We show by modelling that the quench propagation velocity is not constant in HTS coils but it changes during the quench. Due to the large temperature margin between the operation and the current sharing temperatures, the normal zone does not propagate with the temperature front. This means that the temperature will rise in a considerably larger volume when compared to the quenched volume. Thus, the evolution of the temperature distribution below current sharing temperature T cs after the quench onset affects the normal zone propagation velocity in HTS more than in LTS coils. This can be seen as an acceleration of the quench propagation velocities while the quench evolves when margin to T cs is high. In this paper we scrutinize quench propagation in a stack of YBCO cables with an in-house finite element method software which solves the heat diffusion equation. We compute the longitudinal and transverse normal zone propagation velocities at various distances from the hot spot to demonstrate the distance-variation of these velocities. According to the results in our particular simulation case, the longitudinal normal zone propagation velocity is 30 % higher far away from the quench origin compared to its immediate vicinity when T op =4.2 K and T cs =15 K.

A 10 H superconducting MgB 2 coil, suitable for an induction heating application, has been wound, cooled and initially tested. Eight kilometres of MgB 2 tape were first insulated and then wound into double pancake coils before being assembled into the complete coil. The process is described in detail, focusing on hands-on experience gained, especially regarding the different steps in the wet-winding technique used. The insulation, soldering, winding and cooling techniques applied proved practical for manufacturing of the large MgB 2 coil. The complete coil was cooled by conduction cooling to 8 K and tested with a current of 185 A.

In a voltage-current measurement of a superconductor the sample must be kept at a constant temperature. Otherwise, the measured critical currents and n values are distorted. However, when conduction cooling is applied and currents are high keeping a constant temperature is difficult, or even impossible, to achieve. We measured voltage-current characteristics of a conduction-cooled multifilamentary MgB 2 tape in the self-field above 23 K. Simultaneously, the sample temperature was monitored. Due to the losses at the contact resistances, the sample holder and the sample, considerable sample warming was detected already below the critical current. Therefore, at low ramp rates, the critical current and the conductor n value cannot be determined from the measured voltage-current characteristic. However, when the ramp rate was increased, the critical currents and the n values levelled out. The required increase depended on the sample holder. We studied two sample holders and scrutinized the effect of thermal and electrical conductivities on the distorted critical currents and n values.

Superconducting magnets operate at low temperatures, and therefore, even small heat pulses can ruin their stable operation. For example, resistive joints or changes in the operation current generate heat which must be extracted to prevent a quench. In impregnated magnets, the transverse thermal conductivity inside the coil has a vital influence on the heat extraction, and it dominates the 3D quench propagation. In this study, the transverse heat conductivity is measured from the cross-section of a small epoxy impregnated MgB2 coil at temperatures between 10 and 35 K. Finally, the results are analysed and compared with the results of a computational model based on heat conduction equation solved with the finite element method. The results show that effective thermal conductivity is over two orders of magnitude higher in the parallel direction with conductor axis compared to the perpendicular direction. The measured effective thermal conductivities at 20 K parallel to the broad tape face and perpendicular to the broad tape face were 1.55 W/mK and 0.31 W/mK, respectively. The fill factor of the measured coil was 60.5% for the whole conductor.

No abstract