E.P.A. van Lanen

Distinct voltage–current characteristics of Nb3Sn strands with dispersed and collective crack distributions

Two ITER-type Nb3Sn superconductor strands, one prepared with the bronze route and the other with the internal-tin route, were used to investigate the impact of filament cracking on the strands transport properties. Careful mechanical polishing allowed unambiguous identification of the microscopic fractures of filaments caused by axial straining of the strands. After application of high axial tensile strain, densely and uniformly spaced cracks were observed in the bronze strand, while fewer but more correlated cracks occurred in the internal-tin strand. Crack initiation was observed in the bronze strand after an applied tensile strain of more than 0.8%, while for the internal-tin strand cracks were found already in the unloaded specimen, with further crack growth beyond 0.3% applied strain. With the Pacman strain device, the voltage–current characteristics at zero applied strain were measured after several successive applications of incrementally increasing tensile strain. Distinct dissimilarities in the voltage–current characteristics were observed between the dispersed and the collective crack distributions. We also modelled the influence of cracks on the voltage–current characteristics of the two strands by considering two limiting cases of the crack behaviour.

EU contribution to the test and analysis of the ITER poloidal field conductor insert and the central solenoid model coil

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.

Full-scale calculation of the coupling losses in ITER size cable-in-conduit conductors

With the numerical cable model JackPot it is possible to calculate the interstrand coupling losses, generated by a time-changing background and self-field, between all strands in a cable-in-conduit conductor (CICC). For this, the model uses a system of equations in which the mutual inductances between all strand segments are calculated in advance. The model works well for analysing sub-size CICC sections. However, the exponential relationship between the model size and the computation time make it unpractical to simulate full size ITER CICC sections. For this reason, the multi-level fast multipole method (MLFMM) is implemented to control the computation load. For additional efficiency, it is written in a code that runs on graphics processing units, thereby utilizing an efficient low-cost parallel computation technique. A good accuracy is obtained with a considerably fast computation of the mutually induced voltages between all strands. This allows parametric studies on the coupling loss of long lengths of ITER size CICCs with the purpose of optimizing the cable design and to accurately compute the coupling loss for any applied magnetic field scenario

Simulation of Interstrand Coupling Loss in Cable-In-Conduit Conductors With JackPot-AC

Within the framework of the design analysis of ITER PF coil joints, a model is developed that simulates the coupling loss between strands in a cable-in-conduit conductor (CICC). The present version of this model can simulate these losses in a cable section, subjected to any type of time-changing background field. It calculates the trajectories of all strands in the CICC, and uses this as the foundation for the electrical properties of the model, including strand transport properties, saturation and shielding. The simulation results are first compared with measurements on sub-size CICCs with different strand coating, which affects the interstrand resistance. In all but one of these simulations, the coupling loss time constants are lower than the measured values. A better agreement is obtained with the simulation of an ITER PF1 conductor, subjected to Twente Press experiments. For this simulation, only one final stage sub-cable is used, assuming that coupling currents between them is negligible due to the stainless steel wraps around them.

Modeling of current distribution in Nb3Sn multifilamentary strands subjected to bending

In Nb3Sn cable-in-conduit conductors (CICCs), strands follow complex trajectories that result in a periodic bending strain acting on the strands upon electromagnetic loading and thermal contraction. Such a periodic bending strain leads to degradation of the overall transport performance of a CICC. Aiming for a better understanding and quantitative correlation between strand degradation and CICC test results, a detailed strand model is essential in combination with accurate intra-strand resistance data, the spatial filament strain distribution, and the associated filament crack distribution. Our novel numerical strand model is a 3D network of resistors including superconducting filaments, normal matrix elements, and an outer stabilizing shell or inner core. Along the strand length, matrix elements have Ohmic resistance, there is a filament-to-matrix contact resistance (Rfm) between filaments and matrix elements, while superconducting filaments have a power-law voltage‐current (VI) characteristic with critical current (Ic) and an n-value described by the ITER Nb3Sn strain scaling law based on measured strand data. The model simulates the VI characteristic in a periodic bending experiment and provides the associated spatial potential distribution. The VI characteristics representing the low- and high-resistivity limits (LRL and HRL) are identified for periodic and uniform axial bending. The voltage level for the current transfer regime depends on the strand internal resistivities, i.e. the filament-to-matrix contact and the matrix resistivity, the twist pitch and the bending wavelength. The simulation results show good agreement against Ic degradation, as experimentally measured by the TARSIS facility, versus the assessed peak bending strain. In addition we discuss different methods for determining the applied peak bending strain. The model provides a basis to find a practical relationship between a strand’s VI characteristic and the periodic bending strain, as well as a mapping of well-characterized strand performance to that of a full-size CICC. (Some figures may appear in colour only in the online journal)

Validation of a strand-level CICC-joint coupling loss model

Calculating the coupling losses in cable-in-conduit conductor (CICC) joints requires a large amount of numerical effort, which is why the numerical system is often reduced by grouping strands together. However, to better understand the loss behaviour, and eventually the stability mechanism in such joints, a full-sized model working on the level of individual strands is more desirable. For this reason, the numerical cable model JackPot-AC has been expanded to also simulate the coupling losses in a CICC joint. This model has been verified with AC loss measurements on a mock-up joint, which was subjected to an applied harmonic field at different angles. The mock-up joint consisted of two sub-sized CICCs connected by a copper sole. For additional verification the AC loss of one of these conductors and the copper sole was also measured separately. The results of the simulation agree with the measurements, and the model therefore proves to be a useful analytical tool for examining the coupling loss in CICC joints

Numerical analysis of the DC performance of ITER TF samples with different cabling pattern based on resistance measurements on terminations

The DC transport properties of ITER TF conductors are analyzed in the SULTAN facility, but their associated current sharing temperature (Tcs) can give scope for different interpretations. To extract the conductors pure performance during such short sample tests requires a detailed quantitative model such as JackPot, for which as many as possible input parameters are based on experimentally verified data. The crucial interstrand contact resistance data of the TFPRO-2 and TFJA-3 samples, which have a different cabling and joint layout, has recently become available. With only three parameters for JackPots joint model, we were able to find a good match between the measured and simulated interstrand resistances for all terminations, despite their different layout. For Nb3Sn conductors, the axial strain is the only free model parameter left for matching the simulations with SULTAN Tcs tests. We were able to find a match with the measured voltage–temperature characteristic for all but one conductor. The simulation results indicate that the termination layout can be the cause of lower sample performance due to current imbalance. The outcome of this analysis also confirms that transverse load degradation can be significantly mitigated by finding an optimal set of cable twist pitches.

Direct measurements of inter-filament resistance in various multi-filamentary superconducting NbTi and Nb3Sn strands

For a proper characterization of multi-filamentary NbTi and Nb3Sn strands and a better understanding of their performance in short sample tests, as well as for increased understanding of inter-strand current redistribution in cabled conductors, a quantitative knowledge of the inter-filament transverse resistance is essential. In particular, in the case of strain or crack distributions among and along filaments in strain-sensitive superconductors such as Nb3Sn cable-in-conduit conductors, a much better understanding of the voltage?current transition is required as a basis for the analysis of full-size cables.Two particular four-probe voltage?current methods are developed to measure the transverse inter-filament resistance distribution directly, both in well-established and in state-of-the-art superconductors that are presently applied in the ITER, JT-60SA and LHC magnets. To extract values of the filament-to-matrix contact resistance from these direct experiments, some further assumptions are needed. These assumptions are based on FEM simulations and on measurement of the longitudinal strand resistance.An overview is given of a wide range of measurements on various NbTi and Nb3Sn strands, performed at temperatures below 10?K and at various applied magnetic fields. We present the results of the experiments and simulations and demonstrate how the extracted characteristic parameters provide a better insight into the current flow patterns within the strands.

Interstrand Resistance Measurements on the Conductor Terminations of TFPRO2 and JATF3 SULTAN Samples

A number of ITER TF cable-in-conduit samples that have been tested in the SULTAN facility performed significantly below the single strand expectations. Although performance degradation related to transverse electromagnetic loads plays a role in Nb3Sn CICCs, a similar effect can occur when the current in the sample is non-uniformly distributed among the strands, driving part of the strands into saturation already at low currents and temperatures. A sufficiently homogeneous current distribution at low electric field level (less than 10 ¿V/m) requires low interstrand resistances in the sample terminations and/or a homogeneous distribution of contact resistances between strands and the joints copper sleeve. To evaluate the influence of the joint design on qualification testing, we performed post-mortem interstrand resistance measurements on the conductor terminations of two different joint design concepts: the TFPRO2 and JATF3 SULTAN samples. The interstrand resistances and their distribution were determined on a large number of strand combinations. Accordingly, comparisons were made on how the sample performed in SULTAN. From the different design aspects, full solder filling of the terminations appears to have the strongest effect on the interstrand resistance reduction.

Direct Measurement of Inter-Filament Resistance in Superconducting Multifilamentary NbTi and

A quantitative knowledge of inter-filament transverse resistance will allow us to describe current redistribution processes inside strands more accurately. This is particularly important for the analysis of the influence of strain and crack distribution patterns in Nb3Sn filaments on the shape of the voltage-current curves. Several indirect methods are commonly used to assess inter-filament resistance. Here we use a direct method to measure transverse inter-filament resistance and filament-to-matrix contact resistance. Two four-probe voltage-current methods are developed for measurements below 10 K at various background magnetic fields. In addition to FEM (Finite Element Method) simulation, also a new 3D strand model is developed to simulate the current- and voltage distributions. The experimental methods, first results as well as the simulations using the FEM method and new 3D strand model are described.