - Only for beam orbit changes, not for emittance
growth - Studies needed
38Fast beam current decay monitoring
- Very fast beam current monitor, could detect
losses within short time - Measuring proton losses with, say, ?N / ?t
1010 protons - Interlock condition
- if (?N / ?t gt Nthreshold(E) 1010) THEN BEAM
DUMP - ?t could be as short as one turn
- Nthreshold decreases with energy to be always
efficient - For start of LHC operation, when intensity is
limited, resolution should be no problem - Challenge must be fast and accurate
- First discussions with experts - looks promising
39Fast beam current decay monitoring
- PLUS
- Independent method to measure beam losses
- Does not require collimators to have correct
settings - Fast for reduced accuracy (lt 1ms)
- Slow for high accuracy (gt 10ms)
- Limited complexity one instrument
- MINUS
- Needs to be demonstrated if practical
- Setting of thresholds required
- Not sufficient for all LHC operation modes and
for shortest accidental beam losses time
constants - Could be ok for 450 GeV, not for 7 TeV
- Studies needed
40Protection and redundancy at 450 GeV
41Protection and redundancy at 7 TeV
42Conclusions
- Protection for LHC starts before extraction from
SPS - Protection is required during the entire cycle
- Large redundancy in protection
- Availability of the machine due to the complex
protection is an important issue - Large energy stringent protection required - too
few interlocks could lead to severe damage of the
LHC - Unprecedented complexity too conservative
interlocking of the machine protection systems
could prevent efficient LHC exploitation - Initial operation with part of the protection
systems - Commissioning of other protection systems during
initial operation
43Some questions to the workshop..
- Fast beam current decay monitoring
- Fast beam position change monitoring
- What can be achieved?
- Who has experience?
- Where else might such systems be required?
44Acknowledgements
- The presentation is based on the work that was
performed in many groups in several CERN
Departments, as well as collaborators from other
labs (Fermilab, GSI, Protvino, Triumf, .) - Contributions of many colleagues are
acknowledged, in particular for the discussions
in the Machine Protection WG, Collimation WG and
Injection WG - particular thanks for R.Assmann, H.Burkhardt,
E.Carlier, B.Dehning, B.Goddard, E.B.Holzer,
J.B.Jeanneret, V.Kain, B.Puccio, J.Uythoven.
M.Zerlauth
45- Recent question in MPWG Can we dump the beam in
time after a quench of a dipole magnet? - Beam is to be dumped before the current in the
dipoles starts to decay - The sequence of following actions has to be
determined - beam loss causes the magnet to quench
- the voltage builds up and exceeds the threshold
of the quench detector - the quench detector detects the voltage after
some time - the quench detector fires quench heaters or
triggers the energy extraction - at the same time, the PIC is informed
- PIC sends a dump request to the BIC
- the heaters become efficient
- the BIC sends a dump request to the beam dumping
system - the voltage exceeds the diode voltage, and the
current starts to bypass the magnet - the switch opens, and the current in the string
of magnets decays - The Powering Interlock Controller is the only
system sending (direct) beam dump request after
powering failures - Secondary protection with collimators, BLM
(possibly BPM / beam lifetime)
Defined sequence
Whos quickest?
46Hardware configuration (main dipole circuit) and
signal transmission
Quench Loop stretching over the arcs
Beam Dump Request
Beam Interlock Controller
Beam Dump Request
Beam Dumping System
Quenching (dipole) magnet in the arc
47The PIC process times
- Interlock Controller is based on a PLC controlled
process, monitoring and controlling up to 45
electrical circuits (gt200 signals) - For time (beam) critical circuits -gt configurable
hardwired matrix in parallel
PIC Controller (PLC) Process time lt 5ms
Power Converters
Quench Protection System
Power Permit, Powering Failure, Discharge
Requests
Quench Signals, Discharge Requests
For main circuits Hardware Matrix (CPLD) Process
time lt 0.2ms
Beam Dump Requests
48The Timescales
Decay starts somewhat before the complete arc
extinction due to resistive arc // resistor
Diode becomes conductive gt 80ms
Current in dipole circuit
lt 15 µs
lt 100 µs
5..7 ms
T1 EE system reads Quench signal
Mechanic opening arc extinction
T3 Last branch open, arc is extinct Current in
dipole magnets decays
T2 Switch Opening
lt 100µs
0.2 5ms
lt 270µs
lt 15 µs
lt 100 µs
PIC process
T2 PIC reads signal
T3 PIC issues beam dump
T5 Completion of beam dump
T0 Quench signal
T4 Beam dump is received by the BIC
T1 Quench Loop Controller Receives signal
49Conclusions
- Due to the delay in the switch opening of the
13kA energy extraction system, the beam is dumped
before the current in the magnet starts decaying - also true in case of self triggering of EE
- For all other sc magnets the time constraints are
less critical - For time critical signals (mainly main dipole and
quadrupole circuits due to large effects on the
circulating beams), a hardwired matrix within the
PIC can be configured in parallel to the PLC
process - What to do for nc magnets?!