Screening currents increase thermal quench propagation speed in ultra-high-field REBCO magnets

TL;DR

The paper demonstrates that superconducting screening currents significantly accelerate electrothermal quench propagation in ultra-high-field REBCO magnets, a factor often neglected in magnet design. It introduces MEMEP-FD, a coupled electromagnetic-thermal framework that uses field-separation iterations, voltage limitation, and axial symmetry, validated against a PEEC benchmark. In a 32 T insert model, accounting for screening currents leads to rapid, coil-wide quench with markedly lower peak temperatures ($\sim$95 K) compared to scenarios neglecting screening currents ($\sim$300 K). The results, including comparisons between Theva and Fujikura tapes, highlight the critical role of $J_c(B,\theta,T)$ anisotropy and AC losses in drive of quench dynamics, with direct implications for design of NMR, MRI, and fusion magnets where precise quench protection is essential.

Abstract

Superconducting REBCO ($RE$Ba$_2$Cu$_3$O$_{7-x}$, where $RE$ is a rare earth, typically Y, Gd or Eu) electromagnets are useful for many applications like medical magnetic resonace imaging (MRI), nuclear magnetic resonance (NMR) spectroscopy, and magnets for particle accelerators and detectors. REBCO magnets are also the core of many nuclear fusion energy start-ups. In order to avoid permanent damage during operation, magnet design needs to take electro-thermal quench into account, which is due to unavoidable REBCO tape or magnet imperfections. However, most high-field magnet designs do not take superconducting screening currents into account. In this work, we show that it is essential to consider screening currents in magnet design, since they highly speed-up electrothermal quench propagation. Our study is based on detailed numerical modeling, based on the Minimum Electromagnetic Entropy Production (MEMEP) and Finite Differences (MEMEP-FD). Benchmarking with well-established Partial Element Equivalent Circuit (PEEC) model supports the correctness of MEMEP-FD. This work focusses on a 32 T all-superconducting magnet design and we analyze in detail the time evolution of electrothermal quench. Our findings will have an impact in the design of ultra-high-field magnets for NMR or user facilities, and possibly for other kinds of magnets, like those for fusion energy.

Figures (19)

• Figure 1: (Left) Qualitative sketch of the 32 T high-field magnet cross-section. The HTS insert is made of REBCO coated conductor. (Right) Sketch of the HTS tape cross-section with the co-wound metal-insulation layer on the right (not to scale).
• Figure 2: Qualitative behavior of the assumed homogenized $E_\varphi-J_\varphi$ relation of (\ref{['JE']}) for the superconductor-normal composite. The graph also shows the relation assuming if there are only supercurrents or only normal currents (first and second term of (\ref{['JE']}), respectively). Example for $\rho_n=10^{-9}$$\Omega$m and $J_{ce}=1.2\cdot 10^{9}$ A/m$^2$.
• Figure 3: The modeling results from MEMEP-FD and PEEC-S agree. Computations for a double pancake assuming uniform $J_\varphi$ and $J_r$ at each homogenized turn. Each homogenized turn is made of 5 real turns.
• Figure 4: Evolution of the electrothermal quench in the cross-section of the metal-insulated REBCO insert made of Theva APC tape when there are damaged turns at the bottom pancake. The damaged turns are 41 to 50 from the inner radius, which consist in one homogenized turn. Screening currents are taken into account. Quantity $I_r$ is the turn-to-turn radial current.
• Figure 5: Time evolution of the engineering current density, $J_{ce}$, normalized angular density, $J_\varphi/J_{ce}$, and quenched region. For the quenched region, we use the criterion of $|J_\varphi|/J_{ce}\ge 2$, where the heat generation at the quenched section due to angular current is approximately half of the heat generation in normal state or higher (see text).