1926 IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 21, NO. 3, JUNE 2011
Simulation of Interstrand Coupling Loss in
Cable-In-Conduit Conductors With JackPot-AC
E. P. A. van Lanen and A. Nijhuis
Abstract—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 in-
terstrand resistance. In all but one of these simulations, the cou-
pling 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 simula-
tion, only one final stage sub-cable is used, assuming that coupling
currents between them is negligible due to the stainless steel wraps
around them.
Index Terms—CICC, coupling loss model, ITER, supercon-
ducting strands.
I. INTRODUCTION
T
HE ITER tokamak contains six Poloidal Field (PF) coils
to provide shaping and vertical stability of the plasma.
They are designed with NbTi Cable-In-Conduit Conductors
(CICCs), which are wound in double-pancakes, and connected
in series with joints at the outer diameter [1]. Although stability
analysis of these coils concluded that they will operate success-
fully within the specified temperature margin [2], [3], a more
detailed analysis on the coil joints was also recommended [4].
There are essentially two reasons for this: First, the transport
current is not fully transposed in the cable inside the joint, and
close to it, making this region more vulnerable for instability
[5], [6]. The second reason concerns the orientation of the back-
ground field with respect to the joints. The joints are placed
in a position where they are expected to suffer the least from
the longitudinal component of the PF coil’s field. However, the
loss due to other components of the background field, not only
from the same PF coil, but also from other coils in the system,
has not yet been investigated.
Estimating the effect from the current non-uniformity and the
field orientation on the thermal stability of the joints requires
a highly detailed model. Such a model was already developed
Manuscript received August 03, 2010; accepted September 25, 2010. Date of
publication November 09, 2010; date of current version May 27, 2011. This
work was supported by ITER under Service Contract No. UT-ITER/CT/09/
4300000070.
The authors are with the Faculty of Science and Technology, Univer-
sity of Twente, 7500 AE Enschede, the Netherlands (e-mail: e.p.a.van-
lanen@tnw.utwente.nl).
Digital Object Identifier 10.1109/TASC.2010.2082474
Fig. 1. Illustration of the CICC division in Cable Sections (CS), separated by
Cable Intersections (CIS).
earlier to analyze the DC behavior of CICCs [7]. This model
(JackPot) accurately calculates the trajectories of all strands, and
deduces all electrical properties from this. It is fully based on ex-
perimentally determinable conductor properties, such as strand
parametric scaling laws and interstrand contact resistances. It
has been used successfully to explain the differences between
short and long length samples of ITER PF conductors [8], and
is now being expanded with time-dependent components.
This paper describes the first step, in which only a short
length of CICC is considered, which is not connected to a
power source, and placed in a homogeneous background field
that changes in time. This makes it already possible to simulate
coupling loss measurements on cable sections in a solenoid or
dipole magnet, and to perform parametric studies.
II. MODEL DESCRIPTION
Similar to other models [6], [9], [10], JackPot-AC simulates
the CICC as a network of (mutual) inductances, induction
voltage sources and two types of resistors; one that represents
the electrical contacts between strands, and one that involves
the superconducting strand properties (see Fig. 1). The back-
bone of the model is the cabling subroutine. It calculates the
coordinates of the central axes of the strands, which are used to
determine many of the electrical properties [7]. For example,
an interstrand conductance is only present between strands if
they are close enough together.
The strand trajectories are now also used to calculate the mu-
tual inductances. For this, it is assumed that the coupling takes
place between all piecewise linear sections that represent the
strands, including the sections that belong to the same strand.
The calculation of a section’s self inductance is done with the
theory described in [11]. The total self-inductance of a single
1051-8223/$26.00 © 2010 IEEE