Multiscale numerical modelling of REBCO conductors for fusion : multi-tape architectures, variability, AC losses and multiphysics coupling #
Framework: Université de Lorraine / GREEN, PEPR SupraFusion, scientific collaboration envisaged with CEA, including possible measurement campaigns on prototypes at Cadarache.
Keywords: REBCO, fusion, HTS conductors, multi-tape cables, CORC/HFRC, FEM, AC losses, electromagnetic modelling, variability, multiphysics.
Contact :
Kévin Berger : kevin.berger@univ-lorraine.fr
Context and objective #
Superconducting magnets for fusion require conductors able to carry very high currents under high magnetic field, with limited losses, good stability and mechanical robustness compatible with operating constraints. REBCO tapes are now a major route towards high-field magnets, but their assembly into multi-tape conductors carrying several kA to several tens of kA remains an open scientific and technological challenge.
Several architectures are currently investigated worldwide: straight or twisted stacks, CORC/HFRC conductors, Roebel, CroCo, SECAS, TSTC, VIPER or hybrid concepts. These solutions differ in tape arrangement, transposition level, electrical contacts, compactness, AC losses, self-field effects, current redistribution and manufacturing constraints.
The PhD project will build on the experience of GREEN in the modelling of superconducting systems: AC losses, nonlinear E − J laws, self-field effects, finite-element formulations and electromagnetic, thermal or mechanical couplings applied to HTS devices. It will be carried out in the framework of PEPR SupraFusion, with a scientific collaboration envisaged with CEA, in particular on high-current REBCO conductor architectures for fusion. Measurement campaigns may be carried out or exploited at CEA Cadarache on conductor or sub-cable prototypes, depending on sample availability and experimental facilities.
Develop a multiscale numerical modelling workflow to compare different REBCO conductor architectures for fusion, integrating AC losses, current distribution, self-field effects, realistic tape variability and targeted multiphysics couplings.
Scientific challenges #
Comparing several conductor architectures on common grounds #
Multi-tape REBCO conductors are often studied with assumptions specific to each architecture, which makes direct compari- son difficult. The PhD will aim to define numerical benchmark cases to compare several conductor families using common criteria: AC losses, effective critical current, current homogeneity, engineering current density Je , self-field effects, influence of contact and termination resistances, and robustness against defects.
Predicting AC losses and current distribution in nonlinear REBCO conductors #
The calculation of electromagnetic losses is central to the PhD project. In a REBCO conductor, losses depend on the geometry, the local orientation of the tapes, the applied magnetic field, the self-field, electrical contacts and the nonlinear superconducting law.
The current distribution evolves with the current level, field ramps, progressive saturation of some tapes, the local dependence of Ic(B, T, θ), and the onset of weakly dissipative zones. These effects directly influence hysteretic losses, coupling losses and losses associated with current redistribution.
The models developed should allow the evaluation of losses in tapes, coupling losses between tapes or sub-cables, the influence of contact resistances, the localization of dissipative zones and the effect of ramps representative of fusion operating conditions. Particular attention will be paid to non-idealized external magnetic-field scenarios, including irregular ramps, fast variations or non-periodic sequences, for which reduced or hybrid models may complement detailed finite-element calculations.
The models will combine methods adapted to the scale considered : finite-element formulations such as H, H-ϕ, T-A or J-A for the local calculation of magnetic field, E − J law and losses ; equivalent or nonlinear network models to represent contacts, terminations and couplings at conductor scale. This organization will make it possible to build hybrid approaches balancing physical accuracy and computation time.
Defining the appropriate level of description from tape to full conductor #
A fusion conductor may contain several tens or even several hundreds of tapes. A full geometrical representation of all details is rarely compatible with parametric studies. Conversely, excessive homogenization may hide important local phenomena. The PhD will therefore seek to identify the relevant level of description : individual representation of selected tapes, equivalent groups, representative cells, cross-section models or reduced models. The objective is to preserve the dominant physics while keeping computation times compatible with architecture comparison and sensitivity analyses.
Including material, geometrical and assembly variability #
The behaviour of a multi-tape conductor does not depend only on the nominal properties of the tapes. It is influenced by variability at several levels: material properties, geometry, stabilization, electrical contacts, mechanical assembly and manufacturing.
The models may include variations in Ic , n-value, copper thickness, total tape thickness, tape width, positioning, angular misalignment, twist pitch, compaction, contact resistance, termination resistance or assembly quality. Local defects may also be considered : degraded tape, poor contact, non-uniform pressure, imperfect soldering or locally less efficient cooling. The objective will be to identify which sources of variability significantly affect losses, current distribution and robustness, in order to move beyond the modelling of an ideal conductor.
Confronting the models with prototypes and experimental measurements #
The models should be confronted, whenever possible, with representative experimental data. Measurement campaigns may be carried out or exploited with CEA Cadarache on REBCO conductor or sub-cable prototypes. Quantities of interest may include critical current, V − I curves, AC losses, segmented voltages, contact effects, response to current or field ramps, and the influence of heterogeneities.
This model/experiment comparison will help adjust modelling assumptions, identify dominant parameters and propose benchmark cases useful for PEPR SupraFusion.