Bericht vom
Accelerator Reliability Workshop
ARW 2011
A. Denker
Reliability gemäß LEO:
Ausfallsicherheit
Betriebssicherheit
Funktionsfähigkeit
Verlässlichkeit
Zuverlässigkeit
Statistik: Übersicht
 1. 2002 ESRF
2. 2009 Triumf
3. 2011 Kapstadt
 41 Beiträge von Beschleunigern
– 78 Teilnehmer, ohne “Locals”: 1/3 Amerika, 1/5 Asien, knapp ½ Europa
 3 Beiträge aus anderen Einrichtungen
– Kernkraftwerk Koeberg,
– SALT (South African Large Telescope): jwd, Temp.gradienten
– South African Square Kilometer Array Project:
jwd, möglichst ferngesteuert, da bereits ein Handy stört
 Zwei Diskussionsrunden
– Magnete
– Webseite/Forum
Statistik: wer war da
 Lichtquellen:



SLAC, BNL, Diamond, ESRF, PSI, Australian synchroton,
Spring8, SSRF
Spallationsquellen:
PSI, SNS
Zyklotrons:
allg.: NSCL, iThemba, Uppsala, RIKEN, IBA
med.: MGH, Orsay, HZB,
Sonstige:
GSI, HIMAC, INFN, DaLinac, Siemens, CERN, Fermilab
Historischer Überblick: Hardy (ESRF)
Ausblick: Hardy (ESRF)
Ansprüche Spallation: PSI
Theorie und Programme: SNS
Theorie <-> echtes Leben
Trend formale Organisation: GSI
Trend formale Organisation: SLAC
Trend formale Organisation: CERN
Trend formale Organisation: BNL
Trend formale Organisation: NSCL
Automatisierung: Australien
 Lichtquellen:



SLAC, BNL, Diamond, ESRF, PSI, Australian synchroton,
Spring8, SSRF
Spallationsquellen:
PSI, SNS
Zyklotrons:
allg.: NSCL, iThemba, Uppsala, RIKEN, IBA
med.: MGH, Orsay, HZB,
Sonstige:
GSI, HIMAC, INFN, DaLinac, Siemens, CERN, Fermilab
Kontrollsystem: ESRF
e_log: INFN
e_log: http://midas.psi.ch/elog
RF verbessert: RIKEN Ringzyklotron
Erreicht durch
- Vakuum verbessert
- CW – konditionieren…
RF verbessert: DALINAC
Temp.sensor auf RF Regelkarte
Diagnose: Diamond
Remotely Controlled, Mobile, Thermal Imaging Platform
fehlende Diagnose: CERN
Wartung auch für Ersatzteile: Spring 8
Probleme durch Kleinigkeiten: SINAP
Probleme durch Kleinigkeiten: CERN
Umfrage: Magnetfehler
Stromausfälle: iThemba
Stromausfälle: Australien
med. Beschleuniger: MGH: >95%
med. Beschleuniger: Orsay
med. Beschleuniger: Siemens, prev. maintenance
med. Beschleuniger: IBA, Field Replaceable Unit
Reliability unter besonderen Umständen
Reliability unter besonderen Umständen
uptime: 94%
Zusammenfassung
 Erfahrungsberichte aus der Praxis – nicht geschönt
 große Unterschiede zwischen den Beschleunigeranwendungen:
Lichtquellen Kernphysik/Spallation Med. Beschleuniger
Betriebszeit
~ 4500 h
~ 6000 h
stark variierend
Uptime
> 98 %
> 90 %
> 95 %
 die Probleme sind jedoch bei allen ähnlich
 intensiver Austausch, sowohl in Podiumsdiskussionen
– Lüdeke - zentrale Datenbank für Reliability, [email protected]
– Spencer - http://slac.stanford.edu/pubs/slactns/tn04/slac-tn-09-001.pdf

als auch mit den Teilnehmern
viele Ideen für zu Hause mitgebracht
History
 1977: start of cyclotron operation for nuclear physics (VICKSI)
 1995 – 2006: Ionenstrahllabor ISL – laboratory for ion beam
applications
– internal and external (~ 70%) users
– ion energy: eV < Eion< 800 MeV
– research areas:
• materials modification and
ion-solid-interaction
• materials analysis
• medical applications
 since 2007: accelerator operation
for therapy purposes only
Accelerator Layout
2 x 14.5 GHz ECR
sources on 150 kV
platforms
RFQ
5.5 MV
Van-de-Graaff
16 dedicated
target stations
k = 132
Accelerator Performance
cyclotron in operation since 1977
averaged downtime before 1995: 10 %
15
Downtime (%)
start of therapy
10
start of RFQ operation for users
5
0
95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10
Reduction of Downtime
step by step process
addressing all subsystems:
sources, injectors, beam lines, cyclotron, control system
 preventive maintenance
 increased redundancy
 modernisation
 improved diagnosis
 reduction of elements
Preventive Maintenance







regular belt change of Van-de-Graaff
service of rotating parts
cyclic change of spare power supplies used on HV terminal
drying of SF6 gas
cleaning of isolators
service on vacuum pumps: oil, bearings
replacement of water tubes
Modernisation
 new computers for control system
 replacement of old dipole power supplies
 exchange of shunt against transducer regulation in quadrupole


power supplies (gain in stability: factor 10)
discrete rectifiers replaced by complete 3-phase modules
replacement of main coil power supply of cyclotron
side effect: less energy consumption
Redundancy / Reduction of elements
 smaller variety of pumps, vacuum gauges, power supplies….
 whenever possible: spare parts for quick exchange
 low intensity proton beams: no pre-bunching
 no water cooling of deflector plates in beam line dipoles
Improved Diagnostics
 display of accelerator status
Improved Diagnostics
 display of accelerator status
 24 h charts
start of main magnet
overshoot procedure
Improved Diagnostics
 display of accelerator status
 24 h charts
 beam stability programme
ISL  Protons for Therapy (PT)
 11/04: decision to close ISL at the end of 2006
– Post-Docs and technical staff on temporary
positions left
– people were transferred
to other departments
– stop of investments
 9/06: start of planning
operation solely for PT
– reduced man-power
(less beam-time)
– reduction of beam lines,
cables…
– this step:
almost completed
 nevertheless:
maintain reliability
Accelerator Performance
small number of beam time hours:
major events have huge impact on statistics
Downtime (%)
15
10
~ 4500 hours/year
~ 1750 hours/year
5
0
95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10
ISL  PT: Operation Comparison
 changing ion species and
energies
 ~ 15 target stations varying
requirements on focusing
 H, 68 MeV
– cyclotron fixed frequency
– one NMR-probe/dipole
 2 target stations,
identical focusing
– 1/4 of existing beam line system
 34 weeks/year
 3 shifts a day (24/24)
 12 therapy weeks/year
 2 shift operation (6:00 -22:00)
 Thursday: start up and tuning
 Friday: quality control of accelerator
 weekend: standby*
 Monday-Friday: Therapy
* exceptions on weekends:
- experiments
- infants, requiring more than 4 sessions
PT: Reliability
 availability 95 % in 2007
2/3 of downtime due to one
major event: electrostatic injection
– preventive maintenance
– replacement of Ta shields by Ti
(good experience in ECR source)
– after one week: failure
– fault of new ceramics ?
– Ti shields (now Ta again)
– delay of 2 days
 uptime 2008: 98 %
worst case: failure in electricity
supply at 6:00 am
– delay of 2 hours
Accelerator Operation: Reliability
 uptime 2009: 95 %
1/3 of downtime again due to one
major event: water leak in RF
– interruption of therapy week for the
first time since 1998
(110 therapy weeks)
 availability 2010: 95 %
frequent drops in RF
error difficult to find: isolator
problems on tube socket of anode
power supply
Lessons Learned
 turbo pumps on 60 % of rotational speed (standby mode)


increases service intervals about factor 5
analysis of residual gas for water
logging of electricity for failure analysis
cryo pumps on
cryo pumps off
Wish List
 Uninteruptable Power supplies:
– overall solution: too expensive in investment and man power
– thus only for computers of control system
 counter on frequently moved Faraday cups
240
230
220
Voltage [V]
210
200
V1
V2
V3
190
180
170
160
150
140
14:47:12.00
14:47:12.50
14:47:13.00
time
14:47:13.50
ISL  PT: Installation of a Tandetron




further shortening of beam lines
less rooms
reduction of radiation safety
easy and reliable operation:
– no moving parts
– source on “ground potential”
 installation:
Apr. 07:
Oct. 07:
Sep. 08:
Oct. 08:
Mar. 09:
Aug. 10:
Dec. 10:
Jan. 11:
purchased from BAM, start of dismantling
transfer to HMI, installation starts
first beam from source
first beam through tandetron
first beam through cyclotron
acceptance test finished, applied for licence
licence granted
first therapy with tandetron as injector
ISL  PT: Installation of a Tandetron
 start of tandetron beam tests: perfect short term stability
measured on FC behind cyclotron
5
30
4
3
20
2
1
15
0
14:35
14:36
14:37
14:38
14:39
14:40
I FCZ1 (nA)
Std. Dev. (%)
25
ISL  PT: Installation of a Tandetron
 start of tandetron beam tests: perfect short term stability
but long term stability unsatisfactory
17.0
16.5
I FC (nA)
16.0
15.5
15.0
14.5
14.0
14:45
14:50
14:55
15:00
15:05
15:10
15:15
ISL  PT: Installation of a Tandetron
 start of tandetron beam tests: long term stability unsatisfactory
 now: short and long term stability better than 5 %
2.10
Icup (µA)
2.05
2.00
1.95
21:00
01:00
05:00
Time (start 7. Apr. 2011)
stability_QS1!plot110407 JR 08.04.2011
Conclusion
 12 therapy weeks per year
 past years: uptime at least 95 %
225
200
Patients per Year
175
150
125
100
 1661
 1437
 1014
 829
 677
 536
 437
 317
 166
 234
0
 105
25
 30
50
 1227
75
98 99 00 01 02 03 04 05 06 07 08 09 10 11
Conclusion
 12 therapy weeks per year
 past years: uptime at least 95 %
 but: the finest hardware is useless without dedicated personnel
→ sincere thanks to all the people involved
Thank you for your attention!

ISL PT: Installation of a Tandetron - Helmholtz