sns operational experience

SNS Operational Experience

Project X Workshop April 12, 2011

Kevin Jones – on behalf of the SNS team

2 Managed by UT-Battelle for the U.S. Department of Energy SNS Operational Experience – Project X Workshop

SNS Accelerator Complex

The Front-End produces a 60 Hz, 1 ms long chopped H- beam

Accumulator Ring: Compress 1 msec long pulse

to 700 nsec

2.5 MeV

Superconducting LINAC Front-End

RTBT

HEBT

Injection Extraction

RF

Collimators

945 ns

1 ms macropulse

Cur

rent! mini-pulse

Chopper system makes gaps

Current!

1ms

Liquid Hg Target

~1000 MeV

1 ms macropulse 1 ms

<1 µsec

186 MeV

Warm Linac

•  The SNS accelerator is the highest power pulsed hadron linear accelerator

•  Uses superconducting RF for acceleration

•  Storage Ring to compress the 1 ms linac beam into a 1 µs “short-pulse” on the neutron production target


Beam Power History

•  Running at ~ 1 MW for ~ 1.5 years •  Present operational power is dictated by budget allowance

–  Not limited by equipment or beam loss!

Equipment / Availability Limit Operational budget

limited

Bea

m P

ower

(kW

)

1000 1 MW beam power on target achieved in routine operation after 3 years

Ion Source, LEBT

Stripper foil

Target CMS leak

HVCM


SNS has demonstrated reliable operation at ~1 MW


Since 2006 operational performance improvement at SNS has been dramatic


Our ramp-up goals were adjusted to meet user expectations and match budgets

•  Beam power: kept up initially, but leveled off at ~ 1 MW after fall 2009

!"

#!!"

$!!"

%!!"

&!!"

'!!!"

'#!!"

'$!!"

'%!!"()*+!%"

,)-+!%"

./0+!1"

234+!1"

()*+!1"

,)-+!1"

./0+!&"

234+!&"

()*+!&"

,)-+!&"

./0+!5"

234+!5"

()*+!5"

,)-+!5"

./0+'!"

234+'!"

()*+'!"

!"#$%&"#'%!()

"*%+$

,-%

6478/9":;/9"

<-=3/9">)0?;0@/4-)"

!"#!!#$"#$!#%"#%!#&"#&!#'"#'!#

(""#

)*+,"$#

-*.,"$#

/01,"%#

234,"%#

)*+,"%#

-*.,"%#

/01,"&#

234,"&#

)*+,"&#

-*.,"&#

/01,"'#

234,"'#

)*+,"'#

-*.,"'#

/01,("#

234,("#

)*+,("#

!"#$

%&'#()#*()(+,%-.

/%

546708#9:08#

;.<308#=*1>:1?04.*#

(Peak power attained)

•  Availability is a more difficult goal, and stronger driver for operational parameters

•  We could run at higher powers, but the availability may suffer


Initial and ongoing operation revealed system weakness that have been substantially addressed

FY07-FY10 Downtime by group

0

50

100

150

200

250

300

350

400

450E-

HVCM

Ion

Sour

ce RFTa

rget

E-M

agPS

E-ch

oppe

rCo

ntro

lsVa

cuum

Co

olin

g E-

othe

rCr

yo APPr

ot. S

ys.

Fac.

/Mec

h. S

ys. BI

Ops

Mis

c./M

ag/R

S/ES

HNe

ut. I

nst.

System

Hou

rs o

f do

wnt

ime

FY07 FY08 FY09 FY10

System FY07 FY08 FY09 FY10E-HVCM 379.4 421.2 309 240.0Ion Source 394.7 142.2 101.8 113.5RF 162.3 230.2 226.4 126.3Target 140.1 158.9 12 140.5E-MagPS 78.6 162.2 68.1 88.5E-chopper 241.8 50.3 58.3 30.4Controls 161.3 58 90.4 24.1Vacuum 90.2 124 33.7 11.4Cooling 165.2 31.2 27.5 24.3E-other 85.9 45.6 40 36.1Cryo 15.2 4.7 38.9 21.1AP 19 27.3 20.2 7Prot. Sys. 19 9.4 26.4 13.6Fac./Mech. Sys. 5.7 3.1 31.7 2.9BI 15.2 8.4 16.2 0.6Ops 8.8 13.4 6.6 5.1Misc./Mag/RS/ESH 7.1 2 3.8 1.4Neut. Inst. 0.2 2 0 0Total 1989.7 1494.1 1111 886.8


SNS reliability compares favorably with other moderate to high-power facilities

•  Facilities with the fewest long outages have the highest availability

Data compiled at the 2008 ICFA High Brightness workshop, Nashville TN

100

10

1

0.1

< 1 minute

Trip Frequency per day

>1 minute, < 1 hour

>1 hour, < 3 hours

> 3 hours


Linac Activation History

•  Superconducting Linac activation is not increasing, despite significant increase in power and operational hours

•  Beam loss is not a limiting factor (at least for 1 MW beam)


How Much Beam is Lost in the SNS SCL ???

•  We did not know what to expect –  models indicated no loss, but…

•  Activation measurements indicate < 1 W/m in the warm sections between our cryo-modules –  < 10-4 of the beam throughout the superconducting linac

•  Measurements in the 10-5 fractional beam level are difficult –  Loss monitors are quite sensitive, but do not tell you much about why you lost

beam

•  Laser profile device turns out to be a good way to create controlled beam spills of 10-6 beam –  Increases the integrated beam loss about 10% (or we are nominally losing 10-5

throughout the linac)


!"

!#$"

%"

%#$"

$"

$#$"

&"

&#$"

'"

()*+,-./0123""

()*+,-./0124""

()*+,-./012!""

()*+,-./012%""

()*+,-./012$""

()*+,-./012&""

()*+,-./012'""

()*+,-./0125""

()*+,-./0126""

()*+,-./0132""

()*+,-./0133""

()*+,-./0134""

()*+,-./013!""

()*+,-./013%""

()*+,-./013$""

()*+,-./013&""

()*+,-./013'""

()*+,-./0135""

()*+,-./0136""

()*+,-./0142""

()*+,-./0143""

()*+,-./0144""

()*+,-./014!""

()*+,-./014%""

()*+,-./014$""

()*+,-./014&""

()*+,-./014'""

()*+,-./0145""

()*+,-./0146""

()*+,-./01!2""

()*+,-./01!3""

()*+,-./01!4""

()*+,-./01!!""

!"#$%&'()*

+&

,-.&/01%2&!"#$%2&789:.;"

<=>?@AB>;"

(C>>DE"<=>?"

SNS Linac Transverse Lattice: Design vs. Operation

!"

#$"

#%"

#&"

#'"

#!"

%$"

%%"

%&"

%'"

%!"

(()*+,-./0$$""

(()*+,-./0#$%""

(()*+,-./10#$&""

(()*+,-./10#$'""

(()*+,-./10#$!""

(()*+,-./10##$""

(()*+,-./10##%""

(()*+,-./10%$%""

(()*+,-./10%$&""

(()*+,-./10%$'""

(()*+,-./10%$!""

(()*+,-./10%#$""

(()*+,-./10%#%""

(()*+,-./102$%""

(()*+,-./102$&""

(()*+,-./102$'""

(()*+,-./102$!""

(()*+,-./102#$""

(()*+,-./102#%""

(()*+,-./10&$%""

(()*+,-./10&$&""

(()*+,-./10&$'""

(()*+,-./0&$!""

(()*+,-./0&#$""

!"#$%&'()*

+&

,,-&./0%1&!"#$%1&

3456-7" 89:;<=>:7"

•  SCL quads run much lower than design •  Warm linac is run close to design

•  SCL beam loss is significantly lower for the reduced field settings!!

•  Empirically derived

CCL quad fields

!"

#!"

$!"

%!"

&!"

'!!"

'#!"

'$!"

'%!"

'&!"

#!!"

()*+,-./01*2!!"

()*+,-./01*2!#3"

()*+,-./01*2!43"

()*+,-./01*2!$3"

()*+,-./01*2!53"

()*+,-./01*2!%3"

()*+,-./01*2!63"

()*+,-./01*2!&3"

()*+,-./01*2!73"

()*+,-./01*2'!3"

()*+,-./01*2''3"

()*+,-./01*2'#3"

()*+,-./01*2'43"

()*+,-./01*2'$3"

()*+,-./01*2'53"

()*+,-./01*2'%3"

()*+,-./01*2'63"

()*+,-./01*2'&3"

()*+,-./01*2'73"

()*+,-./01*2#!3"

()*+,-./01*2#'3"

()*+,-./01*2##3"

()*+,-./01*2#43"

()*+,-./01*2#$3"

()*+,-./01*2#53"

()*+,-./01*2#%3"

()*+,-./01*2#63"

()*+,-./01*2#&3"

()*+,-./01*2#73"

()*+,-./01*24!3"

()*+,-./01*24'3"

()*+,-./01*24#3"

!"#$%&

'(")&*+,,&-.&/01(23+.#034.5&*-."3&

,89-/:";<=>?@A=:"


Intra-Beam-Scattering Beam Loss V. Lebedev, FNAL

•  Simple estimates indicate this could be a loss contributor at SNS

•  Only an issue for H- beams –  Considered a proton source experiment to

test this loss mechanism –  Now planning on a low energy foil (strip H-

to p) experiment

Collisions between H- in the accelerated bunch can

strip the outer electron

Stripping probability is known :


Linac Beam Loss Situation •  SNS has unexpected beam loss in the SCL

–  OK for 1 MW, not acceptable for 10 MW –  There is a suite of measurement tools available at SNS –  Challenge is to measure the 6-D initial beam distributions down to halo

levels –  And understand measured beam loss

•  We should use the existing machines to understand the nature of this loss

Low loss tune

Before Matching

After Matching 40 deg tail

Longitudinal BSM, S. Aleksandrov


SNS Superconducting Linac

•  160 m, 23 cryo-modules, 81 cavities • Operating at 1 MW, 925 MeV, 60 Hz, 5% beam duty cycle


SNS RF Layout

•  One cavity / klystron: easy, flexible, expensive •  High voltage drive and transmitters are common-mode failure points though

–  Alternatives include hot spare, single power supply / RF source

Beam


Superconducting Cavity Amplitudes Cavity Design

051015202530

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

cavity

E0 (M

V/m

)

First Run

05101520253035

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

cavity

E0 (M

V/m

)

Ring Commissioning Run

05101520253035

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

cavity

E0 (M

V/m

)

0

5

10

15

20

25

30

35

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81

E0 (M

V/m

)

Cavity

Sept. 2008

0

5

10

15

20

25

30

35

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81

E0 (M

V/m

)

Cavity

Sept. 2010

• SCL cavity gradient levels were not what we expected –  We grossly underestimated the gradient variability –  But the SCL is operationally quite flexible !!

•  Make sure there is enough margin in the cavity design gradient


RF System: Independently Powered Cavities

• One klystron per cavity: conservative but robust

81 x 550 kW klystrons


Superconducting Cavity Fault Recovery

•  A cavity fault recovery scheme is developed to adjust downstream cavity setup, to accommodate upstream cavity changes –  Uses a difference technique, with initial beam based measurements –  Successfully demonstrated and used at SNS

•  Could work in < 1 sec if needed

0

200

400

600

800

1000

1200

1400

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81

Cavity

Phas

e C

hang

e(de

g)

Final cavity phase found within 1 degree, output energy within 1 MeV

Cavity 3a turned off

-300

-200

-100

0

100

200

300

400

1 7 13 19 25 31 37 43 49 55 61 67 73 79Cavity

Phas

e C

hang

e (d

eg)

-6

-4

-2

0

2

4

6

Mea

sure

d Er

ror (

deg)

Phase ChangeMeasured Error

Turned on cavity 4a, reduced fields in 11 downstream cavities


High Voltage Power Supply System (HVCM)

•  HVCM used new technology (IGBT) –  Do not assume success with new technology

•  Early on HVCM was a major down time problem –  More robust components have greatly helped –  For extreme reliability applications, need to consider hot-spare,

independent power supply/klystron, etc.

Initial experience – some “fireworks”


Ion Source reliability remains a concern

• At SNS the ion source is rising to the top of the reliability concern list –  Long-term plan is to incorporate a dual source with a

magnetic LEBT for redundancy


Ring experience has been very positive Foil bracket issues

•  World record intensity for protons accumulated in a Ring

Glowing foil at 1 MW

We have not been limited by: - Beam instabilities - Space charge induced beam loss

“Convoy” electron direct impact


Ring Injection: More Difficult than Originally Envisioned

H- beam from Linac Thin

Stripping Foil

To InjectionDump

ThickSecondary Foil

pH0

H-Dipole magnets

H- beam from Linac Thin

Stripping Foil

To InjectionDump

ThickSecondary Foil

pH0

H-Dipole magnets

•  Need to handle clean transport of injected beam, circulating beam, un-stripped H- beam and partially stripped H0 beam –  Not much space –  Careful treatment of beam transport through 3-D fields –  Fair amount of re-work in this area at SNS –  Evaluating laser stripping as a future option


Clean Extraction from the Ring: No Problem • We have only used second stage chopping for the past

~ one year •  1st chopper stage is slow rise time (~100 nsec) LEBT

chopper • We never implemented a planned “Beam-in-Gap” kicker

to clean the gap • We are running a smaller gap than initially planned (up

to 75% beam vs. 68% beam)

time

Cur

rent


Ring Activation History

Pow

er (M

W)

1.0

•  Activation by the injection stripper foil is the highest in the SNS accelerator

•  Close to activation expectations •  ~ Monotonic increase with beam power


Targets, Dumps, Collimators: More trouble than we imagined

•  High power operation requires good understanding and control of primary and waste beams

•  Redundant safety systems – avoid excessive nuisance trips

Direct measurements (beam position, power density, …) are easier than….

…model based extrapolations from upstream measurements

•  Fast beam shut-off systems: –  SNS errant beam to turn-off delay is ~ 20 µs –  Can not buy these systems: custom hardware / software

< 20 µs

Zoom-in on an errant beam wave-form


Summary • SNS is running a ~ MW proton superconducting linac

–  > 5000 hrs/year operation –  Beam loss is not a limitation

•  Although we do see unexpected small loss levels –  Reliability approaching 90%

•  Improving •  Still have many more trips than requirements for other applications

– Can use as a test bed for recovery concepts

• New technologies require shake-out periods • A robust, intelligent control system is essential to success • Customer requirements must be fully understood

sns operational experience

Documents