Performance of Injection Protection Systems V.Kain AB/CO

4/12/2005 Verena Kain, AB-CO 1

Performance of Injection Protection Systems

V.Kain AB/CO

Contents •SPS Extraction – Transfer – LHC Injection •Injection Protection System Principle•Passive Protection•Interlocking•Simulations for Protection Level•ConclusionsInput from B. Goddard, R. Genand, R. Schmidt,J. Wenninger, M. Werner

4/12/2005 Verena Kain, AB-CO 2

SPS Extraction – Transfer – LHC Injection

Beam 2

Beam 1

Injection Process

LSS6/TT60 TI 2 LHC injection LHC SPS

TE

D T

T60

TE

D T

I2

TD

I IR

2

MKE MKI

IR2 BEAM1

Extraction Transfer Injection

TED…beam stopper in transfer lineTDI…injection stopper in injection region

To define the Injection Protection System, SPS Extraction, Transfer and LHC Injection must be treatedtogether.

SPS:MSE…extraction septumMKE…extraction kicker

LHC:MSI…injection septumMKI…injection kicker

BEAM2: LSS4/TT40 – TI 8 – IR 8

4/12/2005 Verena Kain, AB-CO 4

Injection/Extraction Constraints

Inside, damage visible over ~1m (melted steel)

8x1012p+ = ¼ of full batch

Holes in Cu : 450 GeV p+ beam(from 2004 TT40 materials test)

5.3x1012p+ = 1/6 of full batch

• Damage limit:~2x1012p+, ~5 % of injected batch

• Small Aperture:– 7.5 LHC aperture at 450GeV– Tight aperture in transfer line (MSI injection septum ~7)

Full nominal injected LHC batch: 3.3x1013p+, 450 GeV

4/12/2005 Verena Kain, AB-CO 5

What can go wrong: magnet trips, kicker failures, wrong

settings…

• Magnet trips can move trajectory by many in short time (MSE extraction septum: 40 in 1ms)• Kicker erratics, missings, timing etc.• Operator error• Corrupted settings

10 cm~25cm long hole in QTRF chamber

Extraction septum (MSE) trip during high intensity extraction, Oct 2004

Slow failures:10 in > 2-3ms: relying on interlocking

Fast failures:10 in < 2-3ms: interlocking + collimators

Ultra-fast failures: 10 in few s: collimators

Injection Process

4/12/2005 Verena Kain, AB-CO 6

Principle of Machine Protection for Injection

Process

ProtectionHW surveillance:PCS, FCCM, settings monitoring,equipment status

LocalBIC

MasterExtr./Inj.BIC

Failu

re,

Err

or

eff

ect

on:

HW

Beam Collimators, Absorbers

Inj./Extr.Inhibit

ACTIVE

PASSIVE

AvoidanceProcedures to avoid dangerous situationse.g. never inject high intensity beam in empty LHC

Beam Presence Flag → protects against many failures

4/12/2005 Verena Kain, AB-CO 7

Passive Protection for fast and ultra-fast failures

• Protection of LHC aperture and MSI injection septum aperture: – TCDI collimators for failures upstream of injection

regions– “generic” protection system with full phase coverage

• Dedicated collimators for kicker failures:– MKI (LHC) failure: TDI beam stopper + TCLI

collimators are 90° downstream – MKE (SPS) failure: TPSG diluter is 90° downstream

• No dedicated collimators for septum failures– Protection from MSE and MSI failures interlocking

4/12/2005 Verena Kain, AB-CO 8

TCDI Transfer Line Collimators

• Close to LHC and injection septum (last 300m matching section of TLs)

• Robust, based on LHC collimator design (1.2m C jaw)

• FLUKA model of 300m of TI 8 local shield for each TCDI

x’

x

0-60-120 degree collimators

60o120o

LHC apertureto protect at7.5

amax

6.9

• 3 collimators / plane (0-60-120°)

• Setting: 4.5, tolerances: ≤1.4

• 2 motors/ jaw (angular control)

• Protection level 6.9: result from comprehensive Monte-Carlo simulation including all imperfections: beat, mismatch from SPS, tolerances,…

6

4/12/2005 Verena Kain, AB-CO 9

TDI injection stopper, TCLI collimators

• TDI injection stopper:– Protect LHC (especially D1) against MKI kicker failures– 90° downstream of the MKI ~4m long hBN+Al+Cu jaws– local protection of SC LHC magnet D1 with mask -> TCDD (1m, Cu)

• Auxiliary collimators TCLIs – For MKI-TDI phase advance ≠90°, and for flexibility (halo load…)– At nx180°±20° from TDI (1.2m long C jaws)

MKI

Overview, vertical plane: functionality of TDI injection stopper

orbit

4/12/2005 Verena Kain, AB-CO 10

TDIMKI +90˚

TCDD

TCLIBTDI +340˚

TCLIATDI +200˚

KickerMKI

LEFT OF IP2

RIGHT OF IP2 TCLIM

TDI – TCDD - TCLI

Topview

4/12/2005 Verena Kain, AB-CO 11

Interlocking for Extraction – Transfer -Injection

• Segmented Interlocking System, different possible operational modes in a safe way– e.g. Extraction without injection, IF TED (transfer line

beam stopper) in the beam

• Without “LHC injection permit” NO “SPS extraction permit”

• Beam Presence condition for high intensity injection

• Safe Beam Flag: “maskable” interlock signals are ignored

4/12/2005 Verena Kain, AB-CO 12

Interlocking System: linking injection with extraction

LSS6/TT60 TI 2 upstream TI 2 downstream LHC injection LHC SPS

SP

S e

xtra

ct p

erm

it

SPS MKE extract

SPS LSS6MASTER

BIC

LHC IR2 injection

BIC

TI 2 upstream

BIC

TI8

ups

trea

m

TE

D T

T60

TE

D T

I2

SPS MKDdump

SP

S b

eam

per

mit

SPS ringBIC

SP

S r

ing

UP

s

SPS LSS6/ TT60BIC

LSS

6/T

T40

UP

s

SPS safe intensity flag

TD

I IR

2

UP

s

TI 2 downstream

BIC

TI8

dow

nstr

eam

LHC ringBIC

dumpLHC MKD

LHC

bea

m p

erm

it

LHC

inje

ctio

n

UP

s

Includes TI 2 after TED

TI 2 interlocking

LHC

inje

ct p

erm

it

LHC MKIinject

LHC safe parameters

4/12/2005 Verena Kain, AB-CO 13

Protection level simulations to quantify system

performance

• Extensive tracking simulations to check performance

• MKI kicker failure scanned with injection absorber setting

• Full Monte Carlo of single and grouped failures at injection

– 14 magnet and kicker families (SPS extraction, Transfer Line, LHC injection) for LSS4 - TI 8 – IR8

– Full TL + LHC injection region aperture model (~3 km)

– All imperfections and errors included

Safe LHC injection losses on aperture below 5% damage limit during injection

4/12/2005 Verena Kain, AB-CO 14

Injection kicker failure simulation results

• TDI, TCLI collimator setting of 6.8, to guarantee max. 5% above 7.5

• Increasing collimator opening increases risk of damage

N/N0 of particles with amplitudes >7.5 y

6.8

4/12/2005 Verena Kain, AB-CO 15

Single failure tracking Monte Carlo results (1000 seeds per

failure)Family Tolerable

error[k/k0]

required reaction time

[ms]

Covered by

LHC TL

MPLH (LSS4 bumper) 0.185 201.0 TCDI PCS

MKE (LSS4 kicker) 0.125 - TCDI -

MSE (LSS4 septum, =23ms) 0.005 0.1 TCDI FCCM

MBHC (TT40 H bend) 0.005 5.1 TCDI FCCM

MBHA (TT40 H bend) 0.012 31.47 TCDI PCS

MBI (main TI 8 bends) 0.003 2.7 TCDI FCCM

MCIBH (start TI 8 H bend) 0.630 389.0 TCDI PCS

MBIAH (end TI 8 H bend) 0.003 7.9 FCCM FCCM

MBIBV (end TI 8 V bend) 0.003 43.4 PCS PCS

3MCIAV (end TI 8 V bend) 0.183 98.43 PCS TCDI

MSI (LHC injection septum) 0.0035 3.5 FCCM n/a• PCS = standard Power Convertor Surveillance (≥3ms)• FCCM = Fast Current Change Monitor, dedicated new system

Grouped Failures: Powering Scheme for Extraction – Transfer – Injection (beam2)

MBIAV

MBI

MBHA

MSE

BA418kV/660V

18kV/877V

18kV/560V

18kV/100V

MBIAHSR818kV/570V

MBIBVMSIB and A

18kV/400V

MBHC4001QTLF4004MCIBH8040MQID8010MQIF8020MQID8030

18kV/400V

MQIF18kV/660V

MQID18kV/660V

MQIF8720MQID8730MQIF8740MQID8750MQIF8760MQID8770MQIF8780MQID8790MQIF8800MQID8810

18kV/400V

MDSV4002QTRD4003

18kV/400V

QTRF4002QTMD4001

Grouped Failures

MBIAV

MBI

MBHA

MSE

BA418kV/660V

18kV/877V

18kV/560V

18kV/100V

MBIAHSR818kV/570V

MBIBVMSIB and A

18kV/400V

MBHC4001QTLF4004MCIBH8040MQID8010MQIF8020MQID8030

18kV/400V

MQIF18kV/660V

MQID18kV/660V

MQIF8720MQID8730MQIF8740MQID8750MQIF8760MQID8770MQIF8780MQID8790MQIF8800MQID8810

18kV/400V

MDSV4002QTRD4003

18kV/400V

QTRF4002QTMD4001

B

A C

D

E

Grouped failure tracking Monte Carlo results (1000 seeds per

failure)Group Tolerable time

after switch-off [ms]

Covered by

LHC TL

A 1.3 TCDI FCCM on MBHC (0.1% tolerable

error)

B 0.1 TCDI FCCM on MSE

C 15.8 TCDI PCS

D 3.5-5.8, > 20 FCCM on MSI TCDI/PCS

E 4.0-5.4, > 20 FCCM on MBIAH, MSI

FCCM on MBIAH(0.15% tolerable

error)• In some cases grouped failures can be ~ 5 times worse than single failures

• e. g. MBHC• Grouped failures covered with protection for single failures BUT: requires increased performance

4/12/2005 Verena Kain, AB-CO 19

Discussion of results• Fast magnet Current Change Monitor (FCCM) is required

– Specification: I/I=0.1%, reaction time ~50s

• With FCCM at the MSI injection septum LHC protection looks OK

• Transfer Line (TL): FCCMs are proposed for MSE, MBI, MBIAH, MBHC– MKE extraction kicker faults can still cause damage to the TL…

possible solutions being studied.

• All other single failures are covered for Transfer Line & LHC– Transfer Line collimation system (TCDI) gives full

protection from upstream failures– Failures at end of line: slow enough for normal power

converter surveillance or FCCM

• Grouped Failures: can be ~ 5times faster than single failures– Covering single failures also covers grouped failures– But needs increased surveillance system performance

4/12/2005 Verena Kain, AB-CO 20

Fast Current Change Monitor for MSE extraction

septum

• Managed to detect I/I<0.3%, reaction time < 50 s

– larger ripple on test-bench MSE power supply than real circuit (0.17% instead of 0.04%)

• Looks promising, needs to be finalized

FCCM tested for MSE by M. Werner from DESY this month FCCM

MSE test-bench

FCCM measures changes of magnetvoltage: no comparison with reference value

FCCM

4/12/2005 Verena Kain, AB-CO 21

Conclusions

• Tracking simulations extensively used to define protection systems and to determine protection levels– Quantitative results requirements identified, specification of surveillance performance

– LHC “fully protected” with foreseen active and passive protection• Require Fast Current Change Monitor to measure 0.1% I, ~50s

reaction time• FCCM Development (M. Werner) looks promising even for fast

extraction septum• Appropriate interlocking system has been specified

– At present the protection systems cannot fully exclude TL damage• alternative solutions are being studied

• Simulations for total power cut and combined failures (protection device failure + other failure) will be carried out