Present status of the SST-1 project

Present status of the SST-1 project Y.C. Saxena, SST-1 Team Institute for Plasma Research, Bhat, Gandhinagar, India Abstract. SST-1 is a steady state superconducting tokamak used to study the physics of plasma processes in tokamaks under steady state conditions and to gain knowledge about technologies related to steady state tokamak operation. Major subsystems of SST-1 are described and the present status of the SST-1 project is presented. 1. Introduction The steady state superconducting (SC) tokamak SST-1 is under design and fabrication at the Institute for Plasma Research. The objectives of SST-1 include studying the physics of plasma processes in tokamaks under steady state conditions and learning the technologies related to steady state tokamak operation. These studies are expected to contribute to the tokamak physics database for very long pulse operations. The SST-1 [1, 2] tokamak is a large aspect ratio tokamak, configured to run double null diverted plasmas with significant elongation κ and triangularity δ. In the following we give a brief description of the SST-1 tokamak and discuss the present status of the project. 2. The SST-1 machine 2.1. Physics issues for the SST-1 tokamak The specific objective of the SST-1 project is to produce 1000 s elongated double null divertor plasmas. There are several conventional questions in tokamak physics, which will be addressed again in the steady state scenario. Some of these are related to energy, particle and impurity confinement, the effect of impurities and ELMs in steady state on energy confinement, stability limits and their dependence on current drive methods, resistive tearing activities in the presence of RF fields, disruptions and vertical displacement events (VDEs), and thermal instability. In steady state operations non-inductive current drive will sustain the plasma current. Different aspects of current drive, such as different current drive methods and their combinations, current drive efficiency, profile control and bootstrap current, will be studied. An efficient divertor is required for steady state operations with elongated plasma. Various aspects of divertor operation, such as steady state heat and particle removal, erosion and particle recycling, radiative divertors and pumped divertors, will be studied. Advanced tokamak regimes are of prime interest in fusion research. These regimes are characterized by high β N and high bootstrap current, and are generally obtained in high (H mode) and very high (VH mode) confinement modes in plasma with high triangularity, elongation and large negative shear. Although SST-1 is not optimized for advanced tokamak regimes, we propose to attempt some experiments in this direction within the limitations of the machine. 2.2. Machine parameters and features The choice of the parameters is dictated by the technological and physics goals. As this is our first experience with SC coils, we have decided to use an NbTi superconductor at 4.5 K and hence have restricted the toroidal field to 3 T at the plasma centre. Low aspect ratio machines are difficult to design using SC coils due to space restrictions. Furthermore, higher aspect ratios have advantages such as high bootstrap current and better confinement. We have, therefore, opted for a large aspect ratio ( 5) in SST-1. At other tokamaks substantial improvements in confinement (VH mode) and β N with higher triangularity (δ 0.4 0.8) have been observed. Elongation improves the current carrying capacity of the plasma. With elongation in the range κ 1.6 2.0, improvement in β N has been observed. We have, therefore, chosen ranges of κ and δ similar to these ranges. The double null configuration allows for the distribution of power between a larger number of divertor plates, thus reducing the heat load per plate. We have, therefore, selected the double null configuration, with the provision to go to single null operations in future. The machine has a major radius of 1.1 m, a minor radius of 0.20 m, a toroidal field of 3.0 T at the plasma centre and a plasma current of 220 ka. Elongated plasma with elongation in the range Nuclear Fusion, Vol. 40, No. 6 c 2000, IAEA, Vienna 1069

Y.C. Saxena and SST-1 Team Table 1. Typical operating points of SST-1 Phase I (L mode) Phase II (H mode) Basic plasma (Step 1) (Step 2) (Step 3) (Step 4) Low Low High parameters Circular Circular Elongated Elongated power field power ohmic LHCD LHCD LHCD (0.5 MW) (1 MW ) B T (T) 1.5 1.5 1.5 3 3 1.6 3 I p (ka) 110 110 150 220 220 100 220 q (Cylindrical q) 2.5 2.5 4.1 5.7 5.7 6.6 5.7 H factor, τ E/τ ITER89-P 1 1 1 2 2 2 n (m 3 ) 3 10 19 5 10 18 8 10 18 1 10 19 2 10 19 2 10 19 5 10 19 n 0 (m 3 ) 4.5 10 19 7.5 10 18 1.2 10 19 1.5 10 19 3 10 19 3 10 19 7.5 10 19 P aux (MW) 0.17 0.5 1 1 1 5 Z eff 1.7 5 5 1.7 1.7 1.7 1.7 Elongation, κ x 1 1 1.8 1.8 1.8 1.8 1.8 Triangularity, δ x 0 0 0.66 0.66 0.66 0.66 0.66 Elongation, κ 95 1 1 1.6 1.6 1.6 1.6 1.6 Triangularity, δ 95 0 0 0.54 0.54 0.54 0.54 0.54 τ E (ms) 12 10 11 12 26 12 13 T e (kev) 0.24 0.82 1 1.8 2 0.9 1.9 T i (kev) 0.09 0.32 0.4 0.74 0.8 0.36 0.78 T e0 (kev) 0.39 1.37 1.7 3.1 3.3 1.5 3.2 T i0 (kev) 0.16 0.55 0.67 1.2 1.3 0.6 1.3 β (%) 0.18 0.1 0.2 0.12 0.25 0.39 0.61 β N (%, m T/MA) 0.48 0.28 0.4 0.32 0.68 1.3 1.7 β p 0.33 0.19 0.46 0.49 1.1 2.3 2.6 f bs 0.08 0.04 0.1 0.12 0.25 0.55 0.62 BS collisionality 0.57 0.11 0.12 0.07 0.09 0.2 0.14 correction 1.7 1.9 and triangularity in the range 0.4 0.7 can be produced. Hydrogen gas will be used and the plasma discharge duration will be 1000 s. Auxiliary current drive will be based mainly on 1.0 MW of LHCD at 3.7 GHz. Auxiliary heating systems include 1 MW of ICRH at 22 91 MHz, 0.2 MW of ECRH at 84 GHz and an NBI system with peak power of 0.8 MW (at 80 kev) with a variable beam energy in the range 10 80 kev. Superconducting coils for both toroidal field (TF) and poloidal field (PF) are to be deployed in the SST-1 tokamak. An ultrahigh vacuum (UHV) compatible vacuum vessel, placed in the bore of the TF coils, will house the plasma facing components (PFCs). A high vacuum cryostat will enclose all the SC coils and the vacuum vessel. Liquid nitrogen (LN 2 ) cooled thermal shields between the vacuum vessel and the SC coils as well as between the cryostat and the SC coils will reduce the radiation heat load on the SC coils. A normal conductor ohmic transformer system will be provided to initiate the plasma and sustain the current for the initial period. A pair of vertical field coils will be provided for circular plasma equilibrium at the startup stage of the plasma. A set of saddle coils placed inside the vacuum vessel will provide fast vertical control of the plasma, while PF coils are to be used for shape control. Other subsystems include RF systems for pre-ionization, auxiliary current drive and heating, an NBI system for supplementary heating, cryogenic systems at liquid helium (LHe) and LN 2 temperatures, and a chilled water system for heat removal from the various subsystems. A large number of diagnostics for plasma and machine monitoring will be deployed along with a distributed data acquisition and control system. 2.3. Operational scenario SST-1 is to be operated in two phases. Phase I will involve various steps, starting with ohmic circular plasma to full power operations with the divertor 1070 Nuclear Fusion, Vol. 40, No. 6 (2000)

Article: Present status of the SST-1 project configuration. Phase II will involve advanced tokamak operations. Table 1 shows typical operating parameters in various stages of phase I and phase II. The machine operations will commence with a circular, pulsed plasma driven by ohmic field with a pulse length of 1 s. Lower hybrid current drive will be attempted to sustain current in the limiter configuration. Up to 200 kw power will be used at this stage. Step 3 will involve divertor operation at 1.5 T magnetic field. The plasma will be initiated as circular plasma, current drive will then be taken over by LHCD and divertor coils will be brought in on a slow timescale ( 3 4 s) to produce elongated divertor plasma. In the next step of SST-1 operations, the toroidal field will be increased to 3.0 T and long pulse, elongated plasma will be produced. Full design parameters will be attempted and a total of 1 MW auxiliary power will be introduced. Advanced tokamak modes will be explored in phase II of the SST-1 operations. In the normal operating range (phase I), the values of β, β N and β p are low. In order to improve β values with the same auxiliary power of 1 MW, we may operate at low magnetic field. Further improvement of β N and β p canbeachievedbyoperating at lower currents. Operations with VH mode, non-monotonic q profiles and significant bootstrap current will also be attempted after suitable modification of the divertor for enhanced power handling capacity and enhancement of the RF power systems. 3. Magnet system The magnet system comprises the TF coil system, the PF coil system, the ohmic transformer, the vertical field coils and the vertical position control coils. A cross-section of SST-1, indicating various magnets, is shown in Fig. 1. 3.1. Toroidal field coil system The TF coil system design requirements include the production of a 3.0 T magnetic field on the plasma axis with <2% ripple within the plasma volume. The TF assembly should be capable of providing steady state operation and should withstand plasma disruptions and VDEs without quenching. The assembly shall be capable of being cooled downandwarmedupinlessthan15days.thetf coil system consists of 16 modified D shaped TF coils arranged symmetrically around the major axis spaced 22.5 apart. The contour of the D shaped TF coil consists of a straight leg and five arcs. The over- Figure 1. Schematic diagram of the SST-1 tokamak, illustrating various magnetic field coils and vacuum system components. Table 2. Toroidal field coil system Number of coils 16 Shape Modified D Turns per coil 108 Double pancakes per coil 6 Rated current 10 ka Field at plasma axis 3.0 T Maximum field 5.1 T Maximum field ripple <0.35% Bore dimensions (radial) 1190 mm Bore dimensions (vertical) 1746 mm Outer dimensions (radial) 1560 mm Outer dimensions (vertical) 2120 mm Average turn length 5500 mm Centring force per coil 2.73 MN Tension in the coil 90 110 MPa Total inductance 1.12 H Total stored energy 56 MJ Dump time constant 12 s Peak dump voltage ±600 V all dimensions of the TF coils are dictated by the need to have 2% field ripple at the plasma edge, large enough radial ports on the vacuum vessel as to allow radial access for NBI and human access inside the vacuum vessel for assembly of in-vessel components. A total of 17.28 MA T at a peak current of 10 ka per turn will produce a field of 3.0 T at the plasma centre and a maximum field of 5.1 T at the Nuclear Fusion, Vol. 40, No. 6 (2000) 1071

Y.C. Saxena and SST-1 Team Table 3. Poloidal field coils Coil type Number Coil radius Vertical Winding Number of NI max of coils (m) location cross-section turns per coil (MA T) (m) (mm 2 ) PF1 1 0.45 0.0 71 320 80 0.80 PF2 2 0.45 ±0.43 71 163 40 0.40 PF3 2 0.50 ±0.93 136 380 192 1.92 PF4 2 1.72 ±1.03 85 136 40 0.40 PF5 2 2.01 ±0.65 85 136 40 0.40 PF6 2 1.35 ±0.35 100 100 16 0.16 TF conductor. Each of the TF coils will be made up of six double pancakes, each pancake having nine turns. Conventional soldered joints will be used to make inter-double pancake joints. The turn to turn insulation in the pancake will be given by cryogenic grade E glass tapes with half overlaps. NEMA grade G-11CR sheets will separate each of the pancakes. The winding pack of six double pancakes, with a ground wrapping, will be vacuum pressure impregnated (VPI) with epoxy, and then shrink fitted into a stainless steel (SS316L) casing which supports most of the electromagnetic load. All the TF coils are connected in series and are protected against quenching by a suitable dump resistance, switching and sensing system. Typical dump time constant is 12 s and the maximum voltage across the coils at the dump time is ±600 V with respect to ground. The main parameters of the TF coils are summarized in Table 2. 3.2. Poloidal field coil system 3.2.1. Plasma equilibrium and shaping The PF coil system [3] is comprised of nine SC coils (PF1 to PF5) and two normal conductor coils (PF6). A free boundary, axisymmetric, ideal MHD equilibrium model based code has been used for design and optimization of the PF system. The PF6 coils are placed in the bore of the TF coils, inside the vacuum vessel. This is required in order to obtain plasma shapes with high triangularity and is necessitated by the large size of the TF coils and the consequent large distance of the external PF (PF1 PF5) coils which requires a large number of ampere turns in the external coils in the absence of PF6 coils for highly triangular plasma shapes. The PF coils allow for a wide range of elongation and triangularity and support a large range of plasma equilibria. The feasibility of limiter operation during plasma current rampup, double and single null operation at a plasma current of 220 ka, double null operation at a plasma current of 330 ka and various startup scenarios are the design goals for the PF system. The positions of the PF coils were optimized for flexibility and cost. Table 3 summarizes the characteristics of the PF coils. The designed PF system allows for plasma flexibility in elongation in the range 1.7 1.9, triangularity in the range 0.4 0.7, plasma inductance in the range 0.75 1.4 and poloidal β in the range 0.01 0.85, and a slot divertor configuration. Different reference equilibria with κ = 1.8 have been used to obtain currents in the PF coils. The coil currents obtained for various values of l i and β p are given in Table 4. 3.2.2. Ohmic transformer and vertical field coils It is desirable to provide for ohmic startup and current rampup to a certain extent before auxiliary current drives and heating take over to sustain the plasma for a long time. The SST-1 tokamak deploys SC magnets for plasma shaping and equilibrium and the obvious first choice would be the use of these coils for plasma startup and current rampup. As the SC coils are, however, made up of the conductor NbTi, the stability of the SC cable demands that the rate of change of the currents in these coils be limited below a certain level to avoid the quenching of these coils during plasma startup. The total flux storage in the PF system is limited to 0.5 V s. In order to use these coils for plasma startup, with the rate of change of current limited to producing a magnetic perturbation of <2 T/s on the SC conductor, a flux storage of about 3 V s is required to compensate for resistive flux loss for such a slow startup. An ohmic transformer [4], comprised of a central solenoid (TR1) and two pairs of compensation coils (TR2 and TR3) will, therefore, be used for plasma startup and initial current rampup. These coils are made from hollow copper conductors. The TR2 and 1072 Nuclear Fusion, Vol. 40, No. 6 (2000)

Article: Present status of the SST-1 project Table 4. Currents in PF coils for reference equilibria at κ = 1.8 for the double null divertor configuration at I p = 220 ka Current in each coil (MA T) β p l i PF1 PF2 PF3 PF4 PF5 PF6 0.01 0.75 0.232 0.002 1.162 0.005 0.018 0.112 0.84 0.75 0.195 0.011 1.140 0.051 0.044 0.105 0.01 1.00 0.217 0.048 1.217 0.050 0.090 0.075 0.83 1.00 0.181 0.054 1.191 0.102 0.147 0.068 0.01 1.40 0.201 0.087 1.267 0.105 0.191 0.042 0.83 1.40 0.167 0.092 1.236 0.155 0.241 0.038 TR3 coils minimize the magnetic field produced by TR1 in the plasma to values of less than 10 G at full flux storage. The transformer has a storage capacity of 1.4 V s and can be used for producing circular plasma with currents up to 100 ka for almost 1 s. A pair of vertical field (VF) coils will keep this circular plasma in equilibrium during the initial phase. This pair of normal conductor coils is required to establish an equilibrium field on a fast timescale during plasma startup and current rampup in the initial stage. The current drive will be, later, taken over by LHCD and the PF coils will be ramped up at a slow rate, in about 3 s in a pre-programmed way so as to achieve the desired plasma equilibrium and provide the divertor configuration for an elongated triangular plasma. 3.2.3. Feedback control coils Vertically elongated plasma is inherently unstable to perturbations in the vertical direction. In the absence of any conducting structures surrounding the plasma, the instability grows on an Alfvén timescale which is typically a few microseconds. A set of passive stabilizers (described in Section 6), placed inside the vacuum vessel, slow down the growth rate of this vertical instability to =38 s 1 for most unstable equilibria corresponding to κ = 1.9. Further stabilization of the instability is provided by active feedback using a pair of coils placed inside the vacuum vessel. The active feedback coils are toroidal loops of major radius 1.35 m, connected in a saddle configuration and placed at ±0.5 m from the central plane. A maximum of 8 ka turns is required to keep plasma excursions within 0.01 m from the central plane. For long pulse operation it is necessary to maintain the plasma major radius at a nominal value. The radial control requirements have been analysed mainly due to disturbances from minor disturbances, type I and type III ELMs. It is found that the PF6 coils can adequately handle the radial control requirements. PF6 will, therefore, also be used for radial position control in addition to equilibrium and shaping. 3.3. Conductor for SC coils All SC coils will be made using a cable in conduit conductor (CICC) based on NbTi+copper. The choice of NbTi is based on the fact that the operating fields in TF and PF coils are moderate. The CICC will be made up of 135 strands cabled in a 3 3 3 5 cabling pattern and conduited inside an SS304L conduit. The CICC is designed for an operational current of 10 ka at 5 T and 4.5 K, with a critical current of 36 ka. The nominal critical current for each strand at 5 T and 4.5 K is 272 A. The CICC design has been carried out in consultation with the National High Magnetic Field Laboratory, Florida State University, Tallahassee, USA. In the design and optimization of the CICC, various operational constraints have been met. M/S Hitachi Cables Ltd., Tsutchiura, Japan, have fabricated test samples of CICC as per optimized design. The qualification tests carried out on the strands extracted from the sample CICC include critical current measurements at 5 T and 4.2 K, determination of superconductor to normal state transition index n and residual resistivity ratio, measurement of hysteresis losses, copper to non-copper ratio, sharp bend tests and spring-back tests. One strand from each of the first stage triplets was taken for the test on extracted strands. Void fraction and conduit thickness measurements were carried out on the CICC. The results indicate that degradation in the strands due to cabling and jacketing is less than 5% and a critical current 35 ka at 5 T, 4.2 K is ensured. A 600 m test section has been fabricated and preliminary Nuclear Fusion, Vol. 40, No. 6 (2000) 1073

Y.C. Saxena and SST-1 Team tests on model coil fabricated from a test section have been carried out at the Khurchatov Institute, Moscow. The results from these tests [5] indicate suitability of the conductor for TF coils of SST-1. The conductor is also found to be suitable for PF magnets with slower ramp rates than anticipated earlier. Further tests with larger flow rates are envisaged in order to establish the limit on the ramp rates for various current levels in the conductor. 3.4. Support system for SC magnets AC load (kw) not to scale 9.71 8.99 0.72 0.019 0.015 Load on TF and PF coils (o) Load on PF coils (*) Load on TF coils (.) The straight legs of TF coils are wedged to form the inner vault [6]. The outer vault is formed by connecting intercoil structures between the TF coils. These vaults resist the centring force and overturning torque experienced by the TF coil system. There is an insulation break between each of the TF coils. The SC PF coils are supported on the TF coil casings and the intercoil structures, with the coils and structures forming a rigid cold mass of about 30 t at 4.5 K. The TF coils are further supported on a base support system consisting of a ring with 16 cantilevered beams. The ring rests on eight columns that are inside the cryostat and have LN 2 intercepts to minimize the conduction loss at 4.5 K as the cold mass load is transferred from these columns to the main machine support. The main machine support is comprised of 8 columns, supporting the base frame of the cryostat and the cold mass, which are firmly grouted to ground. 3.5. Present status of the magnets The engineering design of the SST-1 SC magnets has been completed. The manufacturing of the CICCs for the TF coils has been completed and the conductors delivered to the contractor for winding. The conductor for the PF magnets is being manufactured at present. The contract for winding of the magnets, both superconducting and resistive, has been awarded to M/S Bharat Heavy Electricals Ltd, Bhopal, India. The design and fabrication of the winding system has been completed. Tests on the insulation system using the VPI process are in progress. The winding of a prototype double pancake using a dummy conductor is likely to commence soon. Tests on hole drilling and stub welding on CICC are in progress and trials on joints are continuing. Tooling and assembly of the winding systems for copper coils is in progress. The coils are scheduled to be delivered in the third quarter of 2000. 0.004 0 0 3 1003 1006 Time (s) not to scale Figure 2. Average AC loss on all coils (TF+PF) (circles), TF coils (chain line) only and PF coils (asterisks) only, during operation of the SST-1 machine. 4. Cryogenic system for SST-1 The SC magnet system (SCMS), consisting of TF and PF coils, in SST-1 has to be maintained at 4.5 K in the presence of steady state heat loads. In addition the pulsed heat loads during plasma operation have to be taken care of by the cooling system so as to maintain the magnets in the SC state. The magnets will be cooled using forced flow of supercritical helium (SHe) through the void space in the CICC. Furthermore the magnets have to be energized from power supplies at room temperature using vapour cooled current leads, which evaporate He at the cold end to gas helium at =300 K at the warm end of the lead. A helium refrigerator/liquefier with cold circulation system for SHe is, therefore, required for this purpose. In order to minimize the heat loads on magnets and support system at 4.5 K, LN 2 shields are provided between the cold mass at 4.5 K and warmer surfaces, for example vacuum vessel and cryostat, at 300 K. An LN 2 storage and distribution system is provided for this purpose. 4.1. Heat loads at 4.5 K The different heat loads experienced by the SC coils can be classified into steady state heat loads (including radiation losses from LN 2 shields and residual gas conduction) and losses during operation. The operational losses include Joule heating, AC losses in the joints and AC losses in the CICC due to the current rampup/down in the coils and current changes in the feedback coils. In addition 1074 Nuclear Fusion, Vol. 40, No. 6 (2000)

Article: Present status of the SST-1 project to these loads there are heat loads arising from the conduction from supports, eddy currents induced in the magnet casings and structures, losses in the transfer lines and bus ducts, and heat loads from cryogenic valves, bayonets and diagnostics inserts. A total steady state heat load =180 W is expected. In addition the SC coils are subjected to pulsed loads during plasma operation. A total pulsed load of =125 kj is expected during one plasma pulse. Such pulses will be repeated every 5000 s with a total of six pulses per day. The magnets will be energized using ten pairs of current leads, one pair for TF coils and nine pairs for SC PF coils. The TF coils will be energized for about 10 h per day, while the PF coils will carry current on average for about two hours per day. This load would on the average evaporate =150 L/h of LHe to helium gas at =300 K, in the current leads of the SC coils. The SC TF and PF coils of SST-1 are subjected to various external disturbances arising from plasma breakdown, startup, current rampup/rampdown in PF coils and current oscillations in feedback coils during 1006 s plasma pulse (normal operation). The PF coil current rampup/rampdown duration is 3 s (maximum ramp rate is 3 ka/s). The pulsed field losses in the winding pack include mainly hysteresis losses due to the parallel and transverse field components and coupling losses due to the transverse field component. The AC losses have been estimated from the energy expressions given in Ref. [7]. Figure 2 shows the average AC loss power deposition on the conductor during current rampup, rampdown in PF coils and 1000 s flat-top operation for all the coils, for TF coils and for PF coils. 4.2. Flow requirements for SHe The SC coils are cooled by forced flow of SHe through the void space in CICC. For this purpose the entire magnet system is divided into several parallel paths. The flow requirements for each of the paths is estimated on the basis of the requirement for stability of the SC coils in the presence of pulsed loads superimposed on the steady state heat loads. A total flow =0.3 kg/s is required to keep the SC temperature well below the current sharing temperature in the presence of peak pulsed loads. 4.3. The LHe plant A closed cycle LHe plant is required to cater for both the steady state and the transient heat loads. Taking into account both the steady and transient loads and considering the uncertainties in the estimates, an LHe plant with a refrigeration capacity of 400 W at 4.5 K and 200 L/h liquefaction at a pressure of 1.2 bar (a) is to be installed. The plant will, in addition, also provide the refrigeration capacity of around 250 W for the heat dissipation in the cold circulation pump. The above total capacity will be achieved with LN 2 pre-cooling. The total capacity of the plant (both refrigeration and liquefaction) would be variable from 50 100%. It will also be possible to operate the plant without LN 2 pre-cooling at the reduced capacity. A cold circulation system, for the flow of SHe through the SC coils in the closed cycle, forms part of the LHe plant. The system will be comprised of a cold pump, heat exchangers and valves. The system will be designed for a nominal flow rate of 0.3 kg/s with a variability of 50 120% of the nominal value. The heat exchangers of the pump circuit will ensure the supply temperature of SHe, to the SC coils, to be 4.5 K. The heat exchanger located on the return side will dissipate the steady state as well as transient heat loads into a buffer dewar. An external dual bed, full flow, on-line purifier with automatic regeneration is provided to remove impurities. In order to confirm the purity of the processed helium gas, the purifier will be equipped with an impurity monitor to detect the above impurities at <1 ppm level. A buffer main control dewar (MCD) is provided in the plant. The purpose of this is to absorb the heat loads generated within the SC coils and return cold helium vapours to the cold box at a constant temperature and pressure. The SHe coming from the SC coils delivers heat to the MCD through a suitably designed heat exchanger before returning to the cold pump. The MCD also serves as a reservoir for supplying LHe to the current leads. The capacity of the MCD will be 2500 L. 4.4. LN 2 distribution system As mentioned above, LN 2 cooled radiation shields (LN 2 shields) are provided between the SC coils and the vacuum vessel as well as between the cryostat and the SC coils. While the cryostat walls will be at room temperature, the vacuum vessel walls will have different temperatures under different operating conditions. During baking the walls will be at 525 K, during wall conditioning the temperature will be 425 K and during plasma operation the temperature will be 325 K. The LN 2 consumption during these three phases will be =1200 L/h, =600 L/h and Nuclear Fusion, Vol. 40, No. 6 (2000) 1075

Y.C. Saxena and SST-1 Team =300 L/h, respectively. In addition, LN 2 at 150 L/h is required for the pre-cooler and purifier of the LHe plant and at 200 L/h for the NBI system. Liquid nitrogen storage tanks of 105 m 3 have been provided. The LN 2 will be purchased commercially and filled in these tanks to replenish the liquid consumed. The gas/vapours from the applications will be released to the atmosphere. Appropriate distribution systems with valves, transfer lines and phase separators are to be provided. 4.5. Present status of the cryogenic system The helium plant is at present being engineered by M/S Air Liquide Cryogenie, Sassenage, France, and is likely to be delivered by the end of 2000. An He gas management system, including high pressure and medium pressure storage vessels and recovery system, has been designed and is being procured indigenously. A recovery compressor and two gas balloons have been procured. Three LN 2 storage tanks, each of 35 m 3 capacity, have been procured and commissioned. The tanks have been manufactured by M/S Inox India Ltd. Prototype current leads, LHe and LN 2 transfer lines and bus ducts are under fabrication, and test systems for these have been designed and fabrication is in progress. 5. Vacuum vessel, cryostat and pumping system The vacuum system (Fig. 1) of SST-1 [8] is comprised of (a) A UHV compatible vacuum vessel, (b) An HV compatible cryostat, (c) LN 2 cooled radiation shields, (d) Pumping system. The vacuum vessel is a fully welded vessel made from sixteen modules, each module consisting of a vessel sector, an interconnecting ring and three ports. The ring sector sits in the bore of the TF coil, while the vessel sector with ports is located between two TF coils. There are sixteen radial ports and 32 vertical ports. The vacuum vessel is designed as an ultrahigh vacuum system with partial pressures for all gases, except hydrogen, less than 1 10 8 torr. Effective wall conditioning techniques are proposed to be used including baking to 525 K, discharge cleaning and boronization. The vessel is made of SS304L material. It has a height of 1.62 m, a midplane width of 1.07 m, a total volume of 16 m 3 and a surface area of 75 m 2. The cryostat is a high vacuum chamber, which encloses the vacuum vessel and the SC magnets. It is a sixteen-sided polygon chamber made of SS304L with a volume of 35 m 3 and surface area of 59 m 2. The base pressure inside the cryostat will be maintained at less than 1 10 5 torr to minimize residual gas conduction losses on the SC magnets. Liquid nitrogen cooled panels (LN 2 panels) are placed between all surfaces having temperatures higher than 85 K and surfaces at 4.5 K in order to reduce the radiation loads on the SC magnets and the cold mass support system. The total surface area of all the panels is 126 m 2. Each of the panels is made up of 8 mm diameter tubes, vacuum brazed to a 1 mm thick SS304L sheet. The cooling method is based on the latent heat of vaporization. During normal operations 10000 L/s pumping speed is required to achieve a base pressure of less than 1 10 8 torr in the vacuum vessel. Two turbomolecular pumps, each of 5000 L/s speed, will be used for this purpose. Two closed cycle cryopumps will be used during wall conditioning of the vacuum vessel. The main gas load from the vacuum vessel is during steady state plasma operation. Sixteen turbomolecular pumps, each with a pumping speed of 5000 L/s at 10 3 T for hydrogen, are to be connected to sixteen pumping lines on the vacuum vessel. The net speed of each pumping line is estimated to be 3900 L/s. The net pumping speed provided for the cryostat using two turbomolecular pumps is 10000 L/s. The vacuum vessel and cryostat will be pumped down from atmospheric pressure to 10 3 torr using two separate Roots pumps of 2000 m 3 /h capacity. 5.1. Present status of the vacuum system The design of the vacuum system has been completed and a fabrication contract for fabrication of the vacuum system, cryostat, LN 2 panels and support structure has been awarded to M/S Bharat Heavy Electrical, Tiruchirappali. Trials on the methodology of fabrication of the vessel sector, without joints, and the ring sector, with one joint, have been conducted. A prototype, comprisingonevesselsectorwithports,tworingsectorsand a one eighth section of the cryostat, is in the initial stage of fabrication. Brazing trials for LN 2 panels are being carried out. Process establishment trials for 1076 Nuclear Fusion, Vol. 40, No. 6 (2000)

Article: Present status of the SST-1 project Z Z Figure 3. Section of SST-1 with PFCs and plasma equilibrium. electropolishing and ultrasonic cleaning are in progress. Turbomolecular pumps, gate valves and Roots pumps have been procured and tested. 6. Plasma facing components The PFC of SST-1 [9], comprised of divertors and baffles, poloidal limiters and passive stabilizers (Fig. 3), is designed to accommodate the envelope of equilibria defined by elongation κ = 1.7 1.9, δ = 0.4 0.7 and plasma internal inductance l i = 0.75 1.40. In long pulse discharges, like the one in SST-1, the design of the plasma facing components has to ensure a steady state heat removal capability. Particle removal in steady state is also a major concern. The inboard and outboard divertor plates are designed for the worst possible heat loads by assuming an in out asymmetry of 1:2 (single null) for the inboard divertor and a 1:4 (double null) asymmetry for the outboard divertor. An up down asymmetry of 1:1.2 has been assumed in order to take into account the relatively high power loads on the divertor plates facing the B drift direction for ions. The normal incident peak heat flux (estimated by assuming a SOL width of 5 mm for the heat flux λ q at the outboard midplane) at the inboard and outboard strike points is 1.6 and 5.6 MW/m 2, respectively. The poloidal inclination of the outboard divertor plates is adjusted so as to have an average heat flux at the strike points of less than the allowed limit of 0.6 MW/m 2. However, the inboard divertor is not inclined to the vacuum vessel for optimum results but is kept parallel to it due to space constraints. However, the average heat flux is within the tolerable limits. The target points of the inboard and outboard divertor plates have been chosen at a distance as large as practicable from the null point. This reduces the electron temperature at the target plate and decreases the impurity influx from the divertor region into the core plasma region. A baffle has been incorporated in the design so as to form a closed divertor configuration that helps to increase the neutral pressure in the divertor region, which improves neutral particle recycling and hence reduces the electron temperature. This geometry also helps in efficient pumping of the divertor region. The design constraints in choosing the slot geometry arise from minimizing the backflow of neutral atoms. The slot width (gap between divertor and baffle) was arrived at by using the neutral particle code DEGAS. These results were confirmed by use of the plasma and neutral transport coupled code B2-EIRENE [10]. The Nuclear Fusion, Vol. 40, No. 6 (2000) 1077

Y.C. Saxena and SST-1 Team baffle is designed assuming an average heat flux of about 0.6 MW/m 2. One of the main objectives of SST-1 is to demonstrate steady state particle removal from the vessel during steady state discharges. The main gas load to the pumping system is during steady state plasma operation and there will be particle flux from the core plasma to the divertor region. With a steady state plasma density of 1 10 19 m 3,plasmavolume of 1.65 m 3 and expected particle confinement time of 12 ms, the particle exhaust expected from the core to the divertor region is 22 torr L/s. As the pressure in the divertor region will be 1 10 3 torr, the pumping speed required for the divertor region is 11 000 L/s. Taking into account the conductance of the baffle slot, wide operating range of plasma density and uncertainty of particle confinement time, the divertor pumping system is designed to provide a pumping speed of 63 000 L/s at 1 10 3 torr pressure for hydrogen gas at the pumping plenum of the vacuum vessel. A pair of poloidal limiters is provided to assist plasma breakdown, current rampup and current rampdown and for the protection of RF antennas and other in-vessel components during normal operation and during VDEs and disruptions. The outboard limiters are made movable in order to protect RF antennas that are movable. The inboard and outboard limiters are designed to handle about 4% of the input power during steady state operation. In addition to this the limiters are expected to take up to about 50% of the input power during plasma current rampup and rampdown phases and during short duration circular plasma operation. The support structure of the outboard limiter is designed to have a radial movement of ±30 mm. On the inboard side, a safety limiter is placed 30 mm away from the seperatrix. The outboard limiter plasma facing surface is designed with appropriate poloidal curvature to avoid interference with divertor operation at highκ and -l i equilibria. The height of the set of limiters is restricted to ±0.145 m, so as to fill the vertical opening between the top and bottom passive stabilizers, and to allow thermal expansion and other required clearances. The horizon of the outboard poloidal limiter plasma facing surface conforms to a circular arc of radius 0.30 m centred at the major radius of 1.03 m on the midplane up to a height of ±0.125 m and for the reference equilibrium this closely approximates the +3 cm flux surface (Fig. 3). For this configuration, a peak heat load of 6.4 MW/m 2 is expected for normal incidence. As we limit the steady state heat removal capability to less than 1 MW/m 2, the limiter surface is shaped/profiled in the toroidal direction. We opted for a semicircular shape with the front edge having a radius of curvature of 1.095 m in the toroidal direction. The toroidal width of the limiter is 0.37 m with appropriate curvature at the ends. For the inboard limiter the peak heat load for normal incidence is estimated to be also 6.4 MW/m 2. To reduce the heat flux to less than 1 MW/m 2,this limiter is also toroidally profiled, with a semicircular shape and a front curvature of 0.278 m in the toroidal direction. The toroidal width of the inboard limiter is 0.22 m with optimum curvature at the edges. Passive stabilizers comprised of conducting structures surrounding the plasma are provided to reduce the growth rate of the vertical instability and hence make active feedback control possible. There are two pairs of stabilizers in SST-1, one on the inboard and the other on the outboard side of the plasma placed above and below the midplane. The stabilizers are located closer to the plasma to have greater mutual coupling with it when the plasma moves from its equilibrium position. High conductivity plates of stabilizers provide a large time constant of the order of 70 ms for the induced currents to decay and thus considerably slow down the growth rate of the plasma motion. To allow plasma startup with an ohmic flux swing, an electrical break is incorporated. The top and the bottom stabilizers are connected in the saddle configuration with a current bridge at the location of this break. The passive stabilizers are designed to handle heat fluxes of 0.25 MW/m 2.The plasma facing surfaces of the stabilizer plates are covered with 20 mm thick graphite tiles. Active cooling is provided to the stabilizers in order to maintain their temperature below 150 C. Isostatically pressed fine grain graphite is chosen as the baseline armour material for the PFC of the SST-1 tokamak. Good thermal conductivity and good mechanical behaviour at elevated temperatures and after exposure to high temperatures for long duration are required for the actively cooled substrate of the divertor and limiter assemblies. In addition to the above properties, passive stabilizer material requires a high electrical conductivity at elevated temperatures. Copper based alloys such as copper zirconium or copper chromium zirconium have, therefore, been selected for this purpose. The PFCs are actively cooled so as to keep the surface temperature of the plasma facing surface less than 1000 C. The PFCs are also designed for baking up to 350 C. 1078 Nuclear Fusion, Vol. 40, No. 6 (2000)

Article: Present status of the SST-1 project The engineering design of various PFCs and their supports has been completed and the fabrication of the various components is under way. 7. Current drive and auxiliary heating 7.1. Lower hybrid current drive (LHCD) The LHCD system would be used as the prime method of steady state plasma current in circular as well as in elongated plasma. The system has been optimized for efficient current drive for 1.5 and 3.0 T operation, for different plasma elongation within 1.7 1.9 and triangularity of 0.4 0.7 for plasma inductance values of 0.75 1.4. Lower hybrid waves with an asymmetric spectrum (N 1.8 4.0) would be launched by changing the phase between adjacent waveguides from 40 to 160 using high power phase shifters through a radial port at 3.7 GHz to drive plasma current during different operating scenarios. Two klystrons would supply 1 MW of CW power to feed a grill consisting of 64 narrow waveguides in two rows placed on the equatorial plane of the LFS radial port. The main components of the system are a low power section, high power amplifier (klystrons), high power transmission line, high power phase shifter (to vary the launched spectrum), high power directional coupler (to monitor forward and reflected power), vacuum window and grill. Each Thomson CSF klystron (TH2103D) operates at 3.7 GHz and delivers 250 kw in two arms of the WR284 waveguide for 1000 s. Power from each of these outputs would be divided between sixteen channels of the transmission line consisting of the above mentioned components. It is to be noted that the power level in each of the channels is around 15 kw and is definitely easy to handle. Finally the power will be coupled to the plasma through the grill interfaced by a vacuum window assembly. 7.2. Ion cyclotron resonance frequency (ICRF) Ion cyclotron heating is chosen to heat the plasma to 1.0 kev, during a pulse length of 1000 s. A 1.5 MW ICRF system would operate at different frequencies between 20 and 92 MHz for different heating scenarios at 1.5 and 3.0 T operation. Conventional design criteria have been implemented in developing the 1.5 MW, 1000 s RF system. Various efficient heating scenarios (e.g. the second harmonic of the majority species at 1.5 and 3.0 T, minority deuterium in hydrogen plasma) would be attempted. The generator is tetrode based and modular. For the last high power stage, tetrode TH526 has been selected which has a wider frequency bandwidth. All the components used in designing the generator have high frequency performance. The RF generator will be placed in the generator room about 90 m away. A 90 m long pressurized 9 in., 50 Ω transmission line would feed the antennas. The RF power is divided by a 3 db hybrid coupler and then by two Tees to obtain equal power at each antenna. The entire transmission line will be pressurized at 3 bar to avoid breakdown. Estimated power losses are 0.6 and 0.25 kw/m in the inner and outer conductor, respectively, increasing the temperature by 116 C (maximum). Hence cooling is mandatory. Forced dry air flow pressurized at 3 bar with 24 m/s flow velocity, through an annular section of 9 in. transmission line or water flow through the inner conductor has been selected. Slow stub matching (in seconds) as well as fast frequency ( 2 ms) matching techniques would be used to optimize matching during the discharge. Reflected power would be replenished by increasing the RF input power as per the feedback signal till slow matching is achieved. Impurity generation by the ICRH antenna is not so severe. The interface would be made of SS304L for better mechanical strength and coated with copper ( 100 µm) for better RF transmission. For 1000 s pulse operation the interface will be actively cooled. 7.3. Electron cyclotron resonance frequency (ECRF) An ECRF system at 200 kw, 84 GHz will be used initially to pre-ionize and start the SST-1 discharges. A gyrotron source capable of delivering 200 kw CW has been chosen. Second harmonic X and O mode launching would be used during 1.5 T operation. Provision for low as well as high field side launch has been made. The output mode of the gyrotron with an internal mode converter is HE 11. A transmission line consisting of a DC break, bellows, mitre bend, polarizer and corrugated waveguide terminating with a barrier window will be used to transmit power from the gyrotron to the tokamak. The total attenuation of the line is 1.1 db. A quasi-optical reflecting mirror system has been designed to steer the microwave beam toroidally. The launcher design for the ECRF system is basically determined by the available space, Nuclear Fusion, Vol. 40, No. 6 (2000) 1079

Y.C. Saxena and SST-1 Team Table 5. Parameters of neutral beam Species Energy Power Power density to plasma Shine-through Focal lengths Divergence (kev) (MW) (MW/m 2 ) (%) (m) (deg) H, D 30 80 0.5 1.7 15 40 <5 7.1 (h), 5.4 (v) <1 required beam width and cross-section in the plasma. An overmoded waveguide is used to transmit the power, which follows quasi-optical transmission with a Gaussian pattern. The low field side launcher consists of two reflectors: a focusing mirror and a plane mirror. For long pulse operation the reflectors require effective cooling. A negligible part of electromagnetic power is absorbed during O mode launch in its first pass. After reflection from the vessel wall the beam is absorbed in the centre. 7.4. Steady state neutral beam injector Neutral beam injectors have demonstrated their capability in providing novel plasma configurations in short pulse confinement devices. These plasmas have demonstrated improved stability and confinement. The role of NBT as a sustaining device for thermal, particle and momentum influx to the plasma have, therefore, assumed prominence among all the options available for heating systems in steady state machines. A steady state NBI system is envisaged for SST-1. For the SST-1 machine the power required in NBI in the low and high density phases is 0.5 and 1.7 MW, respectively. This requirement is accommodated through a tangential injection of the beam, corresponding to maximum absorption at the 0.98 m radius of tangency in the plasma. The beam parameters and their effect on the plasma have been modelled using the 1-D transport code BALDUR. These parameters are summarized in Table 5. The injection is expected to raise the ion temperature to 1 kev, contribute to a fuelling of 4 torr L/s, impart a toroidal momentum of 10 5 m/s and drive a current of 40 ka at the core of the plasma. The power will be delivered from a single beam line and the dynamic range of voltage will be accommodated in a single source [11]. 7.5. Present status of current drive and auxiliary heating systems The engineering design report for the ICRH system on SST-1 has been completed. Fabrication of the amplifiers up to 20 kw and 47.6 MHz has been completed. Regular RF shots with the 20 kw stage of the generator are being generated on the dummy load. Experiments on an antenna test facility, to understand the RF gas breakdown phenomena, with the help of a fast wave antenna have been initiated. The engineering design has been completed for the 1 MW, 3.7 GHz microwave system for LHCD. During the design phase, the detailed layout of the 64 channels of the system in four layers has been finalized. The details of the mechanical structure have been worked out. Prototypes of many of the high power components have been designed, fabricated and tested at low power levels. The 25 W low power solid state microwave system to be used as the input power to the klystrons has been integrated, commissioned and tested successfully. The engineering design report, for an 82.6 GHz ECRH system, has been completed. The final order for the Gyrotron along with the transmission line has been placed. The engineering design of the NBI system has been completed. Research and development activity related to the fabrication of cryopumps, ion sources, etc. is in progress. 8. Diagnostics for SST-1 A large number of plasma diagnostics will be deployed on SST-1. Quantities to be measured include the plasma current, position and shape, β p, density, electron and ion temperatures in the core, edge and divertor regions, impurity concentrations, radiated power, q profile, surface temperatures of the PFC and fluctuations in the basic parameters over a wide frequency range. Shaped plasma requires 2-D measurements. Long pulse operations have implications for the diagnostic techniques. Table 6 gives a summary of the electromagnetic diagnostics to be deployed on SST-1. Other diagnostics for SST-1 include a FIR interferometer and Thomson scattering, ECE, charge exchange, thermography, soft X ray imaging, hard X ray monitoring, visible and VUV spectroscopy and motional stark effect (MSE) diagnostics. The MSE diagnostics will utilize the NBI system of SST-1. 1080 Nuclear Fusion, Vol. 40, No. 6 (2000)

Article: Present status of the SST-1 project Table 6. Electromagnetic diagnostics for SST-1 Diagnostics Number Measurable parameter Parameter range Present status Rogowski coils 4 Plasma current 10 250 ka Under fabrication Mirnov coils 116 B fluctuations Under fabrication Rogowski coils 110 Halo current Under fabrication Magnetic probes 48 pairs Plasma position and shape Under fabrication Saddle loops 16 pairs Locked mode detection Under fabrication Fiber optic current sensors 2 Plasma current 10 250 ka Under procurement Flux loops 24 Loop voltage 0.5 50 V Hall probes 96 Plasma current 10 500 ka Assembly and and position calibration in progress Langmuir probes 96 Divertor plasma density 10 17 10 19 m 3 Design completed The diagnostics are at various stages of design, fabrication and testing. 8.1. SST-1 data acquisition and control system Since SST-1 will be operated under steady state for 1000 s, lossless acquisition of diagnostic data is the prime focus. In order to reduce the large storage requirements it should be possible to have event driven sampling rates. Thus the normal data can be acquired at a slow rate while on a specific event the sampling rate could be increased. It is desirable to be able to view and process data on-line from network nodes. Keeping in view the requirements of a flexible system which will allow selection of any kind and mix of hardware such as CAMAC, VME and VXI, a Grand Interconnect (GI) serial highway system of M/S Kinetic Systems Corp. has been chosen. A pentium processor, with Windows NT as host system, has been selected to take advantage of the benefits of modern technology. The GI ring architecture has attractive features such as a fast 125 Mbit/s fibre optic link, for distances up to 2 km between chassis. It supports CAMAC, VME and VXI on a maximum of 126 I/O chassis. It permits multibuffering and multirate data acquisition with 16 levels of loop nesting. It is based on a single master protocol to maximize the data throughput. Each GI based data acquisition hardware system is connected to a dedicated pentium based server which is responsible for managing the activity, for example, configuring the digitizers, collecting data, providing triggers and distributing events within the ring, catering for the demands originating from the digitizers and storing the data locally. These servers are connected to a campus wide ATM network. A Windows NT platform is used as the operating system. MS-SQL as RDBMS will be used for local data and configuration storage. Application development tools include JAVA, Visual C++, LabWindows/CVI and Visual Basic for GUI. MATLAB will be used for data analysis. 8.2. SST-1 control system The SST-1 control system [12] is based on a distributed control system. For efficient management, the control system has been divided into three groups: the machine control system, the discharge control system and the diagnostics supervisory control. The machine control system will configure and operate the various technical subsystems with a hierarchical control structure. The low level, front end, systems would be based on field bus devices, programmable logic controlers and VME systems. A control area network (CAN) is selected as the default standard for the field bus. Discharge control will be active during discharge and discharge related subsystems will be under the direct control of the discharge control supervisor. The discharge control system will take decisions regarding operational phases and control the real time control system. Real time control systems would be based on the VME standard and will use reflective memory as the standard for real time digital data communication. The diagnostics supervisory control will advise diagnostics control regarding expected plasma parameters and expected diagnostics setting depending on the planned experiment. It will also collect real time information from some of the diagnostics for plasma state analysis pur- Nuclear Fusion, Vol. 40, No. 6 (2000) 1081