ITER Vacuum Vessel Supports

Transcription

1 ITER Vacuum Vessel Supports Intermediate report Contract EFDA ENEA n 07/ Andrea Capriccioli FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 1 of 49

2 Outline Abstract pag.3 1. Forces evaluation pag.4 2. VV supports actual design pag.6 2.a) Forces evaluation; 2.b) Sizes and degrees of freedom; 2.c) Toroidal restraint system; 3. Vertical Supports alternative solutions pag.11 3.a) flexible plates; 3.b) pendulum W7-X style 3.c) connection rods Ignitor style 4. Toroidal Supports alternative solutions pag.25 4.a) pendulum W7-X style.; 4.b) central guide; 4.c) flexible plates. 5. Clamping sleeves Kostyrka pag Conclusions pag.34 References Appendix 1 (AGOM pot bearings) Appendix 2 (Alternative solution: connection rods and flexible plates) pag.35 pag.36 pag.47 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 2 of 49

3 Abstract In the ITER experimental Tokamak reactor, high currents flow in the Vacuum Vessel (VV) during Plasma disruptions. The interaction between these currents and the toroidal-poloidal magnetic field produces high local forces. The direction of the electromagnetic forces acting on the VV can be upward or downward, while horizontal net force arises from the plasma tilting: their combination with the VV dead weight gives resultants both in vertical (up or downward) and horizontal directions. The VV supports have to withstand all these forces, allowing only the slow radial displacements due to the thermal expansion. Other local forces appear around the ports (discontinuities in the VV shell) during the pulse: these last forces have null resultant but produce local moments and ports distortions. The present work includes the evaluation of the reference design (see also the presentation in the VV DR meeting in Cadarache on the 28 th June 2007). As a consequence of the previous activity, alternative proposals will be evaluated: the main differences consist in bidirectional vertical supports (working in up/down directions) and in metallic components (avoiding elastomers or in general organic materials). In this case there are no problems with radiations, temperature, vacuum conditions. In the present work a first identification and a preliminary analysis of the proposed alternative solutions have been performed. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 3 of 49

4 1. Forces evaluation The total downward force on the VV supports, due to the weight, results equal to 90 MN; The electromagnetic upward forces during VDE (III) result [1]: total VDE vertical force : 60 MN (by ampl. fact. 1.1; asymmetry peak = 2 MN) total VDE net horizontal force: 20 MN (by ampl. fact. 1.7; application point 11 m) The force on the VV vertical support n 3 closest to the peak of the vertical force results MN downward due to the vertical forces only. The total net horizontal force results 20 MN, Supp. n 1 20 MN Fz1 n 2 Fz m 12 m 40 n 3 Fz m 9.96 m 13 m Fz4 n 4 Fz5 n 5 90 rotated with respect to the max vertical peak (see the figure on the right side). In this case the max vertical (z) upward contribution, due to the net horizontal force, results: 2*Fz1*12 + 2*Fz2* *Fz3* *Fz4* Fz5*13 = 20*11*1.7 Fz1=0.92 * Fz5 Fz2=0.50 * Fz5 Fz3=0.17 * Fz5 Fz4=0.77 * Fz5 2*Fz1 + 2*Fz2 = 2*Fz3 + 2*Fz4 + Fz5 Fz5 = 6.5 MN Fz3 = 0.17*6.5 MN = 1.1 MN upward. Finally, for the upward disruptions: on each n 3 supports (3+symmetric) the total force results: = 0.3 MN downward and on the n 5 support the total force results : = 4.1 MN upward FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 4 of 49

5 For the downward disruptions we have to add the dead weight and the total vertical force of 72 MN downward [1]. If the same vertical amplification and the peak factors are applied, we obtain a vertical force equal to /9 * = 20.8 MN downward. In this case the application point of the net horizontal force (25 MN) is not clear and it isn t possible to calculate its contribution to the vertical compression. A reference value of 22 MN can be assumed as first approximation. Summarizing, the VV supports must bear vertical forces from 4.1 MN upward to 22 MN downward. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 5 of 49

6 2. VV supports actual design 2.a) Materials TEFLON Teflon is the DuPont trade name (PTFE: Poli Tetra Fluorine Ethylene ) and/or Fiberslip: are material typically used when the radiation dose results lower then rad (DOE Handbook Design Considerations April 1999). In ITER a peak dose to the pad of rad is estimated (to be carefully checked). If higher doses would occur additional shields could be necessary (Teflon/Fiberslip contains Fluorine and under radiation could produce sulphuric acid if Oxygen or Hydrogen are present). NEOPRENE Neoprene is the DuPont Performance Elastomers trade name for a family of synthetic rubbers based on polychloroprene. On the market several types of rubber insulators/ supports are available (see figures next page AGOM company only for example). These kind of bearings can carry very high loads (>50 MN) and are designed for combined vertical plus horizontal loads ( 15% of the max vertical load). Rotations ±0.01 radiant are also allowed. The average contact pressure should be less than 20 N/mm2 with max value 40 MPa. The ITER VV dead weight results about 10 MN per support and a diameter of 800 mm is enough to take only this load. If the VDE downward vertical forces are considered (VDE III Slow) the compression can reaches 22 MN (see Chapter 2. Forces evaluation). In this case the diameter should be 1.5 m. In Appendix 1 the AGOM pot bearing catalog and technical data sheet are reported. Both Teflon and Neoprene could need of an active temperature control. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 6 of 49

7 [1] Design Description Document -DDD15- Vacuum Vessel - September Pag.107 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 7 of 49

8 2.b) Sizes and degrees of freedom The limit seems to be the maximum pressure ( 40 MPa) on the Neoprene: this involves large dimensions of the pot bearings. The first solution could be to double the number of supports. In this last hypothesis, the diameter of each support results close to 1 m (1.5/ 2). About the degrees of freedom, in a cylindrical reference system, the slow radial VV thermal expansion must be guaranteed; all the other displacements (toroidal, axial and fast radial too) should be restrained. The elements in order to assure this task are the VV supports: they should have a bilateral behavior (up/down effect in vertical direction). To obtain a bilateral behavior of the actual vertical supports, groups of ropes without preload were used. But the preload is necessary because, when an upward disruption occurs (4.1 MN upward), a detachment with the consequent impact between VV feet and bearing pads will increase the local pressure on the Teflon and Neoprene pads. This behavior is generally avoided using preloaded systems. In this case it s possible to preload the ropes without influence on the bearing pads; the rough sketch shows an example of possible solutions [3]: INCONEL SA-637 Gr The point is that the ropes preloading is practically equivalent to rigid rods (easier and chipper). ASTM A453 Gr A more efficient system could be the use of dumpers, instead of ropes or rods. The only note regards the absence of a vertical displacement amplification. It should be necessary to verify the pressure transient on the bearing pads during the VDEs and increase the dampers efficiency, with the same system used for the radial restraints (vert.+hor.levers see the figure below). Preloading system: some examples (wedges; thread link) FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 8 of 49

9 Other point is the radial arm configuration: Bearing Radial arm Vertical ropes Horizontal lever Vertical link Vertical lever - the radial arm is subjected to bending (at least a spherical joint is necessary); - the radial arm cooperates with the toroidal restraints against rotations around a local vertical axis. To make free the radial arm from the vertical rotations also, it s necessary a design review linking the support with the horizontal lever through two spherical joints. Dampers The first intermediate conclusion of the previous evaluation is the adequacy of the sliding bearing pads and the neoprene use to the ITER compression loads. The limit seems to be the size necessary to place the bearing pads themselves and the systems to avoid VV vertical displacements during upward VDE. A minor change regards the introduction of two spherical joints in the radial arm connection between VV and horizontal lever. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 9 of 49

10 2.c) Toroidal restraint system In the actual design a couple of toroidal restraints are added under each lower port. This configuration, where the port is placed between two elements, seems to be particularly critical. The main reason is the possibility to reach the seizing in some conditions (port thermal expansion, vertical displacements, influence on the vertical reaction forces, port rotations, etc.). The seizing of only one port can lead the whole VV to lose its central position and to increase the bending stress on the perpendicular ports. Only large tolerances could avoid the seizing probability, leaving the VV free to move radially yet. The possibility to change with continuity the toroidal stiffness of the restraint system (in this case through the vertical displacement of the VV) is an element that can be preserved with other systems, without the disadvantages shown in actual solution. The toroidal restraint systems that seem to fit well with actual geometry are both the central rail and the pendulum solutions. With both the previous systems it s possible to change the stiffness with continuity and a brief explanation will be presented in the second part of the present document (see Chapter 4). FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 10 of 49

11 3. Vertical Supports alternative solutions Three alternative solutions were analyzed, as first approximation (the radial restraint against fast horizontal forces remains necessary for all; only the introduction of two spherical joints between the radial arm and the horizontal lever is required). Flexible plates for vertical and toroidal supports The first alternative solution is the use of flexible plates both as vertical and toroidal restraint. A brief parametric analysis is shown in Tab.1 next page. The first results seem not to exclude the possibility of vertical plates with a total height lower than 1000 mm. Other structural analyses are necessary to take into account all the loads and/or displacements. The second alternative solution is the use of spherical joints both for vertical and toroidal restraint (type W7-X). Some analyses were performed, but other analyses are necessary to verify if 9 vertical supports are enough to support the vertical reference max load (22 MN downward). For the toroidal restraint with spherical joints no analyses were performed (no problems seem to be). Spherical joints for vertical and toroidal supports The third alternative is the use of Connection roads (type Ignitor) for the vertical supports and flexible plates (or central guide) for the toroidal restraint. Some analyses were performed and the results are shown in the following pages. Connection rods and flexible plates or central guides FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 11 of 49

12 3.a) flexible plates FPN FUSTEC Frascati, 10 August 2007 Tab.1 Plates thicknesses [mm] Net height (Total=Net+300 mm) N of plates / total width Max SEQV [MPa] * F=22 MN; rad displ=40 mm Radial Reaction Force [MN] full support (40 mm) 24/ / / 220 (tot: 345) / / / 269 (tot: 422) / / / 295 (tot: 454) / / / 304 (tot: 461) Common tangential length = 950 mm Gap between plates =10 mm NO VDE TANGENTIAL FORCES NO MOMENTS (around Vertical and Toroidal axes) NO Buckilng *UX >< 0 (if UX=0 stress picks are present at the 4 corners: for example ) Tick.2 Tick.1 Loaded elements FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 12 of 49 X Grounded or fixed elements

13 3.b) pendulum W7-X style The pendulum system in the W7-X reactor was possible because the vertical supports constrain only vertical displacements (downward). The W7-X Plasma Vessel is free to move on the horizontal plane, both with the previous plane sliding design and the actual spherical pendulum system. As result of this, an auto centering system (ACS) was possible as the easiest way to leave the PV free to expand, keeping its centre. It s to note that, with the selected vertical supports, a rotation of all the PV around the vertical axis, can be allowed. Another important point is the absence of electromagnetic forces on the PV: the dimensioning of the ACS was performed to balance the forces from the vertical supports friction coeff. and the elastic reaction of the connections (bellows) between PV and Cryostat. In case of not sliding vertical supports, the only free degree of freedom is the radial one (in a cylindrical ref. sys.). This means that a toroidal pendulum system (with the stiffness necessary to withstand the VDE forces) goes in contrast with the vertical ones. For this reason the toroidal pendulum system is compatible only with the spherical vertical supports (or, in general, with sliding supports). The analysis of this last type of supports is not part of the present document but a first dimensioning step is reported in the next pages. Only for example purpose, the figures below show the pendulum system (only for compression forces): Spherical joint: inner pendulum plus external shells (total height remains 780 mm) FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 13 of 49

14 Inner pendulum: the diameter of the two spherical joints is 400 mm and the height of the cylinder connecting the two half spheres, is 210mm. View of the inner spherical side of the lower pendulum support: In the following pages preliminary stress analysis results are reported. The aim of the analysis is only a first evaluation of the stresses level and the reaction friction forces. A further detailed analysis with the real maximum acceptable dimensions should be necessary. Basic point is the right definition of the friction coefficient between the two half parts of each spherical joint, to minimize the reaction friction forces: in this case the conservative value of 0.2 was set. As preliminary result, a double number of supports could be the first feasible solution. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 14 of 49

15 The max SEQV stress results of 738 MPa, with this dimensions: to decrease the maximum value it should be necessary to increase the diameter of the cylindrical connection and the total height of the joint. This last point it s also necessary to insert the other halves of spheres to give a bilateral behavior to the joint itself. The following figures show the system sliding (with a friction factor = 0.2) under the VV dead weight, as vertical load, and a radial displacements of 1 mm. The reaction force due to the friction between inner spheres and outer spherical housings results now about 3 MN (per each vertical support). In this case, the max SEQV results about 370 MPa in the cylindrical connection (see the two figures below). FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 15 of 49

16 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 16 of 49

17 3.c) connection rods Ignitor style A general support scheme, designed for Ignitor reactor, is shown in fig 1 and the cross (a) and longitudinal (b) sections are shown in fig.2: only one sector 174 mm width is visible in the section (b). The total number of sectors will be defined on the basis of the maximum vertical load and materials properties. The distance between the two rotation centres (400 mm) has been defined to have a maximum vertical displacement < 2mm during baking (40 mm max radial displacement [2]). (a) 380 mm 60 (b) 174 mm 60 fig mm φ140 mm mm 220 mm mm fig.2 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 17 of 49

18 The maximum height results of 780 mm that should be available between the pedestrian ring and the green zone of the lower port (see fig. below and the yellow rectangle that schematizes the connection rods vertical supports area). Vertical ropes Radial arm Horizontal lever Bearing Vertical link Vertical lever Dampers An ANSYS model of half sector was performed to evaluate the stresses in the components (forged external shell, connection rod, cylindrical insert). Fig. 3 next page, shows the model with the components and the main geometrical dimensions. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 18 of 49

19 fig The external forged shell is ASTM-A 453 Gr 660 while the connection rod and cylindrical inserts are in INCONEL 718. A total vertical force downward of 22 MN [2] was applied to the vertical support: this load requires 8 Nodes with applied Vertical Force modules (fig.2 shows only Symmetry plane half module). A preliminary structural analysis is shown in the following pages. Symmetry plane ANSYS FEM model: Grounded nodes FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 19 of 49

20 Material Prop.: ASTM-A 453 INCONEL 718 Sy 0.2 =589 MPa (RT) Su=893 MPa (RT) Sy 0.2 =1010 MPa (RT) Su=1246 MPa (RT) Sm = 295 MPa (RT) Sm = 451 MPa (77 K) Sm = 667 MPA (RT) Sm = 863 MPA (77 K) Vertical displacements UY (mm) in CSYS 0 Inserts displacements (x200) [mm] FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 20 of 49

21 von Mises stress (MPa) von Mises stress (MPa) FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 21 of 49

22 von Mises stress (MPa) The zone of max stress (377 MPa) is in the forged shell and (512 MPa) in the connection rod: the results seem to be acceptable compared with only (Pm+Pb) at RT <1.5 Sm = 442 MPa (ASTM-A) and 677 MPa (Inconel). The following figures show the system sliding (with a friction factor=0.2) under the VV dead weight, as vertical load, and a radial displacements of 1 mm. The reaction force due to the friction between connection rod and cylindrical inserts results about 0.84 MN (per each whole vertical support formed by 8 modules). In this case the max SEQV results of 287 MPa in the connection rod (see the two figures below). FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 22 of 49

23 Displacements USUM [mm] von Mises stress [MPa] FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 23 of 49

24 The figure on the right side shows two halves (87 mm width) and the next figure shows the single module 174 mm width. Clearly the division in modules is for calculation purpose only In the real manufacturing the forged shell (part A in fig.1) will be a single forging piece and all the parts (connection rods included) will be drilled together and at the same time n 8 units (1400 mm) FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 24 of 49

25 4. Toroidal Supports alternative solutions 4.a) pendulum W7-X style When the vertical supports allow toroidal displacements, as in actual bearing pads solution, the pendulum restraint system seems to be a really good choice. The figures below show the W7-X Auto Centering System [2]. In the W7-X reactor no electromagnetic forces act on the Vacuum Vessel and the pendulums stiffness is not a critical point. In the figure below it is possible to note the relative high L/D ratio M The system allows Vacuum Vessel vertical displacements and radial thermal expansion with small toroidal rotations of the whole VV itself. The pendulum solution for ITER reactor has to take into account several basic points: - the presence and the entity of the net horizontal force; - the presence of the radial restraint system; - during the disruption event, the opportunity to spread the reaction forces with different weights between radial and tangential ports (to minimize the stress level in the Ports-VV connection). FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 25 of 49

26 At the moment no possibilities there are to perform a detailed design and analysis of a proper toroidal restraint style W7-X. Only a scheme of pendulum with variable axial stiffness is shown in fig 4: Fig.4 It will be possible to change with continuity the axial stiffness at constant pendulum length, if the screws pitches are identical. This could be fixed on one end to the lower port and on the other end to the pedestal ring. 4.b) Central guide A sketch of toroidal restraint with only one central guide (or rail) is shown in the figures: VV Pedestal ring In this case the possibility to reach the seizing is lower than the actual design, and the same toroidal stiffness variability (with VV vertical displacement) is maintained. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 26 of 49

27 4.c) Flexible plates The flexible plates could be used for toroidal restraint 272 mm only; the scheme of the cross section is shown in fig.5: the total number of plates will be defined on the basis of the maximum toroidal load and materials properties. 300 mm 100 mm 60 mm If we consider a net horizontal force of 20 MN (VDE III slow) [1] and the fault of all the radial restraints (worst fault condition), the max tangential force occurs in the two 300 mm toroidal supports closest to the plane perpendicular to the 8 mm axis of the horiz. force. 780 mm 10 mm 100 mm fig. 5 f f Ft3 Toroidal restraint Vacuum Vessel and Ports f Ft1 Ft2 20 MN f Ft4 40 f Ft5 x In the scheme left side, the max force occurs in the toroidal support n 3 (and symmetric). In this hypothetical worst condition the max force results of 7.6 MN and 10 is the total number of plates per each toroidal support. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 27 of 49

28 The figure below show the group of 10 plates 10 mm thick (without the heads in the yellow zones: grounded and loaded parts). The gap between each plate is 8 mm and the total radial width will be about 270 mm (if other 50mm as initial and final heads radial thickness can be assigned). Loaded region 10 mm thick (tangential forces) Lateral flexible regions 10 mm thick Central flexible region 8 mm thick Grounded region 10 mm thick The figure on the right shows the ANSYS FEM model of the single plate in INCONEL: In the figure next page the single plate and its global dimensions are reported. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 28 of 49

29 1400 mm 60 mm loaded 60 mm grounded All the degree of freedom are blocked in the lower part of the plate while the upper part can slides only along the tangential (Z) axis only. No moments around vertical axis have been applied. A preliminary structural analysis of the single plate is shown below: Displacements (x10) [mm] FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 29 of 49

30 SEQV stress middle section [MPa] SEQV stress top position [MPa] FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 30 of 49

31 The stress in the middle of the shell thickness results very low = 201 MPa with a Sm = 451 MPa, while in the shell top position (membrane+bending) there are local values of 852 MPa. In these 4 nodes the high level of stress is due to a stress concentration and a factor 3 can be applied in these points for peak values (moreover the previous results have been obtained for the worst fault conditions, at RT and without proper design rules). During baking, displacements of 40 mm radially and max 2 mm vertically are expected. In this situation each toroidal support (10 plates) reacts with a vertical force of 0.62 MN and a radial one of 0.34 MN. 40 mm 2 mm SEQV top [MPa] FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 31 of 49

32 SEQV mid [MPa] The stress of the plates results mainly from bending: 580 MPa (against 60 MPa in mid sec). This value results less than 1.5 Sm = 677 MPa (at RT). Clearly the initial position of the vertical supports can be set to obtain the connection rods in vertical position during operation (VV temperature =100 C): in this case both at RT and baking temperature (200 C) the max vertical displacement results of 1mm instead of 2. In this last configuration the vertical force decreases to one half and the max stress to 530 MPa. In this first solution the global restraint system consists of vertical supports (connection rods) and toroidal supports (flexible plates): both the systems cooperate to take the vertical force (up or downward) during VDE event. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 32 of 49

33 5. Clamping sleeves Kostyrka A possible alternative to the dumpers (hydraulic or mechanical) are the clamping sleeves. The figure below shows the very simple item with the central axis and the external sleeve. When the hydraulic chamber inside the sleeve is under pressure, the elastic membrane presses on the cylinder, blocking it. During the pulse the VV remains blocked and it can expand after, completely free. For a reference diameter of 100 mm per 200 mm length, each sleeve is able to support 0.3 MN. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 33 of 49

34 6. Conclusions 1. The bearing pad solution, as Vertical support system, results acceptable for the maximum load. The limit is the large size and to double the number of supports seems to be a feasible solution (see Appendix 1). 2. The replacement of the ropes with other system (rods or dampers) doesn t involve complex design review (nevertheless further analyses are necessary). 3. The actual Toroidal restraint system shows potential risks. The proposal of the pendulum system seems to be the right choice: a further detailed analysis is recommended. The other typology (central rail and flexible plates) are feasible if in accordance with the selected vertical support system. 4. The alternative solution for the Vertical supports with connection rods seems to be achievable but the space allocation in the toroidal direction has be verified (see Appendix 2 also). 5. The solution with spherical joints (W7-X type), as Vertical supports, could be analyzed too (high reaction forces by friction) but higher vertical space seems to be necessary. To double the number of vertical supports can be another possibility. 6. The Radial restraint system (with secondary implementations) remains basically unchangeable and common to all the actual and proposed systems. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 34 of 49

35 References [1] Design Description Document - DDD15 - Vacuum Vessel - September 2004 [2] A. Cardella ITER Vacuum Vessel Support System, ITER Vacuum Vessel Design Review Working Group Meeting, Cadarache 28 June 2007 [3] A. Capriccioli ITER VV Vertical Supports, presentation at the Working Group Meeting - Cadarache 28 June 2007 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 35 of 49

36 APPENDIX 1 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 36 of 49

37 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 37 of 49

38 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 38 of 49

39 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 39 of 49

40 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 40 of 49

41 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 41 of 49

42 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 42 of 49

43 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 43 of 49

44 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 44 of 49

45 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 45 of 49

46 FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 46 of 49

47 APPENDIX 2 Alternative solution with connection rods as vertical supports and flexible plates as toroidal restraint In this solution the global restraint system consists of vertical supports (connection rods) and toroidal supports (flexible plates): both the systems cooperate to take the vertical force (up or downward) during the vertical plasma disruption event. The radial support system (radial arm, horizontal and vertical levers, dumpers etc.) has to be translated upward (through the part ocgs) to intercept the radial forces coming from the ports. In the figure below, a rough scheme shows the global system: Radial arm with spherical connection Central rail Horizontal lever VV lower ports level ocgs Pedestal ring level To explain the meaning of the central rail, it s necessary to note that, when a net horizontal force pushes the VV in its same direction, the ports bend and rotations around local vertical axes occur. As a result of that, high stress will appear in the VV-Ports connection zone (see the reinforcement proposal; Cadarache 28/06/2007). FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 47 of 49

48 The following figure shows a rough scheme of the ports distortion: F Both the proposed supports show lability against these rotations (in the actual reference design the moment should be partially balanced by only one -of the two per ports- toroidal supports and by the connection radial arm / horizontal lever). In the present alternative design, the radial arm itself and the vertical translation support have been used to insert a central guide. This guide and the toroidal support help the Ports to take the VV, decreasing both the maximum stress level of the VV-Ports connection and the VV horizontal displacement (see the fig. on the right) and avoiding F the connection rods to support dangerous rotations around a vertical axis. In this case the stiffness ratio (~20% between the toroidal and radial supports stiffness) changes, because the stress passes from the VV-Ports connection to Ports-Supports system. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 48 of 49

49 Easier than the previous scheme and with similar characteristics is the following: the only difference consists in replacing the group of flexible plates with another central guide system (icgs inner central guide system). Clearly, as in the outer central guide system, proper vertical gap will be foreseen between the parts (a) and (b). b icgs a If the guides show evidence of problems with reference to the sliding assurance, two half groups of flexible plates can be used. In this case also, other analyses are necessary to evaluate the effects of horizontal displacements induced by the ports bending, during the VDE. FPN FUSTEC doc. FUSTEC-TVV-CKSUP-001 Rev.0 page 49 of 49