Search

In the framework of the EU-JA Broader Approach agreement, the 20 Toroidal Field (TF) coils (18 coils for the tokamak + 2 spares) produced for the JT-60SA tokamak were qualified in cold conditions and at full current (25.7 kA) in self-field conditions and heated up until quench occurs, along standardized testing procedures, at the Cold Test Facility (CTF) installed at CEA-IRFU (Saclay, France). The analysis of the tests results will be presented using simple dynamic models for the coolant exchange with cable in the steady-state regime. In addition to this approach, the coils operating limits in CTF conditions will be explored, integrating the global statistics of their strands' critical performances. The latter are built by processing TF strand performances data obtained during strand production phase. Those statistical data will be used to calculate the TF coils performances in the CTF configuration. Results will be discussed and compared to the TF coils experimental statistics, followed by a tentative quantified interpretation: The results will be transposed into variation of effective macroscopic parameters such as local heat load on pancakes or temperature margin potential increase. Once this macroscopic study is assessed, it stands as the first step toward more refined future analyses. Finally, using the above method, a first predictive application will be conducted on the JT-60SA tokamak operation configuration.

The toroidal field system of the JT-60SA tokamak is composed of 18 NbTi superconducting coils. Each TF coil is composed of six cable-in-conduit conductor lengths, wound in double-pancakes, and carrying a nominal current of 25.7 kA. These coils are being tested in the single-coil configuration at the so-called cold test facility (CEA/IRFU, Saclay, France). The test program includes for all coils a quench test at nominal current obtained by a progressive increase of the operating temperature at helium inlet up to quench (around 7.5 K). Thanks to the accuracy of the voltage measurements, it has been possible to follow the very early development of the quench at the scale of a few millimeters normal length. The paper presents the evolution of the resistive voltages measured on eight coils before the safety discharge is triggered. In addition, this early quench development over one conductor length was also simulated using the THEA code with relevant boundary conditions. Two different quenches, which occurred on two different coils, were analyzed, one starting on a central pancake (peak magnetic field) and the other starting on a side pancake (warmer conductor due to heat transfer from the casing). The simulations show good agreement with experiments, stressing particularly three phases in the quench development clearly identified in the measured resistive voltages of all coils.

JT-60 SA is a fusion experiment which is jointly constructed by Japan and Europe and which shall contribute to the early realization of fusion energy, by providing support to the operation of ITER, and by addressing key physics issues for ITER and DEMO. In order to achieve these goals, the existing JT-60 U experiment will be upgraded to JT-60 SA by using superconducting coils. The 18 Toroidal Field Coils (TFC) of the JT-60 SA device are provided by European industry and tested in a Cold Test Facility (CTF) at CEA Saclay. At the summer 2017, 14 TFCs will have been tested at the nominal current of 25.7 kA and at a temperature between 5 and 7.5 K. The main objective of these tests is to check the TF coils performances and hence mitigate the fabrication risks. These tests allowed checking a certain number of performances of the TFCs: DC/AC insulation, cooling down characterization, RRR of the conductor, pressure drop in the winding pack, and temperature margin against a quench. This paper will show the testing program progress and give an overview of the main experimental results obtained during these tests. The main performances of each coil will be summarized, analyzed, and discussed in the light of the expected TFCs performances.

In the framework of the European–Japanese project JT-60SA, the quench tests are performed for each of the 18 NbTi superconducting tokamak toroidal field coils (TFC) in the Cold Test Facility at the CEA Saclay, Gif-sur-Yvette, France. While launching these experimental quench tests, the conductor current sharing temperature ( $T_{cs}$ ) is reached by progressively increasing the inlet helium temperature. The quite complex quench dynamics are then observed due to several coupled physical phenomena influencing the quench propagation. In order to better understand the experimental analyses on coils quench behavior, a numerical model has been used, combining the computation code thermal hydraulic and electric analysis of superconducting cables for the longitudinal quench transient modeling along the cable-in-conduit conductor, and an additional interturn thermal coupling model for transverse heat flux modeling. In this paper, several parametric studies will be done, thanks to simplified numerical simulations in order to identify the predominant physical phenomena driving the JT-60SA TFC quench propagation. The analysis focuses on the linear power impact on the quench initiation, the quench propagation dynamics, and the reverse flow effect.

The toroidal field (TF) system of JT-60SA tokamak comprises 18 NbTi superconducting coils. In each TF coil, 6 cable-in-conduit conductor (CICC) lengths are wound into 6 double-pancakes and carry a nominal current of 25.7 kA at a temperature of 4.7 K. Each coil is tested in the cold test facility (CEA Saclay), up to quench. A SuperMagnet (CryoSoft) model has been developed, each of the 12 pancakes being modeled with THEA and cryogenic circuit being modeled with FLOWER. The experiments showed that helium inlet temperature increases until quench triggering at about 7.5 K on C11 and C13, with quench starting on a lateral and on a central pancake, respectively. Each test has been simulated, applying (or not) a realistic heat load from casing to winding pack that was estimated from experimental measurements. A parametric analysis has been performed, considering realistic or null heat flux deposition, variation of friction factor (in fabrication quality range), and CICC critical current density (in strand Jc performance range). This last parameter was found to have the largest impact on the localization of the first quenched pancake (central or lateral).

The Toroidal Field system of the JT-60SA tokamak is composed of 18 NbTi superconducting coils all individually tested in the Cold Test Facility (CTF, CEA/IRFU Saclay, France). The test program includes for each coil a progressive operating temperature increase at nominal current (25.7 kA) up to quench followed by a safety discharge. To account for the heat coming from the casing to the winding pack (WP), a transient test procedure was settled so as to trigger the quench at peak field in the coil; nevertheless the quench location was identified on a side double-pancake in the first two tested coils. We present the CEA analyses for the quench test of the second coil which showed similar behavior as the first one but received a more extensive instrumentation. The coil performance is analyzed in light of the NbTi strand superconducting properties coming from the strand characterization program and of the full-size conductor tests carried out by CEA in the SULTAN facility. These analyses involve cable thermal and electrical modelling developed at CEA and already used in ITER R&D, design and characterization programs.

JT-60SA is a fusion experiment which is jointly constructed by Japan and Europe and which shall contribute to the early realization of fusion energy, by providing support to the operation of ITER, and by addressing key physics issues for ITER and DEMO. In order to achieve these goals, the existing JT-60U experiment will be upgraded to JT-60SA by using superconducting coils. The 18 TF coils of the JT-60SA device are provided by European industry and tested in a Cold Test Facility (CTF) at CEA Saclay. The first coils were tested at the nominal current of 25.7 kA and at a temperature between 5 K and 7.5 K. The main objective of these tests is to check the TF coils performances and hence mitigate the fabrication risks. These first tests allowed checking a certain number of performances of the first TF coils: DC/AC insulation, cooling down characterization, RRR of the conductor, pressure drop in the winding pack and temperature margin against a quench. This paper will give an overview of the main experimental results obtained during these tests. These results will be analyzed and discussed in the light of the expected TF coils performances.

In order to check the performance of the JT-60SA Toroidal Field (TF) coils and hence mitigate their possible fabrication risks, a series of tests have been carried out in the Cold Test Facility (CTF) at CEA Saclay in nominal conditions at 5 K and 25.7 kA. One major test performed is the so called “temperature margin test” during which the inlet helium temperature of the Winding Pack (WP) is controlled to increase progressively to the calculated current sharing temperature of 7.3 K in the high field region to launch the quench. The measurements of voltage, pressure and temperature signals allow us to study the thermo-hydraulic behavior of the JT-60SA coils during a quench. This paper will present the 6 quench tests on 5 TF coils accomplished between February and September 2016, of which the first manufactured coil TFC10 has carried out two quench tests and two pure Fast Discharges without any quenches. It will analyze in particular the temperature and pressure rise in helium, the corresponding physical phenomena after each voltage variation, the four quench dynamic phases, the quench propagation velocity and the dissipated power in the coils.

Previous thermohydraulic studies performed for the ITER TF, CS and PF magnet systems have brought some important information on the detection and consequences of a quench as a function of the initial conditions (deposited energy, heated length). Even if the temperature margin of the Correction Coils is high, their behavior during a quench should also be studied since a quench is likely to be triggered by potential anomalies in joints, ground fault on the instrumentation wires, etc. A model has been developed with the SuperMagnet Code (Bagnasco et al., 2010) for a Side Correction Coil (SCC2) with four pancakes cooled in parallel, each of them represented by a Thea module (with the proper Cable In Conduit Conductor characteristics). All the other coils of the PF cooling loop are hydraulically connected in parallel (top/bottom correction coils and six Poloidal Field Coils) are modeled by Flower modules with equivalent hydraulics properties. The model and the analysis results are presented for five quench initiation cases with/without fast discharge: two quenches initiated by a heat input to the innermost turn of one pancake (case 1 and case 2) and two other quenches initiated at the innermost turns of four pancakes (case 3 and case 4). In the 5th case, the quench is initiated at the middle turn of one pancake. The impact on the cooling circuit, e.g. the exceedance of the opening pressure of the quench relief valves, is detailed in case of an undetected quench (i.e. no discharge of the magnet). Particular attention is also paid to a possible secondary quench detection system based on measured thermohydraulic signals (pressure, temperature and/or helium mass flow rate). The maximum cable temperature achieved in case of a fast current discharge (primary detection by voltage) is compared to the design hot spot criterion of 150 K, which includes the contribution of helium and jacket.

In the framework the design activities conducted in EU for the future fusion demonstration reactor (DEMO), extensive analyses were led in the EUROfusion context, aiming at ultimately designing the DEMO magnets system. The DEMO tokamak configuration was updated in view of improving the operation merits versus the previous design defined in 2015. In the objective of properly addressing tokamaks general magnetic systems, CEA has adapted macroscopic design tools and associated methods to size magnets for pre-dimensionning or optimization purposes. Once pre-dimensioned using inputs from system-level analyses the magnet concepts undergo an evaluation process by detailed analyses (thermohydraulic and mechanical) highly accurate but strongly time consuming. To avoid repeating this sequence, we aimed in this paper at improving pre-dimensioning process reliability for the final benefit of the whole design workflow. The material exposed in this paper is the path to this improvement achieved through a sery of pre-dimensioning parametric cases applied on the DEMO TF system baseline newly issued in 2018 by EUROfusion. The optimization steps are described, showing which design TF system variables are chosen for optimization and what is their contribution to ultimately rend the design as efficient as possible with respect to given merits. In this aim, former detailed finite elements analyses, carried out on the previous DEMO configuration, are taken as reference to support the quantitative choices of the pre-dimensioning method improvement (criteria or input modifications). Furthermore, TF design is also explored on a new aspect: the possibility of including radial plates in the winding pack.

Strain dependence of Nb3Sn superconducting properties is known to be responsible for the degradation of transport current capability of large steel jacketed cable-in-conduit conductors (CICCs). The mechanical deformations of the strands in the cables due to both cooldown after heat treatment and Lorentz force during operation are the main sources of strand-in-cable critical current degradation. The complete modeling of a CICC relies first on the modeling of the single strand with its superconducting filaments and then on the modeling of the strands in the cable. This paper reports on the description of the models and the analyses performed using coupled mechanical and electrical modeling. The key of this modeling lies in its ability to characterize the superconducting properties of the twisted Nb3Sn filaments inside a locally bent strand from a global mechanical deformation map, taking into account possible current redistribution among filaments all along the strand length. In the same way, strands may experience local critical current degradation in a CICC, and current transfer between strands at their contact points must be modeled to predict cable performance. This work can be seen as the first stones of an integrated coupled mechanical–electrical modeling of large CICCs made with strain-sensitive strands.

Coupling losses induced in cable-in-conduit conductors (CICC) when subject to a time-varying magnetic field are a major issue commonly encountered in large fusion tokamaks (e.g., JT-60SA, ITER, DEMO). The knowledge of these losses is crucial to determine the stability of CICC but is yet difficult to achieve analytically (thus in a short computation time) given the specific and complex architecture of these conductors although numerical solutions such as THELMA and JACKPOT already exist. In an attempt to ease the resolution of this problem, we have previously presented a theoretical generic study of a group of elements twisted together (representing a cabling stage of a CICC) and derived the analytical expression of its coupling losses. We have now extended this study to a two cabling stage conductor by establishing an analytical model to calculate its coupling losses as function of its effective features. In a second part, we compare our results to these of THELMA and JACKPOT on geometries representing ITER CS and JT-60SA TF conductors. Finally, we have set up a specific algorithm to reconstruct strand trajectories from X-ray images and have extracted the effective geometrical parameters of a JT-60SA TF conductor. Our next objective is then to extract its effective electrical parameters from interstrand resistivity measurements to be able to compare the coupling losses predicted by our analytical model with those measured within the SULTAN facility.

In the framework of the European fusion program for energy, EUROfusion funds the studies for the future fusion power demonstrator reactor DEMOnstration fusion reactor (DEMO). Commissariat a l'energie atomique et aux energies alternatives (CEA) is involved in the conceptual design of the superconducting conductors for the toroidal field (TF) and central solenoid (CS) magnets. The CEA design proposal corresponds to wind and react Nb3Sn pancake-wound coils using cable-in-conduit conductors cooled at about 4.5 K by forced flow of supercritical helium. This paper presents the latest thermal-hydraulic analyses performed on both TF and CS conductors. Two TF conductor designs have been analyzed, with nominal currents of 111 kA and 88 kA, respectively. The analyses were performed in normal (burn) and off-normal (quench) conditions. Burn simulations focused on the central and lateral pancakes with heat load corresponding to neutron heating. The central pancake is the most critical one regarding the magnetic field, while the lateral one receives a more important heat load from the casing. The influence of case cooling on temperature margin (ΔTma) was analyzed by means of a dedicated 2D Cast3M model. The ΔTma sensitivity to driving parameters such as conductor friction factor correlations was analyzed. The impact of interturn thermal coupling on temperature margin was also assessed. Quench studies were performed on the TF conductor, considering a quench initiated either on the innermost turn or at the middle of hydraulic length. Regarding the CS conductor, burn scenario was led on the latest design featuring a nominal current of 53.7 kA, focusing on the impact of ac losses and of dwell duration.

Cable-in-conduit conductors made of Nb 3 Sn strands, as envisaged in high-field magnets for fusion applications (e.g., the International Thermonuclear Experimental Reactor (ITER) project), have been shown to be prone to degradation of their current transport capability when subject to high electromagnetic forces (coming from the combination of a high current and a high magnetic field), particularly under cyclic loading. Although some optimization of the cable layout can be envisaged to mitigate this effect, the knowledge of the electrical properties of Nb 3 Sn strands under periodic bending (simulating in-cable operating conditions) is generally considered a needed characteristic for magnet designers and users. The French Alternative Energies and Alternative Energies and Atomic Energy Commission (CEA)/Institute for Magnetic Fusion Research (IRFM) has developed a new Versailles Project on Advanced Materials and Standards (VAMAS)-like mandrel in order to test under relevant field and temperature conditions high-field superconducting strands subject to periodic bending provided by the strand deformation under its own Lorentz force. The advantage of this setup is its similarity with the setup used in critical current measurements on a VAMAS mandrel, which makes its implementation in existing facilities quite easy and rather cheap, as well as possible direct comparisons with unbent samples on a classical VAMAS mandrel. This paper presents the design layouts and carefully depicts the assembly of the different tested samples. The first test results obtained on Nb 3 Sn strands are given and discussed.

International audience; Lorsqu'un composite supraconducteur est dans un environnement à champ magnétique variable, des courants dits de couplage sont induits. Ceux-ci, en traversant les parties résistives du composite, génèrent une puissance thermique locale. Le but de la présente contribution est de présenter un modèle permettant de calculer analytiquement et de manière précise le terme source de l'équation de la chaleur associé à cet échauffement à tout instant et en tout point de l'espace. Le modèle est ensuite appliqué à un brin réel et les résultats des simulations sont présentés et commentés.

While the ITER Tokamak is proceeding with major construction and assembly phases, design and R&D activities regarding the next step experimental reactor, the so-called DEMO machine, are underway in Europe. In particular, the magnet systems enter a refined design phase, under the general guidelines of Eurofusion, where winding packs (WP) are defined, conductor designs are proposed, and subscale analyses are performed in order to crosscheck the systems code (PROCESS) predimensioning. This paper summarizes the design iterations led at CEA on the toroidal field (TF) and central solenoid (CS) magnet systems. It presents, in particular, the optimization of the WP architecture with regard to the filling of the available space (TF) and flux maximization (CS). The paper gives the relevant electrical, thermal, and mechanical parameters for each system design, and tries to give insight on the tradeoffs and possible changes that could be implemented, given some margin on fixed systems code inputs. Finally, it also includes the refined magnetic field maps calculations performed to confirm the design operating field and explore other scenario time points that include poloidal field and plasma contributions.

Following the test of the first Central Solenoid (CS) conductor short samples for the International Thermonuclear Experimental Reactor (ITER) in the SULTAN facility, Iter Organization (IO) decided to manufacture and test two alternate samples using four different cable designs. These samples, while using the same Nb3Sn strand, were meant to assess the influence of various cable design parameters on the conductor performance and behavior under mechanical cycling. In particular, the second of these samples, CSIO2, aimed at comparing designs with modified cabling twist pitches sequences. This sample has been tested, and the two legs exhibited very different behaviors. To help understand what could lead to such a difference, these two cables were mechanically modeled using the MULTIFIL code, and the resulting strain map was used as an input into the CEA electrical code CARMEN. This article presents the main data extracted from the mechanical simulation and its use into the electrical modeling of individual strands inside the CICC.

The quench of one of the ITER magnet systems is an irreversible transition of the conductor from superconducting to normal resistive state. The normal zone propagates along the cable-in-conduit conductor, dissipating a large power. The detection has to be fast enough (1 to 2 s) to initiate the dumping of the magnetic energy and avoid irreversible damage of the systems.The experience of CEA is based on the operation of the superconducting tokamak Tore Supra for more than 20 years. In support of ITER, CEA was also very involved in quench detection investigations during these past 3 years.The primary quench detection in ITER is based on voltage detection, the most rapid detection. The very magnetically disturbed environment during a plasma scenario makes the voltage detection particularly difficult, inducing large inductive components across the pulsed coils (10 kV) or coil subcomponents. Voltage compensations therefore have to be designed to discriminate the resistive voltage associated with the que...

In the framework of the JT-60SA project, France and Italy will provide to JAEA 18 Toroidal Field (TF) coils including NbTi cable-in-conduit conductors. During the tokamak operation, these coils could experience a quench, an incidental event corresponding to the irreversible transition from superconducting state to normal resistive state. Starting from a localized disturbance, the normal zone propagates along the conductor and dissipates a large energy due to Joule heating, which can cause irreversible damages. The detection has to be fast enough (a few seconds) to trigger the current discharge, so as to dump the stored magnetic energy into an external resistor. The JT-60SA primary quench detection system will be based on voltage measurements, which are the most rapid technology. The features of the detection system must be adjusted so as to detect the most probable quenches, while avoiding inopportune fast safety discharges. This requires a reliable simulation of the early quench propagation, performed in this study with the Gandalf code. The conductor temperature reached during the current discharge must be kept under a maximal value, according to the hot spot criterion. In the present study, a hot spot criterion temperature of 150 K was taken into account and the role of each conductor component (strands, helium and conduit) was analyzed. The detection parameters were then investigated for different hypotheses regarding the quench initiation.

In case of a magnet quench, the primary quench detection system is classically based on voltage detection. In addition, a secondary quench detection is required and it could rely on signals of thermohydraulic nature. A model based on the coupled GANDALF [1] and FLOWER [2] codes has been developed in previous papers for the Central Solenoid [3] , [4] , and for the Toroidal Field Coils [5] , [6] of ITER. This study focuses on Poloidal Field Coil PF5 as it has the highest magnetic field value along the conductor. Results of some parametric studies are presented and the influence of the initiation parameters of the quench (energy, location and length of heat deposition, time duration for detection). The possibility of a secondary thermohydraulic detection and the main impact on cryogenic system are analyzed and discussed.

The ITER central solenoid (CS) is a highly stressed magnet that must provide 30 000 plasma cycles under the ITER prescribed maximum operating conditions. To verify the performance of the ITER CS conductor in conditions close to those for the ITER CS, the CS insert was built under a USA-Japan collaboration. The insert was tested in the aperture of the CSMC facility in Naka, Japan, during the first half of 2015. A magnetic field of up to 13 T and a transport current of up to 60 kA provided a wide range of parameters to characterize the conductor. The CS insert has been tested under direct and reverse charges, which allowed a wide range of strain variation and provided valuable data for characterization of the CS conductor performance at different strain levels. The CS insert test program had several important goals as follows. 1) Measure the temperature margin of the CS conductor at the relevant ITER CS operational conditions. 2) Study the effects of electromagnetic forces and strain in the cable on the CS conductor performance. 3) Study the effects of the warmup and cooldown cycles on the CS conductor performance. 4) Compare the conductor performance in the CS insert with the performance of the CS conductor in a straight hairpin configuration (hoop strain free) tested in the SULTAN facility. 5) Measure the maximum temperature rise of the cable as a result of quench. The main results of the CS insert testing are presented and discussed.

The DEMO reactor is expected to be the first application of fusion for electricity generation in the near future. To this aim, conceptual design activities are progressing in Europe (EU) under the lead of the EUROfusion Consortium in order to drive on the development of the major tokamak systems. In 2014, the activities carried out by the magnet system project team were focused on the toroidal field (TF) magnet system design and demonstrated major achievements in terms of concept proposals and of consolidated evaluations against design criteria. Several magnet system RD activities were conducted in parallel, together with broad investigations on high temperature superconductor (HTS) technologies. In this paper, we present the outcomes of the work conducted in two areas in the 2014 magnet work program: 1) the EU inductive reactor (called DEMO1) 2014 configuration (power plant operating under inductive regime) was the basis of conceptual design activities, including further optimizations; and 2) the HTS RD activities building upon the consolidated knowledge acquired over the past years. © 2016 IEEE.

While the ITER machine is still under construction, design and RD and, second, the accurate computation of the magnetic field on the superconductor, which determines its transport current capability. The exercise was carried out within the CEA studies for the Eurofusion DEMO project in its 2014 and 2015 versions. The design of the superconducting cable is based on ITER-like conductors to take advantage of the already existing experience gained on this type of conductors through the ITER R&D and qualification programs, whereas the design criteria that are recalled have been agreed within the European fusion community. This paper depicts the method and the results of the dimensioning of a proposal TF winding pack.

The stability analysis [1] of the Poloidal Field Coils of the ITER Project, specially the PF5, PF3 and PF2 coils for scenario #2, normal mode has been performed using the Gandalf [2] code. The conductor stability is assessed in the most severe operating conditions, at the minimal temperature margin time of the 4th pulse. It consists in determining the maximal deposited energy (perturbation in mJ/cm 3 of superconducting strands) which can be absorbed without runaway. This study is performed for three initial different designs: Cu–nonCu ratio of 2.3, 1.6 and 4.4 for PF5, and Cu–nonCu ratio of 2.3, 1.6 and 6.9 for PF3 and PF2. The results are presented as well as a discussion about the phenomena, important parameters and uncertainties. For each type of perturbation, the calculated stability limits are quite similar from one design to another. The low Cu–nonCu ratios do not show significant stability degradation compared to the original ITER design with high ratios.

TFAS1 (toroidal field advanced strand) sample is composed of two cable-in-conduit conductors intended to be used in ITER toroidal field coils. This sample was tested in the SULTAN facility in order to benchmark the performance in years 2005-2006. An early voltage evolution was observed in both conductors during current ramp up, which makes it difficult to assess correctly the conductor performance. In order to improve the voltage behavior, the two terminations and two halves of the joint were filled by solder at CRPP with the aim to achieve a more uniform current distribution in the conductors. Thereafter, the TFAS1 sample was retested in the SULTAN facility. The retest revealed an unexpected increase of resistance after solder filling at the conductor extremities. As a consequence the voltage slope did not disappear, and further reduction of conductor performance was observed. The results of the retest of the TFAS1 sample with solder filled extremities are presented and discussed in this paper.

The JOSEFA (joint subsize experiment facility) experimental test facility, installed at CEA/Cadarache is devoted to perform tests at cryogenic temperature on subsize superconducting conductor and joint samples under parallel or transverse magnetic field. This facility was built in 1993 to investigate the performances of joints of cable-in-conduit conductors at subsize level and further upgraded in the framework of European tasks. The samples of hairpin type using subsize ITER conductors are cooled by a circulation of supercritical helium in a temperature range from 5 to 15 K and tested at a maximum current up to 10 kA. Two different helium bath cooled magnets allow to apply DC or AC transverse magnetic field up to 3.5 T or longitudinal magnetic field up to 7.5 T. A sliding system with a 240 mm stroke on the sample cryostat allows to test separately in the same sample either the conductor or the joint performances. The paper reports on how, through the conductor and joint development tasks, the facility performances were successfully increased and tested. The ITER TFMC joints using Nb/sub 3/Sn conductors were first developed on this facility. The last developments, performed on the ITER PF NbTi conductors and joints proved this facility to be a versatile and useful tool for superconducting magnet developments and showed the interest of possible upgrading to finalize conductor design.

As the main objective of the toroidal field model coil (TFMC) is to demonstrate the industrial feasibility of the ITER TF coils, the TFMC design includes the particular features of these coils, such as the insertion of the conductors inside stainless steel radial plates (see Fig. 1 ). During ITER transients such as the safety discharge of the magnetic system, eddy currents are induced in the plates, generating a power, which is transferred into the conductor helium channels by a long diffusion process through the conductor steel and insulation, and eventually quenches the system. In this case, the delay introduced by this diffusion is a key parameter in the quench evolution. The model of this diffusion is presented. The diffusion equation is added to the thermohydraulics code gandalf, which enables to take into account the helium circulation along the channel. The ITER TFMC experiment offers an unique possibility to check this model. Specific safety discharges were intentionally initiated from 25 kA to study this phenomenon. The results of these experiments are compared with the predictions. The model proves to be quite effective and can be used for ITER. The heat conductivity and the specific heat used in the model are discussed.

The construction and testing of the Toroidal Field Model Coil (TFMC) is part of one of the ITER large R&D projects. The main goal was to demonstrate the feasibility and the mechanical integrity of the design. One of the highlights of the first test phase was to measure the current sharing temperature, T/sub CS/, of the conductor by heating the helium entering from the inlet. Because neither temperature sensors nor voltage taps are positioned inside the coil, only the helium inlet temperature and the voltage along the whole conductor length can be used for the evaluation of T/sub CS/. In addition, an inner pancake joint is located at the inlet in a rather high magnetic field and the peak field region is only about 1.5 m apart from the joint. The determination of the T/sub CS/ relies on the exact knowledge of the thermohydraulics of both the joint and the conductor region. The paper describes and compares the different numerical models used for the evaluation of the T/sub CS/. Nine T/sub CS/ tests at different coil currents were performed, all ending in a quench. The measured T/sub CS/ is in good agreement with the expectations.

Previous quench studies performed for the International Thermonuclear Experimental Reactor (ITER) Toroidal Field (TF) Coils have led to identify two extreme families of quench: first "severe" quenches over long initial lengths in high magnetic field, and second smooth quenches over short initial lengths in low field region. Detailed analyses and results on smooth quench propagation and detectability on one TF Cable In Conduit Conductor (CICC) with a lower propagation velocity are presented here. The influence of the initial quench energy is shown and results of computations with either a Fast Discharge (FD) of the magnet or without (failure of the voltage quench detection system) are reported. The influence of the central spiral of the conductor on the propagation velocity is also detailed. In the cases of a regularly triggered FD, the hot spot temperature criterion of 150 K (with helium and jacket) is fulfilled for an initial quench length of 1 m, whereas this criterion is exceed (Tmax ≈ 200 K) for an extremely short length of 5 cm. These analyses were carried out using both the Supermagnet{trade mark, serif} and Venecia codes and the comparisons of the results are also discussed.

A new kind of superconductor, using the cable-in-conduit concept, is emerging, mainly in the context of fusion activity. At present no large Nb3Sn magnet in the world is operating using this concept. The difficulty of this technology, which has now been studied for 20?years, is that it requires major advances in several interconnected new fields, such as handling a large number (1000) of superconducting strands, high current conductors (50?kA), forced flow cryogenics, Nb3Sn technology, low loss conductors in pulsed operation, high current connections and high voltage insulation (10?kV), as well as demonstration of its economical and industrial feasibility. CEA has been very much involved, during the past ten years, in this development, which took place in the framework of the NET and ITER technological programmes. One major milestone was reached in 1998-1999 with the successful tests by Euratom-CEA of three full size conductor and connection samples in the SULTAN facility in Switzerland.

In the scope of the Engineering Design Activity (EDA) for ITER (International Thermonuclear Experimental Reactor), a R&D task has been launched in 1993 for development, manufacturing and testing of subsize and full size joints, to be used in the layer-to-layer and pancake-to-pancake connections of the ITER coils. A set of design criteria and manufacturing choices is initially agreed as a common basis for the work carried out in parallel by the four ITER parties (Japan, European Union, United States, Russian Federation). The staged work includes variations of layout parameters in subsize samples, technological investigations for manufacturing methods and performance analysis. As a last stage of the task, full size joint samples, manufactured from sections of the ITER conductors, are tested in the PTF and SULTAN test facilities under ITER relevant operating conditions.

International audience; Nb3Sn is now commonly used in the design of high-field large-scale magnets. However, it is a brittle material, the superconducting properties of which degrade under mechanical strain. Both ITER TF and CS magnets make use of Nb3Sn strands in cable-in-conduit conductors. Experiments have been carried out in the TARSIS facility at University of Twente aiming at measuring the strand critical current as a function of periodically applied strain/stress. Until recently, these experiments have given good indications of the strand behavior, but they had not been fully understood because of the lack of an accurate description of the local strain along the tested strand. Furthermore, they cannot be extrapolated directly to a real cable-in-conduit conductor because they do not simulate the differential thermal contraction, which puts the strand under longitudinal compression. Using the mechanical code MULTIFIL developed at Ecole Centrale de Paris, associated with the electrical code CARMEN developed at CEA/IRFM, this paper aims at understanding the mechanisms of the critical current reduction during a TARSIS experiment by coupling the local strain map of the strand to the complex current paths between Nb3Sn filaments. Comparison with experimental results and with analytic limiting cases are presented and discussed.

Two short length samples have been prepared and tested in SULTAN to benchmark the performance of high current density, advanced Nb3Sn strands in the large cable-in-conduit conductors (CICC) for ITER. The cable pattern and jacket layout were identical to the toroidal field model coil conductor (TFMC), tested in 1999. The four conductor sections used strands from OST, EAS, OKSC and OCSI respectively. The Cu:non-Cu ratio was 1 for three of the new strands, compared to 1.5 in the TFMC strand. The conductors with OST and OKSC strands had one Cu wire for two Nb3Sn strands, as in TFMC. In the EAS and OCSI conductors, all the 1080 strands in the cable were Nb3Sn. A dc test under relevant load conditions and a thermal-hydraulic campaign was carried out in SULTAN. The CICC performance was strongly degraded compared to the strand for all the four conductors. The current sharing temperature at the ITER TF operating conditions (jop = 286 A/mm2, B = 11.15 T) was lower than requested by ITER.

The Poloidal Field Conductor Insert (PFCI) of the International Thermonuclear Experimental Reactor (ITER) has been designed in the EU and is being manufactured at Tesla Engineering, UK, in the frame of a Task Agreement with the ITER International Team. Completion of the PFCI is expected at the beginning of 2005. Then, the coil shall be shipped to JAERI Naka, Japan, and inserted into the bore of the ITER Central Solenoid Model Coil, where it should be tested in 2005 to 2006. The PFCI consists of a NbTi dual-channel conductor, almost identical to the ITER PF1 and PF6 design, 45 m long, with a 50 mm thick square stainless steel jacket, wound in a single-layer solenoid. It should carry up to 50 kA in a field of 6 T, and it will be cooled by supercritical He at 4.5 K and 0.6 MPa. An intermediate joint, representative of the ITER PF joints and located at relatively high field, will be an important new item in the test configuration with respect to the previous ITER Insert Coils. The PFCI will be fully instrumented with inductive and resistive heaters, as well as with voltage taps, Hall probes, pick-up coils, temperature sensors, pressure gauges, strain and displacement sensors. The test program will be aimed at DC and pulsed performance assessment of conductor and intermediate joint, AC loss measurement, stability and quench propagation, thermal-hydraulic characterization. Here we give an overview of the preparatory work toward the test, including a review of the coil manufacturing and of the available instrumentation, a discussion of the most likely test program items, and a presentation of the supporting modeling and characterization work performed so far.

A short sample of the NbTi cable-in-conduit conductor (CICC) manufactured for the ITER PF insert coil has been tested in the SULTAN facility at CRPP. The short sample consists of two paired conductor sections, identical except for the sub-cable and outer wraps, which have been removed from one of the sections before jacketing. The test program for conductor and joint includes DC performance, cyclic load and AC loss, with a large number of voltage taps and Hall sensors for current distribution. At high operating current, the DC behavior is well below expectations, with temperature margin lower than specified in the ITER design criteria. The conductor without wraps has higher tolerance to current unbalance. The joint resistance is by far higher than targeted.

The primary quench detection for ITER magnets is classically based on resistive voltage measurements. In addition, a safety related secondary quench detection relying on signals of thermohydraulic nature is required for the TF coils. Although not required by safety, a similar secondary quench detection system has been investigated for the other ITER coils [central solenoid (CS), poloidal field (PF), and correction coil (CC)]. The quenches in the ITER bottom correction coil (BCC3) of either the top pancake (P8) or the whole coil (eight pancakes), without a current fast discharge, are studied using the coupled Gandalf and Flower codes. The minimum quench energy is deposited in the area where the temperature margin is minimal, close to the outlet of the conductor. Because PF5, BCC3, and BCC6 are hydraulically linked in parallel, the influence of a PF5 coil quench without a fast discharge is studied. For all these cases, the cryogenic consequences are determined providing qualitative and quantitative information. Key signals are presented like pressure, temperature, and mass flows in the conductor itself, as well as at the ends of feeders, in the Cold Termination Box, where the sensors are located. The feasibility of the secondary thermohydraulic quench detection is investigated, leading to an associated proposal for the instrumentation (difference of mass flows: inlet-outlet). The main conclusions concerning these signals are discussed.

During a fast discharge of the TF system, eddy currents and high associated heat production in the winding plates can induce a quench by heat transfer. A quasi 2D model based on the coupled GANDALF and FLOWER codes [1] allows the calculation of the radial heat transfer from the plates to the CICC through the conductor insulation as well as hydraulic transients with an external cryogenic circuit. This model is used in three particular cases: 1) Plasma Disruption (PD) with associated heat deposition followed by a Fast Discharge (FD) (with self Joule heating of conductor) with 2 s delay; 2) Quench initiation with Minimum Quench Energy (MQE), followed by a 2 s heat deposition associated with a plasma disruption and then FD; 3) Case where the quench has not been detected earlier by the primary detector (“undetected quench” [2]). The signals regarding temperature, pressure and mass low rate reach significantly high values especially in the conductor, at the inlet and outlet of the helium channel in the feeder at the Cold Termination Box (CTB), where the sensors are located, and at the exhaust cryolines. A feasibility of a secondary detection, provided by thermohydraulics signals is studied.

Works on bending strain in Nb3Sn wires were initiated in support of the 'react-and-wind' technique used to manufacture superconducting coils. More recently, the bending strains of Nb3Sn strands in cable-in-conduit conductors (CICC) under high Lorentz forces have been thought to be partly responsible for the degradation of the conductor performance in terms of critical current and n index, particularly for the international thermonuclear experimental reactor (ITER) conductors. This has led to a new wave of experiments and modelling on this subject. The computation of the current transport capability in an Nb3Sn wire under uniform bending used to be carried out through the so-called Ekin's models, and more recently through numerical simulations with electric networks. The flaws of Ekin's models are that they consider only two extreme cases or limits, namely the so-called long twist pitch (LTP) or short twist pitch (STP) cases, and that these models only allow computation of a value for the critical current without reference to the n index of the superconducting filaments (i.e. this index is implicitly assumed to be infinite). Although the numerical models allow a fine description of the wire under operation and can take into account the filament's n index, they need a refined meshing to be accurate enough and their results may be sensitive to boundary conditions (i.e. current injection in the wire), also general intrinsic parameters cannot be easily identified. In this paper, we propose clearly to go further than Ekin's models by developing, from a homogeneous model and Maxwell's equations, an analytical model to establish the general equation governing the evolution of the electric field inside an Nb3Sn strand under uniform bending (with possible longitudinal strain). Within the usual strand fabrication limits, this equation allows the definition of one single parameter to discriminate the STP and LTP cases. It is also shown that whereas Ekin's LTP model corresponds well to a limiting solution of the problem when the transverse resistivity tends toward zero (or the twist pitch tends towards infinity), Ekin's STP model must be modified (improved) when the filament's n index is finite. Since the general equation cannot be solved analytically, we start from the LTP model and develop a first order correction to be applied when the transverse resistivity, the twist pitch and the filament's n index are finite. Using a simple but realistic law for depicting the strain dependence of the critical current density in the Nb3Sn filaments, we can fully compute the corrected expression and give the result under a general analytical formula for a strand submitted to both bending and compressive/tensile strains. The results are then compared in two different cases with those obtained with the numerical code CARMEN (based on an electrical network) developed at CEA. Last, a semi-empirical formula has been developed to evolve continuously from the LTP limit to the improved STP limit when the transverse resistivity evolves from zero to infinity. The results given by this formula are again compared with the numerical simulations in two different cases. Last, comparisons with experimental results are discussed.

The PFCI is a single-layer solenoid wound from a 45 m long ITER-type NbTi dual-channel cable-in-conduit conductor, designed to be representative of the one currently proposed for the ITER PF1&6 coils. The PFCI, installed in the bore of the ITER central solenoid model coil (CSMC) at JAEA Naka, Japan, and well instrumented from both the thermal hydraulic and the electromagnetic points of view, has been successfully tested in June-August 2008. The test concentrated on DC performance (current sharing temperature and critical current measurements) and AC loss measurements. The results of the analysis of those measurements are reported in the paper, with particular attention to the comparison with the PFCI short sample, which was previously tested in the SULTAN facility. The evolution of the DC performance of the CSMC is also discussed.

In the framework of the DEMOnstration fusion power plant (DEMO) design coordinated by the EUROfusion consortium, a pre-conceptual design of the superconducting magnet system has been developed. For the toroidal field coils (TFCs), three winding pack (WP) options have been proposed; exploring different winding approaches (pancakes vs. layers), and manufacturing techniques (react & wind vs. wind & react Nb3Sn). Thermal-hydraulic and mechanical analyses on the three WPs have produced encouraging results, with some critical issues to be solved in future studies and optimizations. The experimental tests on TF prototype short sample conductors have demonstrated a limited performance degradation with electro-magnetic cycles and significantly lower effective strains than most of the large-size Nb3Sn conductors reported in literature. The toroidal field quench protection circuit has been studied, starting from different topologies and focusing on the most promising one. Two designs are also presented for the central solenoid magnet, with preliminary evaluations on the AC losses during the plasma breakdown. Finally, the design of a TF winding pack based on HTS conductors and the experimental tests on “fusion-relevant” HTS cables are illustrated.

In the framework of the EUROfusion DEMO project, design studies for the tokamak magnet systems are conducted across several European institutions. Superconducting magnets are required to generate the high magnetic fields needed for the plasma confinement and control. The reference conductor design is based on Cable-In-Conduit Conductors (CICC) cooled at cryogenic temperatures by forced circulation of supercritical helium. The proposed Toroidal Field (TF) coils Winding Pack (WP) design should satisfy the design/safety criteria during operation (e.g., minimal temperature margin) and in off-normal conditions (e.g., hotspot temperature). Quenches are studied to ensure that the proposed conductor design and associated quench protection system guarantee the integrity of the magnet; such events are a safety and investment protection issue, and as such deserve high levels of scrutiny. Quench propagation in a coil is a 3-dimensional (3D) problem. For this reason, a transient pseudo-3D modeling tool was developed for coupled thermal and thermo-hydraulic calculation in a tokamak superconducting coil. The coupling tool is based on a 1D model of the cable using the THEA code, considering current distribution, helium flow, thermal conduction in the strands and propagation of the quench along the conductor; the 2D transverse thermal diffusion across turns is modeled using the Cast3M code, considering the conductor jacket and insulation, on a selected set of cross-sections along the coil. The aim of the analysis is to assess the quench behavior of the CEA proposal for DEMO TF coil. The hotspot temperatures as well as the propagation of the normal zone along the conductor length are evaluated in a realistic quench scenario, taking into account the impact of transverse heat diffusion.

In the framework of the EUROfusion DEMO project, studies are conducted in several European institutions for designing the tokamak magnet systems. In order to generate the high magnetic fields required for the plasma confinement and control, the reactor should be equipped with superconducting magnets, the reference design being based on Cable-In-Conduit Conductors (CICC) cooled at cryogenic temperatures by forced circulation of supercritical helium. In order to propose a toroidal field (TF) winding pack (WP) design compatible with DEMO requirements, CEA has developed several tools addressing the different areas related to magnet dimensioning. An accurate calculation of magnetic field along the conductors is provided by the TRAPS code, and conductor design is performed by using an integrated macroscopic home design code based on simplified models accounting for superconducting properties, mechanics and thermal. This multi-physic tool gives a realistic but not assessed design. Indeed it is based on an assumed operating temperature that must be validated with an elaborate code, since it is linked with temperature margin design criterion (1.5 K). A dedicated modelling tool was developed by coupling the THEA code for 1D thermo-hydraulics in cables and the Cast3M code for 2D transverse thermal diffusion in a limited number of coil cross-sections, enhancing the accuracy of the outputs as being a quasi-3D approach. This tool allows a better assessment of the flux exchange between WP and casing, and the modelling of inter-turn and inter-pancake thermal coupling. The coupling methodology is described, as well as its validation on the simulation of a heat exchanger. A calculation was performed on the CEA proposal for DEMO TF coil in a burn (steady state) scenario, and finally providing a realistic assessment of the temperature margin.

In this paper, we discuss the predictive analysis performed in support of the test program of the International Thermonuclear Experimental Reactor (ITER) poloidal field conductor insert (PFCI). A subset of the test program items was considered, with particular emphasis on DC performance and AC losses. The results and implications of the comparison of selected predictions from different laboratories will be presented.

In the framework of the design activities conducted in the EU for dimensioning the future fusion demonstration reactor (DEMO), extensive analyses were carried out within the framework of the EUROfusion consortium, with the objective of defining the design of the DEMO magnets system. To this aim, CEA has developed ad-hoc predimensioning tools and associated methods in order to size the different magnets: toroidal field coils, central solenoid (CS) and poloidal field (PF) coils. Once predimensioned, the magnet concepts undergo a detailed evaluation procedure with more complex tools and methods. As these detailed analyses are time consuming, the whole design process benefits from an accurate and robust predimensioning process. The predimensioning tools described herein address various aspects driving the operational limits of the magnets while energized in the tokamak, related to electromagnetic, thermal, and mechanic phenomena. In this paper, we expose the latest developments in the predimensioning tools and the methods employed for obtaining a rapid and reliable macroscopic semi-analytical representation of the magnets. The assessment of the predimensioning tools on reference configurations (e.g., ITER) is also described. The application of these design tools on the DEMO configuration issued by EUROfusion is presented, and the resulting magnet design proposals are summarized: First, the PF system, composed of six coils, with one PF selected for study here. Second, the CS, with a modular geometry (five modules) and pancake winding. The approach here assesses the resilience of the CS to a fast-transient (breakdown). The outcomes of a sensitivity study on parameters/criteria are discussed, and extended to some tentative recommendations on the design approach for the DEMO PF and CS magnets. For the PF, a design is obtained for the PF5 coil following a dimensioning methodology consistent with the ITER PF system design with the same level of definition of the plasma operating scenario. As for the CS, it is stressed that consideration of the internal field (from shielding) substantially affects temperature margin results.