VIPER: an industrially scalable high-current high-temperature superconductor cable

VIPER cable design allows for a simple joint design and manufacturing process. Joints, which supply current to cables, are significant sources of resistive heat load, historically complex, fragile and challenging to fabricate, and often the cause of cable and application failures. The fabrication complexity for joints in LTS cables can be high [30, 31]. While higher temperature stability margins can relax joint resistance requirements in HTS cables, HTS joints to date have not achieved the simple design, consistent low resistances, and easy fabrication required for commercial applications [16, 32, 33]. Nano-ohm resistances have been achieved, but to do so hundreds of tapes needed to be individually connected per joint, resulting in complex, long, and error-prone fabrication processes that do not scale favorably and make in-field repair and maintenance difficult [34–36].

These challenges are all addressed in VIPER cable. During fabrication, the VPI solder process brings individual HTS tapes into intimate electrical contact with each other, and a roll-formed copper jacket provides a mechanically strong and electrically continuous shell from termination to termination. Thus, the challenge of creating cable joints is reduced to connecting the external copper cable jacket directly to another cable jacket or current lead. As a result, VIPER joints involve minimal manufacturing complexity. For each cable pair, a double saddle is simply clamped between two cable jackets.

An example of a VIPER cable-to-cable joint designed for a cable test in the 'praying-hands' configuration is shown in figure 1. VIPER joints provide a high degree of design flexibility, accommodating, for example, 'shaking hands' style joint configurations and smaller copper saddles. Smaller saddles minimize the distance between cables, reducing joint resistance and increasing current density. Joint fabrication can be completed in several person-hours per joint with basic tooling and produces consistently low resistances. The electrical resistance for four VIPER cable-to-cable joints at operating temperatures and magnetic fields are shown in table 1. These low resistances were maintained over thousands of mechanical loading cycles and multiple thermal cycles. Current was shown to be uniformly distributed within the HTS stack less than 5 cm beyond the joint regions at 77 K, indicating effective current redistribution. The joints were easily disassembled and re-used in a few hours, facilitating access for inspection and maintenance essential in real-world applications. At MIT, a single joint pair was re-used over ten times for cable tests in liquid nitrogen with no change in electrical or mechanical performance. At SULTAN, the joints for the Bravo sample were re-used for the Delta sample and achieved similar (even slightly better) joint resistances as shown in table 1.

In high-field magnets, I×B forces can destroy HTS tapes by driving them into surrounding structural material, by creating high tape stack face pressures, and by creating strain from the circumferential expansion of the magnet [38]. The Alpha pair followed the design shown in figure 1 but had one HTS stack each. It underwent 2000 I×B cycles with 17 kA × 10.9 T = 185 kN m⁻¹ (36 MPa transverse pressure) on the HTS stack at T = 5 K, a reverse I×B cycle, and a cryogenic-to-room-temperature thermal cycle. Critical temperature (measured at I = 7 kA current and T = 30 K) degraded only slightly (2.0% and 3.8% for the two cables) in the first 30 cycles before stabilizing for the remainder of the test.

The Bravo and Charlie pairs were identical to each other and had one HTS stack in each cable. They underwent 1550 (Bravo) and 500 (Charlie) I×B cycles at 35 kA × 10.9 T = 382 kN m⁻¹ (75 MPa transverse pressure) on the HTS stack at 5 K, a reverse I×B cycle, and a thermal cycle (Bravo). The Charlie cables were axially strained to 0.5% using a specially engineered support structure to simulate the hoop strain that occurs in a magnet. To our knowledge, this is the first simultaneous I×B loading and axial strain test in straight superconducting cables at realistic operating conditions, avoiding the need for placing complex 3D test cables inside expensive and rare large-bore magnets.

The critical current evolution for the Bravo and Charlie pairs are shown in figure 3. Critical current in the four cables degraded by between 3.1% and 4.1% in the first 30 cycles then stabilized. The ability to withstand 382 kN m⁻¹ on a single HTS stack with negligible degradation improves on the previous record of 102 kN m⁻¹ by almost a factor of four [11].

The Delta pair (4 HTS stacks each) experienced 150 I×B cycles at 50 kA × 10.9 T = 545 kN m⁻¹ for the cable and 136 kN m⁻¹ (27 MPa transverse pressure) per HTS stack, two thermal cycles, and 24 quench simulation events. Delta cables responded similarly to cycling as Alpha, Bravo, and Charlie with critical current (I_c = 45.5 kA measured at B = 10.9 T and T = 10 K) degradations of 2.4% and 2.5%.

A summary of the electromechanical results is shown in table 1. For all eight cables, performance degradation asymptoted to between 2.0% and 4.1%, independent of cable fabrication specifics, HTS tape manufacturer, I×B loading, axial strain, and operating conditions. Sectioning, polishing, and SEM imaging of the cables after the SULTAN tests, such as a cable from the Charlie pair shown in figure 4, showed no significant macroscopic mechanical damage due to the immobilization of the HTS tapes in the solder.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Microscopic 100 µm lengths of REBCO layers were observed in inter-HTS tape solder voids, and had separated from their substrates. These low-density microscopic defects—due to a combination of fabrication imperfections and mechanical cycling—do not significantly degrade the electrical performance. Electromechanical finite element analysis was performed and suggests that plastic strains in the solder, which concentrate at the corner tips of the HTS stack and approach 1%, likely cause a few percent of the HTS tape to degrade under loading, consistent with experimental observations. The analysis showed that ensuring good HTS structural support through good solder ductility, small solder coefficient of thermal expansion, and strong solder adhesion to the HTS and copper former are the key drivers to eliminate critical current degradation.

The immobilization of the HTS stacks provided by the VPI solder mitigates I_c degradation from large I×B forces in VIPER cables and effectively transfers the force into the surrounding steel support structure. Because of the cable twist pitch, I×B forces applied during SULTAN testing were incident not just normal to the flat of the tape (favorable direction) but in the shear direction (unfavorable direction) as well. No significant delamination of HTS was observed in SEM imaging of cable anywhere along the twist pitch, corroborating the absence of I_c and joint resistance degradation despite up to 2000 I×B cycles in SULTAN tests. This compares favorably to the challenging development of LTS Nb₃Sn cables for fusion magnets, in which substantial efforts were required to overcome the large and continuous I_c degradation observed in SULTAN tests [39], and to other present HTS cable concepts, in which repeated movement of HTS stacks or sub-strands caused unstabilized I_c degradation of 10% to 20% at 2000 cycles under twelve times less I×B force (∼30 kN m⁻¹) and seven times less transverse pressure (∼10 MPa) per HTS stack than was applied to VIPER cables [40]. Because force per stack—not total cable current—is the critical risk retirement metric for eliminating I_c degradation, a VIPER cable with six HTS stacks with roughly six times the current density of the Bravo or Charlie cables can be deployed with minimal additional risk of I_c degradation.

Off-normal heating in a superconducting cable can cause thermal runaway known as quench, which has been a major outstanding issue for insulated HTS magnets, particularly for fusion and accelerator magnets with large stored energy. Quench detection in HTS cables is more challenging than in LTS cables due to their higher thermal stability, which can cause lower normal zone propagation velocities (NZPVs), slower detection times, and higher probability of damage [41, 42]. Quench detection techniques that are fast enough to avoid damage in practical HTS cable-based magnets with high electrical noise and significant energy storage have not been demonstrated to date.

For the Delta cables in SULTAN, NZPVs were measured by using the propagation of a 100 µV cm⁻¹ electric field front on axial voltage tap arrays during quench events with results between 0.04 and 0.42 m s⁻¹ under various operating conditions (B = 10.9 T; T = 10 K, 20 K; operating current I_op = 36.0 kA, 40.5 kA, 45.0 kA, 47.5 kA; I_op/I_c = 0.8, 0.9, 1.0, 1.05). The measured NZPVs were equal to or faster than those measured in individual HTS tapes, which have been found between 0.07 and 0.13 m s⁻¹ for typical operating conditions [41]. VIPER NZPVs approach those in LTS conductors [43], are consistent with analytical calculations [44], and enable a wide optimization for voltage tap-based quench detection systems on VIPER cables.

Figure 5 shows a Delta cable recovering from three increasing-duration 45 W external resistive heater pulses but quenching on a fourth. A detailed multiphysics model successfully captures the superconducting, electrical, thermal, and magnetic behavior of the cables without relying on any free parameters. Figure 6 shows a model-generated cryostability map with a predictive boundary between stability and quench in response to external heating that is in good agreement with measured data. The model confirms that VIPER cryostability results from the favorable combination of high heat capacity and low electrical resistance of the copper former at 20 K, the amount of copper used, and the rapid heat removal pathway provided by the VPI soldering of the HTS and copper former.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

At typical cable operating conditions (B = 10.9 T, T = 20 K, I_c = 31.5 kA, I_op < 31.5 kA, I_op/I_c < 1.0, and helium mass flow rate ${\dot{\mathrm m}}_{\mathrm{HE}}$ = 2.0 g s⁻¹), Delta cables were stable up to 0.8 MW m⁻³ of heating, induced by the external heater running at 57 W for up to 15 s and depositing 0.9 kJ of energy. In response to heater pulses greater than 47 W, the cables recovered from strong current-sharing operation with electric fields in excess of 100 µV cm⁻¹, 100 times greater than the 1 µV cm⁻¹ HTS superconducting-to-normal transition criterion. Quenches in response to external heater pulses were almost exclusively observed at T < 20 K and I_op/I_c > 1.0.

Temperature excursions on the Delta cable were monitored with two fiber optic quench detection technologies: an array of discrete fiber Bragg gratings (FBG) [45] and a distributed single ultra-long FBG (ULFBG) [46]. Each technique uses a variation of monitoring the temperature-dependent spectral response of FBG inscribed into optical fibers. FBG and ULFBG coated optical fibers were embedded in the outer circumference of the Delta cable copper jackets, ensuring good thermal contact and mechanical protection. Figure 5 shows that both technologies demonstrated consistent, unambiguous detection of local temperature excursions of 2 K to 3 K. ULFBG detection times were on the order of 100 ms, oven ten times shorter than time-to-damage from quenches in superconducting fusion magnets [47]. FBG response times to hot spots at the gratings are as rapid but total detection times are dependent on the grating spacing and NZPV.

Both techniques represent improvement over traditional voltage tap approaches in, for example, large-scale fusion magnets, which require multi-second wait times to achieve sufficient signal-to-noise and could result in potentially destructive temperature rises up to 150 K [48]. The two fiber-optic systems are complementary: the ULFBG system is a sectional measurement that provides fast detection times at the expense of positional information; the FBG system provides spatially resolved temperature excursions anywhere in the cable but requires time for the normal zone to propagate to the nearest discrete grating. Both technologies have favorable scaling to the 100 m lengths required for high-field fusion magnets and other applications, requiring manufacturing of longer fibers but utilizing the same equipment and analysis techniques used in the shorter experiments described here.