- Email: [email protected]
Engineering challenges and development of the ITER Blanket System and Divertor
Fusion Engineering and Design 96–97 (2015) 34–41
Contents lists available at ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Engineering challenges and development of the ITER Blanket System and Divertor Mario Merola ∗ , Frederic Escourbiac, Alphonse Rene Raffray, Philippe Chappuis, Takeshi Hirai, Stefan Gicquel, the ITER Blanket and Divertor Sections procuring Domestic Agencies, Blanket Integrated Product Team ITER Organization, Route de Vinon sur Verdon, CS 90 046, 13067 St Paul Lez Durance, France
a r t i c l e
i n f o
Article history: Received 17 September 2014 Received in revised form 29 May 2015 Accepted 10 June 2015 Available online 23 June 2015 Keywords: Blanket Divertor
a b s t r a c t The ITER Blanket System and the Divertor are the main components which directly face the plasma. Being the first physical barrier to the plasma, they have very demanding design requirements, which include accommodating: (1) surface heat flux and neutronic volumetric heating, (2) electromagnetic loads, (3) nuclear shielding function, (4) capability of being assembled and remote-handled, (5) interfaces with other in-vessel components, and (6) high heat flux technologies and complex welded structures in the design. The main functions of the Blanket System have been substantially expanded and it has now also to provide limiting surfaces that define the plasma boundary during startup and shutdown. As regards the Divertor, the ITER Council decided in November 2013 to start the ITER operation with a full-tungsten armour in order to minimize costs and already gain operational experience with tungsten during the nonactive phase of the machine. This paper gives an overview of the design and technology qualification of the Blanket System and the Divertor. © 2015 ITER Organization. Published by Elsevier B.V. All rights reserved.
1. Introduction The ITER Blanket System and the Divertor are the main components which directly face the plasma (Fig. 1). Being the first physical barrier to the plasma, they have very demanding design requirements, which include accommodating: (1) surface heat flux and neutronic volumetric heating, (2) electromagnetic loads, (3) nuclear shielding function, (4) capability of being assembled and remote-handled, (5) interfaces with other in-vessel components, e.g. diagnostics, in-vessel coils, cabling, and (6) high heat flux technologies and complex welded structures in the design. This paper summarizes the main requirements, the design and the related R&D and technology qualification of the Blanket System and the Divertor. An overview of the integration requirements and the remote handling design and R&D can be found elsewhere [1–3]. 2. The Blanket System 2.1. Overview The main functions of the Blanket System are to:
∗ Corresponding author. Tel.: +33 442176936. E-mail address: [email protected] (M. Merola). http://dx.doi.org/10.1016/j.fusengdes.2015.06.045 0920-3796/© 2015 ITER Organization. Published by Elsevier B.V. All rights reserved.
• Constitute the primary interface to the plasma in the main chamber providing a plasma-facing surface compatible with the plasma performance requirements (heat loads, impurity influx) and a limiting surface defining the plasma boundary during limiter operation and plasma start-up/ramp-down. • Contribute in providing neutronic shielding to the Vacuum Vessel (VV) and external vessel components. • Contribute in absorbing radiation and particle heat fluxes from the plasma and from the plasma heating systems. • Provide heat and neutronic shielding to in-vessel diagnostics, such as waveguides, bolometers and in-vessel coils.
Provide passage for the plasma diagnostics, viewing systems, microwave antennas or launchers, neutral beam injectors, the gas and pellet fuelling systems and other minor ancillaries. The Blanket System consists of 440 Blanket Modules (BM) covering ∼600 m2 as illustrated in Fig. 2. A BM comprises two major components: a plasma facing First Wall (FW) panel and a Shield Block (SB). Each BM is attached to the VV through a mechanical attachment system of flexible supports and a system of key pads. Each BM has electrical straps providing electrical connection to the VV to route halo currents appropriately. Cooling water is supplied to the BM by manifolds supported on the VV behind or to the side of the SB and is designed to remove up to 736 MW of thermal power
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
35
Fig. 1. Location of the Blanket System and Divertor in ITER.
from the blanket. The coolant is routed firstly through the FW, and then through the SB. The BMs are segmented into 18 poloidal locations: rows 1–6 are the inboard region, rows 7–10 are the upper region and rows 11–18 are the outboard region [4]. The inboard and upper modules (except BM10) are segmented toroidally into 18 equal modules, and the outboard modules (except BM14 and 15) are segmented into 36 modules. In the upper and equatorial port region (BM10, 14 and 15), the modules are located between ports and therefore segmented into 18 modules. In the Neutral Beam (NB) area, vessel sectors 2, 3 and 4 have a custom segmentation for BM 14 and 15. The NB region with the neutral beam injection ports creates a particular geometry challenge for the blanket design and special modules have been developed. The NB shinethrough also needs to be accommodated by the impacted BMs. The pre-pulse and pulse coolant parameters at the Blanket inlet (that is, at the chimney bulk head) are:
Fig. 3. First Wall panel of Blanket Module 1.
Inlet operating temperature: 70 ◦ C (−5/+5 ◦ C). Inlet pressure: 4.0 MPa (−0.2/+0.6 MPa). Mass flow rate for all 440 wall mounted Blanket Modules: a minimum of 3140 kg/s. The requested measurement accuracy for the pressure, temperature and flow rate of the cooling water system is ±2% of the nominal value. The Blanket shall be baked at 240 ± 10 ◦ C by circulating hot water. During baking conditions, the nominal inlet pressure for the Blanket shall be 4.4 MPa at 10% of the total flow rate. The design of the Blanket System was carried out by the Blanket Integrated Product Team, which include the ITER Organization and six Domestic Agencies (CN, EU, KO, JA, RF and US). This effort culminated in the Blanket Final Design Review held in April 2013 and formally closed in July 2013. The successful achievement of this milestone has allowed the Blanket System to progressively move towards the construction phase. 2.2. The First Wall panels
Fig. 2. The Blanket System.
Following the ITER Design Review of 2007, the main functions of the Blanket System have been substantially expanded and it has now also to provide limiting surfaces that define the plasma boundary during startup and shutdown. This has led to a redefinition of the design heat fluxes and a shaping of the plasma facing surface to avoid the exposure of leading edges. The FW panels, which face the plasma, have become fully remote-handleable parts to ensure the possibility of replacing them in situ in case of damage [5]. The FW is shaped for minimizing the plasma heat load on edges caused by ports, diagnostic openings and assembly gaps between the FW panels. Similarly, shaping would help avoid leading edges in the case of radial misalignment of adjacent FW panels. The design of the FW panel is structured on a strong backing steel beam, oriented in the poloidal direction (as shown in Fig. 3). The beam section is typically 350 mm (toroidal) × 150 mm (radial), over the entire poloidal length of the FW. Elongated plasma facing units, called “fingers” are attached to the beam in the toroidal direction with an overhang for full coverage of the SB. The fingers are constructed using a composition of three main materials, 316L(N)IG austenitic steel for the structure, copper chromium zirconium (CuCrZr) alloy for the heat sink and beryllium as the plasma-facing material (or “armour”).
36
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
Fig. 6. EHF FW semi-prototype manufactured by the Chinese Domestic Agency.
Fig. 4. NHF FW semi-prototype manufactured by the European Domestic Agency (tested to 10,000 cycles @ 2 MW m−2 and 2000 cycles @ 2.5 MW m−2 ).
The FW panel includes one central bolt (to fix the panel onto the SB, two electrical straps, four radial pads and four X-pads (all including the related fixation features and fittings) and two water pipe connection interfaces to the SB. Connections are accessible from the front face, through a recess in the shaping that gives access to a 60 mm wide aperture. The beam is attached to the SB by a central bolt, located deep into the SB. The bolt is preloaded to 140–800 kN to accommodate disruption-induced loads, depending on the BM poloidal location. The bolt preload is reacted by four radial pads, pressing the FW panel onto the SB surface. Four X-pads are embedded into the sides of the beam for reacting the radial torque, and vertical and lateral forces. The central bolt shall require both a Low Friction Coating and an Anti-Seize Coating so that torque values can be minimized for assembly/disassembly processes and the bolt can be removed for replacement of the FW panel after several years of operation. Both bolt and pads are insulated from electrical currents by a 250 m-thick alumina layer. A dedicated qualification programme has been carried out to validate the capability of this coating to resist to the design loads during operation. The heat flux on the FW varies depending on the poloidal location, which led to the development of two types of FW panels: • “Normal Heat Flux (NHF)” panels, designed to accommodate a maximum heat flux of 2 MW m−2 . There are 218 of them and all are procured by the European Domestic Agency [6]. • “Enhanced Heat Flux (EHF)” panels, designed to accommodate a maximum heat flux of 4.7 MW m−2 . There are 222 of them, located on the inboard and outboard equatorial region and top region. 80% of them are procured by the Russian Domestic Agency [7] and 20% of them by the Chinese Domestic Agency. In the NHF panels, the water coolant flows into pipes (made of 316L steel) embedded into the heat sink, whereas in the EHF panels the heat sink is in direct contact with the cooling water and a hypervapotron geometry is used in the FW cooling channels. Relevant mock-ups and prototypes have been manufactured and tested by the three procuring Domestic Agencies up to and beyond the design flux values (Figs. 4–6).
Fig. 7. Representation of a Shield Block.
cut outs required to accommodate all the components located onto the VV including the in-vessel coils, the manifolds and the diagnostics. The steel/water ratio has been optimized with respect to neutron shielding to about 85/15. This ratio is achieved by optimizing the number and size of poloidal cooling channels within the SB. A number of deep slits are machined into the SB to reduce the impact of electromagnetic loads. As an example, a representation of SB 1 is shown in Fig. 7. The front face of the SB has a much higher nuclear heating than the rear side. The position and size of the cooling holes are optimized to achieve proper cooling, acceptable pressure drop and the desirable water/steel ratio. Poloidally drilled holes are the main cooling path in the SB. The basic fabrication method for a SB starts from a single forged stainless steel block (possibly Electro Slag Remelted) and includes deep drilling of holes, welding of the cover plates for the water headers, and machining of the interface surfaces. Full-scale prototypes have been manufactured by the Chinese and Korean procuring Domestic Agencies as illustrated in Fig. 8.
2.3. The Shield Blocks The main function of the SB is to contribute in providing nuclear shielding, to support the FW panels and to provide the connections for the supply of cooling water to the FW panel. There are also many
Fig. 5. EHF FW semi-prototype manufactured by the Russian Domestic Agency (tested to 7500 cycles @ 4.7 MW m−2 and 1500 cycles @ 5.9 MW m−2 ).
Fig. 8. SB full-scale prototype.
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
37
Fig. 9. Flexible cartridge assembly.
2.4. The Blanket Connections The Blanket Module Connections (BMCs) comprise a number of components and sub-assemblies [8]. The two main ones among them are the flexible cartridge assemblies and the electrical strap assemblies. The flexible cartridge assemblies (Fig. 9) are used to mechanically attach the BMs onto the VV and react the radial loads. There are four of them per BM and they are located at the rear side of the SB, where nuclear irradiation is lower. They are custom machined to compensate the manufacturing tolerances. They provide a rigid fixation along the radial direction while accommodating the differential thermal expansion between the BM and the VV along the poloidal and toroidal directions. The cartridge is manufactured from Alloy 718. The threaded region is installed (screwed) into the corresponding VV housing. The thread will be coated with a low friction/anti-seize coating to aid installation and any subsequent maintenance activity. The central bolt is manufactured from 660 steel. The thread is coated with low friction/anti seize coating to reduce the friction, and thus torquing requirements during installation and removal. The conical insert is manufactured from Aluminium Bronze. A locking system (such as a SpiralockTM thread) is used to properly lock the bolt during operation. This component is coated with qualified ceramic layer to electrically insulate the cartridge from the SB. The electrical strap assembly (Fig. 10) is used to control the current paths from the FW to the VV shell (two straps between FW and SB + two straps between SB and VV). It is manufactured by assembling identical copper alloy (CuCrZr) lamellas with CuCrZr interlayers and massive 316L stainless steel blocks at each extremity. The lamellas are joined together including the end blocks, thus
Fig. 11. Regular sector bundles of the Blanket Manifolds.
allowing final custom machining and ensuring good electrical contact between the different components [9]. 2.5. The Blanket Manifolds The BMs are connected to the Tokamak Cooling Water System to transfer the energy deposited on the Blanket System. The cooling water is supplied to the BMs through a piping system, the Blanket Manifolds [10]. The entire system consists of 360 cooling circuits feeding 440 Blanket Modules. There are three sizes of pipes in the Manifold – 48.3, 60.3 and 70 mm connected to individual, twin and triple modules, respectively, in order to keep the water velocity in every circuit at an acceptable level of max 7 m/s. Due to space constraints the pipes are arranged in compact bundles routed vertically through the upper ports, from the top to the bottom of the VV. The ends of the pipes are subsequently routed circumferentially, forming Branch pipes which connect with individual BMs. The simplified manifold system is shown in Fig. 11. For each regular VV sector, it consists of two inboard bundles (providing essentially identical inlets and outlets of approximately 7.5 m in length), two outboard bundles, two upper ports arrays, and connecting branch pipes through coaxial connectors to the BMs. 3. The Divertor 3.1. Overview
Fig. 10. Electrical strap assembly.
The main function of the Divertor is to control the density of both the fuel and impurity species by intercepting the magnetic field lines defining the scrape-off layer outside the last closed flux surface of the plasma, leading the plasma ions to the Divertor target plates where they are neutralized into gas which can then be pumped away by the vacuum system. At the same time, fresh fuel is injected into the plasma chamber. As a consequence, the Divertor is the component in the ITER machine with the highest design steady state heat loads. It also has to house a number of in-vessel
38
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
Fig. 12. 3D view of the Divertor cassette assembly.
diagnostics thus putting additional constraints in the design and integration. The Divertor covers the bottom part of the plasma-facing surface (210 m2 ) and has to remove up to 204 MW of thermal power via a pressurized water flow. The design comprises 54 Divertor cassettes. Each of them includes the Cassette Body (CB, procured by the European Domestic Agency) and three Plasma-Facing Components (PFCs), namely the inner vertical target (IVT, procured by the European Domestic Agency), the outer vertical target (OVT, procured by the Japanese Domestic Agency), and the dome (procured by the Russian Domestic Agency) as shown in Fig. 12. The overall assembly of the PFCs onto the CB is carried out by the European Domestic Agency and the high heat flux (HHF) performance tests and HHF tests during series production of the PFCs by the Russian Domestic Agency. The CB is a welded box structure made of austenitic steel (grade 316L(N)-IG and XM-19). The PFCs are formed by a steel support structure (also made of austenitic steel) onto which the so-called “plasma-facing units” (PFUs) are individually mounted after having been manufactured and tested separately. This strategy separates the “conventional” part of the PFC (the support structure), from the “high tech” part (the PFUs), thus minimizing the manufacturing risks and rejection rate. Similarly to the BMs, the armour is joined onto a CuCrZr heat sink, which is in turn directly actively cooled by the water coolant. Each Divertor CB is connected to the Tokamak Cooling Water System through a pair of radial pipes. The CB supports the PFCs, routes the coolant water to them, and contributes in providing neutron shielding for the VV and magnetic field coils. The PFCs are cooled in series. The IVTs and OVTs are the PFCs that, in their lower part, intercept the magnetic field lines, and therefore shall have to remove the highest heat loads coming from plasma via conduction and radiation during the steady state operation. The upper part of the IVTs and OVTs provides a baffle for neutral particles. The Dome consists in the inner and outer particle reflector plates and Umbrella. The inner and outer particle reflector plates of the Dome, together with the VTs, form a V-shaped channel root, which promotes plasma detachment. The Dome Umbrella part, which is located just below the separatrix, baffles neutrals and protects the CB and diagnostics from direct interaction with the plasma. The Divertor cassettes are fastened to two concentric toroidal rails welded to the VV. The pre-pulse and pulse coolant parameters at the Divertor inlet (that is, at the cassette pipe stubs) are: Inlet operating temperature: 70 ◦ C (−5/+5 ◦ C). Inlet pressure: 4.0 MPa (−0.2/+0.6 MPa). Mass flow rate for all 54 Divertor cassettes: a minimum of 870 kg/s.
The requested measurement accuracy for the pressure, temperature and flow rate of the cooling water system is ±2% of the nominal value. Similarly to the Blanket System, the Divertor shall be baked at 240 ± 10 ◦ C by circulating hot water. During baking conditions, the nominal inlet pressure for the Divertor shall be 4.4 MPa at 10% of the total flow rate. The coolant circuit shall also be designed to allow the bake-out of the Divertor components at 350 ◦ C by circulating gas through the cassette body and PFCs. To achieve 350 ◦ C at the Divertor surface, it is necessary to take into account the heat losses, and therefore the baking system should be designed for a gas inlet temperature of 400 ◦ C. In September 2011, the ITER Organization proposed the possible start of the ITER operation with a full-tungsten armour Divertor in order to minimize costs, via extending its lifetime into the active phase, and already gain operational experience with tungsten during the non-active phase of the machine. In endorsing this proposal, the ITER Council gave a two-year period to develop the design and the required tungsten technologies, as well as to perform the physics programme to enable an informed decision. A task force was set up to run a crash programme to fully develop the full-tungsten Divertor option within the given time frame. This has culminated with the full-tungsten Divertor Final Design Review held in June 2013 and successfully closed in October 2013. This positive outcome, as well as the encouraging results from the physics programme, has resulted in the recommendation of the proposal to start the ITER operation with a full-tungsten armour Divertor by the Scientific and Technical Advisory Committee of the ITER project in October 2013. This has paved the ground for the ITER Council to formally decide along the same direction in November 2013. Since then, the Divertor actions have been focussed in re-directing the on-going procurement activities to the new armour option, to revise the procurement specifications, update the baseline documentation and develop a full-tungsten Divertor schedule. A detailed description of the Divertor design and of the Divertor physics basis can be found elsewhere [11]. This paper will focus on the design novel aspects due to the introduction of the full-tungsten armour. 3.2. Full-tungsten Divertor design To minimize the impact on the already running procurement activities and on the interfaces with other components, the general strategy in considering the new full-tungsten Divertor design has been to avoid, where possible, modifications to the previous baseline based on the carbon-tungsten armour. The main design efforts were aimed at: (1) ensuring thermo-structural integrity, including the neutronic heating during nuclear phase and (2) ensuring complete leading edge protection to avoid melting during steady state and slow transients and to ensure minimization of the damage induced by off-normal events (essentially fast transients) [12]. A full-set of neutronic [13], electromagnetic, thermal and mechanical analysis has confirmed the first point, while the second has been achieved through a combination of global plasma-facing component tilt, the inclusion of toroidal roof shaping for the OVT upper (baffle) region and the introduction of local toroidal chamfers at the monoblock level on the strike point region of the OVT and IVT. Thermal analysis is finding that the nominal 8 mm monoblock thickness will likely need to be reduced down to 6 mm if the monoblock surface is to remain below typical recrystallization temperature (∼1300 ◦ C) under nominal peak loading conditions. According to the ITER Heat and Nuclear Load Specifications, at steady state, the time averaged incident power flux density
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
39
Fig. 13. Tilting scheme of the full-tungsten Divertor PFCs – OVT, IVT and dome.
at an unshaped, axisymmetric Divertor surface is expected to be ∼10 MW m−2 , with an approximate doubling for durations up to 10 s during slow transient events. These heat loads are assumed to be 60% conductive/convective and 40% radiative during steady state loads and conservatively 100% conductive/convective during slow transient loads. The conductive/convective heat flux flows along magnetic field lines and impacts the Divertor targets with angles  ⊥ = 2.7◦ (3.2◦ ) and Â|| = 5.6◦ (3.7◦ ) at the OVT (IVT) strike points in the 15 MA burning reference scenario. To those angle values, 0.5◦ must be added to ⊥ in the analysis due to the global tilt of the Divertor target. To evaluate the shadowing, a poloidal gap width of 0.5 mm and 3 mm is taken between adjacent PFUs and between two halves of the Divertor target, respectively. A toroidal gap between adjacent monoblocks along the coolant axis is similarly taken at 0.5 mm. As far as the radial misalignments are concerned, a maximum step of ±0.3 mm is assumed as worst case. 3.2.1. Tilting The Divertor PFCs are globally tilted when assembled onto the Cassette Body (CB), to ensure that any deviations due to assumed assembly tolerances do not create leading edges with respect to the shallow magnetic field line impact characteristic of the Divertor region (see Fig. 13). By tilting the OVT and IVT, a nominal step between adjacent PFCs (i.e. from cassette to cassette) is implemented on the straight (strike point region) parts of the targets to ensure full shadowing of leading edges taking account the assumed cassette assembly tolerances (nominal step ±2 mm). Similarly, each set of Dome PFUs (dome umbrella and inner and outer reflector plates (IRP and ORP)) has its own tilting scheme. The high (HFS) and low field sides (LFS) of the Dome umbrella are toroidally tilted in opposite directions to account for the possibility of loss of control situations in which the outer strike point falls on the LFS or the inner strike impacts the HFS of the umbrella. The IRP and ORP are also tilted along the axis aligned with the plasma-facing surfaces to protect against loss of control situations in which the strike points fall onto these plates. 3.2.2. Toroidal-roof-shaping at OVT baffle Heat loads on the OVT baffle during downward vertical displacement events (DW VDEs) in limiter-like configurations would cause very severe transient heat loads on exposed edges during full
Fig. 14. Shaping strategy for OVT.
performance phases if efforts were not made to provide component shaping. This is due to the global PFC tilting, which although hiding leading edges from PFC to PFC in the strike point areas, exposes such edges in the baffle regions. This only poses a problem during limiter-like impact, when field lines strike the baffle from both toroidal directions. The very high expected energy flux densities during the DW VDE at high plasma stored energy can cause significant damage to any leading edge, though are not expected to compromise cooling interfaces. To minimize the damage (some melting cannot be avoided if the heat flux densities are as high as predicted), toroidal roof-shaping has been adopted in the Divertor design: “set-backs” are proposed on each toroidal edge of the OVT baffle to completely shadow any direct leading edge in limiter-like configurations, while optimizing the surface heat load spreading over the remaining flux wetted surface [14]. A total of 22 PFUs are mounted on the OVT. As shown in Fig. 14, the OVT toroidal roof shaping consists of the introduction of three chamfered PFUs on the “downstream” side and a single chamfered unit on the “upstream” side. Here, upstream and downstream refer to locations with respect to plasma flow into the Divertor. This asymmetric arrangement is a consequence of the OVT tilting; the upstream side is shadowed by the OVT tilt and does not, therefore, require a large set-back (a single chamfered PFU is sufficient). Between the left and right halves of the OVT (recall that the OVT is constructed in two halves to reduce electromagnetic forces), two
40
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
Fig. 15. Illustration of the transition area in the OVT. The circle indicates a downward-facing poloidal step. (For interpretation of the references to colour in this sentence, the reader is referred to the web version of the article.)
PFUs are chamfered in order to avoid leading edges on either side of the 3 mm larger gap (compared with the 0.5 mm much smaller gap between most neighbouring PFUs). Simulations of plasma scenarios and plasma magnetohydrodynamic equilibrium, carried out using the DINA code, are unable to find unmitigated or mitigated DW VDE impacts on the inner baffle or the high field side of the DO umbrella (where a leading edge in the unfavourable direction for limiter-like impact would be introduced by the tilting). No “global” shaping is thus applied to the IVT or the DO PFUs. 3.2.3. Monoblock toroidal chamfer in strike point area To hide leading edges caused by potential misalignments due to assembly tolerances of neighbouring Plasma-Facing Units (PFUs) in the strike point area of the IVT and OVT, individual monoblock shaping is mandatory. At present, the most optimized solutions are still being examined, including detailed physics calculations of particle orbits during transients and penetrations down gaps between monoblocks. A final design which will offer the most satisfactory compromise for steady state power handling and protection against transient driven melting may be possible adopting a simple solution in which each monoblock from the strike point area is slightly chamfered only in the toroidal direction (1–2 mm). 3.2.4. Transition between target area (monoblock toroidal chamfer) and toroidal roof shaping Since individual monoblocks must be 7 mm chamfered in the baffle area to obtain the required set-back (see above), there is a requirement to ensure a proper transition between the straight part of the target and the baffle for the PFUs which form part of the toroidal roof shaping. This region is away from both the strike point and VDE impact areas so that the expected loads are much lower than elsewhere. In the transition area (Fig. 15), the monoblock thickness and chamfer angle increase gradually from the target (8.0 mm at the higher shoulder and 7.5 mm at the lower) to the baffle (8.0 mm at the lower shoulder and 15 mm at the higher). In the transition region, the poloidal steps of the monoblock thickness always face downward in order not to generate leading edges, taking into account the poloidal projection of the field line incidence angles (see red circle in Fig. 15). The monoblocks in the transition area have 12 mm axial length, with simple chamfer geometry at the plasma-facing surface. This design significantly simplifies monoblock manufacture compared to a more complicated threedimensions profile.
Fig. 16. Japanese mock-ups (armour thickness 7.7 mm; tested to 5000 cycles @ 10 MW m−2 and 1000 cycles @ 20 MW m−2 ).
(1) Technology development and validation: demonstration of the fitness-for-purpose of the proposed technology by means of small-scale tungsten monoblock mock-ups and Plasma-Facing Units support leg joint manufacturing and testing. (2) Full-scale demonstration: demonstration of the technology via full-scale-prototype manufacturing and testing. The small-scale mock-ups and full-scale prototype PFUs must withstand pre-defined HHF tests. For the small-scale mock-ups and straight portions of the vertical targets (target, strike point areas) of the full-scale-prototype PFUs, the following high heat flux testing plan is foreseen. • 5000 cycles at 10 MW m−2 • 300 cycles at 20 MW m−2 The cycle numbers have been derived on the assumption that the first set of the full-tungsten Divertor, installed from the beginning of operations, should survive at least until the end of the first full DT campaign, and are based on the estimated pulse numbers and expected additional heating levels in the H/He, DD and first full DT campaigns contained within the current ITER Research Plan. These “steady state” cycles, together with a great deal of additional information covering loading profiles, transient events, etc., has been combined in a dedicated Heat Load Specification for the full tungsten Divertor. Recent high heat flux testing of both Europeans and Japanese mock-ups has demonstrated that technology is now available on both procuring Domestic Agencies satisfying the design requirements for a Divertor intended to survive to the end of the first DT phase (Figs. 16 and 17). A large macro-crack, originating at the loaded surface and penetrating deep into the material depth was
3.3. Full-tungsten Divertor R&D A full-tungsten Divertor Qualification Programme has been developed by the ITER Organization in consultation with the procuring Domestic Agencies. This programme comprises two stages [15]:
Fig. 17. European mock-up (armour thickness 6 mm for Ansaldo (ASD) and 5.5 mm for Plansee (PLS); tested to 5000 cycles @ 10 MW m−2 and 1000 cycles @ 20 MW m−2 ).
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
occasionally found after many hundreds of cycles at the 20 MW m−2 flux density, but does not compromise the power handling capacity of the mock-ups [16]. This “self-castellation” was never observed at 10 MW m−2 nominal steady-state conditions. As also observed in past tests, self-castellation appeared in the Plansee mockup made out of forged tungsten bars. It should be noted that no selfcastellation appeared in the Ansaldo mockups made out of rolled tungsten plates (Fig. 17). Therefore, it is clear that the tungsten manufacturing process plays a key role in the fatigue performances. To this aim, a tungsten material characterization programme has recently been launched with the ultimate goal to better understand the occurrence of self-castellation and related acceptance criteria for higher performance tungsten materials. This information will be useful for the second Divertor set, to be installed after the first full DT campaign, which is aimed to survive till the end of the planned ITER operation.
Summary and conclusion An overview of the latest developments of the ITER internal components (Blanket and Divertor) has been provided. Relevant mock-ups have been successfully tested beyond the design requirements. The present level of maturity of the design and the successful completion of the qualification programme allows a progressive transition from the design to the procurement phase.
Disclaimer The views and opinions expressed herein do not necessarily reflect those of the ITER Organization.
41
References [1] A. Martin, B. Calcagno, Ph.
Contents lists available at ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Engineering challenges and development of the ITER Blanket System and Divertor Mario Merola ∗ , Frederic Escourbiac, Alphonse Rene Raffray, Philippe Chappuis, Takeshi Hirai, Stefan Gicquel, the ITER Blanket and Divertor Sections procuring Domestic Agencies, Blanket Integrated Product Team ITER Organization, Route de Vinon sur Verdon, CS 90 046, 13067 St Paul Lez Durance, France
a r t i c l e
i n f o
Article history: Received 17 September 2014 Received in revised form 29 May 2015 Accepted 10 June 2015 Available online 23 June 2015 Keywords: Blanket Divertor
a b s t r a c t The ITER Blanket System and the Divertor are the main components which directly face the plasma. Being the first physical barrier to the plasma, they have very demanding design requirements, which include accommodating: (1) surface heat flux and neutronic volumetric heating, (2) electromagnetic loads, (3) nuclear shielding function, (4) capability of being assembled and remote-handled, (5) interfaces with other in-vessel components, and (6) high heat flux technologies and complex welded structures in the design. The main functions of the Blanket System have been substantially expanded and it has now also to provide limiting surfaces that define the plasma boundary during startup and shutdown. As regards the Divertor, the ITER Council decided in November 2013 to start the ITER operation with a full-tungsten armour in order to minimize costs and already gain operational experience with tungsten during the nonactive phase of the machine. This paper gives an overview of the design and technology qualification of the Blanket System and the Divertor. © 2015 ITER Organization. Published by Elsevier B.V. All rights reserved.
1. Introduction The ITER Blanket System and the Divertor are the main components which directly face the plasma (Fig. 1). Being the first physical barrier to the plasma, they have very demanding design requirements, which include accommodating: (1) surface heat flux and neutronic volumetric heating, (2) electromagnetic loads, (3) nuclear shielding function, (4) capability of being assembled and remote-handled, (5) interfaces with other in-vessel components, e.g. diagnostics, in-vessel coils, cabling, and (6) high heat flux technologies and complex welded structures in the design. This paper summarizes the main requirements, the design and the related R&D and technology qualification of the Blanket System and the Divertor. An overview of the integration requirements and the remote handling design and R&D can be found elsewhere [1–3]. 2. The Blanket System 2.1. Overview The main functions of the Blanket System are to:
∗ Corresponding author. Tel.: +33 442176936. E-mail address: [email protected] (M. Merola). http://dx.doi.org/10.1016/j.fusengdes.2015.06.045 0920-3796/© 2015 ITER Organization. Published by Elsevier B.V. All rights reserved.
• Constitute the primary interface to the plasma in the main chamber providing a plasma-facing surface compatible with the plasma performance requirements (heat loads, impurity influx) and a limiting surface defining the plasma boundary during limiter operation and plasma start-up/ramp-down. • Contribute in providing neutronic shielding to the Vacuum Vessel (VV) and external vessel components. • Contribute in absorbing radiation and particle heat fluxes from the plasma and from the plasma heating systems. • Provide heat and neutronic shielding to in-vessel diagnostics, such as waveguides, bolometers and in-vessel coils.
Provide passage for the plasma diagnostics, viewing systems, microwave antennas or launchers, neutral beam injectors, the gas and pellet fuelling systems and other minor ancillaries. The Blanket System consists of 440 Blanket Modules (BM) covering ∼600 m2 as illustrated in Fig. 2. A BM comprises two major components: a plasma facing First Wall (FW) panel and a Shield Block (SB). Each BM is attached to the VV through a mechanical attachment system of flexible supports and a system of key pads. Each BM has electrical straps providing electrical connection to the VV to route halo currents appropriately. Cooling water is supplied to the BM by manifolds supported on the VV behind or to the side of the SB and is designed to remove up to 736 MW of thermal power
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
35
Fig. 1. Location of the Blanket System and Divertor in ITER.
from the blanket. The coolant is routed firstly through the FW, and then through the SB. The BMs are segmented into 18 poloidal locations: rows 1–6 are the inboard region, rows 7–10 are the upper region and rows 11–18 are the outboard region [4]. The inboard and upper modules (except BM10) are segmented toroidally into 18 equal modules, and the outboard modules (except BM14 and 15) are segmented into 36 modules. In the upper and equatorial port region (BM10, 14 and 15), the modules are located between ports and therefore segmented into 18 modules. In the Neutral Beam (NB) area, vessel sectors 2, 3 and 4 have a custom segmentation for BM 14 and 15. The NB region with the neutral beam injection ports creates a particular geometry challenge for the blanket design and special modules have been developed. The NB shinethrough also needs to be accommodated by the impacted BMs. The pre-pulse and pulse coolant parameters at the Blanket inlet (that is, at the chimney bulk head) are:
Fig. 3. First Wall panel of Blanket Module 1.
Inlet operating temperature: 70 ◦ C (−5/+5 ◦ C). Inlet pressure: 4.0 MPa (−0.2/+0.6 MPa). Mass flow rate for all 440 wall mounted Blanket Modules: a minimum of 3140 kg/s. The requested measurement accuracy for the pressure, temperature and flow rate of the cooling water system is ±2% of the nominal value. The Blanket shall be baked at 240 ± 10 ◦ C by circulating hot water. During baking conditions, the nominal inlet pressure for the Blanket shall be 4.4 MPa at 10% of the total flow rate. The design of the Blanket System was carried out by the Blanket Integrated Product Team, which include the ITER Organization and six Domestic Agencies (CN, EU, KO, JA, RF and US). This effort culminated in the Blanket Final Design Review held in April 2013 and formally closed in July 2013. The successful achievement of this milestone has allowed the Blanket System to progressively move towards the construction phase. 2.2. The First Wall panels
Fig. 2. The Blanket System.
Following the ITER Design Review of 2007, the main functions of the Blanket System have been substantially expanded and it has now also to provide limiting surfaces that define the plasma boundary during startup and shutdown. This has led to a redefinition of the design heat fluxes and a shaping of the plasma facing surface to avoid the exposure of leading edges. The FW panels, which face the plasma, have become fully remote-handleable parts to ensure the possibility of replacing them in situ in case of damage [5]. The FW is shaped for minimizing the plasma heat load on edges caused by ports, diagnostic openings and assembly gaps between the FW panels. Similarly, shaping would help avoid leading edges in the case of radial misalignment of adjacent FW panels. The design of the FW panel is structured on a strong backing steel beam, oriented in the poloidal direction (as shown in Fig. 3). The beam section is typically 350 mm (toroidal) × 150 mm (radial), over the entire poloidal length of the FW. Elongated plasma facing units, called “fingers” are attached to the beam in the toroidal direction with an overhang for full coverage of the SB. The fingers are constructed using a composition of three main materials, 316L(N)IG austenitic steel for the structure, copper chromium zirconium (CuCrZr) alloy for the heat sink and beryllium as the plasma-facing material (or “armour”).
36
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
Fig. 6. EHF FW semi-prototype manufactured by the Chinese Domestic Agency.
Fig. 4. NHF FW semi-prototype manufactured by the European Domestic Agency (tested to 10,000 cycles @ 2 MW m−2 and 2000 cycles @ 2.5 MW m−2 ).
The FW panel includes one central bolt (to fix the panel onto the SB, two electrical straps, four radial pads and four X-pads (all including the related fixation features and fittings) and two water pipe connection interfaces to the SB. Connections are accessible from the front face, through a recess in the shaping that gives access to a 60 mm wide aperture. The beam is attached to the SB by a central bolt, located deep into the SB. The bolt is preloaded to 140–800 kN to accommodate disruption-induced loads, depending on the BM poloidal location. The bolt preload is reacted by four radial pads, pressing the FW panel onto the SB surface. Four X-pads are embedded into the sides of the beam for reacting the radial torque, and vertical and lateral forces. The central bolt shall require both a Low Friction Coating and an Anti-Seize Coating so that torque values can be minimized for assembly/disassembly processes and the bolt can be removed for replacement of the FW panel after several years of operation. Both bolt and pads are insulated from electrical currents by a 250 m-thick alumina layer. A dedicated qualification programme has been carried out to validate the capability of this coating to resist to the design loads during operation. The heat flux on the FW varies depending on the poloidal location, which led to the development of two types of FW panels: • “Normal Heat Flux (NHF)” panels, designed to accommodate a maximum heat flux of 2 MW m−2 . There are 218 of them and all are procured by the European Domestic Agency [6]. • “Enhanced Heat Flux (EHF)” panels, designed to accommodate a maximum heat flux of 4.7 MW m−2 . There are 222 of them, located on the inboard and outboard equatorial region and top region. 80% of them are procured by the Russian Domestic Agency [7] and 20% of them by the Chinese Domestic Agency. In the NHF panels, the water coolant flows into pipes (made of 316L steel) embedded into the heat sink, whereas in the EHF panels the heat sink is in direct contact with the cooling water and a hypervapotron geometry is used in the FW cooling channels. Relevant mock-ups and prototypes have been manufactured and tested by the three procuring Domestic Agencies up to and beyond the design flux values (Figs. 4–6).
Fig. 7. Representation of a Shield Block.
cut outs required to accommodate all the components located onto the VV including the in-vessel coils, the manifolds and the diagnostics. The steel/water ratio has been optimized with respect to neutron shielding to about 85/15. This ratio is achieved by optimizing the number and size of poloidal cooling channels within the SB. A number of deep slits are machined into the SB to reduce the impact of electromagnetic loads. As an example, a representation of SB 1 is shown in Fig. 7. The front face of the SB has a much higher nuclear heating than the rear side. The position and size of the cooling holes are optimized to achieve proper cooling, acceptable pressure drop and the desirable water/steel ratio. Poloidally drilled holes are the main cooling path in the SB. The basic fabrication method for a SB starts from a single forged stainless steel block (possibly Electro Slag Remelted) and includes deep drilling of holes, welding of the cover plates for the water headers, and machining of the interface surfaces. Full-scale prototypes have been manufactured by the Chinese and Korean procuring Domestic Agencies as illustrated in Fig. 8.
2.3. The Shield Blocks The main function of the SB is to contribute in providing nuclear shielding, to support the FW panels and to provide the connections for the supply of cooling water to the FW panel. There are also many
Fig. 5. EHF FW semi-prototype manufactured by the Russian Domestic Agency (tested to 7500 cycles @ 4.7 MW m−2 and 1500 cycles @ 5.9 MW m−2 ).
Fig. 8. SB full-scale prototype.
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
37
Fig. 9. Flexible cartridge assembly.
2.4. The Blanket Connections The Blanket Module Connections (BMCs) comprise a number of components and sub-assemblies [8]. The two main ones among them are the flexible cartridge assemblies and the electrical strap assemblies. The flexible cartridge assemblies (Fig. 9) are used to mechanically attach the BMs onto the VV and react the radial loads. There are four of them per BM and they are located at the rear side of the SB, where nuclear irradiation is lower. They are custom machined to compensate the manufacturing tolerances. They provide a rigid fixation along the radial direction while accommodating the differential thermal expansion between the BM and the VV along the poloidal and toroidal directions. The cartridge is manufactured from Alloy 718. The threaded region is installed (screwed) into the corresponding VV housing. The thread will be coated with a low friction/anti-seize coating to aid installation and any subsequent maintenance activity. The central bolt is manufactured from 660 steel. The thread is coated with low friction/anti seize coating to reduce the friction, and thus torquing requirements during installation and removal. The conical insert is manufactured from Aluminium Bronze. A locking system (such as a SpiralockTM thread) is used to properly lock the bolt during operation. This component is coated with qualified ceramic layer to electrically insulate the cartridge from the SB. The electrical strap assembly (Fig. 10) is used to control the current paths from the FW to the VV shell (two straps between FW and SB + two straps between SB and VV). It is manufactured by assembling identical copper alloy (CuCrZr) lamellas with CuCrZr interlayers and massive 316L stainless steel blocks at each extremity. The lamellas are joined together including the end blocks, thus
Fig. 11. Regular sector bundles of the Blanket Manifolds.
allowing final custom machining and ensuring good electrical contact between the different components [9]. 2.5. The Blanket Manifolds The BMs are connected to the Tokamak Cooling Water System to transfer the energy deposited on the Blanket System. The cooling water is supplied to the BMs through a piping system, the Blanket Manifolds [10]. The entire system consists of 360 cooling circuits feeding 440 Blanket Modules. There are three sizes of pipes in the Manifold – 48.3, 60.3 and 70 mm connected to individual, twin and triple modules, respectively, in order to keep the water velocity in every circuit at an acceptable level of max 7 m/s. Due to space constraints the pipes are arranged in compact bundles routed vertically through the upper ports, from the top to the bottom of the VV. The ends of the pipes are subsequently routed circumferentially, forming Branch pipes which connect with individual BMs. The simplified manifold system is shown in Fig. 11. For each regular VV sector, it consists of two inboard bundles (providing essentially identical inlets and outlets of approximately 7.5 m in length), two outboard bundles, two upper ports arrays, and connecting branch pipes through coaxial connectors to the BMs. 3. The Divertor 3.1. Overview
Fig. 10. Electrical strap assembly.
The main function of the Divertor is to control the density of both the fuel and impurity species by intercepting the magnetic field lines defining the scrape-off layer outside the last closed flux surface of the plasma, leading the plasma ions to the Divertor target plates where they are neutralized into gas which can then be pumped away by the vacuum system. At the same time, fresh fuel is injected into the plasma chamber. As a consequence, the Divertor is the component in the ITER machine with the highest design steady state heat loads. It also has to house a number of in-vessel
38
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
Fig. 12. 3D view of the Divertor cassette assembly.
diagnostics thus putting additional constraints in the design and integration. The Divertor covers the bottom part of the plasma-facing surface (210 m2 ) and has to remove up to 204 MW of thermal power via a pressurized water flow. The design comprises 54 Divertor cassettes. Each of them includes the Cassette Body (CB, procured by the European Domestic Agency) and three Plasma-Facing Components (PFCs), namely the inner vertical target (IVT, procured by the European Domestic Agency), the outer vertical target (OVT, procured by the Japanese Domestic Agency), and the dome (procured by the Russian Domestic Agency) as shown in Fig. 12. The overall assembly of the PFCs onto the CB is carried out by the European Domestic Agency and the high heat flux (HHF) performance tests and HHF tests during series production of the PFCs by the Russian Domestic Agency. The CB is a welded box structure made of austenitic steel (grade 316L(N)-IG and XM-19). The PFCs are formed by a steel support structure (also made of austenitic steel) onto which the so-called “plasma-facing units” (PFUs) are individually mounted after having been manufactured and tested separately. This strategy separates the “conventional” part of the PFC (the support structure), from the “high tech” part (the PFUs), thus minimizing the manufacturing risks and rejection rate. Similarly to the BMs, the armour is joined onto a CuCrZr heat sink, which is in turn directly actively cooled by the water coolant. Each Divertor CB is connected to the Tokamak Cooling Water System through a pair of radial pipes. The CB supports the PFCs, routes the coolant water to them, and contributes in providing neutron shielding for the VV and magnetic field coils. The PFCs are cooled in series. The IVTs and OVTs are the PFCs that, in their lower part, intercept the magnetic field lines, and therefore shall have to remove the highest heat loads coming from plasma via conduction and radiation during the steady state operation. The upper part of the IVTs and OVTs provides a baffle for neutral particles. The Dome consists in the inner and outer particle reflector plates and Umbrella. The inner and outer particle reflector plates of the Dome, together with the VTs, form a V-shaped channel root, which promotes plasma detachment. The Dome Umbrella part, which is located just below the separatrix, baffles neutrals and protects the CB and diagnostics from direct interaction with the plasma. The Divertor cassettes are fastened to two concentric toroidal rails welded to the VV. The pre-pulse and pulse coolant parameters at the Divertor inlet (that is, at the cassette pipe stubs) are: Inlet operating temperature: 70 ◦ C (−5/+5 ◦ C). Inlet pressure: 4.0 MPa (−0.2/+0.6 MPa). Mass flow rate for all 54 Divertor cassettes: a minimum of 870 kg/s.
The requested measurement accuracy for the pressure, temperature and flow rate of the cooling water system is ±2% of the nominal value. Similarly to the Blanket System, the Divertor shall be baked at 240 ± 10 ◦ C by circulating hot water. During baking conditions, the nominal inlet pressure for the Divertor shall be 4.4 MPa at 10% of the total flow rate. The coolant circuit shall also be designed to allow the bake-out of the Divertor components at 350 ◦ C by circulating gas through the cassette body and PFCs. To achieve 350 ◦ C at the Divertor surface, it is necessary to take into account the heat losses, and therefore the baking system should be designed for a gas inlet temperature of 400 ◦ C. In September 2011, the ITER Organization proposed the possible start of the ITER operation with a full-tungsten armour Divertor in order to minimize costs, via extending its lifetime into the active phase, and already gain operational experience with tungsten during the non-active phase of the machine. In endorsing this proposal, the ITER Council gave a two-year period to develop the design and the required tungsten technologies, as well as to perform the physics programme to enable an informed decision. A task force was set up to run a crash programme to fully develop the full-tungsten Divertor option within the given time frame. This has culminated with the full-tungsten Divertor Final Design Review held in June 2013 and successfully closed in October 2013. This positive outcome, as well as the encouraging results from the physics programme, has resulted in the recommendation of the proposal to start the ITER operation with a full-tungsten armour Divertor by the Scientific and Technical Advisory Committee of the ITER project in October 2013. This has paved the ground for the ITER Council to formally decide along the same direction in November 2013. Since then, the Divertor actions have been focussed in re-directing the on-going procurement activities to the new armour option, to revise the procurement specifications, update the baseline documentation and develop a full-tungsten Divertor schedule. A detailed description of the Divertor design and of the Divertor physics basis can be found elsewhere [11]. This paper will focus on the design novel aspects due to the introduction of the full-tungsten armour. 3.2. Full-tungsten Divertor design To minimize the impact on the already running procurement activities and on the interfaces with other components, the general strategy in considering the new full-tungsten Divertor design has been to avoid, where possible, modifications to the previous baseline based on the carbon-tungsten armour. The main design efforts were aimed at: (1) ensuring thermo-structural integrity, including the neutronic heating during nuclear phase and (2) ensuring complete leading edge protection to avoid melting during steady state and slow transients and to ensure minimization of the damage induced by off-normal events (essentially fast transients) [12]. A full-set of neutronic [13], electromagnetic, thermal and mechanical analysis has confirmed the first point, while the second has been achieved through a combination of global plasma-facing component tilt, the inclusion of toroidal roof shaping for the OVT upper (baffle) region and the introduction of local toroidal chamfers at the monoblock level on the strike point region of the OVT and IVT. Thermal analysis is finding that the nominal 8 mm monoblock thickness will likely need to be reduced down to 6 mm if the monoblock surface is to remain below typical recrystallization temperature (∼1300 ◦ C) under nominal peak loading conditions. According to the ITER Heat and Nuclear Load Specifications, at steady state, the time averaged incident power flux density
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
39
Fig. 13. Tilting scheme of the full-tungsten Divertor PFCs – OVT, IVT and dome.
at an unshaped, axisymmetric Divertor surface is expected to be ∼10 MW m−2 , with an approximate doubling for durations up to 10 s during slow transient events. These heat loads are assumed to be 60% conductive/convective and 40% radiative during steady state loads and conservatively 100% conductive/convective during slow transient loads. The conductive/convective heat flux flows along magnetic field lines and impacts the Divertor targets with angles  ⊥ = 2.7◦ (3.2◦ ) and Â|| = 5.6◦ (3.7◦ ) at the OVT (IVT) strike points in the 15 MA burning reference scenario. To those angle values, 0.5◦ must be added to ⊥ in the analysis due to the global tilt of the Divertor target. To evaluate the shadowing, a poloidal gap width of 0.5 mm and 3 mm is taken between adjacent PFUs and between two halves of the Divertor target, respectively. A toroidal gap between adjacent monoblocks along the coolant axis is similarly taken at 0.5 mm. As far as the radial misalignments are concerned, a maximum step of ±0.3 mm is assumed as worst case. 3.2.1. Tilting The Divertor PFCs are globally tilted when assembled onto the Cassette Body (CB), to ensure that any deviations due to assumed assembly tolerances do not create leading edges with respect to the shallow magnetic field line impact characteristic of the Divertor region (see Fig. 13). By tilting the OVT and IVT, a nominal step between adjacent PFCs (i.e. from cassette to cassette) is implemented on the straight (strike point region) parts of the targets to ensure full shadowing of leading edges taking account the assumed cassette assembly tolerances (nominal step ±2 mm). Similarly, each set of Dome PFUs (dome umbrella and inner and outer reflector plates (IRP and ORP)) has its own tilting scheme. The high (HFS) and low field sides (LFS) of the Dome umbrella are toroidally tilted in opposite directions to account for the possibility of loss of control situations in which the outer strike point falls on the LFS or the inner strike impacts the HFS of the umbrella. The IRP and ORP are also tilted along the axis aligned with the plasma-facing surfaces to protect against loss of control situations in which the strike points fall onto these plates. 3.2.2. Toroidal-roof-shaping at OVT baffle Heat loads on the OVT baffle during downward vertical displacement events (DW VDEs) in limiter-like configurations would cause very severe transient heat loads on exposed edges during full
Fig. 14. Shaping strategy for OVT.
performance phases if efforts were not made to provide component shaping. This is due to the global PFC tilting, which although hiding leading edges from PFC to PFC in the strike point areas, exposes such edges in the baffle regions. This only poses a problem during limiter-like impact, when field lines strike the baffle from both toroidal directions. The very high expected energy flux densities during the DW VDE at high plasma stored energy can cause significant damage to any leading edge, though are not expected to compromise cooling interfaces. To minimize the damage (some melting cannot be avoided if the heat flux densities are as high as predicted), toroidal roof-shaping has been adopted in the Divertor design: “set-backs” are proposed on each toroidal edge of the OVT baffle to completely shadow any direct leading edge in limiter-like configurations, while optimizing the surface heat load spreading over the remaining flux wetted surface [14]. A total of 22 PFUs are mounted on the OVT. As shown in Fig. 14, the OVT toroidal roof shaping consists of the introduction of three chamfered PFUs on the “downstream” side and a single chamfered unit on the “upstream” side. Here, upstream and downstream refer to locations with respect to plasma flow into the Divertor. This asymmetric arrangement is a consequence of the OVT tilting; the upstream side is shadowed by the OVT tilt and does not, therefore, require a large set-back (a single chamfered PFU is sufficient). Between the left and right halves of the OVT (recall that the OVT is constructed in two halves to reduce electromagnetic forces), two
40
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
Fig. 15. Illustration of the transition area in the OVT. The circle indicates a downward-facing poloidal step. (For interpretation of the references to colour in this sentence, the reader is referred to the web version of the article.)
PFUs are chamfered in order to avoid leading edges on either side of the 3 mm larger gap (compared with the 0.5 mm much smaller gap between most neighbouring PFUs). Simulations of plasma scenarios and plasma magnetohydrodynamic equilibrium, carried out using the DINA code, are unable to find unmitigated or mitigated DW VDE impacts on the inner baffle or the high field side of the DO umbrella (where a leading edge in the unfavourable direction for limiter-like impact would be introduced by the tilting). No “global” shaping is thus applied to the IVT or the DO PFUs. 3.2.3. Monoblock toroidal chamfer in strike point area To hide leading edges caused by potential misalignments due to assembly tolerances of neighbouring Plasma-Facing Units (PFUs) in the strike point area of the IVT and OVT, individual monoblock shaping is mandatory. At present, the most optimized solutions are still being examined, including detailed physics calculations of particle orbits during transients and penetrations down gaps between monoblocks. A final design which will offer the most satisfactory compromise for steady state power handling and protection against transient driven melting may be possible adopting a simple solution in which each monoblock from the strike point area is slightly chamfered only in the toroidal direction (1–2 mm). 3.2.4. Transition between target area (monoblock toroidal chamfer) and toroidal roof shaping Since individual monoblocks must be 7 mm chamfered in the baffle area to obtain the required set-back (see above), there is a requirement to ensure a proper transition between the straight part of the target and the baffle for the PFUs which form part of the toroidal roof shaping. This region is away from both the strike point and VDE impact areas so that the expected loads are much lower than elsewhere. In the transition area (Fig. 15), the monoblock thickness and chamfer angle increase gradually from the target (8.0 mm at the higher shoulder and 7.5 mm at the lower) to the baffle (8.0 mm at the lower shoulder and 15 mm at the higher). In the transition region, the poloidal steps of the monoblock thickness always face downward in order not to generate leading edges, taking into account the poloidal projection of the field line incidence angles (see red circle in Fig. 15). The monoblocks in the transition area have 12 mm axial length, with simple chamfer geometry at the plasma-facing surface. This design significantly simplifies monoblock manufacture compared to a more complicated threedimensions profile.
Fig. 16. Japanese mock-ups (armour thickness 7.7 mm; tested to 5000 cycles @ 10 MW m−2 and 1000 cycles @ 20 MW m−2 ).
(1) Technology development and validation: demonstration of the fitness-for-purpose of the proposed technology by means of small-scale tungsten monoblock mock-ups and Plasma-Facing Units support leg joint manufacturing and testing. (2) Full-scale demonstration: demonstration of the technology via full-scale-prototype manufacturing and testing. The small-scale mock-ups and full-scale prototype PFUs must withstand pre-defined HHF tests. For the small-scale mock-ups and straight portions of the vertical targets (target, strike point areas) of the full-scale-prototype PFUs, the following high heat flux testing plan is foreseen. • 5000 cycles at 10 MW m−2 • 300 cycles at 20 MW m−2 The cycle numbers have been derived on the assumption that the first set of the full-tungsten Divertor, installed from the beginning of operations, should survive at least until the end of the first full DT campaign, and are based on the estimated pulse numbers and expected additional heating levels in the H/He, DD and first full DT campaigns contained within the current ITER Research Plan. These “steady state” cycles, together with a great deal of additional information covering loading profiles, transient events, etc., has been combined in a dedicated Heat Load Specification for the full tungsten Divertor. Recent high heat flux testing of both Europeans and Japanese mock-ups has demonstrated that technology is now available on both procuring Domestic Agencies satisfying the design requirements for a Divertor intended to survive to the end of the first DT phase (Figs. 16 and 17). A large macro-crack, originating at the loaded surface and penetrating deep into the material depth was
3.3. Full-tungsten Divertor R&D A full-tungsten Divertor Qualification Programme has been developed by the ITER Organization in consultation with the procuring Domestic Agencies. This programme comprises two stages [15]:
Fig. 17. European mock-up (armour thickness 6 mm for Ansaldo (ASD) and 5.5 mm for Plansee (PLS); tested to 5000 cycles @ 10 MW m−2 and 1000 cycles @ 20 MW m−2 ).
M. Merola et al. / Fusion Engineering and Design 96–97 (2015) 34–41
occasionally found after many hundreds of cycles at the 20 MW m−2 flux density, but does not compromise the power handling capacity of the mock-ups [16]. This “self-castellation” was never observed at 10 MW m−2 nominal steady-state conditions. As also observed in past tests, self-castellation appeared in the Plansee mockup made out of forged tungsten bars. It should be noted that no selfcastellation appeared in the Ansaldo mockups made out of rolled tungsten plates (Fig. 17). Therefore, it is clear that the tungsten manufacturing process plays a key role in the fatigue performances. To this aim, a tungsten material characterization programme has recently been launched with the ultimate goal to better understand the occurrence of self-castellation and related acceptance criteria for higher performance tungsten materials. This information will be useful for the second Divertor set, to be installed after the first full DT campaign, which is aimed to survive till the end of the planned ITER operation.
Summary and conclusion An overview of the latest developments of the ITER internal components (Blanket and Divertor) has been provided. Relevant mock-ups have been successfully tested beyond the design requirements. The present level of maturity of the design and the successful completion of the qualification programme allows a progressive transition from the design to the procurement phase.
Disclaimer The views and opinions expressed herein do not necessarily reflect those of the ITER Organization.
41
References [1] A. Martin, B. Calcagno, Ph.