Optimization of Nb3Sn SC racetrack coil

1
Optimization of Nb3Sn SC racetrack coil
SMC working group October-29th, 2007
F. Regis Supervisor P.Fessia CERN - AT-MCS
2
Summary

• Review on magnetic work
• 2D Mechanical model
• Set up of the model, materials
  properties, contact elements distribution
• First results
• Conclusion and future steps

3
Review on magnetic model

• A different solution has been proposed to
  decrease Bmax in the coil head (Roxie-coil in
  air)
• A double spacer configuration is introduced to
  lower the induction on the inner pack (layout
  2-2-17),
• The final configuration chosen has two spacers of
  the same dimensions for the two layers (some
  computations have been done with different
  dimensions as well. Drawback layer-jump)

• Iss 13960 A
• Bp 12.94T
• B0 9.65T

• 3D magnetic models in Ansys (MVP and MSP
  formulation)
• Xchecked with Cast3m and VF Opera

4
Mechanical model

• Magnet assembly Plane42 (4nodes element)
• Contact elements Conta171, Targe169
• 1st step assembly (interference between vkey and
  yoke)
• 2nd step cool down at 4.2K
• 3rd step powering _at_ 0.25-0.5-0.75-0.9-1.2 of
  Iss

• Island-coil Bonded
• Horseshoe-coil Bonded
• Coil layers midplane insulation Bonded
• Other contacts standard (pure sliding)
• No friction at the moment
• No interference on horizontal key
• Interference between yoke and vkey

Startup parameters
5
Material properties - I
Ref. CERN EDMS 827451
6
Material properties - II

• 4 Line added for NED SMC. The values derive from
  alloys used in LHC dipoles prototypes.
• emax is 13 at 293 K (15 at 4.2 K) for the
  2014-T651
• emax is 16 at 293 K (31 at 4.2 K) for the
  5083-H321
• 5 Line added for NED SMC. Source ASM tables. 5a
  apex means that the properties are taken for
  Ti6Al4V-ELI sheet, annealed along longitudinal
  direction. 5b apex means that the properties are
  taken for Ti6Al4V-ELI sheet, annealed along
  transverse direction.
• 6 Source E.D. Marquardt, J.P. Le, and Ray
  Radebaugh Cryogenic Materials Properties
  Database
• 7 Average values. The properties are strongly
  dependent on the type of composite (G10 or G11)
  and on the resin volume fraction. The Youngs
  Modulus values are obtained from traction tests.

7
Analysis

• Analysis procedure
• ix 100, 200, 400, 600, 800, 1000
• Check for Stress profile on lower and upper coil
  edges (sx, sy) to keep smax about 150 MPa _at_
  1.2Iss
• Check for Stress profile on pole coil pack to
  assure compression _at_ 1.2Iss
• Check for Stress distribution on assembly
  components (Von Mises criterion)
• The Yoke is made up of Magnetil low carbon iron.
  It shows brittle behavior at cryo temperature so
  the failure criterion adopted is Rankines
  (maximum primary stress)
• The yoke is assumed without any defects

Ref. Design and Optimization of the 12.5 T EFDA
dipole magnet
8
Results Lower layer
Ix600µm
9
Results Upper layer
Ix600µm
10
Results Assembly
Ix600µm
11
Results Assembly-II
Ix600µm
Some remarks

• The stress along Pad-y upper edge shows
  transition traction-compression due to Poissons
  effect
• sy along the coilpack due to hkey strongly
  depends on the different properties between
  island/horse and Nb3Sn

12
Results Shell

• The stress distribution s? on X0 plane changes
  with ix max bending stress position is reversed
  for ixgt400µm
• s? distribution at assembly does not show such a
  reversed trend

13
Results sx,avg
14
Results sx,avg
15
Failure criteria displacements
Yoke Magnetil steel s1,max 307 MPa lt 480 MPa
sx,lim/s Where s1.5 Safety Factor
sx,lim720 (7K)
Shell Al 2014-T651 sVMeqv,max 197 MPa lt 363
MPa sx,lim/s Where s1.5 Safety Factor
sx,lim545 (4.2K)
Pad-X Pad-Y clearance ePx-Py 4mm ?xCD
3.63mm Where ?xCD is computed from ux,px(p2)-
ux,py(p3)
ux,yoke(p2) -0.008mm _at_ CD
16
Yoke shell

• Analysis Criteria
• The stress distribution sx on the island-coil
  side must be around 20 MPa _at_ Iss
• The failure criteria for the yoke and shell must
  be satisfied
• The sxmax on the coil pack must be lower then 150
  MPa
• For this trial study ix600µm
• Iss computed for each configuration of wyoke

• It can be shown that
• To get the required pre-s on the coil pack, the
  bigger the yoke the smaller the outer tube
  radius, being the rigidity of the shell more
  effective then the thermal contraction of the
  shell
• tshell 20mm wyoke 55mm is enough to respect
  all the design constraints, but Pblad 60 MPa
  (can we get it?)
• wyoke 90 mm (SD01), tshell 15mm to assure a
  higher Bp 12.9 T

17
Conclusions

• A 2D mechanical frictionless model has been set
  up
• Four different macros have been set up to post
  process data from Ansys (s and displacements)
• Simulations reveal that an horizontal
  interference of 600 µm could assure the required
  levels of pre-s and smax on coil pack, as well as
  the respect of failure criteria on the magnet
  components
• A study on the sensibility of yoke and shell
  dimensions has been carried out, pointing out
  that using a yoke width of 90 mm could lead to
  the same pre-s but increasing the peak field
• Some more computations have to be done,
  considering a lower pre-stress by the bladders
  and increasing the yoke stiffness considering
  likely defects in the laminates
• The next step will be the introduction of
  friction in the model