Upload
bartholomew-dickerson
View
221
Download
0
Tags:
Embed Size (px)
Citation preview
NLC - The Next Linear Collider Project
James T Volk
May 2002
Report to MAC on Magnets
James T VolkFermilab
May 10, 2002
NLC - The Next Linear Collider Project
Participants
SLAC Seung Rhee, Cherrill Spencer, Jim Spencer
SLAC Magnetic Measurement GroupScott Anderson, Zack Wolf
Fermilab Magnetic Measurement
Joe DiMarco
James T VolkMay 2002
LBNLJose Alonso, Jin-Young Jung
NLC - The Next Linear Collider Project
Recommendations from Oct ‘01 MAC
• Implement proposed upgrade of stability measurement systems
• Perform estimate on radiation damage tolerances allowed by NLC spec for each permanent magnet design
• Consider reducing the number of alternative designs
• Some effort should be devoted to determining if vibrations from LCW can be maintained with in specifications
• When a PM design meets NLC specs has been identified its cost for construction, installation and operation should be compared to an EM
James T VolkMay 2002
NLC - The Next Linear Collider Project
Implement proposed upgrade of stability Measurement Systems
• Due to budget cap there is no money to upgrade the Fermilab system at this time
• SLAC has moved the rotating coil system to a small room from the high bay of the light assembly building. Improved temperature and humidity control
• Both SLAC and FNAL systems are adequate for current needs
May 2002 James T Volk
NLC - The Next Linear Collider Project
Perform estimate on radiation damage tolerances allowed by NLC spec for each PM design
• Still gathering information on expected radiation levels in Main LINAC and Damping Rings.
• Old Measurements from SLC and damping rings inconsistent due to assumptions and calculations So these data are little use to NLC.
• Have been doing measurements in SLC damping ring distinguish different types of particles
• Radiation levels do vary by orders of magnitudes and types depending on location. Need to get better models.
• Different types of radiation have different effects on Permanent Magnets.
James T VolkMay 2002
NLC - The Next Linear Collider Project
Radiation Damage Issues
• SM Cobalt more resistant than ND-Iron– SM Cobalt more expensive– Has longer lived activation products– ND-Iron dependent on manufacturing process– And brick geometry
• Still need more investigation– Both literature– Do Computer modeling (FLUKA)– Network with other groups engaged in this type of work– And Experiments– Need to test actual magnets with real beams
James T VolkMay 2002
NLC - The Next Linear Collider Project
Radiation-Induced Demagnetization(Japanese experience with 200 MeV protons)
• Material Type has large impactRed: N48 High Br (1.4T)
Low Hc (1.15 MA/m)Blue: N32Z Lower Br (1.14 T)
Higher Hc (2.5 MA/m)
• Material Shape has large impact(All samples are discs 10 mm dia)Circle: thickness = 2 mm (Pc = 0.5)Triangle: thickness = 4 mm (Pc = 1.0)Square: thickness = 7 mm (Pc = 2.0)- Higher Permeance coefficient
increases resistance (x 10)
• SmCo is much more resistant than NdFeB
May 2002 James T Volk
NLC - The Next Linear Collider Project
Have Radiation Test Dipole Design
PM Material
PM Material
May 2002 James T Volk
2 inch gap
NLC - The Next Linear Collider Project
Radiation Damage Testing
• First dipole is in FNAL shop expect to build 5-10 dipoles to test various aspects of magnets aging, different radiation fields, different manufacturers
• Have in hand 100 Hitachi ND-Iron magnets to start tests
• Have a space at FNAL LINAC to expose dipoles to 400 MeV neutrons and protons
• Need to find other sources of particles
• Looking for help from University groups– An Ideal project for a small group!
James T VolkMay 2002
NLC - The Next Linear Collider Project
Aging Data From Dexter
Prediction of N36 loss at 60 deg C over 17 years
y = 0.0961Ln(x) + 0.0619
R2 = 0.987
-0.2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 110000 120000 130000 140000 150000
Time in Hours
Flu
x L
oss
%
N36(60C) natural log prediction out to 17 yearsJames T Volk
May 2002
NLC - The Next Linear Collider Project
Consider Reducing the Number of Alternative Designs
• Corner tuner has been eliminated
• Work on the sliding shunt and counter rotating quads has stopped due to budget cap
• New rod tuning mechanism for wedge tuner has been built center stability is encouraging
• Computer models of Wedge and Ring Quad are proceeding
• There is still time to explore models before the CDR is finalized in ‘04
James T VolkMay 2002
NLC - The Next Linear Collider Project
Wedge Quad Rod Turning Mechanism
Tuning Rod
May 2002 James T Volk
Pin slides in slot turning rod
NLC - The Next Linear Collider Project
Photo of the Wedge Quad on the SLAC measuring set-up.
Wedge magnet, secured to V block Rotating coil read-out
Tuning rods rotation mechanism. One rod only connected. Smaller gear wheel turned with wrench.
Aluminum V block, secured to granite table
Cooling fan for readout stand
Heidenhain 0.5µm indicator
May 2002James T Volk
NLC - The Next Linear Collider Project
FNAL SSW Measurements of Wedge Quad
FWSQ001-6 all data
-15
-10
-5
0
5
10
15
20
20.5 21.0 21.5 22.0 22.5 23.0 23.5 24.0 24.5 25.0
YGdl Tesla
Y C
ente
r m
icro
ns
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
James T VolkMay 2002
NLC - The Next Linear Collider Project
SLAC Rotating Coil Data
May 2002 James T Volk
FWSQ001-6 at SLAC
-7.00
-6.00
-5.00
-4.00
-3.00
-2.00
-1.00
0.00
1.00
2.00
17 17.5 18 18.5 19 19.5 20 20.5
Gradient Tesla
cen
ter
shif
t m
icro
ns
X center
Y center
NLC - The Next Linear Collider Project
Finding Tuning Rod Minimum
Integrated strength v. rod angle
y = 9E-05x2 + 0.0003x + 20.96R2 = 0.9974, run8, Rod A
y = 8E-05x2 + 0.0003x + 20.96R2 = 0.9987, run 9 Rod A
Run 13, Rod Ay = 8.28E-05x2 - 3.29E-04x + 2.10E+01
R2 = 9.97E-01
Run 14 Rod By = 1E-04x2 - 3E-06x + 20.953
R2 = 0.9839
Run 15 Rod Cy = 9E-05x2 - 2E-05x + 20.953
R2 = 0.9974
Run 20, Rod Dy = 9E-05x2 + 0.0001x + 20.952
R2 = 0.999820.95
20.955
20.96
20.965
20.97
20.975
-15 -10 -5 0 5 10 15
Rod angle relative to min pos, degrees
Int.
str
eng
th,
Tes
la
Run 8, Rod A Run 9, Rod A Run 13, Rod A Run 14, Rod B Run 15, Rod C Run 20, Rod D
Poly. (Run 8, Rod A) Poly. (Run 9, Rod A) Poly. (Run 13, Rod A) Poly. (Run 14, Rod B) Poly. (Run 15, Rod C) Poly. (Run 20, Rod D)
NLC - The Next Linear Collider Project
PM quad: centers measured repeatedly overnight on SLAC rotating coil set-up
FWSQ001-6, PM Quad Run 109
Measurement Number (each # = 8 min)
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
Y c
ente
r (m
icro
ns)
-20.6
-20.4
-20.2
-20.0
-19.8
-19.6
-19.4
-19.2
-19.0
-18.8
-18.6
-18.4
Even though room temp. which determines the Al block and magnet temp. is very constant (~0.1 °C), the Y center varies by > 1 µm and is ~ correlated with the Al block temp.
X center varies <1 µm and is not correlated with Al Block temperature.
FWSQ001-6, Run 109
Measurement number (each #= 8mins)
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140X
cen
ter
(mic
rons
)
-49.6
-49.4
-49.2
-49.0
-48.8
-48.6
FWSQ001-6, Run 109
Aluminum Block Temperature(C)
21.30 21.32 21.34 21.36 21.38 21.40 21.42
Y C
ente
r (m
icro
ns)
-20.6
-20.4
-20.2
-20.0
-19.8
-19.6
-19.4
-19.2
-19.0
-18.8
-18.6
-18.4FWSQ001-6, Run 109
Aluminum Block Temperature(C)
21.30 21.32 21.34 21.36 21.38 21.40 21.42
X c
ente
r (m
icro
ns)
-49.6
-49.4
-49.2
-49.0
-48.8
-48.6
James T VolkMay 2002
NLC - The Next Linear Collider Project
PM Wedge quad overnight in SLAC rotating coil setup measuring strength
FWSQ001-6 PM Quad
Measurement Number
0 50 100 150 200 250 300 350 400
Inte
grat
ed S
tren
gth
(T)
23.3340
23.3345
23.3350
23.3355
23.3360
23.3365
23.3370
23.3375Run 83
Strength of the PM quad measured repeatedly during ~18 hours. Various parts of apparatus and magnet have their temperatures measured at same time. Room has tightly controlled air temperature, nevertheless parts change temp by fractions of 1 °C. Quad is thermally compensated, else strength would change much more with temp. Slope of RH plot indicates integrated strength changes by 0.01133T per °C
FWSQ001-6 PM Quad
Magnet Body Temperature (C)
21.30 21.35 21.40 21.45 21.50
Inte
grat
ed S
tren
gth
(T)
23.3340
23.3345
23.3350
23.3355
23.3360
23.3365
23.3370
23.3375
Run 83
James T VolkMay 2002
NLC - The Next Linear Collider Project
Studies Using Pandira
FWSQ001-6 turn 1 rod only
-700
-600
-500
-400
-300
-200
-100
0
100
0 50 100 150
angle
X c
ntr
mic
rons
balanced rods
turn 1 rod
Tuning Rod Shift Study
-5.000
-4.000
-3.000
-2.000
-1.000
0.000
1.000
2.000
3.000
4.000
5.000
-0.020 -0.015 -0.010 -0.005 0.000 0.005 0.010 0.015 0.020
Inches rod shfited
Cen
ter
shif
t m
icro
ns
James T VolkMay 2002
f
One Rod off by 5 degrees
-5
0
5
10
15
20
25
30
0 50 100 150 200
Tuner angle
Cen
ter
shif
t m
icor
ns
all rods equal
one rod off
NLC - The Next Linear Collider Project
Further Tests
• Using rotating coil to find true minimum value for each rod
• Make new rods so we can attach potentiometers and get angle read out of rods
• Re-work wedge quad to allow for tuning of pole strength
May 2002 James T Volk
NLC - The Next Linear Collider Project
New data on the electromagnetic quad with improved magnetic measuring set-up.
NLC Prototype EM1
Measurement Number
0 50 100 150 200 250 300 350 400 450 500
Tem
pera
ture
(C
)
19
20
21
22
23
24
25
26
Water into Heater Temp Water into Magnet Temp Magnet Coil Upper Temp Magnet Steel Temp
Run 120 NLC Prototype EM1
Measurement Number
0 50 100 150 200 250 300 350 400 450 500
Tem
pera
ture
(C
)
24.6
24.7
24.8
24.9
25.0
25.1
25.2
Water into Magnet Temp Magnet Coil Upper Temp Magnet Steel Temp
Run 120
The SLAC rotating coil measuring set-up has been improved: a 5kW electric direct immersion water heater with an SCR controller has been put in the incoming cooling water circuit. Variations of ~ 4°C in LCW temp reduced to +/- 0.1 °C .
Above plots show water temperatures during a 2.5 day run at 80 amps.James T VolkMay 2002
NLC - The Next Linear Collider Project
Electromagnet LINAC quad magnetic center measurements during 2.5 day run
NLC Prototype EM1
Measurement Number
0 50 100 150 200 250 300 350 400 450 500
Y c
ente
r (m
icro
ns)
10
11
12
13
14
15
Y Center
Run 120
X and Y centers measured with improved rotating coil setup, each data point took 8 mins.
X varies by < 1µm. Y has unexpected increase of ~3 µm over 10 hr period. Y values very sensitive to various apparatus temperatures, typically the magnet core steel.
NLC Prototype EM1
Measurement Number
0 50 100 150 200 250 300 350 400 450 500
X c
ente
r (m
icro
ns)
-14.0
-13.8
-13.6
-13.4
-13.2
-13.0
-12.8
-12.6
-12.4
X Center
Run 120
James T VolkMay 2002
NLC - The Next Linear Collider Project
Electromagnetic quad data from 2.5 day run at 80 amps constant
NLC Prototype EM1
Magnet steel temp deg C
24.6 24.7 24.8 24.9 25.0
Wat
er in
to m
agne
t te
mp
Deg
C
24.6
24.7
24.8
24.9
25.0
Run 120
Magnet steel core temp tracks incoming LCW temp. Y center correlates with magnet steel temp, but in this run does not explain 3 µm increase in Y. Have not found culprit, e.g. see no correlation with (very stable) magnet current. Suspect Y has NOT really moved but our set-up created that illusion. Will continue to understand/improve set-up.
NLC Prototype EM1
Magnet Steel Temperature (C)
24.75 24.80 24.85 24.90 24.95 25.00
Y C
en
ter
(mic
rons
)
10
11
12
13
14
15Run 120
NLC Prototype EM1
Magnet Current, amps
80.192 80.193 80.194 80.195 80.196 80.197 80.198 80.199
Y C
ente
r (m
icro
ns)
10
11
12
13
14
15Run 120
James T VolkMay 2002
NLC - The Next Linear Collider Project
Determining if Vibrations from LCW can be Maintained with in Specifications
• Andrei Seryi has plots more in his Saturday Talk
James T VolkMay 2002
NLC - The Next Linear Collider Project
PM vs EM cost for Construction, Installation and
Operation
• SLAC and Stanford engineering school working methodologies for cost estimation and reliability calculations.
• Still plenty of time between now ‘04 to explore and refine designs
James T VolkMay 2002
NLC - The Next Linear Collider Project
Developing Methodologies to Estimate Overall Magnet Costs- including repair costs
• Use real failure data of magnets from similar accelerator to predict how often there will be failures of the magnets in the NLC. Estimate for a particular failure scenario how much it will cost to fix the failure– accounting for all aspects of the repair and the lost opportunity costs (for those failures happening during operations, when the accelerator is un-useable– have to ascribe an opportunity cost per hour).
• For real failure data: took all the magnet failures that brought down any beamline in SLAC during 5 year period. Using downtime reporting database we scrutinized every one of 75 failures as to its cause, type of magnet, length of time to detect and repair.
• Calculated number of SLAC magnets running during any period and how many hours they ran to calculate mean time between failures.
• Calculate availability values and extrapolate them to known number of NLC magnets running for 30 years. Estimate cost to fix all failures.
James T VolkMay 2002
NLC - The Next Linear Collider Project
SLAC Downtime (Avg)
0
5
10
15
20
25
30
Small Medium Large
Magnet Size
Number of Failures
0
10
20
30
40
50
60
SmallNon-Water Cooled
Medium Large
Magnet Size
Oc
curr
en
ce
SLAC Downtime
0
20
40
60
80
100
120
140
160
180
200
Small Medium Large
Magnet Size
Ho
urs
Magnet Failures at SLAC: Jan. 97 – Dec. 2001
Dow
ntim
e (h
r)
James T VolkMay 2002
NLC - The Next Linear Collider Project
Failure Frequency
0
5
10
15
20
25
30
35
Insulation Water Leak WaterBlockage
HumanError
Connector Other
Failure Type
Eve
nts
Events Min Max AvgInsulation 29 0.2 27.2 8.82Water Leak 22 1 180 16.44Water Blockage 5 0.5 7.5 3.92Human Error 5 0.7 6 2.5Connector 3 1 3.2 1.733Other 12 0.9 10.2 5.8
0
5
10
15
20
25
30
35
40
180
Insulation Water Leak Water
Blockage
Human Error Connector Other
Failure Type
Do
wn
tim
e (h
r)
SLAC Downtime by Failure Type
James T VolkMay 2002
NLC - The Next Linear Collider Project
Water Cooled Magnets (Med & Large) Failures
By Run Time (Medium and Large Size Magnets)
Date Line Run Hour Ma gnets Magne t Hours # Fai lure s MTBF TR MTTR Availa bil ity 1 Ma g PPM2/4/97 - 4/30/97 Linac/BSY 1547 520 804440 1 804440.0 0.2 0.20 0.999999751 0.2
HER 181 1200 217200 5/1/97 - 6/8/98 SLC 8828 2104 18574112 32 580441.0 469.5 14.67 0.999974724 25.3
HER 918 1200 1101600 7/10/98 - 7/31/98 HER&LER 575 2433 1398975 2 699487.5 9 4.50 0.999993567 6.410/30/98 - 12/15/98 HER&LER 1040 2433 2530320 6 421720.0 40.1 6.68 0.999984152 15.81/15/99 - 2/22/99 HER&LER 844 2433 2053452 4 513363.0 15.6 3.90 0.999992403 7.62/24/99 - 5/1/99 Linac 1461 520 759720 2 379860.0 26.1 13.05 0.999965646 34.45/1/99 - 11/29/99 HER&LER 4797 2433 11671101 7 1667300.1 65.65 9.38 0.999994375 5.61/12/00 - 10/31/00 HER&LER 6624 2433 16116192 7 2302313.1 34.6 4.94 0.999997853 2.1
BSY/FFTB 2196 198 434808 BSY/A-Line 630 248 156240
1/10/01 - 12/31/01 HER&LER 7411 2433 18030963 7 2575851.9 37.9 5.41 0.999997898 2.1BSY/FFTB(e+) 2795 509 1422655 2 711327.5 3.05 1.53 0.999997856 2.1BSY/A-Line 820 248 203360
Average 75,475,138 70 1,078,216 701.70 10.02 0.9999907030 9.3
Ma gnets PPMNLC 4362 Availability 0.960257533 39,742.47 SLC 2104 Availability 0.980629065 19,370.93
Fore ca st for NLCOperation Hr/yr 6480 Expected Downt ime 257.5 hr/yearOccurrence/yr 25.7
James T VolkMay 2002
NLC - The Next Linear Collider Project
Water Cooled Magnet Problems
Sub System: Water Cooled Magnets (Med & Large)
A subset of the failure modes analysed in detail using real failure data from SLAC magnets
Orig
in
Det
ectio
n P
hase
Re-
occu
ring
Min
Fre
quen
cty/
Occ
urre
nce
Max
Min
Det
ectio
n Ti
me
Hrs
Max
Min
Fix
ing
Tim
e H
rs
Max
Min
Del
ay T
ime
Max
Min
Rec
over
y H
rs
Max
Min
Qua
ntity
Mag
s
Max
Labo
r C
ost $
Mat
eria
l Cos
t $
Opp
ortu
nity
Cos
t $
Inadequate water pressure Des DR 1 0.05 1 8 0 1 58 Too many loads on water circuit Oper Oper 30 0.01 0.5 4 4.5 1 174 6,750 Conducter Sclerosis (hole gets too small) Oper Oper 30 0.5 1 8 9 1 17,400 675,000 Water passage is blocked due to foreign object Oper Oper 30 0.4 1 2 4 7 5 1 7,200 300,000 Damaged (crimped) coil Inst TR 1 4 0.5 2 0 1 1,200 Water sprayed onto the coil Oper Oper 30 0.7 2 2 8.4 180 10 1 26,376 1,092,000
0 - Poor design of jumpers- short Des D/P 1 0.2 1 8 0 40 8,968 Bad Installation (Bolts not tight) leads to short inst TR 1 4 0.5 0.8 2.5 6 0 1 1,480
0 - Loose jumpers -poor design, lead to overheating Mfg Test 1 0.011 1 8 0 40 493 Loose jumpers- bad installation, lead to overheating Mfg Test 1 4 0.5 2.5 0 1 1,480
0 -
Origin-when the failure is initiated Detection- when the failure is detectedDes: Design DR: Design ReviewMfg: Manufacturing Proto: PrototypeInstall: Installation Test: Testing in labOper: Operation TR: Test Run of System
Oper: Operation
Scenario--magnet gets turned off
James T VolkMay 2002
NLC - The Next Linear Collider Project
LBNL & SLAC work on designing magnets (PMs and EMs) for the damping rings
• Main Damping Ring lattices have been published with detailed requirements on all magnets
• Have 2-D model of DR quadrupoles and transport line dipoles. The Nd Iron style magnets are of reasonable size
• Investigated the Nd Iron quads, with rotating rods to generate the +/-10% adjustability, in more detail to see if they could meet all the requirements.
• Jin-Young Jung (LBNL) used TOSCA to make a 3-D model of damping ring magnets
• Validated the new 3-D model of the Neo quad by simulating it as an infinitely long magnet similar to the PANDIRA code. Poletip fields predicted by the 2 codes agreed to within 0.2%, so TOSCA model good.
James T VolkMay 2002
NLC - The Next Linear Collider Project
Results from 3-D model of PM DR quad
• DR magnets have to have a “C” shape – allow for the extra wide vacuum chamber to extend towards the outer
edge of the ring and be
– capable of accepting the high amounts of synchrotron radiation.
• The TOSCA 3-D model of a 2cm radius Neo quad was run with a 25 cm effective length. The poletip field was 10% less than in the 2-D PANDIRA model. – Decrease due to flux loss
– End Plates do not help reduce this
James T VolkMay 2002
NLC - The Next Linear Collider Project
TOSCA model of ¼ Neo quad with a steel end plate
James T VolkMay 2002
NLC - The Next Linear Collider Project
NLC DR quad as an electromagnet:TOSCA model
• POISSON 2-D model was made of an electromagnet with same steel poletip shape as the pm. It satisfied the DR quad requirements.
• To allay concerns about the size of fringe field out of the end faces a 3-D TOSCA model was made. The coils were carefully modeled in 8 sections to look exactly like the real ones would.The model universe included air out to where the gradient dipole would be => fringe field < 2 gauss
James T VolkMay 2002
NLC - The Next Linear Collider Project
NLC DR Gradient Dipole: TOSCA design. Studying end effects and fringe fields
• To allay concerns about the size of fringe field out of the end faces a 3-D TOSCA model was made. The coils were carefully modeled to look exactly like the real ones would.The model universe included air out to where the nearest quad would be => fringe field < 3 gauss
• TBD: calculate the multipole harmonics in 3D and check field quality is met.
James T VolkMay 2002
NLC - The Next Linear Collider Project
Conclusions
• Slow but steady progress being made on LINAC quads
• We are quantifying the radiation fields in the Main Linac and Damping Rings
• First tests of radiation hardness are underway more ready to begin
• Working on Cost estimates and failure mode analysis
• Working on magnets for Damping Rings
• Looking at magnets for final focus
May 2002 James T Volk