J. García, F. Toral (CIEMAT) P. Fessia (CERN) January, 26 th 2015

 Requirements & specifications.  Latest design presented.  Mechanical analysis of the iron ring magnetic design.  New magnetic optimization.  Conclusions & decisions to be made.  Next steps. 2

Requirements & specifications 3

4 Magnet configurationCombined dipole (Operation in X-Y square) Minimum free aperture150 mm Integrated field2.5 Tm Baseline field for each dipole2.1 T Magnetic length1.2 m Physical length1.505 m Working temperature1.9 K Nominal current< 2500 A Field quality< 10 units (1e-4) * Fringe field40 mT * Iron geometryMQXF iron holes Strand parameters Cu:Sc1.75 Strand diameter0.48 mm Metal section0.181 mm 2 Nº of filaments2300 Filament diam.6.0 µm I(5T,4.2K) A Jc A/mm 2 Cable parameters No of strands18 Metal area3.257 mm 2 Cable thickness0.845 mm Cable width4.370 mm Cable area3.692 mm 2 Metal fraction0.882 Key-stone angle0.67 deg Inner Thickness0.819 mm Outer Thickness mm Insulation: Fibre glass sleeve This method provides stiffer coils for an easier double layer winding. Need validation test of a suitable binder: Ceramic binder is not applicable (only in cured state). Some PVA studies are ongoing but compatibility with impregnation has to be demonstrated. Ceramic binder substitute (suggestion by CTD) is being tested by CERN. *To be discussed at this meeting

Latest design presented (November 2014) 5

6 Two main problems:  Saturation at nominal current for both dipoles causes the increasing of sextupoles: ◦ ∆b3= 37 units ◦ ∆a3= -25 units.  Large radial deformations (causing possible poor training). Two solutions proposed:  New magnetic design using an iron ring placed further away.  Larger collars in order to increase the stiffness of the assembly and make them self-supporting. Worst case: I ID & I OD variation from 10% to 100%

7 Current increase from previous design:  I ID =1885 A (prev A)  I OD =1870 A (prev A)  Greater forces! Saturation multipoles:  b3= 17 (prev. 37)  a3= 18 (prev. 24) Ansys simulation using the new forces (108% I N ) Worst case: I ID =100% I OD = Var. from 10% to100%

Mechanical analysis of the iron ring magnetic design 8

9 Any of the actions taken to increase the stiffness of the assembly decreased the radial displacements of the outer coils. Ellipticity

10 However, the action of an external shell or increasing the outer collars thickness do not reach the inner coils, given the assembly gap between inner collars and outer coils. GAP Outer pressure Magnetic forces (Both dipoles at 108% I N ) The only way to decrease the inner dipole ellipticity is to increase inner collars thickness.

11 OD Max Ellipticity= mm (prev mm) ID Max Ellipticity= mm (prev mm) Geometrical changes: Inner collar thickness increased in 6 mm (21 mm -> 27 mm). Outer collar thickness increased in 2 mm (31 mm -> 33 mm). Outer dipole aperture increased from 218 mm to 230 mm. Outer collar diameter increased from 300 mm to 316 mm. It is difficult to put a limit on the radial displacements. The main goal has been to make it as stiff as possible while keeping reasonable dimensions. It has to be highlighted also that each mm of additional steel has less effect than the previous one.

13 AssemblyCooldownLoad All these results were obtained using the inner Titanium tube. Possible re- evaluation considered. 100% I N Low radial tension at spacers edges, caused by friction Low azimuthal tension caused by friction

New magnetic optimization 14

Inner coil (ID) & Outer Coil (OD) parameters UnitsWhole ironIron ringOnly cryostat Nominal field 100% (ID)T2.11 Nominal field 100% (OD)T2.11 Nominal Field (Modulus, 100% ID & 100% OD)T2.99 Nominal current (ID)A Nominal current (OD)A Coil peak field (Modulus, 100% ID & 100% OD)T3.94 (ID)4.18 (ID)4.25 (ID) Working point (Modulus, 100% ID & 100% OD)% Max fringe field, 20 mm out of the cryostat (Modulus, 100% ID + 100% OD) mT Sextupoles variation ∆b3 / ∆a3 (Worst case)units37 / / 142 / 5 Iron yoke Inner Diam.mm Iron yoke Outer Diam.mm610 - Torque10 5 Nm/m Mean stress at the coil and collar nose interface MPa Using the new geometry: Outer dipole aperture = 230 mm. Collars outer diameter = 316 mm.

16

17 o Other iron geometries were tested, with similar or worse field quality results: o The difficulty comes because the field changes in two ways depending on the powering scenario: Orientation and intensity. o One of these variations could be assumed optimizing the iron geometry. o Both of them combined make any optimized geometry useless when the powering scenario changes, because symmetry conditions are lost.

18 o It is not possible to screen the field below 40 mT without the whole iron, a lot of magnetic material would be necessary. o Adding a warm screen and measuring the fringe field 20 mm from its surface: Rin= 467 mm, Thickness = 10mm, Max Fringe field = 131 mT. Rin= 467 mm, Thickness = 50mm, Max Fringe field = 31 mT. o Other magnetic materials like mumetal would not be a solution, since they have higher permeability, but lower saturation field. o The main problem is that we deal with a dipole: the dipole field decays as 1/r 2 compared to quadrupole field which decays as 1/r 3. o Therefore, we should choose between: Higher fringe field or Higher variation of the multipoles with the current (when magnetic materials shield the dipole field).

19 o Field quality achieved but peak field still high. o Several approaches being evaluated. Currently checking how splitting blocks could help. o Coil ends length ≅ 200 mm. o Coils length ≅ 1400 mm. o Total magnet length = 1505 mm (Current layout).

Conclusions & decisions to be made 20

21  Mechanical design: ◦ Whichever the mechanical design used, the radial displacements in the inner dipole can only be contained using the inner collar. ◦ The thickness of the inner collar will depend on the current in the inner coil and the radial displacement allowed for the inner coil.  Magnetic design: ◦ It is not achievable to meet both fringe field and field quality requirements simultaneously.

22  Inner & Outer dipole apertures.  Iron geometry (if any).

23  February ‘15: 3D Magnetic Optimization  March ’15: Winding Machine.  April ‘15: Short mechanical model.  October ‘15: First coil.  CERN: Binder tests.

24

Annexes 25

26  As a combined dipole that requires a square range of operation in X and Y axis, a large torque arises when both coils are powered.  Due to the expected radiation dose a solution based on mechanical clamping is required to mechanically fix the coils and guarantee the magnet performance.  Other major challenges:  Mechanical model ◦ Powered with 108% of nominal current for sizing purposes. ◦ Material properties used: MaterialE [Gpa]u [-]CTE [K -1 ] Coils & spacers (Impregnated)NbTi+Cu400, *10 -5 CollarsStainlessSteel K / K 0, *10 -5 Inner tubeTi1300, *10 -5 Radial inward forces at the inner dipoleInner titanium tube Large azimuthal displacements of the coilsAzimuthal interference at collar noses Large radial deformations of the assemblyLarge self-supporting collars ChallengeSolution proposed

27  Comparing the main challenges in the case of MCBXFB design for both designs: ◦ The torque between nested dipoles:  It will be the same for both designs. ◦ The separation of the pole turn from the collar nose due to the Lorentz forces:  It does not happen in the CCT dipole (azimuthal forces support).  In the pure cosine-theta it can be overcome with a small interference between the collar nose and the pole turn of the coil. ◦ The elliptical deformation of the support structure under Lorentz forces:  In a CCT dipole the outer formers should hold the outwards radial forces coming from the inner layers, which complicates significantly the assembly and fabrication.  The assembly of two nested collared coils is not easy, but seems more affordable.  The CCT configuration has not been broadly used up to now, so other open questions are: ◦ The handling of the axial repulsive forces between layers. ◦ The influence of the cable positioning accuracy on the field quality. ◦ The training of a large and high field superconducting dipole. ◦ The protection of the magnet in case of quench. ◦ Formers materials to be used (insulation, stiffness and easily machining required). ◦ Coils impregnation.

28 dl R θ d=2Rcosθ I IC -I IC B OC J[A/m]=J 0 cosθ T R R h F F * * Linear Iron

29  2D Ansys Workbench model.  0.5-mm-thick shell elements at the collars.  1-mm-thick shell elements for the rest of the assembly.  Load steps.  t=0-1: Contact offset (pre-stress).  t=1-2: Assembly cooldown.  t=2-3: EM forces (exported from Maxwell, 108% Nominal current).  Convergence/stability challenges  No symmetry boundary conditions can be used. DOF more difficult to constrain.  Many parts involved and linked by contact. Frictional contacts showed better performance that frictionless ones.  Techniques used to achieve convergence:  Adding extra boundary conditions.  Tuning up contact settings at problematic zones (Stabilization dumping factor, Normal stiffness, ramped effects...).

30  Cable, wedges and inter-layer insulation: glass fibre sleeve impregnated with binder treatment as hardener (PVA to be studied).  Wedges: machined from ETP copper.  End spacers: 3D printed in stainless steel.  Ground insulation: Polyimide sheets  Vacuum impregnated coils, radiation hard resin (cyanate-ester blend).  Collars: Machined by EDM in stainless steel.  Iron: To be evaluated.  Connection plate: Hard radiation resistant composite, like Ultem.  End plates: Stainless steel.  Inner pipe: Titanium grade 2 if grade 5 is not available.

31  Customized winding machine lent by CERN ◦ New beam: 2.5 m long. ◦ Electromagnetic brake. ◦ Horizontal spool axis.  Winding process ◦ Stainless steel mandrel protected with a polyimide sheet. ◦ Binder impregnation and curation. ◦ Outer layer will be wound on top of the inner one with an intermediate glassfiber sheet for extra protection. ◦ Vacuum impregnation with hard radiation resin.

32  Collars placed around the coils with a vertical press (custom tooling required).  Layer of protection between both dipoles, likely a glass fibre sheet  Innermost turn of the coils will be protected by a stainless steel sheet from the collar nose sliding.  Iron laminations around the coil assembly.

Protection 33

34  Rutherford cable is modeled as a monolithic wire with the same metallic area, discarding the voids or internal volumes filled with resin.  The wedges are not modeled.  Quench origin is placed at the innermost turn, although it is not where the peak field is placed when both coils are powered.  A uniform magnetic field is assumed in the wires, equal to the peak field.

35

36

37

38