Mike Seidel PSI Reliability Workshop CERN June 22 1 Outline Suited Circular Accelerator Concepts requirements for ADS accelerators cyclotrons rapid cycling synchrotrons FFAG Generics on failure probability trip rates ID: 778139
Download The PPT/PDF document "Reliability challenges for circular ADS ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Reliability challenges for circular ADS drivers
Mike Seidel, PSI
Reliability Workshop CERN, June 22
1
Slide2OutlineSuited Circular Accelerator Conceptsrequirements for ADS accelerators
cyclotrons, rapid cycling synchrotrons, FFAGGenerics on failure probability / trip ratesredundancy in accelerators; AND vs OR fault logics; failure and survival probability measured trip statistics at PSIC
omments on FFAG, J-PARC RCS, PSI cyclotronNew cyclotron ideasH2+ Daedalus, stacked/flux coupled cyclotrons, reverse bend cyc.DiscussionPro’s and Con’s of Circular concepts
Reliability Workshop CERN, June 22M.Seidel, PSI
Slide3Requirements for ADS Acceleratorsenergy: flat optimum
1.2GeV; however 0.8 .. 2.5GeV under discussion power: 2...10MW; Ptherm = P
beam G/(1-k)low losses: 1W/m; PSI: 100W at critical locationreliability & stability: 0.01…0.1 trips per day(!)
efficiency: as best as possible, =Pbeam/Pgrid= 20…30%
cost: as low as possible; optimize for series production; modern nuclear power plant: (5B€)
Reliability Workshop CERN, June 22
M.Seidel, PSI
Slide4High
Intensity
Accelerator Landscape
2.4mA = 1.416MW
PSI
SNS, Sep 2013
(
today
target
limits
intensity
)
Reliability Workshop CERN, June 22
M.Seidel
, PSI
J-PARC RCS
ramping
up
!
1MW
routine
operation
expected
2016
Slide5reliability, trip performancetodays trip performance of accelerators is orders of magnitude worse than desired for ADS, e.g 10…100d
-1 achieved vs. 0.01…0.1d-1 desired by reactor expertsin recent years discussions took place and requirements were somewhat relaxed:
fatigue failure of fuel elements is not seen extremely critical anymoreshort trips (few seconds) can be accepted; shorter than thermal time constantsaccelerator and reactor developers must find compromises, my impression: reactor community in general not very flexible due to strict safety rules; e.g. studying fast reactor startup, bridging trips
IBR-2 pulsed reactor
demonstrates fast cycling
even without technical reasons,
for industrial power production (and consumption) reliability is very
important
!
Reliability Workshop CERN, June 22
M.Seidel, PSI
Slide6SuitedAccelerator Concepts?
Linear
Accelerators
Cyclic Accelerators
normalconducting
linac
superconducting
linac
e
lectrostatic
(E limited!)
p
ulsed
CW:
very
low
gradient
CW
possible
c
ooling
power!
cost
!
Synchrotron
[rapid
cycling
]
Cyclotron
Synchro-Cyclotron
(
cycling
)
Other: Betatron,
Microtron
FFAG
CW
possible
beam
dynamics
!
extraction
!
pulsed
p
ower limited (1MW?)
compact
cycling
/ CW
questionable
no
demonstrator
exotic(laser, plasma, diel. efficiency!)
Reliability Workshop CERN, June 22
M.Seidel, PSI
Slide7classification of circular acceleratorsbending radius
bending field vs.
time
bending field vs. radius
RF frequency vs. time
operation mode
(pulsed/CW)
comment
betatron
induction
microtron
varying
h
classical cyclotron
simple,
but limited
E
k
isochronous cyclotron
suited for high power!
synchro
- cyclotron
higher
E
k
, but low P
FFAG
strong focusing!
a.g.
synchrotron
high
E
k
Reliability Workshop CERN, June 22
M.Seidel, PSI
Slide8reliabilitycalculation in a nutshell
Reliability Workshop CERN, June 22M.Seidel, PSI
example: lightbulbs (MTBF=1000h)N(t) number
of surviving lightbulbsN0 original number of
funct. bulbs
(t)
fractional
failures
per time
constant
, i.e.
no
aging
/
history
(
Markov
process
)
simple, but
for
some
applications
too
simple!
mathematical
treatment
:
S
(>t)
survival probabilityF(>t) failure prob. (1-S(t))f(t) failure
density function =F‘(t)(t) hazard function
beyond
=const: Weibull
distribution
Slide9durationstatistics for un-interrupted run
periodsReliability Workshop CERN, June 22
M.Seidel, PSI
PSI run data analyzed:integrated
distribution of
un-interrupted
beam
times
double log
scale
How many runs per day with duration longer than t.
total trips per day
Slide10modelling with exponential and Weibull distributionsReliability Workshop CERN, June 22
M.Seidel, PSI
Slide11faulttopology of an accelerator
Reliability Workshop CERN, June 22M.Seidel, PSI
&
1
&
2
system
A
system
A´
system
B
system
B‘
system
C
system
C‘
system
C‘‘
system
C‘‘‘
classical
simple
redundancy
multiple
redundancy
beam on
clearly redundancy is desirable
whether redundancy is possible with major subsystems depends on the choice of accelerator concept
Slide12reliability, conceptsnumerical example:tube: MTBF=5000h; MTTR=8h
Linac with 80 tubes, accepting 0 fault: MTBF
eff = 62hLinac with 80 tubes, accepting 1(k=2) fault: MTBFeff = 1.074h
Linac with 80 tubes, accepting 2 faults: MTBFeff = 26.067h
cyclotron with 4 tubes, accepting
0 faults:
MTBF
eff
=
1.250h
binomial distribution,
B
p
= incomplete Beta Function
Reliability Workshop CERN, June 22
M.Seidel, PSI
circular
accelerator
linear
accelerator
only
few
resonators
/
magnets
redundancy
difficult
redundancy
wrt
.
resonators possible
Slide13distribution of trip durationD. Vandeplassche, Proc. IPAC
012
PSI analysis of trip-periods
Reliability Workshop CERN, June 22
double
logarithmic
:
power
law
is
straight
line
.
PSI
PSI
total trips per
day
Slide14next: circular concepts compared…
Reliability Workshop CERN, June 22
M.Seidel, PSIcomplexity
= less availability (?)
isochronous cyclotron: continuous, nothing cycles
FFAG: pulsed, RF cycling
synchrotron: pulsed, magnets + RF cycling
Slide15[investigation towards CW operation:
S.Machida
, FFAG, EMMA, serpentine acceleration]Fixed Field Alternating
Gadient Accelerator ?Reliability Workshop CERN, June 22
M.Seidel, PSI
strong
focusing
, large E
acceptance
,
magnets
fixed
in time
RF must
be
cycled
faster
than synchrotron, but less
intensity than CW
cyclotron
not
ramping
the
magnets
should
be
an
advantage
for
reliability
otherwise same arguments as for RCS hold
cyclic injection/extraction potentially
difficulthigh intensity: not demonstrated
[EMMA]
Slide16Reliability Workshop CERN, June 22
M.Seidel, PSI
[M.Shirakata, J-PARC]f
RF: 0.94MHz – 1.67MHz
Slide17Outline of the J-PARC RCS
Circumference
348.333 m
Superperiodicity
3
Harmonic number
2
Number of bunches
2
Injection
Multi-turn,
Charge-exchange
Injection energy
181 MeV
Injection period
0.5 ms (307 turns)
Injection peak current
30 mA
Extraction energy
3 GeV
Repetition rate
25 Hz
Particles per pulse
5 x 10
13
Output beam power
600 kW
Transition gamma
9.14 GeV
Number of dipoles
24
quadrupoles
60 (7 families)
sextupoles
18 (3 families)
steerings
52
RF cavities
12
Now the RCS is in the final beam commissioning phase
aiming for the design output beam power of 1 MW.
Recently
t
he hardware improvement of the injector linac
has been completed.
⇒
400 MeV in 2013
⇒
8.3 x 10
13
⇒
1 MW
400 MeV H-
3GeV
proton
MLF
: Material and Life Science
E
xperimental
F
acility
MR
: 50-GeV Main
R
ing Synchrotron
⇒
50 mA
in 2014
[
H.Hotchi
, IPAC2015]
Slide18Earthquake
300 kW
Hg-target
replacement
Incident at
Hadron Facility
532 kW
300 kW
as of 3
rd
of June 2015
〜
560 kW
~
10
months interruption
due to the earthquake
593 kW
~
1
month interruption
d
ue to
the fire in MLF
Beam Power History at MLF
Interruption due a trouble of Hg-target
500
kW
400
kW
[
T.Koseki
]
Slide19J-PARC RapidCycling Synchrotron [RCS]
1.01MW avg beampower achieved (Hotchi
, IPAC15); still high losses, not routinehigh average availability: 90% (!)drives neutron source (mercury target); i.e. ADS like application thus J-PARC RCS has demonstrated a respectable high intensity performance!
Reliability Workshop CERN, June 22M.Seidel, PSI
one would think the rapid cycling of the magnets would cause specific problems, more than in an CW accelerator; but nothing outstanding was reported and 90% was achieved, on par with SNS and PSI.
typical failures from the past (taken from
M.Shirakata
presentation
)
LINAC HV DC supply failure (2012: 59h, 2013:103h)
bending magnet supply (resonant circuit, 2014: 115h, oil pump)
radiation accident caused 23 months interruption (Au target got full charge in 5ms instead 2s)
not specific to RCS
Japan’s great earthquake: recovered in 286 days!!
not specific to RCS
possibly related
to cycling
Slide20PSI Ring Cyclotron
8 Sector Magnets:
1 T
Magnet weight:
~280 tons
4 Accelerator Cavities:
860 kV (1.2 MV)
1 Flat-Top Resonator
150 MHz
Accelerator frequency:
50.63 MHz
harmonic number:
6
kinetic beam energy:
72
590 MeV
beam current max.:
2.4 mA
extraction orbit radius:
4.5 m
outer diameter:
15 m
RF efficiency Grid/Beam
0.90
0.640.55 = 32%
rel. losses @ 2.2mA:
~1..2
10
-4
transmitted power:
0.32 MW/Res.
Reliability
Workshop CERN, June 22
Slide21TypicalTrip Causes for Cyclotrons
Reliability Workshop CERN, June 22
electrostatic elements: high voltage breakdowns due to plasma discharges; stray electrons collected on insulators; presence of intense proton beam in vicinity of electrodes and insulators; comparably poor vacuum of 10-6mbarloss tuning: depends critically on complete accelerator chain incl. source; low tail density very sensitive to all parameters
cavities & RF systems: breakdown, multipacting, MW level amplifier chains (eg. 50MHz)
see presentation by
J.Grillenberger
on concrete examples
Slide22critical: injection/extraction schemesdeflecting element should affect just one turn, not neighboured
turn critical, cause of lossesoften used: electrostatic deflectors with thin electrodesalternative: charge exchange, stripping foil; accelerate H
- or H2+ to extract protons (problem: significant probability for unwanted loss of electron; Lorentz dissociation: B-field low, scattering: vacuum 10
-8mbar)
0
-
HV
foil
extraction electrode
placed between turns
extraction by charge exchange in
foil
eg
.: H
-
H
+
H
2
+
2H
+
binding
energies
H
-
H
2
+
0.75eV
15eV
Reliability Workshop CERN, June 22
M.Seidel, PSI
Slide23extraction profile measured at PSI Ring Cyclotrondynamic range:
f
actor
2.000 in
particle density
red
:
tracking
simulation
[OPAL]
black
:
measurement
position of extraction septum
d=50µm
turn
numbers
from
simulation
[
Y.Bi
et al]
Reliability Workshop CERN, June 22
M.Seidel, PSI
Slide24proposed cyclotrons I: H2+ Daedalus cyclotron [neutrino source]
[
L.Calabretta, A.Calanna et al]Reliability Workshop CERN, June 22M.Seidel, PSI
purpose: pulsed high power beam for neutrino production
800MeV kin. energy5MW avg. beam power
note:
complex extraction path
binding
energies
H
-
H
2
+
0.75eV
15eV
Slide25proposed
Cycl
. II: H
2
+
AIMA Cyclotron w reverse bend and multiple 60keV injection [
P.Mandrillon
]
The reverse valley B-field concept avoids the internal loop
(cf. DAEdALUS extraction) for the stripped proton beam from H2+
.
Vacc
=150kV
Vacc
=165kV
Reliability Workshop CERN, June 22
25
Slide26Texas A&M University
Two
Stages Cyclotron: 100 MeV SF
injector
+ 800 MeV SF booster.
Stack
of 3 Cyclotrons in //
Booster: 12 Flux
coupled
stack
of
dipole
magnet
sectors
10
Superconducting
100 MHz RF
cavities
providing
a 20 MeV
Energy
Gain/
turn
multiple power
couplers
per
cavity
Large
turn
separation
allowing
to insert SF
beam
transport
channels
made of Panofsky
Qpoles
(G=6T/m)
proposed
cycl
. III: TEXAS
A&M: 800 MeV SUPERCONDUCTING
STRONG-FOCUSING CYCLOTRON
[
P.McIntyre
, Texas A&M]
Reliability Workshop CERN, June 22
26
recently
: DOE
awards
stewardship funding for idea of strong focusing channels in cyclotrons
Slide27Summary –Reliability of Circular Accelerators
Reliability Workshop CERN, June 22
M.Seidel, PSI
reliability
of high intensity circular accelerators today is
around 90%
(PSI, J-PARC), on par with the
s.c.
linac
and accumulator of SNS;
trip rates are 10…100d
-1
for PSI
, at least three orders of magnitudes worse than desired for ADS
implementation of
redundancy difficult
in circular accelerators; on the upside circular acceleration is an economic concept;
injection/extraction
elements
in cyclotrons are critical devices; their reliability could be improved by certain measures when considering
ADS
cyclotrons with CW
operation should have best stability; next is FFAG with ramping of RF; rapid cycling synchrotron needs magnet ramping; nevertheless high reliability demonstrated by J-PARC
personal remark on choice of technology
:
today a
s.c.
linac
is a straightforward solution for a multi-MW facility;
however, despite of it‘s greater beam dynamics complexity and intrinsically higher susceptibility for trips, an
optimized cyclotron
, built in series can be very cost effective and could reach high availability as well.
Slide28t
hank
you for
the
attention
!
Reliability Workshop CERN, June 22
M.Seidel, PSI