- Email: [email protected]
Commissioning and initial operation of KSTAR superconducting tokamak
Fusion Engineering and Design 84 (2009) 344–350
Contents lists available at ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Commissioning and initial operation of KSTAR superconducting tokamak Yeong-Kook Oh a,∗ , W.C. Kim a , K.R. Park a , M.K. Park a , H.L. Yang a , Y.S. Kim a , Y. Chu a , Y.O. Kim a , J.G. Bak a , E.N. Baang a , S.W. Yoon a , S.H. Hahn a , H.J. Lee a , S.H. Park a , K.H. Kim a , J. Hong a , S.H. Baek a , M.K. Kim a , T.G. Lee a , S.I. Lee a , Y.S. Bae a , H. Yonekawa a , J.H. Choi a , I.S. Hwang a , Y.J. Kim a , K.W. Cho a , Y.M. Park a , J.Y. Kim a , J.H. Lee a , J.S. Bak a , M. Kwon a , G.S. Lee a , J.G. Kwak b , H.S. Ahn c , M.L. Walker d , D.A. Humphreys d , J.A. Leuer d , A. Hyatt d , G. Jackson d , D. Mueller e , D.P. Ivanov f a
National Fusion Research Institute (NFRI), Daejeon, Republic of Korea Korea Atomic Energy Research Institute (KAERI), Daejeon, Republic of Korea c POSCON Cooperation, Kyungbuk, Republic of Korea d General Atomics (GA), San Diego, CA, USA e Princeton Plasma Physics Laboratory (PPPL), Princeton, NJ, USA f Nuclear Fusion Institute (NFI), RRCKI, Moscow, Russia b
a r t i c l e
i n f o
Article history: Available online 8 February 2009 Keywords: KSTAR Commissioning Cool-down Superconducting magnet Initial operation First plasma Preionization
a b s t r a c t The commissioning and the initial operation for the first plasma in the KSTAR device have been accomplished successfully without any severe failure preventing the device operation and plasma experiments. The commissioning is classified into four steps: vacuum commissioning, cryogenic cool-down commissioning, magnet system commissioning, and plasma discharge.Vacuum commissioning commenced after completion of the tokamak and basic ancillary systems construction. Base pressure of the vacuum vessel was about 3 × 10−6 Pa and that of the cryostat about 2.7 × 10−4 Pa, and both levels meet the KSTAR requirements to start the cool-down operation. All the SC magnets were cooled down by a 9 kW rated cryogenic helium facility and reached the base temperature of 4.5 K in a month. The performance test of the superconducting magnet showed that the joint resistances were below 3 n and the resistance to ground after cool-down was over 1 G. An ac loss test of each PF coil made by applying a dc biased sinusoidal current showed that the coupling loss was within the KSTAR requirement with the coupling loss time constant less than 35 ms for both Nb3 Sn and NbTi magnets. All the superconducting magnets operated in stable without quench for long-time dc operation and with synchronized pulse operation by the plasma control system (PCS). By using an 84 GHz ECH system, second harmonic ECH assisted plasma discharges were produced successfully with loop voltage of less than 3 V. By the real-time feedback control, operation of 100 kA plasma current with pulse length up to 865 ms was achieved, which also meet the first plasma target of 100 kA and 100 ms. The KSTAR device will be operated to meet the missions of steady-state and high-beta achievement by system upgrades and collaborative researches. © 2008 Elsevier B.V. All rights reserved.
1. Introduction The commissioning of the Korea superconducting tokamak advanced research (KSTAR) device has been accomplished successfully after completion of tokamak construction as shown in Fig. 1. The KSTAR device has missions to develop a steady-statecapable, advanced superconducting tokamak and to establish a scientific and technological basis for an attractive fusion reactor [1]. To contribute in world-wide fusion research and reactor engineering design, KSTAR will adopt the most outstanding research results
∗ Corresponding author. Tel.: +82 42 870 1751; fax: +82 42 870 1619. E-mail address: [email protected] (Y.-K. Oh). 0920-3796/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.fusengdes.2008.12.099
from the present devices and will exploit high performance steadystate operation which will provide the core technology for ITER and future reactor designs. The designed parameters of KSTAR, as shown in Table 1, are major radius 1.8 m, minor radius 0.5 m, toroidal field 3.5 T, and plasma current 2 MA with a strongly shaped plasma crosssection. The cross-sectional view of the KSTAR device is as Fig. 2. The specific features in the KSTAR design are installation of the fully superconducting magnets, especially the first tokamak device adopting Nb3 Sn superconductor for both TF and PF magnets [2]. The target of the first year operation is the completion of the device commissioning including the first plasma achievement with plasma current over 100 kA and duration longer than 100 ms [3]. The commissioning campaign consists of a set of performancetesting activities which verified sub-system operation and overall
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
345
Fig. 3. Picture of the KSTAR central control room during the SC magnet commissioning. Fig. 1. Bird-eye view of the KSTAR device after the completion of its construction. For the initial operation of the tokamak, basic ancillary systems have been installed such as the vacuum pumping systems, helium distribution system, ECH, and ICRH systems and diagnostic systems. Table 1 Key parameter comparison between KSTAR and ITER devices.
Major radius, R0 Minor radius, a Plasma current, Ip Elongation, k Triangularity, ı Toroidal field, B0 Pulse length Plasma volume Plasma cross-section Plasma shape Normalized beta Plasma fuel Superconductor Auxiliary heating/CD
KSTAR
ITER
1.8 m 0.5 m 2.0 MA 2.0 0.8 3.5 T 300 s 17.8 m3 1.6 m2 DN, SN 5.0 H, D Nb3 Sn, NbTi 28 MW
6.2 m 2.0 m 15 (17) MA 1.7 0.33 5.3 T 400 s 830 m3 22 m2 SN 1.8 (2.5) H, D, T Nb3 Sn, NbTi 73 (110) MW
system integration. The objectives of the commissioning are to demonstrate that the performance tests and system checks are in accordance with the design specifications, that they meet the performance criteria, and to identify any issue that would
Fig. 4. Overall sequence of the KSTAR commissioning and the first plasma operation.
prevent device operation and plasma experiments. Due to the limited access into the tokamak experimental hall during operation, almost all of the control actions were conducted in the main control room as shown in Fig. 3. In KSTAR, the commissioning was divided into four steps: (i) vacuum commissioning including leak detection, (ii) cryogenic cool-down commissioning, (iii) magnet system commissioning, and (iv) plasma discharge commissioning. The commissioning sequence including the annual plan, daily operation, and plasma discharge sequence is summarized in Fig. 4 and the achieved commissioning activities are listed in Table 2. In this paper, the overall commissioning results are introduced and some engineering achievements are described also, including superconducting magnet performance tests and preionization Table 2 Key dates and progress of the KSTAR commissioning.
Fig. 2. Elevation view of the KSTAR device.
Key dates
Commissioning progress
September 14, 2007 March 31, 2008 April 3, 2008
Tokamak construction completed Vacuum commissioning completed Magnet system cool-down started (using 9 kW helium refrigerator) SC phase transition detected (TF coil at 18 K) Cool-down completed (5 K, SHe 600 g/s) Joint resistance and coil insulation measured TF coil commissioning completed (15 kA, 8 h) First ICRH discharge test under TF field only SC coil commissioning completed TF, PF, ECH, gas, diagnostics synchronization test PF null-field measurement and Ohmic plasma started Plasma current over 100 kA achieved 1st plasma announcement (pulse length over 800 ms) Warm up started
April 23, 2008 April 26, 2008 May 5, 2008 May 12, 2008 May 13,2008 May 27, 2008 May 30, 2008 June 3, 2008 June 13, 2008 July 15, 2008 July 20, 2008
346
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
technique for plasma discharges by the second harmonic electron cyclotron heating (ECH) system. 2. Vacuum commissioning The vacuum commissioning phase consists of inspection of the vacuum condition and determination of the leak rate to guarantee long-time reliable operation. The vacuum system in the KSTAR has two separate volumes. The primary plasma vacuum is vacuum vessel and the secondary vacuum is within the cryostat for thermal isolation of the superconducting magnets [4]. The vacuum vessel has a volume of 100 m3 and a pumping system with capacity of 42,400 l/s. The achieved base pressure inside the vacuum vessel was approximately 3 × 10−6 Pa before baking. To reduce the outgas from the in-vessel components, the vessel was baked up to 100 ◦ C and was glow discharge cleaned (GDC) for long period using hydrogen and helium gas. A piezoelectric valve for gas puffing has been installed and is controlled by the plasma control system (PCS). The cryostat for the thermal isolation of the superconducting magnet has many internal components such as insulation supporters and multi-layered insulation. For the evacuation of the cryostat, there are seven sets of turbo pumps with capacity of 16,900 l/s. A base pressure at room temperature of 2.7 × 10−4 Pa was achieved. After the superconducting magnet cool-down the base pressure of the cryostat was changed to about 2.5 × 10−6 Pa. To measure the leak rate from the helium circuits inside the cryostat, a new method of leak detection was adopted. The helium partial pressure was measured by an RGA which was installed at the cryostat pumping duct after closing all the evacuation control valves. Most of the gases except helium are trapped on the magnet surfaces. The measured total helium leak rate in the cryostat was 8.9 × 10−9 Pa m3 /s. Good vacuum performance is a result of repeated detailed inspection of leaks and leak repair activities. 3. Cool-down commissioning The objectives of the cool-down commissioning phase are to cool the cryogenic components down to the operating temperature and to assess the performance of the system consists of a 9 kW rated helium refrigeration system (HRS) and a helium distribution system (HDS) which is installed between the HRS and KSTAR device. After the individual tests of the HRS and HDS, the magnet system cool-down was started on 2 April 2008 [5]. A major control requirement during the cool-down was to keep the temperature distribution differences between the cryogenic components within 50 K to prevent mechanical deformation due to unbalanced thermal contraction. The temperature differences between the TF coils and TF structure was also minimized by parallel helium supplies to prevent the de-lamination or fracture of the insulation layers between each TF coil and structure due to excessive thermal contraction of the coils. The thermal shield was cooled down simultaneously with the magnet system to prevent the condensing of vapors on the thermal shield which can cause degradation of the emissivity of the silver coating on the thermal shield surface. At the final stage of the cool-down, two sets of cold circulators were operated to supply the supercritical helium into the TF and PF magnet systems up to 600 g/s in total. The cool-down commissioning was completed in 1 month as shown in Fig. 5. To monitor the thermal contraction of the superconducting magnets, several displacement sensors were installed at four locations on the gravity support in the radial and vertical directions. The measured radial contraction from the 297 to 11 K was about 7.8 mm inward and the values are in agreement with the finite element (FE) method calculation result. The flow rate deviation between each member of a set of four TF coils in a quadrant was less than 10% and the deviation was kept constant in spite of variations of the total
Fig. 5. The temperature variation of the superconducting magnet system during the cool-down commissioning. The temperature distribution was controlled to be within 50 K for minimum structural deformation.
flow rate. The 10% flow deviation was consistent with the results of individual coil tests at room temperature. The reliability of the sensors which were installed inside the cryostat was also tested. After cool-down, 16 sensors malfunctioned among the set of 857 sensors which consists of 379 temperature sensors, 244 strain sensors, 206 voltage taps for quench detection, and other sensors including displacement sensors, Hall probes, and acoustic emission (AE) sensors [6]. The AE sensors were prepared to monitor the mechanical movement of the superconducting buslines and the frictional motion of the structure interface. However, due to an inadequate selection of the instruments, there was so much high-frequency noise that meaningful detection became very difficult. To improve data acquisition reliability, low-pass filters and an improved data acquisition system will be installed for the next machine campaign. 4. SC magnet commissioning The objectives of the superconducting magnet commissioning were to investigate whether all the superconducting magnets and interfaces could stably operate at the cryogenic temperature and under high current operation conditions and to check the controllability of the coil currents and magnetic field configurations for plasma discharges. There are two kinds of the superconducting material in the KSTAR magnets, one is Nb3 Sn superconductor which used in the 16 TF coils and 5 pairs of PF coils (PF1 ∼ PF5). The Nb3 Sn superconductors generally have a larger temperature margin at higher field and fast field variation conditions. NbTi superconductor is used in the 2 pairs of large PF coils (PF6 and PF7) as a consequence of their lower field operational requirements [7]. The several interface structures on the superconducting magnets consist of the current leads, SC buslines, and several joints [8]. There are two kinds of joints in KSTAR. The first type of joint is a compact strand-to-strand (STS) type joint for the TF coil terminals, and the second type of joint is a lap-type joint [9]. All the conductor surfaces inside the cryostat were covered with glass fiber reinforced plastic (GFRP) to provide electrical insulation for the several kV operation voltage conditions required for plasma operation [10]. 4.1. SC magnet inspection Several magnet qualification tests were performed before and after magnet cool-down. Prior to magnet system cool-down field distribution measurements were made inside the vacuum vessel at room temperature. For these measurements several arrays of Hall
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
probes were installed inside vacuum vessel, and a current of about 100 A was applied to the TF or PF coils. The results showed that the vertical field from the PF coil currents without the TF energized showed large differences compared to linear calculations of the field from the PF current alone. The difference is a result of magnetization of the Incoloy908 jacket material used in the Nb3 Sn magnets and produces a remnant field in the range of a few tens of gauss depending on the current history [11]. Furthermore, all Nb3 Sn magnets showed a higher inductance at current levels below approximately 500 A from the non-linear behavior of the Incoloy908 material. Finite element methods were used to investigate these phenomena and modifications in the PF coil currents were developed to compensate for these effects. As standard procedures for coil inspections prior to the high current operation, superconducting phase transition measurements, joint resistance measurements, and coil insulation tests were conducted. Each coil resistance was monitored during cool-down. The relative resistance ratios (RRR) of all coils were over 100 and this meant that the copper in the superconducting cable could act as a stabilizer in case of coil abnormal operation. The superconducting phase transition also occurred at the temperature of about 18 K for the Nb3 Sn coils and about 10 K for the NbTi coils. These results show that all the coils were superconducting and far from quench at the normal 4.5 K operating temperature. For the joint resistance measurements, a standalone power supply with a current rating of 1 kA was used. The resistance was calculated by slope measurements of the signals based on an applied current variation from + 1 to −1 kA. For the TF joints, additional measurements were conducted using the TF power supply in the range of 10 kA. To eliminate the power supply’s control signal noise in the measurement, the resistance was measured after the converter turned off and the coil was in a persistent operation mode with slow current decay through the bypass diodes in the power supply. The measured results showed that the joint resistance was less than 3 n. The dielectric strength of each coil was measured before and after cool-down. The insulation was stable under a 6 kV test voltage. 4.2. TF magnet operation test The TF magnet power supply uses an insulated gate bipolar transistor (IGBT) type convertor and has a rating of 40 kA. Since the IGBT convertor could not apply negative voltage, an additional slow discharge resistor was prepared for the current discharge of the TF coil after plasma experiments. For the initial operation the current of the TF power supply was tested up to 15 kA. As a protection system of the superconducting magnets, a quench detection system was developed. In the event of a quench occurring in the magnets, the quench detection system triggers a quench protection circuit inside the power supply. The TF system was tested by increasing the current level in steps. Following step testing the TF system operated stably for 8 h at 15 kA. The thermo-hydraulic and magnetic parameters also were measured for the operating conditions. Fig. 6 shows the TF magnet operation features. The current was maintained at 15 kA for 2 h and discharged using a slow discharge system. The magnet operated stably with the temperature rising less than 0.1 K during the current charge and discharge periods. The field inside the vacuum vessel measured by movable Hall probes was accurate within 0.5% compared to calculations. The measured mechanical strain of the TF structure at a TF current of 15 kA was about 210 and equivalent stress was about 42 MPa. From this result, the stress at the rated current of 35 kA was estimated to be about 160 MPa. If the 93 MPa stress measured during the cool-down is added, then the total stress is expected to be about 253 MPa at 35 kA. This is much less than the material operational criteria of 700 MPa and in accordance with the magnet design specification. The quench detection
347
Fig. 6. The TF magnets operated stably in electrical, thermal, and mechanical aspects under the long-pulse operation up to 15 kA.
system also operated reliably. During the current charging and discharging time, the detected signal was less than 25 mV and well below the trigger condition of 50 mV and 2 s [12]. 4.3. PF magnet operation test Prior to the PF power supply operation with PF coil connections, each power supply was tested by connection to a dummy copper coil. The dummy coil had an inductance of 29 mH and resistance of 50 m. The current control was done using the PCS system. The control items were the reference current waveform control, fast current variation using the blip resistor insertion system (BRIS) operation, and the mutual inductance effects from two power supply operation. After each power supply test using the dummy coil, a power supply test with the superconducting coils was conducted. Major test items were reference current controls, BRIS operation tests, adjustment of the current ramping speed, step response tests, and protection mode tests. During the power supply control tests, a magnetic diagnostic signal detection and analysis was also conducted. To analyze the Incoloy908 effect on the field with currents in the PF1 and PF7 coils, the magnetic measurements were compared with and without the TF current. The measurement results showed that the remnant field was about 40 Gauss in the case of zero TF current. But with 15 kA TF current, the remnant field was in the range of a few Gauss. After the individual coil tests, an integrated test for 7 PF coils was conducted. In the initial operation, all the power supplies were operated in a unipolar condition. Fig. 7 shows one of the power supply control signals for the plasma discharge. The overall current ramp-up was started about 7 s before the blip phase (BRIS active phase), and the blip duration was about 150 ms. 4.4. AC loss measurement There are two kinds of ac losses, one is the hysteresis loss which depends on the material type and the other is the coupling loss which depends on the inter-strand contact resistance and cabling. In an environment of slowly varying currents or fields, the hysteresis loss is dominant, but in an environment of high-frequency currents or high field strengths, the coupling loss can be dominant. For KSTAR operation at currents up to 2 MA and long pulses up to 300 s, the ac losses can be a very important factor in deciding the operational capability. Measurements of the ac losses in the KSTAR PF magnet system are difficult because of lack of installed sensors.
348
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
even though it has NbTi conductor. These results could come from the adoption of chrome coating on all three kinds of strands, Nb3 Sn, NbTi, and copper. These results show that the KSTAR construction has advantages for various operational capabilities especially high speed PF6 coil operation as in the first plasma operation. The superconducting magnet commissioning was completed without any severe problems in spite of the fact that individual cool-down tests were not conducted prior to assembly. 5. Plasma commissioning
Fig. 7. Representative current waveform of the PF coils. Seven pairs of PF coils were charged for field null formation and this was followed by blip and feedback current control using the plasma control system.
The conventional measuring method using single pulse operation is not accurate in KSTAR. When the tests are at a low current of around 500 A, there could be uncertainties due to power supply controls or magnetic material effects. So a new approach to determine coil ac losses was adopted for the KSTAR CS model coil [13,14]. Fig. 8 shows the current waveform for the PF1 ac loss. An ac current of 0.5 kA is applied for 600 s with a dc biasing current of 2 kA. The frequency was changed from 0.1 to 0.2 Hz. The figure shows that thermal parameters reached a steady-state condition after 350 s. The comparison of the coupling loss of PF1 was as follows, the coupling loss time constant, n, by trapezoidal method was about 62.5 ms at the initial operation time and about 50 ms after 2 months operation. It could be expected that the mechanical detachment between strands from many pulses resulted in the reduced coupling loss. When compared with sinusoidal method, the measured time constant was about 35 ms. In the trapezoidal test, the error could come from the uncertainty at low current. Another interesting result is that the coupling time constant of the PF6 coil is also about 33 ms
Fig. 8. The ac loss measurement of the PF1 coil under the steady-state operation condition by applying the dc biased sinusoidal current with 2 kA dc bias and 0.5 kA ac current for various frequencies from dc to 0.2 Hz.
The main purpose of the plasma operation commissioning is for the final verification of the machine configurations as a plasmaexperiment capable device through the synchronized operation of all essential components. Other purposes are the development of feasible startup scenarios under strict hardware limitations and low-voltage breakdown achievement using ECH system. The installed hardware and operation environment of KSTAR in the commissioning phase had more limitations than the designed environment. To confirm that the device can create a tokamak plasma, the operation goal for first plasma was sustainment of a 100 kA hydrogen plasma for 100 ms. The commissioning process of the first plasma began with the check of the magnetic configuration of the TF and PF magnets by measurements of the magnetic fields inside the vacuum vessel with the same initial magnetization (IM) conditions of the field magnets. Since the Incoloy908 material used in the Nb3 Sn magnets introduced a certain amount of uncertainty into the magnetic measurements, detailed calculations of the IM scenarios for the startup were performed based on both the experimental information from the movable Hall-probe measurements and a 2-D FE model for Incoloy908 magnetization. To insure lowvoltage startup, ECH preionization was successfully utilized [15]. ECH consists of the second harmonic resonance with microwave power from the 500 kW, 84 GHz gyrotron. Experiments seeking the best preionization conditions were performed under various prefill pressures and IM states of the PF coil currents, and provided synchronization of PF field null configuration with gas injection and ECH for the plasma startup. This kind of “dress rehearsal” for the first plasma enabled the ECH preionization to be reliably generated at the 2nd harmonic gyro-frequency resonance in the 1.5 T TF field, in the optimum gas pressure range. Fig. 4 shows the main sequence of events for a typical plasma discharge. Experiments on the initial field null configurations also have been performed for two different types of IM scenarios. First we tried to maximize the flux swing and the size of the field null region using a “conventional” PF coil current distribution with currents decreasing magnitude between the inside and outside of the machine. The scenario produced plasma up to 100 kA in shot #794. To maximize flux swing capacity of the machine, a unique IM scenario, called the “dipole mode”, was tested and developed. In this mode the outer two coil pairs (PF6 and PF7) have approximately equal and opposite currents and provide better flux and capability with the commissioning phase power and current constraints. This mode has a smaller field null than the conventional mode but can have a higher plasma current ramp rate and more degrees of freedom in the outer PFs to regulate the radial force balance. All the successful startups were essentially ECH-assisted with low loop voltages of 2–3.5 V. Using the dipole scenario, higher current ramping speeds around 1 MA/s were obtained and maximum plasma flat top times were obtained. Development of the plasma control system with the DIII-D team provided flexible control of the coil system and provided for operation of 100 kA plasma current and with pulse length up to 865 ms using real-time feedback of the plasma current (Ip ) and radial position (Rp ) [16]. Experiments were carried out to control plasma current down to the 50 kA level at the end of the shot in order to
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
349
Fig. 9. Results from the first plasma discharges: (a) an Ohmic plasma discharge image, (b) a field null configuration at the initial magnetization, and (c) plasma parameters for two plasma shots; a conventional mode feed-forward plasma (#794) and a dipole mode feedback controlled plasma (#1216).
minimize the effects of vertical disruption events (VDE) which were inevitable when the volt-seconds had been exhausted. Fig. 9 shows one of the results of a plasma pulse utilizing the plasma feedback mentioned above. At a TF current of 14 kA, an ECH resonance layer formed at R = 1.7 m. Under the IM conditions of the dipole scenario, the blip resistors were used for 150 ms, after which the Ip and Rp feedback algorithm was turned on for 600 ms. The ECH was set to start at −30 ms before the PF blip and lasted for 330 ms with a modulated ECH power of 350 kW, aimed 10 cm below the midplane. A second gas puff was added to suppress a kind of MHD instability which occasionally occurs in the burn-through phase. Additional experiments on the optimization of the heating devices have been performed. The ECH incident angle dependency and power threshold have been investigated and several shots on ion cyclotron range of frequency (ICRF) heating efficiency research were performed, utilizing various diagnostics such as the electron cyclotron emission (ECE) radiometer, interferometer, visible spectroscopy, and 10 kHz filter-scopes. Effects of wall conditioning were investigated by performing between-shot ICRF cleaning and glow discharge cleaning with H or He gas at night. It seems that 2 h of night GDC produced wall conditions allowing reliable discharges in the morning; however, there are many uncertainties in analyzing the effects of the wall conditioning method for reproducible plasma discharges. Warm-up of the device starting on 20 July 2008 has been performed after completing all experiments. To keep a uniform temperature distribution, forced helium circulation using the helium refrigerator was adopted until the coil temperature reached around 200 K. At that temperature the helium refrigerator was turned off and natural warm up was allowed.
6. Conclusions The successful commissioning and the first plasma achievement in KSTAR validate the design, engineering and construction aspects of the first major Nb3 Sn superconducting tokamak. One of the remarkable aspects of the KSTAR commissioning is that all commissioning phases were achieved without major problems and within the schedule. KSTAR commissioning showed reliable operational results for the Nb3 Sn superconducting magnet engineering. This is a meaningful benchmark for future SC fusion devices like ITER because they will utilize the same type of Nb3 Sn superconductor. Another major achievement is the successful low-voltage startup achievement by adopting the second harmonic ECH operation. Using this technology a loop voltage of less than 3 V produced reasonable plasmas. The KSTAR device will be operated to demonstrate the steadystate operation capability of high performance advanced tokamak modes. The major research plan for the initial operation phase is to establish the basic operation skills of superconducting tokamak operation through Ohmic, L-, and H-mode plasma experiments for relatively short-pulse lengths up to 20. For this research, in-vessel systems such as limiters, divertors, passive plates, in-vessel control coils, and heating systems will be upgraded in steps. This work is supported by the Ministry of the Education, Science, and Technology. Acknowledgement The authors would like to thank all the KSTAR participants from Korea and from the international collaboration partners who made
350
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
efforts for the successful commissioning during and the first plasma operation. References [1] G.S. Lee, et al., Design and construction of the KSTAR tokamak, Nuclear Fusion 41 (2001) 1515. [2] J.S. Bak, et al., Status of the KSTAR tokamak construction, Fusion Engineering and Design 81 (2006) 2315. [3] Y.K. Oh, et al., Completion of the KSTAR construction and its role as ITER pilot device, Fusion Engineering and Design 83 (2008) 804. [4] H.L. Yang, et al., Progress in the assembly of the KSTAR tokamak, in: Proceedings of the 20th IAEA FEC, Portugal, 2004, FT/P7–17. [5] Y.S. Kim, et al. Commissioning results of the KSTAR cryogenic system, ASC2008 conference, Chicago, submitted for publication to IEEE Transactions on Applied Superconductivity (2009). [6] Y.O. Kim, Local instruments of the KSTAR tokamak, Journal of the Korean Physics Society 49 (2006) S236. [7] K. Kim, et al., Status of the KSTAR superconducting magnet system development, Nuclear Fusion 45 (2005) 783.
[8] Y.M. Park, et al., Engineering design status of the KSTAR SC bus-line, IEEE Transactions on Applied Superconductivity 14 (2004) 1770. [9] K.R. Park, et al., Joint development for KSTAR CS coil, Journal of the Korean Physics Society 49 (2006) S238. [10] W. Chung, et al., Electrical insulation of KSTAR magnet lead, Journal of the Korean Physics Society vol.49 (2006) S228. [11] H. Yonekawa, et al., KSTAR magnetic field measurement at room temperature and remnant field evaluation, ASC2008 conference, Chicago, submitted for publication to IEEE Transactions on Applied Superconductivity (2009). [12] Y. Chu, et al., Quench detection based on voltage measurement for the KSTAR superconducting coils, ASC2008 conference, Chicago, submitted for publication to IEEE Transactions on Applied Superconductivity (2009). [13] Y.K. Oh, et al., Status of the KSTAR superconducting coil fabrication and test, in: Proceedings of the fourth IAEA TM on SSO, 2005 (POS-FT7). [14] S. Lee, et al., AC loss characteristics of the KSTAR CSMC estimated by pulse test, IEEE Transactions on Applied Superconductivity 16 (2006) 771. [15] Y.S. Bae, et al., Status of KSTAR electron cyclotron heating system, Fusion Science and Technology 52 (2007) 321. [16] S. Hahn, et al., Plasma control system for the day-one operation of KSTAR tokamak, 25th SOFT, submitted for publication, P3.23, this issue.
Contents lists available at ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Commissioning and initial operation of KSTAR superconducting tokamak Yeong-Kook Oh a,∗ , W.C. Kim a , K.R. Park a , M.K. Park a , H.L. Yang a , Y.S. Kim a , Y. Chu a , Y.O. Kim a , J.G. Bak a , E.N. Baang a , S.W. Yoon a , S.H. Hahn a , H.J. Lee a , S.H. Park a , K.H. Kim a , J. Hong a , S.H. Baek a , M.K. Kim a , T.G. Lee a , S.I. Lee a , Y.S. Bae a , H. Yonekawa a , J.H. Choi a , I.S. Hwang a , Y.J. Kim a , K.W. Cho a , Y.M. Park a , J.Y. Kim a , J.H. Lee a , J.S. Bak a , M. Kwon a , G.S. Lee a , J.G. Kwak b , H.S. Ahn c , M.L. Walker d , D.A. Humphreys d , J.A. Leuer d , A. Hyatt d , G. Jackson d , D. Mueller e , D.P. Ivanov f a
National Fusion Research Institute (NFRI), Daejeon, Republic of Korea Korea Atomic Energy Research Institute (KAERI), Daejeon, Republic of Korea c POSCON Cooperation, Kyungbuk, Republic of Korea d General Atomics (GA), San Diego, CA, USA e Princeton Plasma Physics Laboratory (PPPL), Princeton, NJ, USA f Nuclear Fusion Institute (NFI), RRCKI, Moscow, Russia b
a r t i c l e
i n f o
Article history: Available online 8 February 2009 Keywords: KSTAR Commissioning Cool-down Superconducting magnet Initial operation First plasma Preionization
a b s t r a c t The commissioning and the initial operation for the first plasma in the KSTAR device have been accomplished successfully without any severe failure preventing the device operation and plasma experiments. The commissioning is classified into four steps: vacuum commissioning, cryogenic cool-down commissioning, magnet system commissioning, and plasma discharge.Vacuum commissioning commenced after completion of the tokamak and basic ancillary systems construction. Base pressure of the vacuum vessel was about 3 × 10−6 Pa and that of the cryostat about 2.7 × 10−4 Pa, and both levels meet the KSTAR requirements to start the cool-down operation. All the SC magnets were cooled down by a 9 kW rated cryogenic helium facility and reached the base temperature of 4.5 K in a month. The performance test of the superconducting magnet showed that the joint resistances were below 3 n and the resistance to ground after cool-down was over 1 G. An ac loss test of each PF coil made by applying a dc biased sinusoidal current showed that the coupling loss was within the KSTAR requirement with the coupling loss time constant less than 35 ms for both Nb3 Sn and NbTi magnets. All the superconducting magnets operated in stable without quench for long-time dc operation and with synchronized pulse operation by the plasma control system (PCS). By using an 84 GHz ECH system, second harmonic ECH assisted plasma discharges were produced successfully with loop voltage of less than 3 V. By the real-time feedback control, operation of 100 kA plasma current with pulse length up to 865 ms was achieved, which also meet the first plasma target of 100 kA and 100 ms. The KSTAR device will be operated to meet the missions of steady-state and high-beta achievement by system upgrades and collaborative researches. © 2008 Elsevier B.V. All rights reserved.
1. Introduction The commissioning of the Korea superconducting tokamak advanced research (KSTAR) device has been accomplished successfully after completion of tokamak construction as shown in Fig. 1. The KSTAR device has missions to develop a steady-statecapable, advanced superconducting tokamak and to establish a scientific and technological basis for an attractive fusion reactor [1]. To contribute in world-wide fusion research and reactor engineering design, KSTAR will adopt the most outstanding research results
∗ Corresponding author. Tel.: +82 42 870 1751; fax: +82 42 870 1619. E-mail address: [email protected] (Y.-K. Oh). 0920-3796/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.fusengdes.2008.12.099
from the present devices and will exploit high performance steadystate operation which will provide the core technology for ITER and future reactor designs. The designed parameters of KSTAR, as shown in Table 1, are major radius 1.8 m, minor radius 0.5 m, toroidal field 3.5 T, and plasma current 2 MA with a strongly shaped plasma crosssection. The cross-sectional view of the KSTAR device is as Fig. 2. The specific features in the KSTAR design are installation of the fully superconducting magnets, especially the first tokamak device adopting Nb3 Sn superconductor for both TF and PF magnets [2]. The target of the first year operation is the completion of the device commissioning including the first plasma achievement with plasma current over 100 kA and duration longer than 100 ms [3]. The commissioning campaign consists of a set of performancetesting activities which verified sub-system operation and overall
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
345
Fig. 3. Picture of the KSTAR central control room during the SC magnet commissioning. Fig. 1. Bird-eye view of the KSTAR device after the completion of its construction. For the initial operation of the tokamak, basic ancillary systems have been installed such as the vacuum pumping systems, helium distribution system, ECH, and ICRH systems and diagnostic systems. Table 1 Key parameter comparison between KSTAR and ITER devices.
Major radius, R0 Minor radius, a Plasma current, Ip Elongation, k Triangularity, ı Toroidal field, B0 Pulse length Plasma volume Plasma cross-section Plasma shape Normalized beta Plasma fuel Superconductor Auxiliary heating/CD
KSTAR
ITER
1.8 m 0.5 m 2.0 MA 2.0 0.8 3.5 T 300 s 17.8 m3 1.6 m2 DN, SN 5.0 H, D Nb3 Sn, NbTi 28 MW
6.2 m 2.0 m 15 (17) MA 1.7 0.33 5.3 T 400 s 830 m3 22 m2 SN 1.8 (2.5) H, D, T Nb3 Sn, NbTi 73 (110) MW
system integration. The objectives of the commissioning are to demonstrate that the performance tests and system checks are in accordance with the design specifications, that they meet the performance criteria, and to identify any issue that would
Fig. 4. Overall sequence of the KSTAR commissioning and the first plasma operation.
prevent device operation and plasma experiments. Due to the limited access into the tokamak experimental hall during operation, almost all of the control actions were conducted in the main control room as shown in Fig. 3. In KSTAR, the commissioning was divided into four steps: (i) vacuum commissioning including leak detection, (ii) cryogenic cool-down commissioning, (iii) magnet system commissioning, and (iv) plasma discharge commissioning. The commissioning sequence including the annual plan, daily operation, and plasma discharge sequence is summarized in Fig. 4 and the achieved commissioning activities are listed in Table 2. In this paper, the overall commissioning results are introduced and some engineering achievements are described also, including superconducting magnet performance tests and preionization Table 2 Key dates and progress of the KSTAR commissioning.
Fig. 2. Elevation view of the KSTAR device.
Key dates
Commissioning progress
September 14, 2007 March 31, 2008 April 3, 2008
Tokamak construction completed Vacuum commissioning completed Magnet system cool-down started (using 9 kW helium refrigerator) SC phase transition detected (TF coil at 18 K) Cool-down completed (5 K, SHe 600 g/s) Joint resistance and coil insulation measured TF coil commissioning completed (15 kA, 8 h) First ICRH discharge test under TF field only SC coil commissioning completed TF, PF, ECH, gas, diagnostics synchronization test PF null-field measurement and Ohmic plasma started Plasma current over 100 kA achieved 1st plasma announcement (pulse length over 800 ms) Warm up started
April 23, 2008 April 26, 2008 May 5, 2008 May 12, 2008 May 13,2008 May 27, 2008 May 30, 2008 June 3, 2008 June 13, 2008 July 15, 2008 July 20, 2008
346
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
technique for plasma discharges by the second harmonic electron cyclotron heating (ECH) system. 2. Vacuum commissioning The vacuum commissioning phase consists of inspection of the vacuum condition and determination of the leak rate to guarantee long-time reliable operation. The vacuum system in the KSTAR has two separate volumes. The primary plasma vacuum is vacuum vessel and the secondary vacuum is within the cryostat for thermal isolation of the superconducting magnets [4]. The vacuum vessel has a volume of 100 m3 and a pumping system with capacity of 42,400 l/s. The achieved base pressure inside the vacuum vessel was approximately 3 × 10−6 Pa before baking. To reduce the outgas from the in-vessel components, the vessel was baked up to 100 ◦ C and was glow discharge cleaned (GDC) for long period using hydrogen and helium gas. A piezoelectric valve for gas puffing has been installed and is controlled by the plasma control system (PCS). The cryostat for the thermal isolation of the superconducting magnet has many internal components such as insulation supporters and multi-layered insulation. For the evacuation of the cryostat, there are seven sets of turbo pumps with capacity of 16,900 l/s. A base pressure at room temperature of 2.7 × 10−4 Pa was achieved. After the superconducting magnet cool-down the base pressure of the cryostat was changed to about 2.5 × 10−6 Pa. To measure the leak rate from the helium circuits inside the cryostat, a new method of leak detection was adopted. The helium partial pressure was measured by an RGA which was installed at the cryostat pumping duct after closing all the evacuation control valves. Most of the gases except helium are trapped on the magnet surfaces. The measured total helium leak rate in the cryostat was 8.9 × 10−9 Pa m3 /s. Good vacuum performance is a result of repeated detailed inspection of leaks and leak repair activities. 3. Cool-down commissioning The objectives of the cool-down commissioning phase are to cool the cryogenic components down to the operating temperature and to assess the performance of the system consists of a 9 kW rated helium refrigeration system (HRS) and a helium distribution system (HDS) which is installed between the HRS and KSTAR device. After the individual tests of the HRS and HDS, the magnet system cool-down was started on 2 April 2008 [5]. A major control requirement during the cool-down was to keep the temperature distribution differences between the cryogenic components within 50 K to prevent mechanical deformation due to unbalanced thermal contraction. The temperature differences between the TF coils and TF structure was also minimized by parallel helium supplies to prevent the de-lamination or fracture of the insulation layers between each TF coil and structure due to excessive thermal contraction of the coils. The thermal shield was cooled down simultaneously with the magnet system to prevent the condensing of vapors on the thermal shield which can cause degradation of the emissivity of the silver coating on the thermal shield surface. At the final stage of the cool-down, two sets of cold circulators were operated to supply the supercritical helium into the TF and PF magnet systems up to 600 g/s in total. The cool-down commissioning was completed in 1 month as shown in Fig. 5. To monitor the thermal contraction of the superconducting magnets, several displacement sensors were installed at four locations on the gravity support in the radial and vertical directions. The measured radial contraction from the 297 to 11 K was about 7.8 mm inward and the values are in agreement with the finite element (FE) method calculation result. The flow rate deviation between each member of a set of four TF coils in a quadrant was less than 10% and the deviation was kept constant in spite of variations of the total
Fig. 5. The temperature variation of the superconducting magnet system during the cool-down commissioning. The temperature distribution was controlled to be within 50 K for minimum structural deformation.
flow rate. The 10% flow deviation was consistent with the results of individual coil tests at room temperature. The reliability of the sensors which were installed inside the cryostat was also tested. After cool-down, 16 sensors malfunctioned among the set of 857 sensors which consists of 379 temperature sensors, 244 strain sensors, 206 voltage taps for quench detection, and other sensors including displacement sensors, Hall probes, and acoustic emission (AE) sensors [6]. The AE sensors were prepared to monitor the mechanical movement of the superconducting buslines and the frictional motion of the structure interface. However, due to an inadequate selection of the instruments, there was so much high-frequency noise that meaningful detection became very difficult. To improve data acquisition reliability, low-pass filters and an improved data acquisition system will be installed for the next machine campaign. 4. SC magnet commissioning The objectives of the superconducting magnet commissioning were to investigate whether all the superconducting magnets and interfaces could stably operate at the cryogenic temperature and under high current operation conditions and to check the controllability of the coil currents and magnetic field configurations for plasma discharges. There are two kinds of the superconducting material in the KSTAR magnets, one is Nb3 Sn superconductor which used in the 16 TF coils and 5 pairs of PF coils (PF1 ∼ PF5). The Nb3 Sn superconductors generally have a larger temperature margin at higher field and fast field variation conditions. NbTi superconductor is used in the 2 pairs of large PF coils (PF6 and PF7) as a consequence of their lower field operational requirements [7]. The several interface structures on the superconducting magnets consist of the current leads, SC buslines, and several joints [8]. There are two kinds of joints in KSTAR. The first type of joint is a compact strand-to-strand (STS) type joint for the TF coil terminals, and the second type of joint is a lap-type joint [9]. All the conductor surfaces inside the cryostat were covered with glass fiber reinforced plastic (GFRP) to provide electrical insulation for the several kV operation voltage conditions required for plasma operation [10]. 4.1. SC magnet inspection Several magnet qualification tests were performed before and after magnet cool-down. Prior to magnet system cool-down field distribution measurements were made inside the vacuum vessel at room temperature. For these measurements several arrays of Hall
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
probes were installed inside vacuum vessel, and a current of about 100 A was applied to the TF or PF coils. The results showed that the vertical field from the PF coil currents without the TF energized showed large differences compared to linear calculations of the field from the PF current alone. The difference is a result of magnetization of the Incoloy908 jacket material used in the Nb3 Sn magnets and produces a remnant field in the range of a few tens of gauss depending on the current history [11]. Furthermore, all Nb3 Sn magnets showed a higher inductance at current levels below approximately 500 A from the non-linear behavior of the Incoloy908 material. Finite element methods were used to investigate these phenomena and modifications in the PF coil currents were developed to compensate for these effects. As standard procedures for coil inspections prior to the high current operation, superconducting phase transition measurements, joint resistance measurements, and coil insulation tests were conducted. Each coil resistance was monitored during cool-down. The relative resistance ratios (RRR) of all coils were over 100 and this meant that the copper in the superconducting cable could act as a stabilizer in case of coil abnormal operation. The superconducting phase transition also occurred at the temperature of about 18 K for the Nb3 Sn coils and about 10 K for the NbTi coils. These results show that all the coils were superconducting and far from quench at the normal 4.5 K operating temperature. For the joint resistance measurements, a standalone power supply with a current rating of 1 kA was used. The resistance was calculated by slope measurements of the signals based on an applied current variation from + 1 to −1 kA. For the TF joints, additional measurements were conducted using the TF power supply in the range of 10 kA. To eliminate the power supply’s control signal noise in the measurement, the resistance was measured after the converter turned off and the coil was in a persistent operation mode with slow current decay through the bypass diodes in the power supply. The measured results showed that the joint resistance was less than 3 n. The dielectric strength of each coil was measured before and after cool-down. The insulation was stable under a 6 kV test voltage. 4.2. TF magnet operation test The TF magnet power supply uses an insulated gate bipolar transistor (IGBT) type convertor and has a rating of 40 kA. Since the IGBT convertor could not apply negative voltage, an additional slow discharge resistor was prepared for the current discharge of the TF coil after plasma experiments. For the initial operation the current of the TF power supply was tested up to 15 kA. As a protection system of the superconducting magnets, a quench detection system was developed. In the event of a quench occurring in the magnets, the quench detection system triggers a quench protection circuit inside the power supply. The TF system was tested by increasing the current level in steps. Following step testing the TF system operated stably for 8 h at 15 kA. The thermo-hydraulic and magnetic parameters also were measured for the operating conditions. Fig. 6 shows the TF magnet operation features. The current was maintained at 15 kA for 2 h and discharged using a slow discharge system. The magnet operated stably with the temperature rising less than 0.1 K during the current charge and discharge periods. The field inside the vacuum vessel measured by movable Hall probes was accurate within 0.5% compared to calculations. The measured mechanical strain of the TF structure at a TF current of 15 kA was about 210 and equivalent stress was about 42 MPa. From this result, the stress at the rated current of 35 kA was estimated to be about 160 MPa. If the 93 MPa stress measured during the cool-down is added, then the total stress is expected to be about 253 MPa at 35 kA. This is much less than the material operational criteria of 700 MPa and in accordance with the magnet design specification. The quench detection
347
Fig. 6. The TF magnets operated stably in electrical, thermal, and mechanical aspects under the long-pulse operation up to 15 kA.
system also operated reliably. During the current charging and discharging time, the detected signal was less than 25 mV and well below the trigger condition of 50 mV and 2 s [12]. 4.3. PF magnet operation test Prior to the PF power supply operation with PF coil connections, each power supply was tested by connection to a dummy copper coil. The dummy coil had an inductance of 29 mH and resistance of 50 m. The current control was done using the PCS system. The control items were the reference current waveform control, fast current variation using the blip resistor insertion system (BRIS) operation, and the mutual inductance effects from two power supply operation. After each power supply test using the dummy coil, a power supply test with the superconducting coils was conducted. Major test items were reference current controls, BRIS operation tests, adjustment of the current ramping speed, step response tests, and protection mode tests. During the power supply control tests, a magnetic diagnostic signal detection and analysis was also conducted. To analyze the Incoloy908 effect on the field with currents in the PF1 and PF7 coils, the magnetic measurements were compared with and without the TF current. The measurement results showed that the remnant field was about 40 Gauss in the case of zero TF current. But with 15 kA TF current, the remnant field was in the range of a few Gauss. After the individual coil tests, an integrated test for 7 PF coils was conducted. In the initial operation, all the power supplies were operated in a unipolar condition. Fig. 7 shows one of the power supply control signals for the plasma discharge. The overall current ramp-up was started about 7 s before the blip phase (BRIS active phase), and the blip duration was about 150 ms. 4.4. AC loss measurement There are two kinds of ac losses, one is the hysteresis loss which depends on the material type and the other is the coupling loss which depends on the inter-strand contact resistance and cabling. In an environment of slowly varying currents or fields, the hysteresis loss is dominant, but in an environment of high-frequency currents or high field strengths, the coupling loss can be dominant. For KSTAR operation at currents up to 2 MA and long pulses up to 300 s, the ac losses can be a very important factor in deciding the operational capability. Measurements of the ac losses in the KSTAR PF magnet system are difficult because of lack of installed sensors.
348
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
even though it has NbTi conductor. These results could come from the adoption of chrome coating on all three kinds of strands, Nb3 Sn, NbTi, and copper. These results show that the KSTAR construction has advantages for various operational capabilities especially high speed PF6 coil operation as in the first plasma operation. The superconducting magnet commissioning was completed without any severe problems in spite of the fact that individual cool-down tests were not conducted prior to assembly. 5. Plasma commissioning
Fig. 7. Representative current waveform of the PF coils. Seven pairs of PF coils were charged for field null formation and this was followed by blip and feedback current control using the plasma control system.
The conventional measuring method using single pulse operation is not accurate in KSTAR. When the tests are at a low current of around 500 A, there could be uncertainties due to power supply controls or magnetic material effects. So a new approach to determine coil ac losses was adopted for the KSTAR CS model coil [13,14]. Fig. 8 shows the current waveform for the PF1 ac loss. An ac current of 0.5 kA is applied for 600 s with a dc biasing current of 2 kA. The frequency was changed from 0.1 to 0.2 Hz. The figure shows that thermal parameters reached a steady-state condition after 350 s. The comparison of the coupling loss of PF1 was as follows, the coupling loss time constant, n, by trapezoidal method was about 62.5 ms at the initial operation time and about 50 ms after 2 months operation. It could be expected that the mechanical detachment between strands from many pulses resulted in the reduced coupling loss. When compared with sinusoidal method, the measured time constant was about 35 ms. In the trapezoidal test, the error could come from the uncertainty at low current. Another interesting result is that the coupling time constant of the PF6 coil is also about 33 ms
Fig. 8. The ac loss measurement of the PF1 coil under the steady-state operation condition by applying the dc biased sinusoidal current with 2 kA dc bias and 0.5 kA ac current for various frequencies from dc to 0.2 Hz.
The main purpose of the plasma operation commissioning is for the final verification of the machine configurations as a plasmaexperiment capable device through the synchronized operation of all essential components. Other purposes are the development of feasible startup scenarios under strict hardware limitations and low-voltage breakdown achievement using ECH system. The installed hardware and operation environment of KSTAR in the commissioning phase had more limitations than the designed environment. To confirm that the device can create a tokamak plasma, the operation goal for first plasma was sustainment of a 100 kA hydrogen plasma for 100 ms. The commissioning process of the first plasma began with the check of the magnetic configuration of the TF and PF magnets by measurements of the magnetic fields inside the vacuum vessel with the same initial magnetization (IM) conditions of the field magnets. Since the Incoloy908 material used in the Nb3 Sn magnets introduced a certain amount of uncertainty into the magnetic measurements, detailed calculations of the IM scenarios for the startup were performed based on both the experimental information from the movable Hall-probe measurements and a 2-D FE model for Incoloy908 magnetization. To insure lowvoltage startup, ECH preionization was successfully utilized [15]. ECH consists of the second harmonic resonance with microwave power from the 500 kW, 84 GHz gyrotron. Experiments seeking the best preionization conditions were performed under various prefill pressures and IM states of the PF coil currents, and provided synchronization of PF field null configuration with gas injection and ECH for the plasma startup. This kind of “dress rehearsal” for the first plasma enabled the ECH preionization to be reliably generated at the 2nd harmonic gyro-frequency resonance in the 1.5 T TF field, in the optimum gas pressure range. Fig. 4 shows the main sequence of events for a typical plasma discharge. Experiments on the initial field null configurations also have been performed for two different types of IM scenarios. First we tried to maximize the flux swing and the size of the field null region using a “conventional” PF coil current distribution with currents decreasing magnitude between the inside and outside of the machine. The scenario produced plasma up to 100 kA in shot #794. To maximize flux swing capacity of the machine, a unique IM scenario, called the “dipole mode”, was tested and developed. In this mode the outer two coil pairs (PF6 and PF7) have approximately equal and opposite currents and provide better flux and capability with the commissioning phase power and current constraints. This mode has a smaller field null than the conventional mode but can have a higher plasma current ramp rate and more degrees of freedom in the outer PFs to regulate the radial force balance. All the successful startups were essentially ECH-assisted with low loop voltages of 2–3.5 V. Using the dipole scenario, higher current ramping speeds around 1 MA/s were obtained and maximum plasma flat top times were obtained. Development of the plasma control system with the DIII-D team provided flexible control of the coil system and provided for operation of 100 kA plasma current and with pulse length up to 865 ms using real-time feedback of the plasma current (Ip ) and radial position (Rp ) [16]. Experiments were carried out to control plasma current down to the 50 kA level at the end of the shot in order to
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
349
Fig. 9. Results from the first plasma discharges: (a) an Ohmic plasma discharge image, (b) a field null configuration at the initial magnetization, and (c) plasma parameters for two plasma shots; a conventional mode feed-forward plasma (#794) and a dipole mode feedback controlled plasma (#1216).
minimize the effects of vertical disruption events (VDE) which were inevitable when the volt-seconds had been exhausted. Fig. 9 shows one of the results of a plasma pulse utilizing the plasma feedback mentioned above. At a TF current of 14 kA, an ECH resonance layer formed at R = 1.7 m. Under the IM conditions of the dipole scenario, the blip resistors were used for 150 ms, after which the Ip and Rp feedback algorithm was turned on for 600 ms. The ECH was set to start at −30 ms before the PF blip and lasted for 330 ms with a modulated ECH power of 350 kW, aimed 10 cm below the midplane. A second gas puff was added to suppress a kind of MHD instability which occasionally occurs in the burn-through phase. Additional experiments on the optimization of the heating devices have been performed. The ECH incident angle dependency and power threshold have been investigated and several shots on ion cyclotron range of frequency (ICRF) heating efficiency research were performed, utilizing various diagnostics such as the electron cyclotron emission (ECE) radiometer, interferometer, visible spectroscopy, and 10 kHz filter-scopes. Effects of wall conditioning were investigated by performing between-shot ICRF cleaning and glow discharge cleaning with H or He gas at night. It seems that 2 h of night GDC produced wall conditions allowing reliable discharges in the morning; however, there are many uncertainties in analyzing the effects of the wall conditioning method for reproducible plasma discharges. Warm-up of the device starting on 20 July 2008 has been performed after completing all experiments. To keep a uniform temperature distribution, forced helium circulation using the helium refrigerator was adopted until the coil temperature reached around 200 K. At that temperature the helium refrigerator was turned off and natural warm up was allowed.
6. Conclusions The successful commissioning and the first plasma achievement in KSTAR validate the design, engineering and construction aspects of the first major Nb3 Sn superconducting tokamak. One of the remarkable aspects of the KSTAR commissioning is that all commissioning phases were achieved without major problems and within the schedule. KSTAR commissioning showed reliable operational results for the Nb3 Sn superconducting magnet engineering. This is a meaningful benchmark for future SC fusion devices like ITER because they will utilize the same type of Nb3 Sn superconductor. Another major achievement is the successful low-voltage startup achievement by adopting the second harmonic ECH operation. Using this technology a loop voltage of less than 3 V produced reasonable plasmas. The KSTAR device will be operated to demonstrate the steadystate operation capability of high performance advanced tokamak modes. The major research plan for the initial operation phase is to establish the basic operation skills of superconducting tokamak operation through Ohmic, L-, and H-mode plasma experiments for relatively short-pulse lengths up to 20. For this research, in-vessel systems such as limiters, divertors, passive plates, in-vessel control coils, and heating systems will be upgraded in steps. This work is supported by the Ministry of the Education, Science, and Technology. Acknowledgement The authors would like to thank all the KSTAR participants from Korea and from the international collaboration partners who made
350
Y.-K. Oh et al. / Fusion Engineering and Design 84 (2009) 344–350
efforts for the successful commissioning during and the first plasma operation. References [1] G.S. Lee, et al., Design and construction of the KSTAR tokamak, Nuclear Fusion 41 (2001) 1515. [2] J.S. Bak, et al., Status of the KSTAR tokamak construction, Fusion Engineering and Design 81 (2006) 2315. [3] Y.K. Oh, et al., Completion of the KSTAR construction and its role as ITER pilot device, Fusion Engineering and Design 83 (2008) 804. [4] H.L. Yang, et al., Progress in the assembly of the KSTAR tokamak, in: Proceedings of the 20th IAEA FEC, Portugal, 2004, FT/P7–17. [5] Y.S. Kim, et al. Commissioning results of the KSTAR cryogenic system, ASC2008 conference, Chicago, submitted for publication to IEEE Transactions on Applied Superconductivity (2009). [6] Y.O. Kim, Local instruments of the KSTAR tokamak, Journal of the Korean Physics Society 49 (2006) S236. [7] K. Kim, et al., Status of the KSTAR superconducting magnet system development, Nuclear Fusion 45 (2005) 783.
[8] Y.M. Park, et al., Engineering design status of the KSTAR SC bus-line, IEEE Transactions on Applied Superconductivity 14 (2004) 1770. [9] K.R. Park, et al., Joint development for KSTAR CS coil, Journal of the Korean Physics Society 49 (2006) S238. [10] W. Chung, et al., Electrical insulation of KSTAR magnet lead, Journal of the Korean Physics Society vol.49 (2006) S228. [11] H. Yonekawa, et al., KSTAR magnetic field measurement at room temperature and remnant field evaluation, ASC2008 conference, Chicago, submitted for publication to IEEE Transactions on Applied Superconductivity (2009). [12] Y. Chu, et al., Quench detection based on voltage measurement for the KSTAR superconducting coils, ASC2008 conference, Chicago, submitted for publication to IEEE Transactions on Applied Superconductivity (2009). [13] Y.K. Oh, et al., Status of the KSTAR superconducting coil fabrication and test, in: Proceedings of the fourth IAEA TM on SSO, 2005 (POS-FT7). [14] S. Lee, et al., AC loss characteristics of the KSTAR CSMC estimated by pulse test, IEEE Transactions on Applied Superconductivity 16 (2006) 771. [15] Y.S. Bae, et al., Status of KSTAR electron cyclotron heating system, Fusion Science and Technology 52 (2007) 321. [16] S. Hahn, et al., Plasma control system for the day-one operation of KSTAR tokamak, 25th SOFT, submitted for publication, P3.23, this issue.