accidental and continuous beam losses protection of the accelerator from beam losses

CAS October 2013 R.Schmidt

● Accidental and continuous beam losses ● Protection of the accelerator from beam losses

Machine Protection

Rüdiger Schmidt, CERN and ESS CAS Accelerator School

October 2013 - Trondheim

CERN

Rüdiger Schmidt CAS Trondheim 2013 page 3

Protection from Energy and Power

● Risks come from Energy stored in a system (Joule), and Power when operating a system (Watt)

• “Very powerful accelerator” … the power flow needs to be controlled● An uncontrolled release of the energy, or an uncontrolled power

flow can lead to unwanted consequences• Damage of equipment and loss of time for operation• For particle beams, activation of equipment

● This is true for all systems, in particular for complex systems such as accelerators

• For the RF system, power converters, magnet system …• For particle beams

This lecture on Machine Protection is focused on preventing damage caused by particle beams

CERN


Content

● Different accelerator concepts: Examples for ESS and LHC ● Hazards and Risks

● Accidental (uncontrolled) beam losses and consequences● Accidental (uncontrolled) beam and probability

● Machine Protection Systems• For high energy proton synchrotrons (LHC)• For high power accelerators (ESS)

● Some principles for protection systems

CERN


LHC pp and ions7 TeV/c – up to now 4 TeV/c 26.8 km circumferenceEnergy stored in one beam 362 MJ

SwitzerlandLake Geneva LHC Accelerator

(100 m down)

SPS Accelerator

CMS, TOTEM

ALICE

LHCb

ATLAS

CERN Proton collider LHC – 362 MJ stored in one beam

CERN


LHC pp and ions7 TeV/c – up to now 4 TeV/c 26.8 km CircumferenceEnergy stored in one beam 362 MJ

SwitzerlandLake Geneva LHC Accelerator

(100 m down)

SPS Accelerator

CMS, TOTEM

ALICE

LHCb

ATLAS

CERN Proton collider LHC – 362 MJ stored in one beam

If something goes wrong, the beam energy has to be safely deposited

CERN


ESS Lund / Sweden – 5 MW beam power

Power of 5000 kW

Drift tube linac with

4 tanks

Low energy beam

transport

Medium energy beam

transport

Super-conducting cavities High energy beam transport

• Operating with protons

• Operation with beam pulses at a frequency of 14 Hz

• Pulse length of 2.86 ms

• Average power of 5 MW

• Peak power of 125 MW

RFQ352.2 MHz

75 keV 3 MeV 78 MeV 200 MeV 628 MeV 2500 MeV

Source LEBT RFQ MEBT DTL Spokes High βMedium β HEBT & Upgrade Target

2.4 m 4.0 m 3.6 m 32.4 m 58.5 m 113.9 m 227.9 m

352.21 MHz 704.42 MHz

As an example for a high

intensity linear accelerator

(similar to SNS and J-PARC)

~ 500 m

CERN


Power of 5000 kW

Drift tube linac with

4 tanks

Low energy beam

transport

Medium energy beam

transport

Super-conducting cavities High energy beam transport

• Operating with protons

• Operation with beam pulses at a frequency of 14 Hz

• Pulse length of 2.86 ms

• Average power of 5 MW

• Peak power of 125 MW

RFQ352.2 MHz

75 keV 3 MeV 78 MeV 200 MeV 628 MeV 2500 MeV

Source LEBT RFQ MEBT DTL Spokes High βMedium β HEBT & Upgrade Target

2.4 m 4.0 m 3.6 m 32.4 m 58.5 m 113.9 m 227.9 m

352.21 MHz 704.42 MHz

As an example for a high

intensity linear accelerator

(similar to SNS and J-PARC)

~ 500 m

ESS Lund / Sweden – 5 MW beam power

If something goes wrong, injection has to be stopped

CERN


Energy stored in beam and magnet system

10 100 1000 100000.01

0.10

1.00

10.00

100.00

1000.00

10000.00

Momentum [GeV/c]

Ener

gy s

tore

d in

the

beam

[MJ] LHC 7.0 TeV

LHC at injection

ISR

SNSLEP2

LHC ions

TEVATRON

SPSppbar

SPS transfer to LHC

RHIC proton

LHC energy in magnets

LHC 4.0 TeV

SPS material test

Factor~200

CERN


What does it mean ……… MJoule ?

360 MJ: the energy stored in one LHC beam corresponds approximately to…

• 90 kg of TNT

• 8 litres of gasoline

• 15 kg of chocolateIt matters most how easy

and fast the energy is released !!

The energy of an 200 m long fast train at 155 km/hour corresponds to the energy of 360 MJ stored in one LHC beam.

• 860 litres H2O from 0 0C to 100 0C

CERN


Consequences of a release of 600 MJ at LHC

Arcing in the interconnection

53 magnets had to be repaired

The 2008 LHC accident happened during test runs without beam.A magnet interconnect was defect and the circuit opened. An electrical arc provoked a He pressure wave damaging ~600 m of LHC, polluting the beam vacuum over more than 2 km.

Over-pressure

Magnet displacement

CERN


Energy and Power density

Many accelerators operate with high beam intensity and/or energy ● For synchrotrons and storage rings, the energy stored in the beam

increased with time (from ISR to LHC)● For linear accelerators and fast cycling machines, the beam power

increases

The emittance becomes smaller (down to a beam size of nanometer)● This is important today, and even more relevant for future

projects, with increased beam power / energy density (W/mm2 or J/mm2 ) and increasingly complex machines

Even a small amount of energy can lead to some (limited) damage● Can be an issue for sensitive equipment

CERN


Hazards and Risks

CERN


Hazard and Risk for accelerators

● Hazard: a situation that poses a level of threat to the accelerator. Hazards are dormant or potential, with only a theoretical risk of damage. Once a hazard becomes "active“: incident / accident. Consequences and possibility of an incident interact together to create RISK, can be quantified:

RISK = Consequences Probability ∙

Related to accelerators● Consequences of an uncontrolled beam loss● Probability of an uncontrolled beam loss● The higher the RISK, the more Protection is required

CERN


Example for ESS● Bending magnet in an accelerator deflecting the beam● Assume that the power supply for the bend in HEBT-S2 fails and

the magnets stops deflecting the beam• Probability: MTBF for power supply is 100000 hours = 15 years

● The beam is not deflected and hits the vacuum chamber• Consequences: what is expected to happen? Damage of magnet, vacuum

pipe, possibly pollution of superconducting cavities● Detect failure and stop beam

5 MW Beam

~ 160 m following the sc cavities

HEBT-S2

CERN


Example for LHC: SPS, transfer line and LHC

1 km

Beam is accelerated in SPS to 450 GeV (288 bunches, stored energy of 3 MJ)Beam is transferred from SPS to LHCBeam is accelerated in LHC to high energy (stored energy of 362 MJ)

Transfer line 3 km

LHC

SPS6911 m

450 GeV 3 MJ transfer to LHC

IR8

Fast extraction kicker

Injection kicker

Transfer line

Injection kicker

IR2 Fast extraction kicker

CERN


Protection at injection

LHC circulating beam

Circulating beam in LHC

LHC vacuum chamber

Transfer line vacuum chamber

CERN



Beam injected from SPS and transfer line


Beam from SPS

Injection Kicker

LHC injected beam

CERN



Kicker failure (no kick)


Beam from SPS

Injection Kicker

Major damage to sc magnets, vacuum pipes,

possibly LHCb / Alice experiments

CERN



Beam absorbers take beam in case of kicker misfiringTransfer line collimators ensure that incoming beam trajectory is ok


Beam from SPS

Injection Kicker Set of transfer line

collimators (TCDI) ~5σ

Injection absorber(TDI) ~7σ

phase advance 900

CERN



Beam absorbers take beam in case of kicker misfiring on circulating beam


Injection Kicker


Circulating beam – kicked out

phase advance 900


Set of transfer line collimators (TCDI)

~5σ

This type of kicker failure happened several times:

protection worked

CERN


(Accidental) beam lossand consequences

CERN


Beam losses and consequences

● Charged particles moving through matter interact with the electrons of atoms in the material, exciting or ionizing the atoms => energy loss of traveling particle described by Bethe-Bloch formula.

● If the particle energy is high enough, particle losses lead to particle cascades in materials, increasing the deposited energy

• the maximum energy deposition can be deep in the material at the maximum of the hadron / electromagnetic shower

● The energy deposition leads to a temperature increase• material can vaporise, melt, deform or lose its mechanical properties• risk to damage sensitive equipment for less than one kJ, risk for damage

of any structure for some MJ (depends on beam size)• superconducting magnets could quench (beam loss of ~mJ to J)• superconducting cavities performance degradation by some 10 J• activation of material, risk for hand-on-maintenance

CERN


Energy loss: example for one proton in iron (stainless steel, copper very similar)

Low energy few MeV, beam transport, RFQ for many machines

SNS - ESS 1 – 3 GeV

LHC 7 TeV

From Bethe-Bloch formula.

CERN



● Proton beam travels through a thin window of thickness ● Assume a beam area of 4 , with rms beam sizes (Gaussian beams)● Assume a homogenous beam distribution● The energy deposition can be calculated, mass and specific heat are known● The temperature can be calculated (rather good approximation), assuming a fast loss

and no cooling

CERN


Heating of material with low energy protons (3 MeV)

Temperature increase in the material: dTFeNp dEdxFe

cFe_spec Fbeam Fe

Temperature increase for a proton beam impacting on a Fe target:

Beam size: h 1.00 mm and v 1.00 mm

Iron specific heat: cFe_spec 440J

kg K

Iron specific weight: Fe 7860kg

m3

Energy loss per proton/mm: dEdxFe 56.696MeVmm

Number of protons: Np 1.16 1012

Energy of the proton: Ep 0.003 GeV

Temperature increase: dTFe 763 K

CERN


Heating of material with high energy protons(> GeV)

Nuclear inelastic interactions (hadronic shower)• Creation of pions when going through matter• Causes electromagnetic shower through decays of

pions• Exponential increase in number of created particles• Final energy deposition to large fraction done by

large number of electromagnetic particles• Scales roughly with total energy of incident particle• Energy deposition maximum deep in the material• Energy deposition is a function of the particle type,

its momentum and parameters of the material (atomic number, density, specific heat)

• No straightforward expression to calculate energy deposition

• Calculation by codes, such as FLUKA, GEANT or MARS

http://williamson-labs.com/ltoc/cbr-tech.htm

CERN


Maximum energy deposition in the proton cascade (one proton): Emax_C 2.0 10 6Jkg

Specific heat of graphite is cC_spec 710.6000 1kg

JK

To heat 1 kg graphite by, say, by T 1500K , one needs: cC_spec T 1 kg 1.07 106 J

Number of protons to deposit this energy is: cC_spec T

Emax_C5.33 1011

Maximum energy deposition in the proton cascade (one proton): Emax_Cu 1.5 10 5Jkg

Specific heat of copper is cCu_spec 384.5600 1kg

JK

To heat 1 kg copper by, say, by T 500K , one needs: cCu_spec T 1 kg 1.92 105 J

Number of protons to deposit this energy is: cCu_spec T

Emax_Cu1.28 1010 Copper

graphite

Damage of a pencil 7 TeV proton beam (LHC)co

pper

grap

hite

CERN



● Calculate the response of the material (deformation, melting, …) to beam impact (mechanical codes such as ANSYS, hydrodynamic codes such as BIG2 and others)

● Beams at very low energy have limited power…. however, the energy deposition is very high, and can lead to (limited) damage in case of beam impact

• issue at the initial stage of an accelerator, after the source, low energy beam transport and RFQ

• limited impact (e.g. damaging the RFQ) might lead to long downtime, depending on spare situation

● Beams at very high energy can have a tremendous damage potential

• for LHC, damage of metals for ~1010 protons• one LHC bunch has about 1.5 10∙ 11 protons, in total up to 2808 bunches• in case of catastrophic beam loss, possibly damage beyond repair

CERN


Controlled SPS experiment● 81012 protons clear damage● beam size σx/y = 1.1mm/0.6mmabove damage limit for copper stainless steel no damage● 21012 protons below damage limit for copper

6 cm25 cm

• 0.1 % of the full LHC 7 TeV beams• factor of three below the energy in a

bunch train injected into LHC• damage limit ~200 kJoule

V.Kain et al

A B D C

SPS experiment: Damage with 450 GeV protons

CERN


Vacuum chamber in SPS extraction line, 2004

● 450 GeV protons, 2 MJ beam in 2004 ● Failure of a septum magnet● Cut of 25 cm length, groove of 70 cm● Condensed drops of steel on other side of

the vacuum chamber● Vacuum chamber and magnet needed to

be replaced

CERN


Collimator in Tevatron after, 2003

● A Roman pot (movable device) moved into the beam

● Particle showers from the Roman pot quenched superconducting magnets

● The beam moved by 0.005 mm/turn, and touched a collimator jaw surface after about 300 turns

● The entire beam was lost, mostly on the collimator

Observation of HERA tungsten collimators: grooves on the surface when opening the vacuum chamber were observed. No impact on operation.

CERN


Beam Current Monitors (BCM) measure current pulse at different locations along the linac.

About 16 µsec of beam lost in the superconducting part of linac

680 µs of beam before sc linac

664 µs of beam after sc linac

16 µs of beam lost in the sc linac

Beam energy in 16 µsEnd of DTL = 30 J End of CCL = 66 JEnd of SCL = 350 J

Beam losses in SNS linac

M.Plum / C.Peters

CERN


Beam loss with low energy deposition

● Beam might hit surface of HV system (RFQ, kicker magnets, cavities)

● Surfaces with HV, after beam loss performance degradation might appear (not possible to operate at the same voltage, increased probability of arcing, …)

● SNS: errant beam losses led to a degradation of the performance of superconducting cavity

• Bam losses likely to be caused by problems in ion source, low energy beam transfer and normal conducting linac

• Cavity gradient needs to be lowered, conditioning after warm-up helps in most cases

• Energy of beam losses is about 100 J • Damage mechanisms not fully understood, it is assumed that some beam

hitting the cavity desorbs gas or particulates (=small particles) creating an environment for arcing M.Plum / C.Peters

CERN


Accidental beam loss and probability

CERN


Beam losses mechanisms

In accelerators, particles are lost due to a variety of reasons: beam gas interaction, losses from collisions, losses of the beam halo, …

● Continuous beam losses are inherent during the operation of accelerators

• Taken into account during the design of the accelerator● Accidental beam losses are due to a multitude of failures

mechanisms● The number of possible failures leading to accidental beam losses

is (nearly) infinite

CERN


Continuous beam losses: Collimation prevents too high beam losses around the accelerator (beam cleaning)

A collimation system is a (very complex) system with (massive) material blocks close to the beam installed in an accelerator to capture halo particles

Such system is also called (beam) Cleaning System

Accidental beam losses: “Machine Protection” protects equipment from damage, activation and downtime Machine protection includes a large variety of systems, including collimators (or beam absorbers) to capture mis-steered beam

Machine Protection

Beam Cleaning

Beam losses, machine protection and collimation

CERN


Regular and irregular operation

Failures during operation

Beam losses due to failures, timescale from nanoseconds to secondsMachine protection systems

CollimatorsBeam absorbers

Regular operation

Many accelerator systemsContinuous beam losses

Collimators for beam cleaning Collimators for halo scraping

Collimators to prevent ion-induced desorption

CERN


Continuous beam losses: Collimation

Continuous beam with a power of 1 MW and more (SNS, JPARC, PSI)• A loss of 1% corresponds to 10 kW – not to be lost along the beam line to

avoid activation of material, heating, quenching, …• Assume a length of 200 m: 50 W/m, not acceptable• Plans for accelerators of 5 MW (ESS), 10 MW and more

Limitation of beam losses is in order of 1 W/m to avoid activation and still allow hands-on maintenance

• Avoid beam losses – as far as possible• Define the aperture by collimators• Capture continuous particle losses with collimators at specific locations

LHC stored beam with an energy of 360 MJ• Assume lifetime of 10 minutes corresponds to beam loss of 500 kW, not

to be lost in superconducting magnets• Reduce losses by four orders of magnitude

….but also: capture fast accidental beam losses

CERN


RF contacts for guiding image currents

Beam spot

2 mm

View of a two sided collimator

for LHC

about 100 collimators are installed in LHC

Ralph Assmann, CERN

length about 120 cm

CERN


Accidental beam losses: Machine Protection

Single-passage beam loss in the accelerator complex (ns - s) • transfer lines between accelerators or from an accelerator to a target

station (target for secondary particle production, beam dump block)• failures of kicker magnets (injection, extraction, special kicker magnets,

for example for diagnostics)• failures in linear accelerators, in particular due to RF systems• too small beam size at a target station

Very fast beam loss (ms)• e.g. multi turn beam losses in circular accelerators• due to a large number of possible failures, mostly in the magnet

powering system, with a typical time constant of ~1 ms to many seconds

Fast beam loss (some 10 ms to seconds)Slow beam loss (many seconds)

CERN


Classification of failures

● Type of the failure• hardware failure (power converter trip, magnet quench, AC distribution

failure such as thunderstorm, object in vacuum chamber, vacuum leak, RF trip, kicker magnet misfires, .…)

• controls failure (wrong data, wrong magnet current function, trigger problem, timing system, feedback failure, ..)

• operational failure (chromaticity / tune / orbit wrong values, …)• beam instability (due to too high beam / bunch current / e-clouds)

● Parameters for the failure• time constant for beam loss• probability for the failure• damage potential

defined as risk

CERN


Probability of a failure leading to beam loss

● Experience from LHC (…..the most complex accelerator)• When the beam are colliding, the optimum length of a store is in the order

of 10-15 hours, then ended by operation• Most fills (~70 %) are ended by failures, the machine protection systems

detect the failure and dump the beams• MTBF of about 6 h

● Other large accelerators (SNS, plans for ESS, synchrotron light sources)

• MTBF between 20 h and up to several 100 h (…. more accurate numbers are appreciated)

● At high power accelerators, most failures would lead to damage if not mitigated = > the machine protection system is an essential part of the accelerator

CERN


Machine Protection

CERN


Example for Active Protection - Traffic

● A monitor detects a dangerous situation

● An action is triggered

● The energy stored in the system is safely dissipated

CERN


Example for Passive Protection

• The monitor fails to detect a dangerous situation

• The reaction time is too short

• Active protection not possible – passive protection by bumper, air bag, safety belts

CERN


Strategy for protection and related systems

● Avoid that a specific failure can happen

CERN


Strategy for protection and related systems

● Avoid that a specific failure can happen● Detect failure at hardware level and stop beam operation● Detect initial consequences of failure with beam instrumentation

….before it is too late…● Stop beam operation

• inhibit injection• extract beam into beam dump block• stop beam by beam absorber / collimator

● Elements in the protection systems• equipment monitoring and beam monitoring• beam dump (fast kicker magnet and absorber block)• chopper to stop the beam in the low energy part• collimators and beam absorbers• beam interlock systems linking different systems

CERN


Beam instrumentation for machine protection

● Beam Loss Monitors• stop beam operation in case of too high beam losses• monitor beam losses around the accelerator (full coverage!)• could be fast and/or slow (LHC down to 40 s)

● Beam Position Monitors • ensuring that the beam has the correct position• in general, the beam should be centred in the aperture

● Beam Current Transformers• if the current difference between two locations of the accelerator is too

high (=beam lost somewhere): stop beam operation• if the beam lifetime is too short: dump beam

● Beam Size Monitors• if beam size is too small could be dangerous for windows, targets, …

CERN


• Ionization chambers to detect beam losses:• Reaction time ~ ½ turn (40 s)• Very large dynamic range (> 106)

• There are ~3600 chambers distributed over the ring to detect abnormal beam losses and if necessary trigger a beam abort !

LHC Beam Loss Monitors

CERN


Layout of beam dump system in IR6

51

LHC Layouteight arcs (sectors)

eight long straight section (about 700 m long)

IR6: Beam dumping system

IR4: RF + Beam instrumentation

IR5:CMS

IR1: ATLAS

IR8: LHC-BIR2: ALICE

InjectionInjection

IR3: Moment Beam Clearing (warm)

IR7: Betatron Beam Cleaning (warm)

Beam dump blocks

Detection of beam losses with >3600

monitors around LHC

Signal to kicker magnet

Beams from SPS

CERN


LHC: Continuous beam losses during collisions

CMS Experiment

ATLAS Experiment

LHC Experiment

ALICE Experiment

MomentumCleaning

RF and BI

Beamdump

Betatron Cleaning

CERN


LHC: Accidental beam losses during collisions

CMS Experiment

ATLAS Experiment

LHC Experiment

ALICE Experiment

MomentumCleaning

RF and BI

Beamdump

Betatron Cleaning

CERN


LHC: Accidental beam losses during collisions

CERN


Beam 2

Beam dump block

Kicker magnets to paint (dilute)

the beam

about 700 m

about 500 m

15 fast ‘kicker’ magnets deflect the beam to the

outside

When it is time to get rid of the beams (also in case of emergency!), the beams are ‘kicked’ out of the ring by a system of kicker magnetsd send into a dump block !

Septum magnets deflect the

extracted beam vertically

quadrupoles

The 3 s gap in the beam gives the kicker time to

reach full field.

Ultra-high reliable system !!

R.Schmidt HASCO 2013 55

CERN Layout of LHC beam dumping system in IR6

CERN


Beam dumping system line for LHC

CERN


LHC Beam dump

●Screen in front of the beam dump block

●Each light dot shows the passage of one proton bunch traversing the screen

●Each proton bunch has a different trajectory, to better distribute the energy across a large volume

50 cm

CERN


High power accelerators …

● Operate with beam power of 1 MW and more

● SNS – 1 MW, PSI cyclotron – 1.3 MW, ESS – planned for 5 MW, FRIB (ions) – planned for 0.4 MW

● ESS (4 % duty cycle): in case of an uncontrolled beam loss during 1 ms, the deposited energy is up to 130 kJ, for 1 s it is up to 5 MJ

● It is required to inhibit the beam after detecting uncontrolled beam loss – how fast?

● The delay between detection and “beam off” to be considered

CERN


Example for ESS

source

dT = dT_detect failure + dT_transmit signal + dT_inhibit source + dT_beam off

inhibit beam interlock signal

Example:

After the DTL normal conducting linac, the proton energy is 78 MeV. In case of a beam size of 2 mm radius, melting would start after about 200 µs.

Inhibiting beam should be in about 10% of this time.

L.Tchelidze

Tim

e to

mel

ting

poin

t

CERN


Some design principles for protection systems

● Failsafe design• detect internal faults• possibility for remote testing, for example between two runs• if the protection system does not work, better stop operation rather than

damage equipment● Critical equipment should be redundant (possibly diverse)● Critical processes not by software (no operating system)

• no remote changes of most critical parameters● Demonstrate safety / availability / reliability

• use established methods to analyse critical systems and to predict failure rate

● Managing interlocks• disabling of interlocks is common practice (keep track !)• LHC: masking of some interlocks possible for low intensity / low energy

beams

CERN


Accelerators that require protection systems I

● Hadron synchrotrons with large stored energy in the beam• Colliders using protons / antiprotons (TEVATRON, HERA, LHC)• Synchrotrons accelerating beams for fixed target experiments (SPS)

● High power accelerators (e.g. spallation sources) with beam power of some 10 kW to above 1 MW

• Risk of damage and activation• Spallation sources, up to (and above) 1 MW quasi-continuous beam

power (SNS, ISIS, PSI cyclotron, JPARC, and in the future ESS, FRIB, MYRRHA and IFMIF)

● Synchrotron light sources with high intensity beams and secondary photon beams

● Energy recovery linacs• Example of Daresbury prototype: one bunch train cannot damage

equipment, but in case of beam loss next train must not leave the (injector) station

CERN


Accelerators that require protection systems II

● Linear colliders / accelerators with very high beam power densities due to small beam size

• High average power in linear accelerators: FLASH 90 kW, European XFEL 600 kW, JLab FEL 1.5 MW, ILC 11 MW

• One beam pulse can lead already to damage• “any time interval large enough to allow a substantial change in the beam

trajectory of component alignment (~fraction of a second), pilot beam must be used to prove the integrity” from NLC paper 1999

● Medical accelerators: prevent too high dose to patient• Low intensity, but techniques for protection are similar

● Very short high current bunches: beam induces image currents that can damage the environment (bellows, beam instruments, cavities, …)

CERN


For future high intensity machines

Machine protection should always start during the design phase of an accelerators

● Particle tracking • to establish loss distribution with realistic failure modes• accurate aperture model required

● Calculations of the particle shower (FLUKA, GEANT, …)• energy deposition in materials• activation of materials • accurate 3-d description of accelerator components (and possibly the

tunnel) required● Coupling between particle tracking and shower calculations● From the design, provide 3-d model of all components

CERN


Summary

Machine protection● is not equal to equipment protection● requires the understanding of many different type of failures that

could lead to beam loss● requires comprehensive understanding of all aspects of the

accelerator (accelerator physics, operation, equipment, instrumentation, functional safety)

● touches many aspects of accelerator construction and operation● includes many systems● is becoming increasingly important for future projects, with

increased beam power / energy density (W/mm2 or J/mm2 ) and increasingly complex machines

● I find it a fascinating topic ……… at least until nothing breaks

CERN


Acknowledgements to many colleagues from CERN and to the authors of the listed papers

● R.F.Koontz, Multiple Beam Pulse of the SLAC Injector, PAC 1967● R.Bacher et al., The HERA Quench Protection System, a Status Report, EPAC 1996● C.Adolphsen et al., The Next Linear Collider Machine Protection System, PAC 1999● M.C.Ross et al., Single Pulse Damage in Copper, LINAC 2000● C.Sibley, Machine Protection Strategies for High Power Accelerators, PAC 2003● C.Sibley, The SNS Machine Protection System: Early Commissioning Results and Future Plans, PAC

2005● S.R.Buckley and R.J.Smith, Monitoring and Machine Protection Designs for the Daresbury

Laboratory Energy Recovery Linac Prototype, EPAC 2006● L.Fröhlich et al., First Operation of the FLASH Machine Protection System with long Bunch Trains,

LINAC 2006● L.Fröhlich et al., First Experience with the Machine Protection System of FLASH, FEL 2006● N.V.Mokhov et al., Beam Induced Damage to the TEVATRON Components and what has been done

about it, HB2006● M.Werner and K.Wittenburg, Very fast Beam Losses at HERA, and what has been done about it,

HB2006● S.Henderson, Status of the Spallation Neutron Source: Machine and Experiments, PAC 2007● H.Yoshikawa et al., Current Status of the Control System for J-PARC Accelerator Complex, ICALEPCS

2007● L.Froehlich, Machine Protection for FLASH and the European XFEL, DESY PhD Thesis 2009● A.C.Mezger, Control and protection aspects of the megawatt proton accelerator at PSI, HB2010 ● Y.Zhang, D.Stout, J.Wei, ANALYSIS OF BEAM DAMAGE TO FRIB DRIVER LINAC, SRF 2012

CERN


CERN and LHC● R.B.Appleby et. al., Beam-related machine protection for the CERN Large Hadron Collider

experiments, Phys. Rev. ST Accel. Beams 13, 061002 (2010)● R.Schmidt et al., Protection of the CERN Large Hadron Collider, New Journal of Physics 8 (2006) 290● R.Schmidt, Machine Protection, CERN CAS 2008 Dourdan on Beam Diagnostics● N.Tahir et al., Simulations of the Full Impact of the LHC Beam on Solid Copper and Graphite Targets,

IPAC 2010, Kyoto, Japan, 23 - 28 May 2010Theses● Verena Kain, Machine Protection and Beam Quality during the LHC Injection Process, CERN-THESIS-

2005-047● G.Guaglio, Reliability of the Beam Loss Monitors System for the Large Hadron Collider at CERN /,

CERN-THESIS-2006-012 PCCF-T-0509● Benjamin Todd, A Beam Interlock System for CERN High Energy Accelerators, CERN-THESIS-2007-019● A. Gomez Alonso, Redundancy of the LHC machine protection systems in case of magnet failures /

CERN-THESIS-2009-023● Sigrid Wagner, LHC Machine Protection System: Method for Balancing Machine Safety and Beam

Availability /, CERN-THESIS-2010-215● Roderik Bruce, Beam loss mechanisms in relativistic heavy-ion colliders, CERN-THESIS-2010-030


CERN


● L.Tchelidze, In how long the ESS beam pulse would start melting steel/copper accelerating components? ESS AD Technical Note, ESS/AD/0031, 2012

● Conference reports in JACOW, keywords: machine protection, beam loss


CERN


Example for LHCCollimation and Machine Protection during

operation

Assume that two 100 MJoule beams (=25 kg TNT) are circulating with the speed of light through the 56 mm diameter vacuum chamber and 2 mm wide collimators

1. Suddenly the AC distribution for CERN fails – no power!2. An object falls into the beam3. The betatron tune is driven right onto a 1/3 order resonance

CERN


LHC from injection to collisions

3.5 TeV / 100 MJoule

0.45 TeV / 13 MJoule

Energy ramp

Luminosity: start collisions

Injection of 1380 bunches per beam About 2 hours

CERN


Orbit for last 1000 turns before power cut

CERN


Example for power radiated during particle collisions for LHC

Rate of collision:

Power in collision products:

Assume LHC operating at 7 TeV with a luminosity of:

Total cross section for pp collision of 110 mBarn:

Power in collision products per experiment:

• Some fraction of the protons are deflected by a small angle and remain in the vacuum chamber

• Some fraction hits close-by equipment

Continuous beam losses

CERN


Total power cut atLHC - 18 August 2011, 11:45

CERN


1. Suddenly the AC distribution for CERN fails – no power for LHC!

CERN


\\cern.ch\dfs\Users\r\rudi\Documents\ConferencesWorkshops\SCHOOLS\CAS\CAS2011\UFO-slideshow.pptx

UFO at LHC

CERN


LHC from injection to collisions: beam loss

CERN


…zoom - going into collisions

CERN


Beam cleaning system captures beam losses

● In case protons are lost because of low lifetime ● In case of protons are lose when colliding beams, and scattering

of protons during the collisions that would be lost around the LHC● In case of protons outside the RF bucket – losing slowly energy –

are captured by collimators in the Momentum Cleaning Insertion

Questions● How to stop high energy protons?● Why so many collimators?● Why carbon composite or graphite used for most collimators?

CERN


Collimator material

● Metal absorbers would be destroyed● Other materials for injection absorber preferred, graphite or

boron nitride for the injection absorber● In case of a partial kick (can happen), the beam would travel

further to the next collimators in the cleaning insertions

P.Sievers / A.Ferrari / V. Vlachoudis7 TeV, 21012 protons

• For collimators close to the beam, metal absorbers would be destroyed

• Other materials for collimators close to the beam are preferred (carbon – carbon)

CERN


Collimation

● For a circular accelerator, the transverse distribution of beams is in general Gaussian, or close to Gaussian (beams can have non-Gaussian tails)

● In general, particles in these tails cause problems when they might touch the aperture

• Background• Quenches in magnets (for accelerators with sc magnets)• For high intensity machines, possible damage of components

● Nearly all particles that are in the centre go first through the tails before getting lost (except those that do a inelastic collision with gas molecules)

● Tails are scraped away using collimators

CERN


Phase space and collimation

x’

x

x’

Starting with a Gaussian beam profile

CERN


Phase space and collimation

x’

x

x’

Collimator outside the beam

CERN


Phase space and collimation: multi turn

x’x’

x

Collimator driven into the beam tail: loss of particles

CERN


Phase space and collimation: multi turn

x’x’

x

Collimator again outside the beam – beam size reduction (for proton synchrotrons)

CERN


Phase space and collimation: single turn

x’x’

x

Collimator in a transfer line or linac: cuts only part of the beam

CERN



x’

x

x’

Collimator in a transfer line or linac: cuts only part of the beam

CERN



x’x’

x

Collimator in a transfer line or linac: several collimators are required

CERN



x’x’

x

Collimator in a transfer line or linac: several collimators are required …. at different betatron phases

90 degrees further down

CERN


Gaussian beam not collimated

4 3 2 1 0 1 2 3 40

3 10 4

6 10 4

9 10 4

1.2 10 3

1.5 10 3

1.8 10 3

2.1 10 3

2.4 10 3

2.7 10 3

3 10 30.003

0

g.h t( )

44 t

CERN


Gaussian beam collimated at 4 sigma

4 3 2 1 0 1 2 3 40

3 10 4

6 10 4

9 10 4

1.2 10 3

1.5 10 3

1.8 10 3

2.1 10 3

2.4 10 3

2.7 10 3

3 10 30.003

0

g.h t( )

44 t

N = 0.999 (number of protons) L = 0.999 (luminosity)

CERN



4 3 2 1 0 1 2 3 40

3 10 4

6 10 4

9 10 4

1.2 10 3

1.5 10 3

1.8 10 3

2.1 10 3

2.4 10 3

2.7 10 3

3 10 30.003

0

g.h t( )

44 t

N= 0.987L = 0.992

CERN



4 3 2 1 0 1 2 3 40

3 10 4

6 10 4

9 10 4

1.2 10 3

1.5 10 3

1.8 10 3

2.1 10 3

2.4 10 3

2.7 10 3

3 10 30.003

0

g.h t( )

44 t

N= 0.863L = 0.866

CERN


Collimation: why so many?

Answer A:● For a transfer line or a linear accelerator, many collimators are

required to take out particles at all phasesAnswer B:● Cite: “It is not possible to stop a high energy proton, it is only

possible to make them mad”● Collimators cannot stop a high energy proton● The particle impact on a collimator jaw is very small, in the order

of microns or even less● Particles scatter….. depends on particle type, energy and impact

on collimator jaw● Staged collimation system in a ring and in a transfer line

CERN


Betatron beam cleaning

Cold aperture

Cleaning insertion Arc(s) IP

Circulating beam

Illustration drawing

Arc(s)

Primarycollimator

Secondarycollimators

Tertiary beam halo + hadronic showers

Shower absorbers

Tertiarycollimators

SCTriplet

CERN


Measurement: 500kJ losses at primary collimators (loss rate: 9.1e11 p/s) – IR7

Daniel Wollmann 94

TCP: ~505 kJ

Q8L7: ~335 J

Q11L7: ~35 J

Q19L7: ~4.7 J

Q8L7:h ~ 6.7e-4

Lower limit:RqLdil ~ 1.22e9 p/s (with cresp= 2 )

Lost energy over 1 s

CERN


Film from Alessandro

CERN


173 bunches grazing incident on injection absorber

Upstream of IP2

Beam 1

Downstream of IP2

Beam 1

Insertion losses: 3 magnets quenched (D1.L2, MQX.L2, D2.R2)

TDI

Losses starting at TDI, no injection loss signature only circulating beam kicked by MKI

In comparison to flashover event of April 18th in P8 (LMC 20/04/11), cleaner in arc less magnet quenches (3), ALICE SDD permanent effect, open MCSOX.3L2 circuit

C.Bracco

CAS October 2013 R.Schmidt 97

Is protection required?


Protection for beam transfer from SPS to LHC

A signal “extraction permit” is required to extract beam from SPS and another signal “injection permit“ to inject beam into LHC

• After extraction the trajectory is determined by the magnet fields: safe beam transfer and injection relies on correct settings – orbit bump around extraction point in SPS during extraction with tight

tolerances verified with BPMs– correct magnet currents (slow pulsing magnets, fast pulsing magnets)– position of vacuum valves, beam screens,… must all be OUT– energy of SPS, transfer line and LHC must match– LHC must be ready to accept beam

• Verifying correct settings just before extraction and injection

• The kicker must fire at the correct time with the correct strength• Position of collimators and beam absorbers in SPS, transfer line and

LHC injection region to protect from misfiring


Case studies

The principles of machine protection are illustrated with examples from different accelerators


Example: SNS

• normal conducting linac• superconducting linac• accumulator ring• transfer lines• target station• beam power on target 1.4 MW• beam pulse length 1 ms• repetition rate 60 Hz

• (more or less) continuous beam to above 1 MW – the deposited energy is proportional to the time of exposure– the risk (possible damage) increases with time

• Protection by detecting the failure and stopping injection and acceleration


SNS damage limits

• Damage of a copper cavity: Time to reach the thermal stress limit for copper assuming a beam size of 2 mm, a current of 36 mA and an energy density of 62 J/gm as maximum permitted energy deposition (from C.Sibley, PAC 2003)

• The SNS MP system uses inputs from BLMs, beam current monitors, RF, power supplies, vacuum system, kickers, etc.

102

Radiation Damage to Undulator Magnets

• Nd2Fe14B magnets lose magnetizationwhen irradiated

• literature: relative demagnetization rate 10−8/Gy (gammas) — 10−4/Gy (fast neutrons)

hybrid magnet structure

Nd2Fe14Bpermanent magnets

soft magnetic pole pieces

Lars Froehlich, DESY and Uni Hamburg, Machine Protection Machine Protection for FLASH and the European XFEL

103

Conclusion

• Superconducting linacs can transport dangerously powerful beams

• Permanent magnet undulators are among the most vulnerable components

• Beam losses must be controlled tightly (FLASH design: 3∙10−8)

• Dark current can be problematic• Good passive & active protection

is required

• FLASH machine protection system is fully functional & reliable

• XFEL machine protection system will be more complex, but concepts & first prototypes are readyLars Froehlich, DESY and Uni Hamburg, Machine Protection Machine Protection for FLASH and the

European XFEL


Machine Protection during all phases of operation

• The LHC is the first accelerator with the intensity of the injected beam already far above threshold for damage, protection during the injection process is mandatory

• At 7 TeV, fast beam loss with an intensity of about 5% of one single “nominal bunch” could damage equipment (e.g. superconducting coils)

• The only component that can stand a loss of the full beam is the beam dump block - all other components would be damaged

• The LHC beams must ALWAYS be extracted into the beam dump blocks– at the end of a fill– in case of failure

• During powering, about 10 GJ is stored in the superconducting magnets, quench protection and powering interlocks must be operational long before starting beam operation


Multiturn beam losses

Consequence of a magnet powering failure– Closed orbit grows and moves everywhere the ring or downstream

the linac (follows free betatron oscillation with one kick) – Beam size explodes – Can happen very fast (for example, if a normal conducting magnet

trips or after a magnet quench)– Can be detected around the entire accelerator

Local orbit bump– Can be generated due to BPM offset– Needs several magnets to fail and cannot happen very fast– Might be detected only locally

• Protection: Detect failure and dump beam• Detection by equipment monitoring and beam monitoring


Failure of normal conducting magnets

After about 13 turns 3·109 protons touch collimator, about 6 turns later 1011 protons touch collimator

V.Kain / O.Bruning

“Dump beam” level

1011 protons at collimator


Fast Magnet Current change Monitors (initial development for HERA, upgrade for LHC in collaboration with DESY)

• Several FMCMs are installed on critical magnets• Tested using steep reference changes to trigger FMCM. The trigger threshold and

the magnet current (resolution one ms)• Beam tests confirmed these results

Reference

PC current

time (ms)

I (A)

FMCM trigger 0.1% drop !

time (ms)

I (A)

10 ms

FMCM triggers @ 3984.4 <103


Principle of LHC / SPS Beam Interlock Systems

BIS

LHC Dump kicker

Beam ‘Permit’

User permitsignals

• ‘User systems’ survey equipment or beam parameters, detect failures and send a hardwired signal to the beam interlock system (user permit)

• The BIS combines user permits and produces beam permit• The beam permit is a hardwired signal to injection / extraction

kickers :• • LHC ring: absence of beam permit dump triggered !• LHC injection: absence of beam permit no injection !• SPS: absence of beam permit no extraction !

Hardware links /systems, fully redundant

SPS Extraction kicker

LHC Injection kicker

SPS Dump kicker


Machine Protection and Controls

• Software Interlock Systems (SIS) provides additional protection for complex but also less critical conditions– Surveillance of magnet currents to avoid certain failures (local bumps)

that would reduce the aperture– The reaction time of those systems will be at the level of a few

seconds– The systems rely entirely on the computer network, databases, etc –

clearly not as safe as HW systems!

• Sequencer: program to execute defined procedures – To execute defined well-tested procedures for beam operation

• Logging and PM systems: recording of data – continuous logging and for transients (beam dump, quench, …) – Very important to understand what happened


P.Sievers / A.Ferrari / V. Vlachoudis

Beryllium

Accidental kick by the beam dump kicker at 7 TeV part of beam touches collimators (about 21012 from 31014 )


Target length [cm]

vaporisation

melting

N.Tahir (GSI) et al.

Copper target

2 m

Energy density [GeV/cm3] on target axis

2808 bunches7 TeV 350 MJoule

Full LHC beam deflected into copper target



Injection absorbers(TCLI) ~7σ

n·180 +/- 20 degrees

Beam absorbers take beam in case of kicker wrong strength


Beam from SPS

Injection Kicker Set of transfer line

collimators (TCDI) ~5σ


Circulating beam – kicked out

Injection kicker – wrong strength

phase advance 900




Injection Kicker

Injection absorberTDI ~7σ

Injection absorbersTCLI ~7σ

Only when beam is circulating in the LHC, injection of high intensity beam is permitted – verification of LHC magnet settings

Probe Beam: Replacing low intensity beam by a full batch from SPS

Set of transfer line collimators TCDI

~5σ

Beam from SPS


Active and passive protection

Start operation with low intensity beam (“pilot beam”)

Active protection• Detect failure• Turn off the beam as fast as possible (e.g. source, RF, …)• Only permit beam injection into the next part of the

accelerator complex in case of positive confirmation that all parameters are within predefined limits

• Abort the beam from a storage ring / accumulator ring

Passive protection• Install collimators and beam absorbers, in particular if active

protection is not possible


Active protection

Monitoring of the beam detects a failure and allows to switch off the beam before damage

• Stored beam in a circular accelerator– multi turn beam losses– monitor beam losses, and dump the beam if losses exceed threshold

• “Continuous” beam in linacs of other accelerators– continuous: if the time constant for a failure is such that the source

can be switched off in time

• There is a large number of possible failures, mostly in the magnet powering system, with a typical time constant of ms to many seconds


Cold aperture

Primary beam halo

Primarycollimator

Secondarycollimators

Tertiary beam halo + hadronic showersSecondary beam halo

+ hadronic showers

Shower absorbers

Cleaning insertion

Tertiarycollimators

SCTriplet

Arc(s) IP

Protection devices

Circulating beamIllustrative scheme


Failure of a kicker magnet

1 km

Extraction kicker magnet:• wrong pulse strength• wrong timingInjection kicker magnet:• wrong pulse strength• wrong timing

Transfer line

LHC

SPS6911 m

450 GeV / 400 GeV3 MJ

Acceleration cycle of 10 s

CNGS Target

IR8

Switching magnetFast extraction

kicker

Injection kicker

Transfer line

Injection kicker



Failure in the transfer line (magnet, other element)

1 km

Wrong setting of magnetsObject in the transfer line

blocks beam passage

Transfer line

LHC

SPS6911 m

450 GeV / 400 GeV3 MJ

Acceleration cycle of 10 s

CNGS Target

IR8

Switching magnetFast extraction

kicker

Injection kicker

Transfer line

Injection kicker



M.Jonker

Beam damage capabilities

CERN


Accidental beam losses: Risks and protection

● Protection is required since there is some risk● Risk = probability of an accident (in number of accidents per year)

consequences (in Euro, downtime, radiation dose to people)● Probability of an accidental beam loss

• What are the failure modes the lead to beam loss into equipment (there is an practical infinite number of mechanisms to lose the beam)?

• What is the probability for the most likely failures?● Consequences of an accidental beam loss

• Damage to equipment• Downtime of the accelerator for repair (spare parts available?)• Activation of material, might lead to downtime since access to

equipment is delayed● The higher the risk, the more protection becomes important

CERN


Damage of LHC during the 2008 accident

Accidental release of an energy of 600 MJoule stored in the magnet system - no beam in LHC

CERN


Energy loss: example for protons in Aluminium

CERN


Temperature increase in the material: dTFeNp dEdxFe

cFe_spec Fbeam Fe

Temperature increase for a proton beam impacting on a Fe target:

Beam size: h mm and v mm

Iron specific heat: cFe_specJ

kg K

Iron specific weight: Fekg

m3

Energy loss per proton/mm: dEdxFeMeVmm

Number of protons: Np

Energy of the proton: Ep GeV

Temperature increase: dTFe

Heating of material with low energy protons

CERN


Strategy for Machine Protection

Beam Interlock System

Collimator and Beam Absorbers

• Early detection of failures for equipment acting on beams and generating beam abort request (inhibit injection, dump beam).

• Active monitoring of the beams detects abnormal beam conditions and generates beam abort request.

• Reliable transmission of beam abort requests to beam stop system. Active signal required for operation, absence of signal is considered as beam abort request and injection inhibit.

• Reliable operation of beam abort system.

• Passive protection by beam absorbers and collimators for specific failure cases.

CERN


LHC: Strategy for machine protection● Definition of aperture by collimators. Beam Cleaning System

Beam Loss MonitorsOther Beam Monitors

Beam Interlock System

Powering Interlocks Fast Magnet Current

change Monitor

Beam Dumping System

Collimator and Beam Absorbers

• Early detection of failures for equipment acting on beams generates dump request, possibly before the beam is affected.

• Active monitoring of the beams detects abnormal beam conditions and generates beam dump requests down to a single machine turn.

• Reliable transmission of beam dump requests to beam dumping system. Active signal required for operation, absence of signal is considered as beam dump request and injection inhibit.

• Reliable operation of beam dumping system for dump requests or internal faults, safely extract the beams onto the external dump blocks.

• Passive protection by beam absorbers and collimators for specific failure cases.

CERN



● Equipment becomes activated due to beam losses (acceptable is ~1 W/m, but must be “As Low As Reasonably Achievable – ALARA”)

CERN


Energy deposition and temperature increase

● SPS at 450 GeV, number of protons per cycle about ● Energy deposition in vacuum chamber (stainless steel)

2.2 MeV/mm● Assuming a beam size of 1 mm, the temperature increase is

expected to be about 1020 C

CERN



● ESS at 3 MeV, number of protons per 14 Hz pulse is about ● Assume we consider the space between RFQ and MEBT, with a

proton energy of 3 MeV● Energy deposition in vacuum chamber (stainless steel)

59.7 MeV/mm● Assuming a beam size of 3 mm, one full beam pulse that is

deflected to the wall would be far above damage limit● If the source is stopped after 30 , the temperature would increase

to about 600 C

CERN



● There is no straightforward expression for the energy deposition● The energy deposition is a function of the particle type, its

momentum, and the parameters of the material (atomic number, density, specific heat)

● Programs such as FLUKA, MARS, GEANT and others are being used for the calculation of energy deposition and activation

● Other programs are used to calculate the response of the material (deformation, melting, …) to beam impact (mechanical codes such as ANSYS, hydrodynamic codes such as BIG2 and others)

Question: what is dangerous (stored beam energy, beam power)?

CERN


What parameters are relevant?

● Momentum of the particle ● Particle type

• Activation is mainly an issue for hadron accelerators

● Time structure of beam● Energy stored in the beam

• one MJoule can heat and melt 1.5 kg of copper

• one MJoule corresponds to the energy stored in 0.25 kg of TNT

● Beam power• one MWatt during one second

corresponds to a MJoule● Beam size● Beam power / energy density

(MJoule/mm2, MWatt/mm2)

Consequences for an accident with 360 MJ beam can be catastrophic (LHC damage beyond repair)

Delicate components can be damaged already with some Joule (e.g. RFQ)Machine protection becomes very important if beam energy > 0.1 – 1 MJ

accidental and continuous beam losses protection of the accelerator from beam losses

Documents

beam energy

beam pulses

mw beam powerpower

mw beam powerif

msaverage power

mwpeak power

ess cas accelerator

power converters