Interpretation of conduit voltage measurements on the poloidal field insert sample using the CUDI–CICC numerical code Y. Ilyin * , A. Nijhuis, H.H.J. ten Kate Low Temperature Division, Faculty of Science and Technology, University of Twente, P.O. Box 217, 7500AE Enschede, The Netherlands Abstract The results of simulations with the CUDI–CICC code on the poloidal field insert sample (PFIS) tested in the SULTAN test facility are presented. The interpretations are based on current distribution analysis from self-field measurements with Hall sensor arrays and cur- rent sharing measurements. The possible variation in voltage–current (VI) curves among the sub-cables in the PFIS caused by the cable self-field and joint non-uniformity is a principal issue for the evaluation of the strand-to-cable performance. The cable transverse volt- ages, initiated by current non-uniformity in a cabled conductor, can affect the shape of the longitudinal VI curve. The basic cable data are obtained by petal-to-conduit contact resistance and conduit resistivity measurements under transverse cyclic load in the Twente Cryo- genic Cable Press. We assessed the impact of the conduit, the sub-cable resistive barrier wraps and the location of the voltage taps on the measured voltage, and some comparisons were carried out with experimental runs on the PFIS. The outcome confirms that the transverse voltages, caused by current redistribution, do affect the shape of the VI transition. This effect seems particularly important in short sam- ple tests. Moreover, it is demonstrated that the location of the voltage taps can affect the experimentally obtained VI curve and solutions for most accurate measurement of the VI are discussed in the paper. The numerical model, CUDI–CICC covers the final cabling stage of inter-petal interactions only, by which the possible role of intra-petal non-uniformities is ignored, although recognized as potentially relevant. Ó 2006 Published by Elsevier Ltd. Keywords: Superconductors; NbTi cable-in-conduit conductors; Voltage–current characteristics; Voltage measurements; ITER model coils; Numerical codes 1. Introduction The multistrand cable-in-conduit conductors (CICC) for the international thermonuclear experimental reactor (ITER) consist of more than 1000 superconducting strands with a strand diameter of about 0.8 mm [1]. The conduc- tors have a void fraction of 33–36% and are cabled by twisting the strands in several stages. One of the NbTi type CICCs (considered throughout the paper) aimed for the ITER poloidal field (PF) coils is designed to carry 45 kA in a magnetic field of 6 T and temperature of 5 K during the normal mode of the magnet system operation [1,2]. The axial electric field along a section of such a conduc- tor in current sharing mode is a measure for the perfor- mance of magnet coils or short samples for comparison to the single strand properties. However, the current in the cable cross-section is not homogeneously distributed due to the intrinsic non-uniformity of the joints, where all strands are connected to a current supply but with nat- ural variations in electrical resistance. A non-uniform dis- tribution of the transport current between the strands or strand bundles of a cable causes an earlier voltage rise in the cable voltage–current (VI) or voltage–temperature (VT) characteristic compared to what is measured on a sin- gle strand. The cable self-field effect, causing a considerable magnetic field gradient over the conductor cross-section, must also be taken into account. This was demonstrated in [3] for the case of the central solenoid insert coil (CSIC). 0011-2275/$ - see front matter Ó 2006 Published by Elsevier Ltd. doi:10.1016/j.cryogenics.2006.01.008 * Corresponding author. Fax: +31 53 489 1099. E-mail addresses: y.ilyin@tnw.utwente.nl, ilyin@tnw.utwente.nl (Y. Ilyin). www.elsevier.com/locate/cryogenics Cryogenics 46 (2006) 517–529