The ITER superconducting magnet system consists of 18 toroidal field (TF) coils, six central solenoid (CS) modules, six poloidal field (PF) coils and 18 correction coils. The National Institutes for Quantum and Radiological Science and Technology (QST), serving as the Japan Domestic Agency (JADA) in the ITER project, is responsible for the procurement of nine TF coils. Manufacturing of the first TF coil began in October 2013 and was completed in January 2020. While the basic technology for TF coil manufacturing has been developed through large-scale qualification trials, many additional issues were found when developing the first TF coil. Accordingly, these technical issues resolved and the TF coil was successfully completed. The technical issues and their solutions are reported in these special issues.

Double-pancake (DP) of ITER Toroidal Field (TF) coil consists of a cable-in-conduit conductor with Nb3Sn strands and a radial plate (RP). In the TF coil winding manufacturing process, a significant technical issue that was considered is that the difference of length between the heat-treated conductor and the RP grove must be controlled within ±0.023% to insert the conductor into the RP. This technical issue was solved by developing a highly accurate winding system and an RP assembly process to adjust the groove length. However, RP assembly was not able to begin before the winding is heat-treated, because the RP groove length is adjusted to the heat-treated conductor length in the assembly process developed. Therefore, it was difficult to satisfy the schedule required by ITER using this original manufacturing process. To solve this issue, an accuracy prediction method for the heat-treated conductor length was developed in which a highly accurate manufacturing process is used, and RP assembly can proceed in parallel with the winding manufacturing process. Applying this optimized process, both the required accuracy of the winding and the scheduling requirements of ITER are successfully satisfied for the series production of TF coil windings.

The ITER Toroidal Field (TF) coil radial plate (RP) is the main structure of a double pancake (DP), and a TF conductor is inserted into the RP groove and affixed using cover plates (CPs). Since the RP and CPs are used at around 4 K and should sustain huge electromagnetic force, full-austenite stainless steel (SS) is used. Furthermore, high-power laser-beam welding (LBW) is applied for welding during RP assembly and RP-CP welding with the aim of minimizing welding deformation and achieving very tight dimensional tolerances. In addition, cold-drawing is applied in CP fabrication for high production efficiency. Combining full-austenite SS and LBW normally generates welding defects and cold-drawing deteriorates fracture toughness. These technical issues have been overcome by introducing the following technical developments. A 75-mm-thick high-power LBW is obtained without defect by optimizing the chemical composition of the RP base metal. A yield stress (YS) of 900 MPa and fracture toughness (K_IC) of 180 MPam^0.5 can be achieved for both the base metal and weld joint at 4 K. In addition, the cold-drawing process for straight CP was optimized to achieve the required YS and KIC through process control and intermediate relevant mechanical testing. Furthermore, optimal LBW conditions for wide-gap weld joints, such as 0.5 mm and 0.7 mm, were developed for RP-CP and CP-CP welding, respectively. Applying these techniques to the fabrication process during RP, CP, and CP welding, optimized manufacturing procedures have been successfully developed to achieve the technical requirements. In addition, these fabrication procedures are well rationalized to satisfy schedule requirements in ITER. Accordingly, series production of RP and CP, and CP welding has commenced, and is proceeding. As of May, 2020, 61 RPs and 50 CP welding out of 63 have been successfully completed.

The National Institutes for Quantum and Radiological Science and Technology (QST) is responsible for manufacturingnine ITER TF coil winding packs (WPs). QST has been proceeding WP manufacturing, the procedure for which has been developed through qualification trials and the authors’ experiences. However, since the WP is a huge superconducting coil that has never been fabricated so far, technical challenges newly arose when manufacturing the first WP, such as tight WP dimension tolerance, current center line (CCL) position control and non-destructive examination (NDE) for joint performance. In addition, unexpected discharge happened at instrumentation wires going through the ground insulation layer after the WP cold test. The authors resolved these problems by clarifying the root causes and developing new techniques through additional qualification trials. As a result, WP fabrication is now in the series production phase. In fact, six WPs have been fabricated and manufacturing of the remaining three WPs is on-going.

We fabricated YBa2Cu3Oy (YBCO)coated conductors (CCs) 0.4 - 8.6 μm in thickness on Hastelloy metal substrates with oxide-textured layers, including IBAD-MgO, which were heated by a self-heating technique (SHT) using the pulsed laser deposition method. The substrate temperature (T_s) measured with a thermocouple or pyrometer was feedback-controlled (FBC) during deposition. It was possible to suppress a-axis grains even in a thick when using the SHT. The critical current (I_c) of the YBCO CCs with a thickness of 4.9 μm reached 1080 A/cm-w at 77 K and in a self-field. In addition, 3.0 - 9.0 vol.%BHO-doped YBCO CCs with a thickness of 3.0 μm were fabricated using pyrometer FBC. The I_c in a magnetic field at 77 K was optimized by adding 3.0 vol.%BHO.

Large Helical Device (LHD) has been in operation since 1998. Propagation of a normal zone has been observed 26 times in a pair of helical coils, named H1 and H2, of the LHD during the 21 years of operation. The only fourth propagation resulted in quick discharge due to the imbalance voltage higher than the preset value of 0.2 V, whereas the propagation in the other cases stopped within a few seconds. Each the coil is divided into three blocks, named H-I, H-M, and H-O, from the inside. Since the conductor of the helical coils consists of NbTi/Cu strands, a pure aluminum stabilizer clad with a Cu-2%Ni layer, and a copper sheath, the current center shifts from the superconducting wires to the pure aluminum stabilizer at the normal zone. Therefore, imbalance voltages between H1 and H2 are induced in all the blocks during propagation of a normal zone. The crosssectional position of the conductor in which the normal zone propagates can be estimated from the difference of the imbalance voltages among the blocks. In 2001, pickup coils were installed along the helical coils by the pitch of 30 degree of the poloidal angle in order to detect the position of a propagating normal zone. The pickup coils detect the change in magnetic field by a shift of current center at the normal zone. The position and velocity of propagating normal zones were detected successfully 15 times after the 10^th propagation. Most of the normal zones were induced at the bottom of the coils, and all of them propagated to one side, which is downstream of the transport current, with recovery on the opposite side. As the results of investigation of all the data, normal zones are considered to have been induced in the conductor in the first or last turn of the first or second layer of the H-I block. Therefore, normal zones should be induced at the position under the worst cooling condition with large disturbance due to slippage of the conductor against the helical coil case.