The yoke halves are clamped together by & 540 mm O.D., 4.95 mm thick stainless steel shell which is pretensioned at room temperature by weld shrinkage and ...
Fermi National Accelerator
Laboratory
Mechanical Behavior of Fermilab-Built 1.5 m Model SSC Collider Dipoles J. Strait, R. Bossert, J. Carson, K. Coulter, S. Delchamps, T. Jaffery, W. Kinney, W. Koska, S. Gourlay, M. Lamm, and M. Wake Fermi National Accelerator Laboratory P.O. Box 500, Batavia, Illinois 60510
R. Sims and M. Winters SSC Laboratory 2550 Beckleymeade Aue.,Dallas,
Texas 75237
September 1991 * Presented at the 12th International
a
Operated
by Universities
Resesrch
Aswciatlon
Conference on Magnet Technology, Leningrad,
Inc. under Contract
No. DE-ACOZ-76CHC3000
USSR, June 23-28, 1991.
with the United Staler
Department
of Energy
Disclaimer
This report was prepared as an account of work sponsored by an agency of the United States Gouernment. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefullness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation. OTfavoring by the United States Government or any agency thereof The views and opinions of authors expressed herein do not necessarily state OPreflect those of the United States Government or any agency thereof
Mechanical
Behavior of Fermilab-Built
1.5 m Model SSC Collider
Dipoles
J. Strait, R. Bossert, J. Carson, K.J. Coulter, S. Delchamps, T.S. Jaffery, W. Kinney, W. Koska, S. Gourlay, M.J. Lamm, M. Wake Fermi
National
Accelerator
Laboratory,
P.O. Box 500, Batavia,
IL 60510, USA
R. Sims and M. Winters Superconducting Super Collider Laboratory, Fermilab/SSCL Magnet Project Fermi National Accelerator Laboratory, P.O. Box 500, Batavia, IL 60510, US AbrlractSeveral model SSC collider (50 mm aperture, 1.5 m magnetic
dipole magneto length) have been built and tested at Fermilab. These magneta are lnatramented with 6tr.h gmge, to mea~nre rtreeaes in the coil, the cold man, shell, and the coil end clamp aneembly. Measurements are made of these qnantltlen during assembly, cooldown, excitation, and wrrmnp. Additional mechanicel meannrementa are made OP magnet sub-agsembliea during mannfaetnring. Data from these meamnrements are prcaented and compared with expectations fkom the design e&nlations.
I. INTRODUCTION
This paper describes the mechanical behavior of 1.5 m magnetic length 60 mm aperture SSC collider dipole magnets that have been built and tested at Fermilab, and comparea the behavior with that expected from the mechanical design c&ulations(l,Z]. The mechanical design of the 2D cram section is folly discussed in [I] and of the magnet end in [a]. The vertically split yoke contacta the collars near the horizontal mid-plane under all conditions to provide support against deilection under the IxB force. The yoke halves are clamped together by & 540 mm O.D., 4.95 mm thick stainless steel shell which is pretensioned at room temperature by weld shrinkage and whose pretension grows with cooldown. A collet assembly, consisting of a four piece GlO inanlator and a stainless steel (DSAS21 and 525) or aluminum (DSAS24 and beyond) outer cylinder, in used to clamp the coil end. Axial force is transmitted from the coil to the shell through a S5 mm thick end plate which is prestressed against the end clamp by four set-screws. Measurements of important sub-assembly dimensions are made during manufacturing. Strain gauge transducers[S] measure the coil arimuthal prestreas and the force between the end clamp and On some magnets strain ganges have end plate. been placed on the shell and on the end &mp. Data are presented from six model magnets denoted DSASZl-325 and DSAS27. DSAS21 and 525 several
M.nurcdpt mcclved Jnnc 24, 1991. U.S. DcPartment of Em,
Work
Sopported
have been fully tested and DSAS24 ia currently under cold test. DSAS22 and 527 have been used for assembly experiments and have not been cold tested. DSAS25 developed a turn-to-turn short during assembly which cleared upon disassembly and could not be repaired. DSAS26 is currently being assembled and ia not discussed. II. ASSEMBLY
DATA
The inner and outer coil prestresa at important pointa during assembly and tenting are shown in Table I. The collars are designed to poaition the conductors (LLI defined by the magnetic deaign[4] without the ose of pole shims. This requires that the coils be molded to the correct aise to achieve prestress in the desired ranges of 55-55 MPa and 4070 MPa in the inner and outer coils respectively. However, pole shims were applied to several of these magnets by adding or removing a layer of Kapton ground insulation at the pole aa shown in Table I. The measured relation between coil+ahi size and prestress haa been used to determine[S] the correct
Table I POLE SHIMS AND COIL PRESTRESS. THE UPPER (LOWER) ENTRY IN EACH PAIR CORRESPONDS TO THE INNER (OUTER) COIL. DSA321
322
32.3
324
316
32,
0 0
0.13 -0.13
0 0
0.13 -0.13
0.13 0
0 0
62 91
71 2,
4, ,6
81 66
*I 61
67 9*
Before shall weldlnl(
69 86
-
44 ‘13
6, 5s
Nter she” weldinS
63 3s
-
53 75
a 61
Before cooldom
62 66
.
52 72
,6 60
AIter cooldown
43 ‘I2
-
22 62
43 62
B=6.,
1* 61
-
4 62
22 40
Pale Shim (mm) Pceltrem (MPa) AIt. couar ~syln&
by Ibe
1.
T
All coil r&e for the long 50 mm dipoles. except DSASZS have the inner prestress witkii the target window but the outer coil prestresa tends ta be systematically higher than is required. Stress relaxation occura, especially for the higher prestress magnet DSAS24, between collar keying and shell welding and between welding and cooldown. Collar detlections with prwtresa are determined by comparing meaaured collared coil diametera with those on ~1 collar pack assembled with keys but no coil. The relation between the vertical deflection and the inner-outer average prestress ia plotted in Fig. 1 for four of the six magnets. (One wa.~ meaanred before adequate procedures for collar diameter measurements were inatituted and is not plotted. Another wan keyed by closing the collara only far enough to start the tapered keys and then driving them in with a force of 500 N/mm. Thin magnet cannot be compared with the others due to the possibility of additional inelastic deflectiona from the tapered key insertion.) The deflectiona are somewhat larger than predicted by fmite element calcnlationa[6] but comparable to those assumed in the design c&ulations[l]. Finite element ealcnlations[6] predict 8 much smaller horizontal than vertical deflection with prestress. The measured horizontal radial deflectiona lie between -0.0s and +0.04 mm. Thin is a wider range than expected and there in no obvious relation between the deflection and prestreas. Because the deflections are determined relative to one nndeflected collar pack, this varistion may represent the currently achieved tolerance on the assembly of the spot-welded pairs of collars[7]. This ir currently under investigation. When the shell is welded the magnet is placed in a press which compresses the yoke aboot the collared coil, cauc.es the shell to conform to the yoke, molded
magnets
and holds the entire wembly in & straight, twist-free state which is then maintained after the two shell halves are welded. Weld shrinkage causes an azimnthal stress to be developed in the shell which maintains the yoke-collar compression after the press in opened. Fig. 2 shows the DSAS24 prestress aa a function of yoking press load before shell welding. The dashed lines show the final prertresaer after The final welding and removal from the press. prestress is the same aa at LL press load of 2.2 kN/mm, implying an asimuthal ahell tension of about 220 MPa. This is comparable to the 175200 MPa assumed in the mechanical design. III. COLD TEST DATA
Tests at liquid helium temperature were performed in a vertical dewar in the Fermilab R&D test facility at Lab 2. The atrain ganges messnring coil atresa and end force were calibrated both at 293 K and 4 K and therefore measure absolute stress at 4 K. The sero-strain resistances of the gauges on the magnet shell and the end clamp were not Therefore these gauges measure calibrated at 4 K. only the changes in strain at 4 K. Two magnets have been cold tested and a third is currently under test. The prestreas lossea with cooldown (Table I) are lQ-SS (S-14) MPa in the inner (outer) coils, changes comparable to those in vertically-aplit yoke 40 mm model magnets and somewhat smaller than in horizontally-aplit models[g]. The set ~~crewn which transfer the axial load from the coil end to the end plate are preloaded at room temperature to S-S kN. The load changed with cooldown by -1 kN, 0 kN, and +I kN in DSAS21, 525, and 524 respectively. The ssimuthsl coil stress at the pole ia plotted aa a function of magnet current squared (proportional
0.15
T
0.10
E 2 -3 0.05
0.00
501 0
40
20 Cdl
stress
60
60
0
(UP@
FIs. 1. Colhr vertlcsl radial deflectton vs. pra.trc,,. The dashed and w”d he. M the relstlonr expected lrom “nita elemsnt cdeulstlons and amumed in the mschanlcal deliof the masuet respectively.
1000 Press
2000 Load
(N/mm)
Fill. 2. DSA324 co11 prcatraw v.. yoklns press load before shall welding. The dashed Une. show the flus1 prastrslsc. after the ehcll vu welded and the prem .,u opened.
2.
The SSC operating current to IxB) in Fig. S. correafonding to 20 TeV is 6.6 kA (6.7 T) or 44 kA . The pole stress remains positive up to well above 7 kA (7 T) in DSAS21 and 524, but the DSAS2S inner coil unloads near the operating current. The outer coil stress remains high through We conclude, the full current range in all cbaes. therefore, that the initial room temperature prestresses in the outer coils (see Table I) are considerably higher than is required, while the DSAS23 inner preatress is marginally tca low. The force between the coil end clamp and the end plate is shown r~s a function of magnet excitation in Fig. 4. The 4-10 kN end-force change from 0 to the operating current id a small fraction of the total LxB force of 110 kN. Most of the axial force in, M in the 40 mm magneta[Q], transferred to the shell via coil-collar-yoke-shell friction. The slope with current squared was initially much smaller in DSAS23 than DSAS21. DSAS2S was observed[lO] to quench occasionally in the return end on the down-ramp at about 4 kA. It wm conjectured that the down-ramp quenching and the small end-force slope might be related and the performance could be improved by increasing the end preload. A thermal cycle was performed to allow the end set wrews to be tightened to a 10 kN load. The end force slope was then comparable to that in DSAS21, but the down-ramp quenching remained. DSAS2S ia currently being disassembled to investigate the cause(a) of this behavior. Strain gauges were mounted on the DSASZS shell to measure azimuthal stress at five azimuthal locations at the magnet center. Axial atrain is slao mecawed at one of the wimuthal locations. Under excitation, the IxB force in the body of the magnet induces a horizontal elliptical deflection, which can be observed by the changes in asimuthal stress in the shell: negative (compressive) near the pole (90’) and positive (extensive) near the mid-plane (0’). If the shell tension ware inauflicient to keep the vertical
(Magnet Current) 2 &A’ 1 Fig. 1. End force ~8. current aquared.
yoke parting-plane gap closed, the gauges near the pole would show a positive stress change. The axial l.xB force ia transferred from the coil to the shell via friction and the end loading set-screws. Therefore both axial and asimuthal stresses change with excitation and the measured strains must be corrected This correction is done 58’ for the P&son effect. from the mid-plane where there are gauges measuring The correction is applied strain in both directions. at the other angular positions by assuming the axial stress to be ssimuthally uniform. The data are plotted versus current squared in Fig. 5. The pattern with azimuth is qualitatively r.8 expected: compressive near the pole (79’) and progressively more extensive nearer the mid-plane. It ia not undemtood why the 0’ atress is amaller than that at 29’. The stress nearest the yoke parting plane is negative over the whole range measured indicating that the yoke gap remains closed up to at 1eaJt 7.7 kA (about 7.6 T). The axial stress change to the operating current, 11 MPa, multiplied by the cross sectional area of the ahell gives an axial force of 57 kN. the remainder of the IxB force ia
a0
0 0
20 40 (Magnet Current) 2 &A2 )
Fir. 3. Coil strem vs. msmet
cmrent
0
60
20 40 (Magnet Current) * (kA2)
60
Fig. 5. Aaimuthal (A) and axial (s) atress change in tba DSA323 cold maas she,, VI. current aquared. The numbam in panathesem are the distances in degnea from the x-c&.
squared.
3.
friction and not by compressive loading of the coil end, 5) The yoke parting plane gap remains cloned to well above the operating point, and 6) end clamp deflections are 50.05 mm under excitation. Isauea atill to be resolved include 1) The larger than expected spread in horizontal radius of the collared coil, 2) The significant difference in end-force slope with excitstian among the magneta tested, and 5) The possibility that long magneta may behave diierently in some reaper&, particularly in their axial mechanics.
balanced by stretching of the coil. Messurements[Q] on long 40 mm magnets indicate that it takes about 1 m to fully transmit the axial force to the shell. Here the magnet center in only 0.50 m from the end of the yoke. Strain gauges are alao mounted on the DSAS2S return end clamp outer cylinder S2 mm from the end of the collared portion of the coil. This ia roughly centered in the 76 mm of coil straight section inside the end clamp. The data with excitation are plotted in Fig. 6. As in the shell, these data are consistent with L horinoatally ellipticsl deflection under excitation. If the deflection is modelled aa a purely elliptical distortion of a 99 mm O.D, 17.6 mm wall cylinder the measured attains correspond to a roughly 0.05 mm increase in horizontal radius. This agrees within about a factor of two with finite element calculationa[ll]. The axial atrains are much smaller than the azimuthal strains. At most a few microatrsin are expected from the measured end force shown in Fig. 4. The apparent atraina most likely reflect imperfect compensation of the magneto-resistance of the gauges 4OOj’
.“““”
..”
““‘..”
ACKNOWLEDGMENTS
We would like to thank the staffa of the Fermilsb Superconducting Magnet Production group and of the Superconducting Magnet R&D Facility at Lab 2 whose dedication and care made tbi work possible. REFERENCES
14
.f
Fr.d.c.,
PI
I3)
e ; L
j+
I41
Awl
-j$zEJ ‘,,; y.j x
20
0
(Magnet Current) 2
4b
t
ISI
6b
&AZ )
PI
Fig. e.
Asimotb.1 (A) and axi.., (s) strain chanS. in tb. return end clamp outer cylindsr of DSA323 vs. current qnlrad. Tb. nombar, in parenthas., u. tb. d1.t.t.c.. in dema fmm tb. x-.xi..
IV.
I. Strait et al., ‘%f.ehsnic.l d..ign of Lb. ZD cr..,~etion of tb. SSC collider dip& magnet,” ,abm,tt.d t. th. 1991 IEEE Partlcl. Accelerator Conferenc., San
[71
CONCLUSIONS I31
Three 1.5 m model SSC collider dipoles have been built and tested at Fermilab. Mechsnical measurements from these and several other partially sssembled msgneta agree well with design expectations. The important results include 1) The collar vertical deflection with preestress and the shell azimuthal ten&n agree quantitatively with design expectations, 2) Coil prestreasea within the desired range have been achieved without pole shims, 5) Preatress loss with cooldorn compares favorably with 40 mm models, 4) Most of the axial IxB force ia transmitted to the shell via collsr-yoke-ahell
PI PI
IllI 4.
CA,
Ma,
69,
1991.
9. D.lchampn cl al., YSC collldar dlpol. m&gn.t and mcebsalcsl duti.” auhmitted to th. 1991 IEEE Putt& Acc.1.ratc.r Coni.r.nc., San Frmcl.co, CA, M.y (I-P, 1991. CL. Goaddt st ..I., “M.uur.m.nt of ht.r,,al kc.. in sup.rconductlnS aceahrator magmt. with strdn gaug. tran.duc.n:’ IEE Yk.n.. h&m. vol. 2S, pp. 1.63.1468, Much 1939. RC. Gopta et al., “SSC 60 mm dipol. cm,. wtfon.” subm1tt.d t. the 3rd lntematlond lnduaial Symposium GA, Msrcb 13-l& 1991. on Sopm Co”‘d.r, AU&da, hf. Wake et .I., “T&m of 1.6 meter mod.1 59 mm SSC collider dipole. at F.rmi1.h.” .ubmitt.d to th. IEEE Partlcl. Accslemtor Confarenc., San Fran&c., CA, May 6.9, 1991. J. Kerby, “M.ch.oic~l Analysis of th. V.rtlc.lly Split Yak. 69 mm SSC Dipole. ” Fermilab Tcehnierl Support Section internal note TS-SSC 91-001, DK. 30, 1990. A. Dar.d et .I.. “Statrar of ,-cm spcttm., I,-m.,onS SSC dlpol. masnet R&D proSqsm at BNL put I: magnat a.wmbly,” .obmitt.d ta the 3rd International lndu.tti Symposium on the SupercoIUder, Atlanta, GA, March 13-16, 1991. W. Korh et al., “T..t, o, (0 mm SSC dlpo,, mod., mqn.t. with rertlc~1Iy split yak...” ,ubm,tt.d to th. IEEE Particle Aeeclerator Conferam., San Fran&co, CA, Ma, B-9, 1991. 1. Strait et al., “Tests of full wale SSC R&D d1p.l. II).g..ts,” JEEE Tmn.. bfbph vol. 16, pp. 1466-146*, Much 1939. Id. Wake, et .I., “Qosncb h.bav1.r of 1.6 m mod.1 SSC coUid.r dip.,. magnet. at Fermilab.” wbmitted t. the 12th Interaatloasl Confercnc. on MaSnet T.cbnoloSy, Lsningrd, USSR, June 24-28, 1991. “Finite element aaslymis of aed cellat J. Ksrhy, mdsnbm,” TS-SSC 90-086, Nor. 8, 1990.
