Post on 30-Jul-2020
transcript
ORNL is managed by UT-Battelle
for the US Department of Energy
1.4 MW Power
Ramp-up and
Reliable Operation
J. Galambos
SNS AAC Review
March 24-25, 2015
2 1.4 MW
SNS: Designed for 1.4 MW
…
The SNS Project Parameter List
3 1.4 MW
Early Power Ramp-up Expectations
• Mason / Holtkamp White Paper, 2002
1.4 MW: mid 2009
4 1.4 MW
1.4 MW Beam Power at SNS: The Reality
• First neutron production at 1.4 MW: June 27, 2014
– Sustained power increase during the 2014 run, increasing power from ~ 1. MW to 1.4 MW
1 hour demo, Sept. 2013
1 day operation June 2014
Unable to run 1.4 MW
Target Limited Beam Power
5 1.4 MW
1.4 MW Operation: The Final Assault
FY 14a Run
– Power ramp-up approach:
• Adiabatic
• Most effective use of what works
– Plan is to operate at 1.3-1.4 MW from now on
Target paranoia
850 kW
1.4Spare target deliveries
1.2
1.0
0.8
1.4
0.6
0.4
Beam
Pow
er
(MW
)
0.2
0.
Oct. 2013
June2014
Jan. 2014
6 1.4 MW
Stable 1.4 MW Operation
• 1.4 MW operation ~24 hr. scale existence proof
– Foil works
– No instability
– RF supports full beam loading
0
200
400
600
800
1000
1200
1400
1600
6/25/1
40:00
6/25/1
412:0
0
6/26/1
40:00
6/26/1
412:0
0
6/27/1
40:00
6/27/1
412:0
0
Beam
Power(kW)
24 hrs
1.38 MW1.41 MW
Water pump outage
7 1.4 MW
Macro-pulse
Structure
(made by the
High power RF
– 60 Hz)
Mini-pulse
Structure
(made by the
choppers ,
1 MHz)
645 ns 300 ns
945 ns
1ms
16.7ms
~1000 mini-pulses per macro-pulse
Some Background: Accelerator Time
Structure
chopped fraction
unchoppedfraction
Chopped gap is to provide an extraction gap in the ring
8 1.4 MW
Proton Beam Power
• Not many knobs:
– Beam energy (Ebeam)
– Rep-rate
– Beam current (Imacro-pulse)
– Macro-pulse length (tmacro-pulse)
– Chopping fraction (fun-chopped)
Pbeam = Ebeam ´Qpulse ´ rep-rate
Qpulse =tmacro-pulse ´ Imacro-pulse ´ fun-chopped
, where
9 1.4 MW
Beam Energy History
• Never run 1 GeV production
• Beam energy constant for last 1.5 years
• Plan to increase to 1 GeV in 1-2 years: plasma processing
800
820
840
860
880
900
920
940
960
980
1000
9/5/05
10/10/0611/14/0712/18/081/22/10
2/26/11
4/1/12
5/6/13
6/10/14
Beam
Energy(M
eV)
Fix cryo-components
Errant beam degradation
Errant beam control, spare cryo-module
10 1.4 MW
Duty Factor History
• Duty factor (=rep-rate x pulse length) is a driver on equipment stress
– Slow increase in pulse length since initial operations
– Final push in 2014
0
10
20
30
40
50
60
70
0
200
400
600
800
1000
1200
10/10/06
6/17/07
2/22/08
10/29/08
7/6/09
3/13/10
11/18/10
7/26/11
4/1/12
12/7/12
8/14/13
4/21/14
RepRate(Hz)
Beam
PulseLength(
ms)
PulseLength
RepRate
RF at full duty factor
11 1.4 MW
Ion Source History: RFQ Output
(Courtesy M. Stockli)
• RFQ output current fairly constant last 2 years
0
5
10
15
20
25
30
35
40
45
50
6/2/06
10/15/07
2/26/09
7/11/10
11/23/11
4/6/13
8/19/14
Beam
Current(m
A)
Source 3
Source 2
Source 4
Degraded RFQ
12 1.4 MW
RFQ Transmission is Low
(Courtesy M. Stockli)
• Reduced RFQ transmission 2012-present
– 60-75%
• Input current 45-55 mA
RFQ operated at higher power
13 1.4 MW
Historical RFQ Performance
• Nominal: operate with ~ 600 kW, 90% transmission
Courtesy of Sasha Aleksandrov 690 kW
620 kW
760 kW
Tra
nsm
issio
n (
%)
Historical power scan
RFQ RF Power
575 kW
90% transmission
14 1.4 MW
Getting the most out of a wounded RFQ
(Courtesy A. Shishlo)
• Running at higher RFQ power gives more current…– But we are significantly below the expected transmission
– Running at 600+ kW is not always possible (Champion)
– New RFQ is a key step in path to reliable 1.4 MW operation
old set-
point
Set-point for 1.4
MW operation, not
always attainableRF
Q O
utp
ut
Curr
ent
(a.u
.)
Normal output should be near saturation and close to
90% transmission
Transmission typically 65-75% (Stockli)
15 1.4 MW
Keys for Reduced Chopping Fraction
(courtesy R. Saethre)
• Fast, reliable clean single stage chopping (LEBT)
• Clean up Ring extraction kicker jitter
Beam chopping circa 2009
slower
slow
After thermal isolation of trigger
electronics
Building
temperature2 days
Extraction-kicker timing drift
16 1.4 MW
Beam Current: Better Chopping
~20 % gap
< 100 ns rise time
Time (us)
Curr
ent
(AU
)
• Chopping quality for 1.4 MW operation
– Smaller gaps than previously used
– High quality LEBT chopping
Ion source waveform
1 ms
17 1.4 MW
1.4 MW: The Final Assault
• Improved chopping covered up other deficiencies
• We need to provide margin in beam energy, current and pulse length to permit reliable, steady 1.4 MW operation
1.4 MW
Design
1.08 MW
Operation
(March 2014)
1.4 MW
Operation
(June 2014)
Energy (GeV) 1.0 0.94 0.94
Rep rate (Hz) 60 60 60
Macro-pulse length (ms) 1.0 0.87 0.97
RFQ output beam current
(mA)
38 32 35
“un-chopped” fraction 0.68 0.72 0.78
18 1.4 MW
2013 AAC Recommendations
• 30. Establish an accelerator operations and development strategy based on beam tests that can be safely performed during the restrictions on operational power, followed by exploration of high power limits as soon as a sufficient spares queue is established.
• 31. Retain flexibility to respond to development outcomes by adjusting to new points in (E, I, L) space if the primary plan does not pan out.
• 32. Socialize the above strategy with the neutron user community.
– Have explored 1.4 MW operation space
– Power increases / decreases were communicated to users
19 1.4 MW
We Need Margin for Reliable 1.4 MW
Operation
• We can run 1.4 MW “when the stars are aligned”
• Areas that will provide “power margin”
– Spare RFQ needs testing with beam: 15-20% increase in transmission (Champion, Aleksandrov)
– Smart Chopping: 5-10% (Plum)
– Beam energy: plasma processing: 7% increase in beam energy (Kim/Doleans)
– HVCM (modulator) development: ~5% from pulse flattening (Anderson)
• Some additional areas for attention
– Injection area: Foil changer, foil holders, electron catcher (Plum)
– Targets (Abercrombie, Galambos)
20 1.4 MW
Smart Chopping
• Implementing chopper controller changes this summer
– 10% increase in “average un-chopped” fraction may be possible
1 turn
Relatively large un-chopped flattop most of injection
Longitudinal “tricks” to recover big gap at extraction time
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 200 400 600 800 1000 1200
ChoppingFrac
on
Time(ms)
GeneralizedRampedChoppingPa ern
Ramp-
up
Flat-top
Ramp-down
Spin-cycle
New features
21 1.4 MW
Beam Loss / Machine Activation at ~1.4 MW
• Linac beam loss
– Compare beam loss along the SCL for different powers
– Track historical activation trend
• Average residual activation along the SCL at 30 cm
• Ring beam loss
– Compare beam loss at injection (dominate loss) for different powers
– Track historical activation of injection region
• Hottest spot down-stream of foil, at 30cm
22 1.4 MW
Superconducting Linac Beam Loss History
• Big drop in losses with focusing strength reduction in early 2009
• Modest benefit since
BLMs along the SCL
23 1.4 MW
SC Linac Activation: Not Horrible
(Courtesy C. Peters)
• Fairly steady activation level since 2010
24 1.4 MW
0
10
20
30
40
50
60
70
80
90
100
Ring_Diag:BLM_A11c
Ring_Diag:BLM_A13b
Ring_Diag:BLM_B04
Ring_Diag:BLM_B07
Ring_Diag:BLM_B09c
Ring_Diag:BLM_BMov02
Ring_Diag:BLM_C01
Ring_Diag:BLM_C05
Ring_Diag:BLM_C08
Ring_Diag:BLM_C11b
Ring_Diag:BLM_D02
Ring_Diag:BLM_DMov01
Ring_Diag:BLM_D09
Ring_Diag:BLM_D11d
Ring_Diag:BLM_D13
Ring_Diag:BLM_A03
Ring_Diag:BLM_A06
Ring_Diag:BLM_A11a
RingB
eam
Loss(Rad
/C)
Dec.2008,650kW
Sept.2009,1MW
June2014,1.4MW
Ring Beam Loss History
• Ring loss not getting worse – a bit of improvement
injection area
BLMs along the Ring
collimation
extraction
25 1.4 MW
Ring Injection Activation (Peak)
(Courtesy C. Peters)
• No obvious jump in activation downstream from the foil
26 1.4 MW
Summary
• SNS was designed to operate at 1.4 MW
– 1.4 MW was achieved, albeit not as designed
– Minimal operational margin though
• Margin is possible with
– Proper RFQ behavior (new RFQ)
– Smart chopping
– Plasma processing
– Modulator flattop