5 MAST 5.1 MAST OPERATIONS

Transcription

1 5.1 MAST OPERATIONS INTRODUCTION The Mega Amp Spherical Tokamak (MAST), which began operation in 2000, is designed to study high temperature, low aspect ratio, highly elongated (κ > 2) plasmas. A key feature of spherical tokamaks (STs) is their low aspect ratio (the ratio of major radius R, to minor radius a) which enables, simultaneously, high plasma stability and good energy confinement. The high natural plasma elongation and strong plasma shaping provide a large plasma current carrying capability, allowing high performance operation at relatively low toroidal magnetic field. The mission of MAST is: to explore the long term potential of the spherical tokamak as a fusion component test facility (CTF) and/or ST power plant; to advance key tokamak physics for optimal exploitation of ITER and DEMO design optimization; to provide unique insight into underlying tokamak physics. Many U.K. universities and international collaborators are involved in the MAST programme. Indeed, experiments are often led by international collaborators, university researchers and/or PhD students. The MAST control room is equipped with remote participation facilities. There is strong coupling between theoretical and experimental activities and a wide range of sophisticated codes is available for planning and interpretation of experiments. Since 2007, the MAST programme has been guided by an international Programme Advisory Committee (PAC). In the longer term, a major upgrade to MAST is planned, the first stage of which will be completed by Figure 5.1: MAST. 5.1

2 5.1.2 MAST PARAMETERS The poloidal cross-section of the MAST plasma, and its current carrying capability (> 1MA), are similar to that of major conventional tokamaks such as ASDEX Upgrade in Germany and DIII-D in the U.S.A. The MAST load assembly comprises a large cylindrical vacuum vessel (height 4.4m, diameter 4m) with a removable centre column and internal poloidal field coils. This is a very adaptable arrangement that allows considerable flexibility over plasma and divertor configurations, enables enhancements to be more easily implemented, offers outstanding diagnostic access and plasma imaging opportunities, and allows innovative plasma start-up techniques to be deployed. High purity plasma conditions are ensured by high temperature baking, periodic boronisation and inter-shot helium glow discharge cleaning. MAST is equipped with two long pulse, high power, neutral beam injectors (NBI) and a high power 28GHz microwave system in the electron cyclotron frequency range (collaboration with Oak Ridge National Laboratory, U.S.A), for plasma start-up studies. MAST benefits from a digital plasma control system and a range of plasma fuelling techniques, including a cryogenic pellet injector. In addition, a fast gas valve is available for disruption mitigation studies (collaboration with FZJ Germany). There are external coils for error field compensation and internal coil arrays for both control of edge localised modes (ELMs) and controlled excitation of Toroidal Alfvén Eigenmode (TAE) instabilities. A divertor science facility enhances the capability to conduct controlled investigations of plasma-material interactions. The main MAST parameters are summarised in Table 5.1. Design Achieved Minor & Major radii a, R (m) 0.65, , 0.85 Elongation κ Minimum aspect ratio A Plasma Current I p (MA) Toroidal Field B φ0 (T) at R Neutral Beam Heating P NBI (MW) Pulse length (s) Table 5.1: MAST parameters. MAST is particularly renowned for its comprehensive array of advanced, and in some cases world leading, diagnostics including: very high temporal and spatial resolution ( ~ ion Larmor radius ~1cm) kinetic diagnostics (Thomson scattering, charge exchange recombination spectroscopy, Motional Stark Effect (MSE) etc.); 2D beam emission spectroscopy; a fast ion deuterium-alpha (FIDA) system; a collimated neutron detector array; a synthetic aperture microwave imaging system; a fast edge Doppler spectroscopy system; visible and infra-red imaging diagnostics; magnetic measurements up to 5MHz; retarding field energy analyzers and a wide range of electrical probes. 5.2

3 5.1.3 RECENT TECHNICAL DEVELOPMENTS MAST experiments in were able to exploit a large number of new technical developments. These recent enhancements include: Installation of additional ELM control coils; Upgrade of neutral beam control and instrumentation systems; Upgrade of the plasma control system (PCS); Installation of a controllable high-field-side gas fuelling system; Installation of a spark gap impurity injector (with Dublin City University); Improvements to the plasma disruption mitigation system (with FZ Jülich); Improvements to the Toroidal Alfvén Eigenmode (TAE) excitation system; Development of event triggering systems based on Field Programmable Gate Arrays (FPGA) technology; Diagnostic developments, viz., o Collimated neutron detector (4 chords) with scanning capability (with Uppsala University) o Fast ion D-alpha (FIDA) system o Synthetic Aperture Microwave Imaging System (with University of York) o 2D Beam Emission Spectroscopy (BES) system (with RMKI Hungary) o Edge Doppler spectroscopy upgrade to enable high frequency measurements o Retarding Field Energy Analyzers (with Liverpool University). The additional 6 ELM control coils, which bring to 18 the total number of coils installed (6 in the upper row and 12 in the lower row) has allowed higher toroidal mode number (n=4 and n=6) magnetic perturbations to be applied and also allowed improved alignment of the perturbations with the plasma equilibrium. The upgrades to the neutral beam and plasma control systems have significantly improved operational reliability and in the latter case provided a suitable platform for further development of real time plasma shape control using the rtefit code (real-time equilibrium reconstruction). The disruption mitigation system, which utilises a fast gas valve on loan from FZJ Germany, was modified to allow higher pressure operation and a wider range of gases to be injected in the campaign. The Toroidal Alfvén Eigenmode (TAE) system, which is used to excite TAE modes in the plasma in order to study their damping, has been upgraded to maximise the operational limits of the drive circuit and increase detection capabilities. A new FPGA generator provides a flexible drive waveform allowing sweeping of the excitation frequency, predistortion of the waveform to compensate for cross-over distortion of the linear class AB amplifiers, and the possibility of gating the excitation. Furthermore, an order of magnitude increase in excitation current has also been achieved. The upgraded system now delivers an RMS current in excess of 10A over the TAE frequency range, kHz, by means of a resonant-transformer matching network which is tuned to a particular frequency with a FWHM of ±30kHz around the peak frequency at maximum amplifier power. The response of the plasma is measured using a selection of magnetic coils, including the high frequency 5.3

4 Mirnov coils dedicated to the measurement of the fluctuations in the range 100kHz-4MHz. Sophisticated methods were developed to resolve small plasma responses from the direct drive signal by (1) gating amplifier signals (at khz rates) through the pre-programming of the FPGA generator and (2) critically damping the drive signals in under 1μs with switched snubber circuits linked to the matching networks (Figure 5.2). Gating Drive Response Figure 5.2: Critically damped TAE drive signal (upper trace) and response of an outboard Mirnov coil (lower trace) when a neighbouring TAE coil is set up to be resonant (no plasma). In the presence of plasma the damping circuit enables the plasma response to be measured in the absence of the much larger drive signal which would otherwise dominate. During the present reporting period, there has been a diversification in the use of Field Programmable Gate Arrays (FPGAs) for diagnostics signal processing and real time feedback as well as for control instruments: A device using an FPGA is used to determine the phase and amplitude of Neoclassical Tearing Modes (NTMs) from Mirnov coil signals and has been developed to provide feedback signals to the Plasma Control System (PCS). When the NTM amplitude reaches a certain threshold, PCS adjusts the plasma vertical position causing the plasma to transition from H-mode to L-mode and the NTM amplitude to reduce. Once the NTM amplitude decays, a further signal is given to PCS to adjust the plasma vertical position restoring H-mode operation. In this way early plasma termination can be avoided; The capabilities of the FPGA based TAE signal generator (see above) were extended to allow more complex excitation waveforms (pre-distorted and square wave amplitude modulated to measure damping times); 5.4

5 A scalar/timer is under development to determine CCD frame synchronisation with respect to the MAST clock (the generic design also allowing applications as a counter scalar to replace obsolete CAMAC devices); A fast timer has been prototyped interfacing via Modbus to machine control PLCs to provide the timing triggers to the various coil pulsed power supplies; A vertical feedback system is also under development to provide non-linear control of the plasma position. All these systems adopt a common architecture, using a processor embedded on the FPGA running linux, allowing the control interface to be standardised. As more groups become interested in exploiting FPGAs for diverse applications (NBI, power supplies, machine protection, machine control, diagnostics), a training course provided through the Xilinx University Program has been planned for May 2012 to raise the capabilities of CCFE staff in the area of FPGA programming. The new spark gap impurity injector (Figure 5.3), developed in collaboration with Dublin City University, was successfully deployed in the MAST divertor science facility to inject carbon into the plasma scrape-off layer to study impurity transport. Langmuir probe Electrodes Figure 5.3: Spark gap impurity injector (developed in collaboration with Dublin City University). The diagnostic developments listed above have significantly improved our capability to study (see Chapter 4) plasma turbulence (beam emission spectroscopy), fast ion behaviour (neutron camera and FIDA), pedestal physics (edge Doppler spectroscopy and microwave imaging) and plasma exhaust (retarding field energy analyzers). All these new diagnostics provided important new data in The synthetic aperture microwave imaging system provides information on the edge plasma current density by measurement of electron Bernstein wave (EBW) emission from the plasma. However, it can also be used in active mode to generate velocity maps of the plasma turbulence. This is achieved by launching a probing signal into the plasma and measuring the Doppler shift of the back-scattered signal MAST OPERATIONS Despite a number of early technical problems the experimental campaign, during which there were approximately 3000 tokamak pulses, was one 5.5

6 of the most successful and productive MAST campaigns to date. All new and upgraded plant and diagnostics were successfully deployed and significant progress was made on all high priority physics experiments (Chapter 4). The physics campaign was completed in January In parallel with physics studies, engineering tests were also carried out in support of the MAST Upgrade programme. For example, a modified design of toroidal field (TF) coil sliding joint was installed and tested on MAST. Tests were also carried out on the neutral beam getter pump systems to determine their suitability for longer pulse operation on MAST Upgrade. The 60GHz high power microwave plant, originally procured for the COMPASS-D project and subsequently deployed on MAST was decommissioned during the last year; major components are being shipped to Huazhong University in China where they will be installed on J-TEXT. At the end of the physics campaign, the MAST vessel was vented to atmospheric pressure in order to carry out essential diagnostic maintenance. This was followed by a period of testing and conditioning of the neutral beam injection systems. This endeavour exploited recent improvements, including a new fast data acquisition system, and was very successful. Both injectors have been operated at powers up to ~ 2.2MW and operation of one of the injectors at full design voltage (75kV) has now been demonstrated. The primary challenge remaining is to increase the beam current further. Since the transmission line snubbers were upgraded on both injectors, there have not been any major deconditioning events. At the MAST Research Forum in March 2012, physics progress was reviewed and forward plans were discussed with international collaborators and UK university partners FORWARD PLANS The final MAST experimental campaign before the device is upgraded significantly (see below) will run from November 2012 until the Summer Prior to this campaign, essential power supply maintenance is being carried out and additional MAST Upgrade type TF coil sliding joints will be installed for testing. The high power 28GHz microwave plant is being refurbished and recommissioned in preparation for electron Bernstein wave assisted plasma startup experiments which will be carried out in collaboration with international partners from the U.S.A. and Japan. The magnetic configuration of one of the neutral beam sources will be changed to a so-called chequerboard configuration. The second source will retain a supercusp configuration, allowing a direct comparison of the two configurations in order to identify the optimum configuration for the sources installed on MAST Upgrade. We will also implement a beam power notching system which will enable the beam power to be removed for a selectable duration in response to a pre-programmed signal or a real-time trigger in response to a plasma event. The results of tests on this system will be used to optimise the design of a more sophisticated real-time neutral beam power control system for MAST upgrade. Diagnostic plans include tests of coherence imaging techniques for measuring divertor flows on MAST in collaboration with Durham University and ANU (Australian National University) and tests of a proton detector in collaboration with Florida State University. 5.6

7 5.2 MAST UPGRADE PROJECT As discussed in the 20010/11 Annual Report, the Research Councils UK (RCUK) 20 year review of the UK Fusion Programme endorsed the Stage 1 for the MAST-Upgrade programme, which would be capable of providing exacting studies of ITER physics and DEMO issues, and would represent a major step forward in addressing the physics validation issues for a Spherical Tokamak Component Test Facility (ST-CTF). In December 2009, EPSRC decided that an extra 20M of funding should be provided for implementation of the Upgrade, and the CCFE Executive agreed to find the remaining 11M to fund the project to completion of Stage 1 in April 2015 from savings in the research programme. The Stage 1 project was subsequently launched formally in July Early in 2011/12 concerns arose that the agreed scope of the project required more funds than were available. The reasons for this were a combination of budget estimates from industry being higher than the costs assumed in the original project planning process and further elaboration of the planning showing that the manpower estimates, especially those needed for the Shutdown to install the upgrade, were too low. A unique feature of MAST-U, with high international impact, is its ability to test innovative divertor concepts including the Super-X. These concepts could play a critical role in reducing the power density of the exhaust from a DEMO or CTF device to within the anticipated engineering limits of divertor components. The innovative divertor concepts will also allow different regimes to be tested against models of divertor behavior, which should lead to a better understanding of the complex, interacting physics issues and perhaps ultimately improve the predictive capability of the models. Therefore it was essential that any re-phasing of the MAST-U project retained in its initial phase the ability to explore effectively the physics of innovative divertors. Following an internal assessment of options and reviews by the MAST Programme Advisory Committee and the Fusion Advisory Board, it was agreed with EPSRC that the implementation of the third beamline and cryoplant should be deferred until sufficient funds are available. In addition it will be necessary to minimize the movement of the existing beamlines and diagnostics to reduce the amount of effort needed during the Shutdown. The intent remains to implement as much as possible of the original scope using the existing funds and to complete the MAST Upgrade Programme as soon as additional funding become available. The estimates obtained from industry for the main components required showed significant scatter and therefore the priority this year has been to issue tenders for these components so that firm quotations from industry can be obtained to reduce the residual uncertainty in the budget. 5.7

8 5.2.1 MAST-U LOAD ASSEMBLY A cross-section of the MAST-U load assembly is shown in Fig: 5.1. Figure 5.1: Cross section of MAST after the Upgrade. Many of the present MAST machine systems will be retained after the Upgrade but the centre column assembly will be replaced in order to deliver the required enhanced performance (i.e. increased toroidal field strength and plasma duration) and most of the in-vacuum components will be changed to implement the coils and protection tiles needed for the new divertor configuration. The recent change of project scope has not significantly affected the work to be done because the target remains to implement all of the originally planned major changes to the load assembly in the MAST-U Shutdown to minimise the number and duration of future interventions. There remain some minor uncertainties regarding the internal ELM coils but these will be resolved shortly. The need to accelerate progress to meet the agreed spending profile for the project has required changes to the planning of load assembly activities so that some items will be on-site at least 6 months earlier than was necessary to meet the original assembly programme and consequently a lot of design work had to be brought forward. This will have the beneficial effect of reducing the risk that slippage on manufacturing contracts could adversely affect the Shutdown activities. It will also make it easier to undertake pre-assembly trials of the components in advance of the Shutdown to identify any problems and to allow review of the installation procedures, tooling and training requirements. However, this acceleration of procurements has increased the level of resources needed by the Load Assembly activities in the next year. 5.8

9 A Centre column and sliding joints The upgrade requires a completely new centre column assembly, which comprises the central part of the toroidal field (TF) coil, a higher flux solenoid and three new coils (the Pc coil and a pair of Px coils) to improve plasma shape control. The Pc coil is located on a tube which sits within a central hole of the TF conductor assembly. Unlike the present MAST design, the solenoid will no longer be captive on the TF conductor assembly and hence the two components can be manufactured concurrently. A major element of the new machine design has been designing the central part of the TF coil and sliding joints to be compatible with the higher current in the TF coil and the longer pulse length. The sliding joint design has been optimised to minimise the peaking of the current density at the end closest to the mid-plane of the machine and hence to minimise the temperature peaking. The final design of the sliding joint is shown in Fig 5.2. The designs now meet the requirements specifications provided the sliding joint resistance is no worse that the presently expected values. Feltmetal TM joint material Cut-out in copper body to optimise current density distribution on joint region Figure 5.2: Detailed design of the MAST Upgrade sliding joints. The conductors forming the central part of the upgraded TF coil are subject to a torque due to the radial component of the solenoid magnetic field at its ends. The ends of the rod are prevented from twisting by a crown arrangement which features a castellated joint keyed into the ends of the TF central conductors (see Fig 5.3). The crown is connected via a torsion ring to the vacuum vessel thus acting as a torsional restrain. The stresses in these components will be much higher than in the present MAST design. Optimisation of the design in this region has necessitated extensive non-linear finite element analysis to find a satisfactory solution. To minimise thermal stresses it has been found necessary to pre-cool the end crowns as well as the central TF conductors to minimise temperature differentials through the duration of the pulse. This has led to some changes in the cooling scheme. The final refinements of the design of the TF centre 5.9

10 conductor assembly, which is now on order, will take account of feedback from the Supplier, particularly regarding the final optimisation of the electrical insulation at the end of the rod between the copper conductors and the stainless steel castellated rings. Figure 5.3: Torsion ring & crown arrangement at the end of the centre rod (applies at both ends but only one shown above). The region at the centre of the top and bottom of the central region of the load assembly is congested and has required extensive optimisation of the designs of the various components that have to fit in the space available. This is now complete and has allowed the route of the tails of the central solenoid, Px and Pc coil tails to be finalised, thus enabling the tender for the these coils to be issued. B Solenoid Despite extensive optimisations, the design analysis of the new central solenoid has shown that the conductor will need to be operated very close to the plastic region (with some local plasticity actually occurring) to achieve the specified flux swing. The operational life of the solenoid is predicted to meet requirements but the assumed material properties will need to be achieved by the conductor used for manufacture; the conductor supplier has agreed to optimise their processes to deliver the conductor with the required mechanical properties. Given the rapid reduction in fatigue life as the full performance of the solenoid is approached, operation at full design performance will be limited to 3,000 pulses but in addition >20,000 pulses will be possible at 90% of full performance. Operation at full solenoid performance will be reserved for experiments that require it and the machine protection control arrangements will enforce the agreed limits for each experiment. C Divertor coils Seven pairs of new in-vacuum coils are needed to provide the conventional and Super-X modes of divertor in MAST-U. The coils will be enclosed in stainless steel vacuum cans following vacuum impregnation of the water-cooled copper conductors and the gap between the coil and the can will then be filled with resin to form a rigid mechanical assembly. The coils will be supported within the MAST vacuum vessel using flexible supports to accommodate the differential thermal movement between the coils and the vacuum vessel both during operation and when the vacuum vessel is baked. 5.10

11 The specification of the required current waveforms of these coils has been finalised by supplementing previous work with new analysis using thousands of examples of plasma equilibria (magnetic field geometries). This work gave useful insight into the mode shapes that can be used to control the plasma. The flux plots show the shape of an underlying Super-X equilibrium (Fig. 5.4a), and a shape that can be used to control the amount of the scrape-off layer that passes into the divertor chambers (Fig. 5.4b). This work demonstrated that in some circumstances the required current waveforms would lead to greater heating of some of the coils than previously specified. Consequently the size of the conductors used has been revised where needed to achieve the required levels of performance. Figure 5.4: Contours of poloidal magnetic flux in Webers. The left hand figure (a) shows a Super-X type equilibrium. The right hand figure (b) represents the control field used to expand or compress the scrape-off-layer and divertor flux. The scheme design reviews of the divertor coils have been completed and a contract has been placed for their manufacture. The detailed design review awaits final discussions with the supplier regarding their preferred approach to some of the manufacturing issues. D Vertical stability Following the upgrade, plasmas having somewhat higher elongation than presently used on MAST will be explored since these give better performance and ultimately would allow demonstration of plasma configurations akin to those desired in the CCFE ST-CTF concept. These plasmas will be more vertically unstable and as a result passive stabilising plates will be required to minimise the reactive power requirement of the associated feedback amplifier. Over the last year modelling has demonstrated the benefit of increasing the number and size of these stabilisation plates. However, incorporating this into the design is challenging because of potential interference with the existing ELM/TAE coils and neutral beam path; the design is being refined to strike the optimal balance 5.11

12 between these. The design of the fast acting radial field coils (the replacement of the existing stainless steel canned MAST P6 coils) will be finalised when the passive stabilisation plate locations are agreed but the original concept of cablein-pipe with a low number of turns is still assumed. E Cryopumps Part of the ultimate scope of the upgrade of MAST is to introduce a strongly pumped divertor with a Super-X capability using a liquid-helium cooled cryopump to achieve sustained high pumping speed with need for only infrequent regenerations. The Cryoplant design and tender package are complete but the contract has not been tendered because of the reduced initial project scope. However, the cryopumps will be installed in the MAST-U Shutdown because they cannot be easily installed later. A range of geometries have been analysed to optimise pumping speed, cryogenic heat loads and operating temperature. The pumping surface will be bare stainless steel pipes cooled by supercritical helium. For optimum performance, the baffles of the radiation screen will be made of copper. F Research and development activities The centre rod, solenoid, Px and Pc coils for MAST-U require cyanate ester resin to be used for the impregnation to achieve the required strength at the high temperatures necessary to meet the performance requirements. A research and development programme was started in 2011 to provide the data needed for the design and to develop sufficient experience of what can be achieved by industry to guide the design decisions. Experimental data is being acquired to support and demonstrate compliance is such areas as bond strength, performance at higher temperatures, fatigue life and crack propagation. In the past year numerous research and development programs have been initiated and some completed. Figure 5.5: Failure mode under cyclic loads for cyanate ester to copper bond. 5.12

13 Samples to test the fatigue life of the bond of cyanate ester resin to copper and to stainless steel have been prepared by the resin manufacturer (see Figure 5.5 for a tested specimen). The performance of the resin to copper bond has, following improvements, achieved consistent results at the level required. However, the fatigue testing of the cyanate ester to stainless steel bond has shown poor results. The reason is thought to be control of the alignment of the samples and insufficient roughening of the stainless steel surface prior to bonding. Improvements to the manufacturing process have started to address these issues. Tests have also been performed to quantify the expected crack energy release rates for cyanate ester resin to copper bonds (see Figure 5.6). This information has been used in modelling and predicting the rate of crack propagation and hence to confirm that the required operational life of the coils will be achieved. Figure 5.6: Typical Double Cantilever Beam (DCB) specimen for crack energy release rate measurements. Two sliding joints incorporating the key features of the new MAST-U design have been installed on MAST. These confirm (within the unavoidable statistical scatter) that the joint resistance is lower than the original joint design, as assumed in the MAST-U design. A further 4 new joints will be installed in 2012 to improve the statistics of the resistance measurements. However, the limits on the I 2 t specification of the coil and sliding joints cannot be certain until operations after the upgrade of MAST commence when the resistance of the sliding joints can be confirmed. A testing rig has been built to allow the selection of material from batches of Feltmetal TM sheets having the best properties for the sliding joints. This is expected to reduce the statistical scatter in joint resistance, which is important because the overall performance possible is determined by the joint having the highest resistance. 5.13

14 Figure 5.7: Section of the coil housing manufactured with the spinning process (a lower and an upper section welded together and subsequently sectioned shown above). Figure 5.8: Upper section of a coil housing manufactured with the explosion forming process. A number of prototypes of the 316L stainless steel housings for the new in-vessel coils have been made to compare different manufacturing techniques. These have shown that it would be possible to form the two halves of the can by spinning stainless steel sheet over a former, as was used when manufacturing the existing MAST coils (see Figure 5.7). However, the preferred approach is to 5.14

15 use explosive forming of coil housings (see Figure 5.8) because of lower cost and better achieved overall tolerances. The plastic deformation involved in both manufacturing methods increases the magnetic permeability of the stainless steel above the maximum acceptable level but subsequent heat treatment at 900 C has proven sufficient to reduce the permeability to very low levels POWER SUPPLIES A contract has been placed for the new 133kA TF power supply which comprises a 24-pulse rectifier employing four cast resin transformers connected to air cooled thyristor bridges with their outputs wired in parallel through air-cored filter inductors to ensure effective sharing of current and fault current limiting. The power supply will deliver the required maximum current rise time of 1s and under normal operating conditions deliver a 2s flat top well within the thermal rating of the TF coil. To minimise the supplier s work on site the rectifier equipment and switchgear will be delivered, having undergone factory testing, housed in weatherproof shelters of welded-steel construction. A separate contract has been placed for the 8 individual 4 quadrant power conversion units needed for the divertor and Px coils. The DC power to feed these convertors will be obtained by using the existing power supply that currently powers the MAST TF coils; this will need modification of its control system to allow voltage rather than current control. The output current of the convertor units is matched to the specific requirements of each coil (ranging from 10kA to 4kA). The stringent divertor physics requirements to reduce voltage ripple at the coils to a maximum of 69mVs on each switching cycle has also been exceeded for all converters. Each unit consists of multiple H bridge inverters in parallel, capacitors, chokes, a crowbar, local controls etc. as illustrated below. By varying the number of parallel H-bridges depending on the output current required it has been possible to standardise the semiconductor and other components, thereby minimising the number of spares required. The units will be supplied in standard cubicles and factory tested before delivery. Figure 5.9: Preliminary layout for a 10kA Divertor power supply unit. 5.15

16 5.2.3 NEUTRAL BEAM SYSTEM The changes to the MAST-U project scope involve deferring the originally planned third beam line. Consequently initial operations of MAST after the upgrade will use the existing two beamlines. The existing beam line at Sector 6 will remain in exactly the same position as it is now but it will be necessary to install a new beam dump (to withstand the power that is not absorbed in the plasma) because of the changes being made inside the MAST vacuum vessel. The second beam line will also stay at its present toroidal location (Sector 8) but will be moved so that it is injecting 65cm off the median plan of the machine and the direction of injection will be changed so that the beam tangency radius is 0.8m rather than the current value of 0.7m. This will involve modification of the MAST vacuum vessel to provide the necessary port and a new support frame to change the height of the beam. Preparations have continued to ensure that it will easily be possible to install additional neutral beam sources when funding permits. As originally planned this will involve installing a double beam box at Sector 2. The upper beam dump required by the off-axis source will therefore be installed during the Shutdown to minimise the in-vessel work needed when this third source is added to the machine DIAGNOSTICS Work on the MAST-U diagnostics has concentrated on assessing the impact of the upgrade on the existing diagnostic suite with the aim to: Minimise the cost of moving diagnostics from their current locations whilst retaining or enhancing the existing diagnostic capability; Maximise the diagnostic capabilities around the super-x divertor; Ensure that changes to diagnostics are compatible with the future upgrade of NBI systems and associated changes to the MAST biological shield wall. Designs for the new high field (Centre Column) magnetic diagnostics and target probes are well advanced, along with the scheme design for the layout of diagnostics within the super-x divertor. Figure 5.10 shows the lines of sight for the divertor bolometer array and the region that the divertor reciprocating probe system can sample. Figure 5.10: Lines of sight for the divertor bolometer array (in yellow) and sampling region of the divertor reciprocating probe system (in green). 5.16

17 5.2.5 MACHINE CONTROL AND PROTECTION The high mechanical and thermal stresses in the new coils mean that only ~3000 full performance pulses will be possible. The fatigue life increases rapidly as the mechanical stress is reduced so operation at ~10% below full performance will permit about an order of magnitude more pulses. Consequently operation at full performance will be treated as a scare resource and will only be allowed when justified by the requirements of the scientific programme. The operational limits will be implemented using an intelligent, hierarchical machine protection system. Much of the existing control and pulse sequencing infrastructure on MAST will need to be replaced to permit reliable operation of MAST following the upgrade. A strategy has been formulated for replacing the obsolete hardware used for machine control and supervision on MAST that allows reuse of all existing software and hence minimises the effort that will be required. Prototypes of new fast timers and event time logging units based on FPGAs have also been produced and successfully tested MACHINE AREA MODIFICATIONS Following the upgrade MAST will have longer pulse lengths and ultimately will have twice the present neutral beam heating power. The plasma temperature is also expected to increase. Hence the number of neutrons produced will increase significantly. Consequently it has been necessary to reassess the adequacy of the shielding provided by the thick concrete shielding walls and roof of the MAST machine using the MCNP neutronics code. This has shown that it is necessary to increase the thickness of the east and south walls from 1200mm to 1600mm to improve their shielding ability since potentially occupied area are adjacent to them. The thickness of part of the west wall will also need to be increased. The labyrinths shielding the access points to the MAST area will also need to be improved. For the initial operation the thickness of the roof is adequate but the shielding at the edge of the roof will need to be improved by extending the height of the outer part of the wall to be at least 200mm higher than its original height. The roof thickness will be increased when additional neutral beams are added PROCUREMENT During the year procurement of the components needed for MAST-U has started. Effort has focused on the components having a long lead-time. In some cases the procurement of material or sub-assemblies required has been separated out from the main procurement contract so that CCFE can be directly responsible for the Quality Assurance before the items are passed on to the supplier of the main components. The only significant delay compared with the plan is contract for the manufacture of the air side coils (the solenoid, Px and Pc coils), which will be issued approximately 3 months late. This does not impact the project time scales as the projected delivery date is still before when they are actually needed. 5.17

18 Component Contract placed Expected Completion Centre Tube Dec 2011 Dec 2012 Centre Rod Conductors Jan 2012 May 2012 Centre Rod Assembly Mar 2012 June 2013 In-vacuum coil cans Feb 2012 Oct 2012 In-vacuum coil conductors Mar 2012 July 2012 In-vacuum coils Mar 2012 Oct 2013 Airside Coil Conductors May 2012 Oct 2012 Bulk Graphite Material Jan 2012 May 2012 Bulk CFC Material Mar 2012 Jul 2013 Divertor Coil Power Supplies Toroidal Field Coil Power Supplies Mar 2012 Sept 2013 Mar 2012 Sept T Crane Jan 2012 Aug PREPARATIONS FOR THE SHUTDOWN During the Shutdown it will be necessary to remove the MAST Load Assembly from the existing machine area to allow the strip-down and assembly work to proceed. Subsequently the upgraded assemblies will be craned back to permit MAST to be reassembled in its current location. The maximum weight of the subassemblies greatly exceeds the capacity of the existing 15T building crane. To minimise costs a dedicated local demountable 50T crane will be installed for MAST-U. This will in part use the existing steel beams that support the MAST shielding roof to support the crane rails. This will result in a limited lifting height and consequently it will be necessary to create a 6m x 5.45m opening in the east shielding wall of the MAST area to allow the outer cylinder of the MAST Load Assembly to be removed and replaced. This opening will be closed afterwards using bespoke reinforced concrete beams. The strategy for the stripping down and rebuilding of MAST is nearing completion and the design of the tooling required is being refined. Advantage will be taken of the modular nature of the MAST vacuum vessel, which permits the assembly work to be split into parallel activities on the end plates, the outer cylinder and the centre tube. Clearance of the areas in which the MAST machine is located is needed for the assembly work within the Shutdown. This work is already well advanced. 5.18