1. Irfu.cea.fr ATLAS-NSW 4 TH MMM WORKSHOP THERMO-MECHANICAL SIMULATIONS FOR THE MM WEDGES (TDR OPTIONS – SCREWED QUADRUPLETS) 24-25 of June 2013 Patrick PONSOT & Patrick GRAFFIN for the CEA-Saclay-Irfu group: F.Bauer, P.Daniel-Thomas, E.Ferrer-Ribas, J.Galan, W.Gamache, A.Giganon, P-F.Giraud, S.Hassani, S.Herlant, S.Hervé, F.Jeanneau, H.LeProvost, O.Meunier, A.Peyaud, Ph.Schune

2. OUTLINE 2013/06/24-25 | PAGE 2 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  As reminder, the status presented during the MMM Saclay workshop (on the 18th of April 2013)  New results about thermo-mechanical simulations with 3 screwed modules on the spacer (tension of the mesh has been included,…)  New results about thermo-mechanical simulations with full size wedge screwed on the spacer (segmented skins ?)  Conclusion (2 parts) The design of M4 prototype before these new results Statement for the design of the quadruplets: Modifications of the prototypes ?

3. TO KEEP IN MIND ! 2013/06/24-25 | PAGE 3 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN In all cases the thermo-mechanical simulations are made with simplified modeling. The results must be considered as predictive tendencies and not as absolute values.  Material used for all the simulations: Component PCBHoneycombStructure Material FR4AluminumG10Aluminum Density (Kg/m 3 )1950129.719002770 Young modulus (MPa)1700029992400071000 Poisson’s ratio0.370.30,3 Coefficient of thermal expansion α.K -1 15x10 -6 23x10 -6 13x10 -6 23x10 -6 Thermal conductivity (W.m -1.K -1 )0,31500,3150

4. AS REMINDER 2013/06/24-25 | PAGE 4 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  Status on the end of April : Presented at Saclay during the 3 rd MMM workshop The modeling has been improved !

5. NEWS ON 2x3 SCREWED MODULES 2013/06/24-25 | PAGE 5 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  Geometry, boundary and loading conditions 2x3 modules + spacer 3 kinematic mounts on the spacer Self weight + Inclined at 0.704° + temperature gradient in Z direction ΔT=2°C The tension of the mesh is included: 15N/cm

6. NEWS ON 2x3 SCREWED MODULES 2013/06/24-25 | PAGE 6 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  2x3 modules with aluminum frames and aluminum honeycombs, only screwed on their external frame on an aluminum spacer (FR4 skins 0.5mm, no sliding around the screws) Displacement for the modules Sector 5 ~44µm ~94µm ~16µm Global 3D Only 16 microns out of the plane, if we can guarantee the full contact with the spacer (planarity)

7. NEWS ON 2x3 SCREWED MODULES 2013/06/24-25 | PAGE 7 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  2x3 modules with aluminum frames and aluminum honeycombs, only screwed on their external frame on an aluminum spacer (FR4 skins 0.5mm, no sliding for the screws ) Stresses for the modules Sector 5 The increasing of the stresses up to ~111 MPa is localized around the support (it should be solved with kinematic mounts) ~15 MPa ~111 MPa Only ~2 MPa on the skins The increasing of the stresses up to 15MPa is localized around the screws. The detailed design should take in account this effect. Higher will be the stiffness of the quadruplets, higher will be the stresses around the fixations

8. NEWS ON 2x3 SCREWED MODULES 2013/06/24-25 | PAGE 8 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  2x3 modules with aluminum frames and aluminum honeycombs, only screwed on their external frame on an aluminum spacer (FR4 skins 0.5mm, no sliding for the screws) Statement: Addition of the tension of the mesh at 15N/cm is possible without increasing of the displacements Remark: In all cases the displacement out of the plane is very dependent of the symmetry and planarity of the components (full contact between the spacer and the modules reduced the Z displacement) To keep under control the stresses, we must use the detailed design for the quadruplet according to its evolution inside the MM community New questions for MM community: With mechanical prototypes, do we want to test also the screwing of the modules on the spacer frame? Can we measure the stresses with a temperature gradient? M4

9. NEWS ON 2 FULL SIZE WEDGES 2013/06/24-25 | PAGE 9 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN Sector 5  Geometry, boundary and loading conditions 2 Full size wedges + spacer 3 kinematic mounts on the spacer Self weight, Inclined at 0.704°, temperature gradient in Z direction ΔT=2°C The tension of the mesh is included: 15N/cm Modification of the boundary conditions for the screwing (2 DoF have been added)

10. NEWS ON 2 FULL SIZE WEDGES 2013/06/24-25 | PAGE 10 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN Sector 5  Full size wedges with aluminum frames and aluminum honeycombs, only screwed on the external frame of an aluminum spacer (FR4 skins 0.5mm, with 2 DoF for the screws) Displacements ~55µm ~105µm ~37µm Global 3D 37 microns out of the plane, if we can guarantee the full contact with the spacer (planarity)

11. NEWS ON 2 FULL SIZE WEDGES 2013/06/24-25 | PAGE 11 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN Sector 5  Full size wedges with aluminum frames and aluminum honeycombs, only screwed on the external frame of an aluminum spacer (FR4 skins 0.5mm, with 2 DoF for the screws) Stresses Only ~1.5 MPa on the skins Increasing of the stresses up to ~37 MPa, is localized around the support The stresses for the skins have been decreased a lot with additional DoF on the screws.

12. WITH SEGMENTED SKINS 2013/06/24-25 | PAGE 12 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN Sector 5  Geometry, boundary and loading conditions 2 Full size wedges + spacer 3 kinematic mounts on the spacer Self weight, Inclined at 0.704°, temperature gradient in Z direction ΔT=2°C The tension of the mesh is included: 15N/cm Modification of the boundary condition for the screwing (2 DoF have been added)

13. WITH SEGMENTED SKINS 2013/06/24-25 | PAGE 13 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN Sector 5  Full size wedges with aluminum frames and aluminum honeycombs, only screwed on the external frame of an aluminum spacer (FR4 segmented skins 0.5mm, with 2 DoF for the screws) Displacement ~38µm ~86µm ~30µm Global 3D 30 microns out of the plane, but we found “unsticking” in the central part of the spacer

14. WITH SEGMENTED SKINS 2013/06/24-25 | PAGE 14 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN Sector 5  Full size wedges with aluminum frames and aluminum honeycombs, only screwed on the external frame of an aluminum spacer (FR4 segmented skins 0.5mm, with 2 DoF for the screws) Behavior of the spacer Out of plane ~30µm Global 3D ~28 MPa is localized around the supports Same behavior as the IP wedge

15. WITH SEGMENTED SKINS 2013/06/24-25 | PAGE 15 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN Sector 5  Full size wedges with aluminum frames and aluminum honeycombs, only screwed on the external frame of an aluminum spacer (FR4 segmented skins 0.5mm, with 2 DoF for the screws) Displacement of the 2 wedges ~22µm~10µm Displacement out of the plane of the 2 wedges are not the same: 10 microns compare to 22 microns (Nota: All the simulations are done without the spacers inside the drift gap) HO wedgeIP wedge

16. WITH SEGMENTED SKINS 2013/06/24-25 | PAGE 16 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN Sector 5  Full size wedges with aluminum frames and aluminum honeycombs, only screwed on the external frame of an aluminum spacer (FR4 segmented skins 0.5mm, with 2 DoF for the screws) Stresses Only ~1.2 MPa on the skins ~23 MPa on the screws Only ~4.5 MPa on the rectangular tubes which linked the segmented skins (20mmx10mmx1mm)

17. NEWS ON 2 FULL SIZE WEDGES 2013/06/24-25 | PAGE 17 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  Full size wedges with aluminum frames and aluminum honeycombs, only screwed on the external frame of an aluminum spacer (FR4 skins 0.5mm, with 2 DoF for the screws) Statement: Addition of the tension of the mesh at 15N/cm is possible without increasing of the displacements Stresses for the skins can be decreased a lot if we have more DoF around the screws (friction factor is not still taken in account?) Segmentation of the skins “introduced” different displacement for the 2 wedges. We should learn a lot of things with the mechanical prototypes (e.g. result with the spacers inside the drift gap) Remark: In all cases the displacement out of the plane is very dependent of the symmetry and planarity of the components Full size skinsSegmented skins Displacement out of plane of the quadruplets (  m) 37HO quadruplet 10 IP quadruplet 22 Stresses for the skins1.51.2 Stresses for the screws (MPa)3723

18. Only one panel - Segmented skins ? 2013/06/24-25 | PAGE 18 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  Only thermal behavior (temperature gradient on Z) for only one full size panel with aluminum frame and aluminum honeycomb, (FR4 skins 0.5mm) Difference about displacement is less than 5 microns

19. Only one panel - Segmented skins ? 2013/06/24-25 | PAGE 19 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  Only thermal behavior for only a full size panel with aluminum frame and aluminum honeycomb, (FR4 skins 0.5mm) Difference about stresses: Structures: 1,9 MPa to 2,3 MPa Skins: 0,4 MPa for both

20. CHECKING WITH SECTOR 7 (45°) 2013/06/24-25 | PAGE 20 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN Sector 7  Full size wedges with aluminum frames and aluminum honeycombs, only screwed on the external frame of an aluminum spacer (FR4 skins 0.5mm, with 2 DoF for the screws) ~1,5MPa ~37µm As for sector 5, only ~37 microns out of the plane, but we found also an “unsticking” at the center of the spacer Global ~100µm ~20µm

21. CHECKING WITH SECTOR 7 (45°) 2013/06/24-25 | PAGE 21 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  Full size wedges with aluminum frames and aluminum honeycombs, only screwed on the external frame of an aluminum spacer (FR4 segmented skins 0.5mm, with 2 DoF for the screws) Sector 7 ~1,2MPa ~30µm As for sector 5, only ~30 microns out of the plane, but we found “unsticking” on all the internal bars of the spacer Global ~75µm ~20µm

22. Before conclusion : Status of the design for M4 prototype 2013/06/24-25 | PAGE 22 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  Where we stand before this workshop ! 2 doublets and 3 drift panels made with 10mm aluminum honeycombs 3 kinematic mounts are directly fixed on the central drift panel (4 screws M6) As in the simulation, all is glued inside the quadruplet and we have spacers inside the drift gap (~40 ceramic pillars) Nota: During the production, the mesh can be dismounted and replaced (the frame of the mesh is screwed on the drift panel) A tension of 15N/cm is applied on the mesh The sealing can be tested by putting pressure up to 6mbars

23. Before conclusion : Status of the design for M4 prototype 2013/06/24-25 | PAGE 23 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  Preliminary drawings have been prepared !  But before to launch the manufacturing we need to know exactly the criteria for the qualification of the mechanical prototypes (assembly method with a spacer?)

24. CONCLUSION 1/3 2013/06/24-25 | PAGE 24 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  Statement for the design and the fixation of the quadruplets The screwing of the quadruplets on the spacer can work if we are able to provide DoF around the screws without lost of the positioning (fixed point, sliding between the quadruplets and the spacer, alignment system) Segmentation of the skins “introduced” different deformation of the IP and HO wedges. A high rigidity and planarity of the spacer are required to avoid to amplify this phenomena (Spacer made of stainless steel ?) Simulations were conservative for the deformation out of the plan (equivalent Young modulus 17GPa) but… We measured higher stiffness for the panel (+30%, see talk from Fabien, “news from Saclay”), the stresses around the fixation could be increased (tbc) Sector 5 with mesh tension and aluminum spacer 3 modules (without DoF for the screws) Full size skins (with 2 DoF for the screws) Segmented skins (with 2 DoF for the screws) Displacement out of plane of the quadruplets (  m) 1637HO quadruplet 10 IP quadruplet 22 Stresses for the skins (MPa)2 (without DoF)1.51.2 Stresses for the supports (MPa)111 (without DoF)3728

25. CONCLUSION 2/3 2013/06/24-25 | PAGE 25 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  Thermal simulations demonstrated that the main issue is the assembly of the quadruplets on the spacer Kinematic mounts have been rejected (expansive, complex, risk to lose the absolute positioning after installation…etc.) Gluing of the quadruplets on the spacer is the backup option (risk to lose a full sector in case of failure during the assembly) Screwing is possible if we have under control the displacements (positioning and alignment)

26. CONCLUSION 3/3 2013/06/24-25 | PAGE 26 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN  Parameters to have under control the displacements with screwed quadruplets All the frames inside a quadruplets should have the same displacements (alignment is only outside) : With spacers inside the drift gap, gluing of all the components (inside a quadruplet) is the best option. The local stresses around the supports and fixations must be applied only on the frames (no contact between screws and skins) A fix point is needed, but all other fixations should allow the sliding (clearance around the screws) Planarity of the spacer and quadruplet (and symmetry of the structures) is required to avoid amplification of the bending (rotation of the frames) The material of the spacer can be adapted to reduce the gap between coefficient of thermal expansion (mass, cost, …): to be checked with new simulations

27. DSM Irfu SIS/LCAP (PC N°12, Bt 123) Patrick PONSOT Commissariat à l’énergie atomique et aux énergies alternatives Centre de Saclay | 91191 Gif-sur-Yvette Cedex T. +33 (0)1 69 08 79 30 | F. +33 (0)1 69 08 89 47 Etablissement public à caractère industriel et commercial | RCS Paris B 775 685 019 2013/06/24-25 Thank you for your attention ! 4th MMM workshop - CEA-Saclay/DSM/Irfu - Patrick PONSOT & Patrick GRAFFIN | PAGE 27