谷貝 剛

Abstract We have demonstrated an advanced superconducting power conditioning system, in which a superconducting magnetic energy storage (SMES) device, a generator based on a fuel cell (FC), and an electrolyzer are used to compensate for electricity fluctuations over a wide frequency range, combined with a liquid hydrogen storage system to both cool the SMES and provide pure hydrogen gas to the FC and other gas-dependent systems. To manufacture the coils for the SMES, we used MgB₂, whose critical temperature is below the boiling temperature of hydrogen. We developed a 10 kJ SMES coil system indirectly cooled by liquid hydrogen using thermosyphon passive heat exchange to isolate the flammable hydrogen from the electrical components. We performed a successful demonstration of this system for both DC and AC currents ramped at different rates. In the present study, we use computer simulations involving heat balance equations to evaluate the stability of the system. The results obtained are expected to lead to the design of future large-capacity energy storage systems, such as the MJ class, which offer comparable performance to conventional NbTi SMES devices.

The operating temperature of a high-temperature superconducting (HTS) coil is desired to be 15-40 K from the viewpoint of specific heat of the material and the operating current. Development of relatively small-capacity coils with gas helium cooling or refrigerator cooling in this temperature range is underway. Here, refrigerant cooling is desirable for large-capacity coils and liquid hydrogen, which has 20.7 K atmospheric pressure saturation temperature, is a candidate refrigerant. However, due to the difficulty of handling hydrogen, there have been few studies on the HTS coil cooled by liquid hydrogen. And in the HTS coil, the heat generated in the normal conducting region causes a chain of temperature rise and decrease of the critical current, which leads to an irreversible increase in coil temperature beyond the balance with the cooling conditions, and so-called thermal runaway is a problem. In this study, we report on the observation of thermal runaway phenomena in Bi2223 coils under liquid hydrogen immersion cooling. The coils were energized twice at saturated condition of 950 kPa pressure and 30 K temperature. We observed thermal runaway in both energizations and the coil tap voltage before thermal runaway was several tens of mV. From this test, it could be said that the liquid hydrogen cooling has a potential to protect HTS coils.

The critical heat flux in liquid hydrogen is ten times higher than that in liquid helium and is approximately half of that in liquid nitrogen. Since the resistivity of pure metal such as copper or silver at 20 K is less than one-hundredth of that at 300 K, HTS magnets immersed in liquid hydrogen are expected to satisfy the fully cyostable condition or to be stable against high resistive heat generation enough for quench detection at a practical current density. In order to examine cryostability of HTS magnets in liquid hydrogen, a pool-cooled Bi2223 magnet with a 5 T magnetic field at 20 K has been designed, fabricated and tested in liquid nitrogen prior to excitation tests in liquid hydrogen. The magnet consists of six outer double pancake coils with the inner diameter of 0.20 m and four inner double pancake coils with the outer diameter of 0.16 m. The resistive voltage to initiate thermal runaway in the coil as-sembly in liquid nitrogen was higher than 1 V that is sufficient high for quench detection.

Microstructural control of Nb3Sn wire is quite important for improving the critical current density (J c). Elemental doping is an effective method to improve the microstructure. In previous work, we have shown that Zn doping of Nb3Sn wire suppresses Kirkendall voids formation, and promotes the Nb3Sn layer formation. As an effective method to dope with Zn, we have previously proposed a Nb/Cu-Ti/Sn-Zn diffusion-pair structural internal tin Nb3Sn wire. In this study, we investigated the effect of Zn doping on the microstructure and the superconducting characteristics of this system, using a single Nb/Cu-Ti/Sn-Zn structural diffusion couple. The Sn composition in the Nb3Sn layer of the Zn doped specimen is slightly slower than that of the non-Zn doped one. The microstructural analysis was carried out by Electron Back Scattered Diffraction (EBSD), Energy Dispersive X-ray Spectroscopy (EDX) and digital image analysis. The average grain size was decreased with Zn doping, leading to an improvement in layer Jc in a low magnetic field. The thickness of the Nb3Sn layer in the Zn doped specimen is also thinner than that in the non-Zn doped one. In the diffusion-pair structure used in this work, increasing of the Zn content in the Sn-Zn core sacrifices the Sn content in the core. A decrease of the Sn content would account for the smaller Nb3Sn thickness in the Zn-doped specimen.

The influence of parent Nb-alloy grain morphology on the layer formation of Nb3Sn and its pinning characteristics was clarified using a fundamental method. Nb-Ta-Hf was used in this work, given its high recovery temperature of more than 700 degrees C, which is higher than the Nb3Sn-phase formation temperature. First, a Cu/Nb-4at%Ta-1at%Hf single-core composite wire (outer: Cu) was prepared. Subsequently, two pieces were cut off, one of which was intermediately annealed at the recrystallization temperature. Thereafter, Sn was attached onto the surface of both the samples by electroplating, and the samples were heat-treated for Nb3Sn layer formation. Evidently, a significantly finer grain morphology appeared on the as-drawn fine parent Nb-alloy layer, leading to a much better pinning force property. (c) 2021 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

The superconducting Magnetic Energy Storage (SEMS) application still has a great potential to stabilize the utility grid when the uncontrollable power generation from renewable sources increases and power flows change rapidly due to the broad introduction of high-speed response semiconductor switching devices. Along with the development of liquid hydrogen supply chain, the SMES system using MgB₂ conductors also attracts great attention at this point. Although the MgB₂ wires which have critical temperature of around 39 K have been commercially available with more affordable prices, their bending strain sensitivity is an issue to be solved for fabricating large-scale conductors and coils. The experience of constructing a 10-kJ SMES system using Bi2223 tapes and the successful demonstration of compensating very fast electric power fluctuations in the previous project will help us to develop a larger-scale MgB₂ SMES system by investigating conductor and coil design while considering its bending strain sensitivity and mechanism of critical current deterioration to maximize its performance as one of the most promising energy storage devices, following the movement toward a CO₂-free environment.

Our group has developed a coil using MgB2 wires for SMES. In this paper, a prototype coil using a MgB2 Rutherford-Type stranded conductor was fabricated based on react-And-wind (R&amp W) method. In the R&amp W method, a Rutherford-Type conductor in which nine MgB2 wires were wound at pitch of 450 mm around a copper former were manufactured using reacted wires (performed by Columbus Superconductors SpA), and then coiling was performed. The coil was cooled by conduction cooling and the I-V properties were evaluated under magnetic field. As a result, in the R&amp W method, critical current of a coil was degraded, since making a coil by hand would cause the strain beyond the scope of the assumption which was the marginally allowable bending strain. This result suggests the coil for SMES proposed in ASPCS is difficult to react before twisting, and Rutherford-Type conductors should be fabricated before reacting, which means that coil processing should be performed based on the React after making stranded conductors and Wind method, or the wind-And-react method.

Addition of Zn to a Cu matrix during Cu-Zn/Sn interdiffusion reactions at 400 degrees C leads to the formation of a solid ternary Cu-Zn-Sn phase, beta-CuZn, at the outermost reaction layer next to the porous epsilon phase. The use of a brass matrix considerably suppresses void formation and promotes homogeneous outward Sn diffusion in pre-annealing, prior to Nb3Sn formation. There are exactly three paths for Ti doping in the internal tin process, i.e. doping to Sn cores, Nb filaments, and Cu matrix. Ti doping to Sn cores causes a Ti-rich layer formation at the boundary of the Nb filament pack; however, no Ti-rich layers are formed as a result of small Ti doping to the matrix and to Nb filaments. The absence of Ti-rich layers is believed to contribute to a smooth Sn diffusion and suppression of void growth. Atom probe tomography measurements reveal that Ti doping to Sn cores leads to a more inhomogeneous Ti distribution near the grain boundary and a larger variation of the grain boundary thickness than doping to Nb filaments, which may contribute to a better Sn grain boundary diffusion. It is concluded that Ti doping to the matrix, instead of doping to Sn cores, might be more effective in maintaining better growth kinetics of Nb3Sn.

Nb3Sn superconducting wires are widely used for high field applications. It is also expected to play a vital role for high field application in the future, such as future circular collider (FCC). Further improvement in J(c) performance is required to realize FCC. We have been studying the effect of element addition into Cu matrix in the internal tin (IT) processed Nb3Sn conductor. In our previous brass method IT wire, where Ti is doped to central Sn cores, Ti is accumulated between the sub-elements and is inhomogeneously distributed across the wire after the heat treatment. To solve this problem, we tried Ti doping to Nb cores on the brass matrix IT wire. We prepared the specimens using Nb-0, 1, 1.54 at%Ti rod as Nb core. Nb3Sn grain morphology and diffusion reaction behavior of these specimens were compared and we investigated the correlation between the microstructure and the superconducting properties on the specimens with Ti doping to Nb cores.

MgB2 superconductors are promising candidates for application to devices such as Superconducting Magnetic Energy Storage, and generators. To apply MgB2 conductors to such devices, the current capacity of a conductor must be in the kiloampere range. Meanwhile, because the current capacity of a MgB2 conductor is typically approximately 100 A at 5 T and 4.2 K, multiple stranded cables are required. One candidate is the Rutherfordtype cable. During the fabrication of Rutherford cables, strands are deformed by large bending strains at edge corners and indented at flat parts from the pressure of roller dies to maintain the cable shape. It is important to understand how critical current degrades during the fabrication of Rutherford-type cables. To optimize the strand transposition length, three types of Rutherford cables were fabricated and the critical current degradation depending on bending and indented strainswas measured. Moreover, to investigate the degradation, inner structures of the strands were observed using micro-focused X-ray computed tomography and an electron probe micro analyzer.

The MgB2 coil production technology obtaining 30 kJ stored energy the investigation about the SMES coil consists of 600 A, 1.7-T Rutherford-type conductors made of commercially-available MgB2 wires. Due to strain sensitivity before/after heat treatment for MgB2 production, the proper designs of the large-scale twisted conductors both in wind and react, react and wind methods are needed, choosing optimized twist pitches and cable compaction factors. To demonstrate the SMES coil performance, we have been carried out the test campaign of conductors and small prototype coils in various temperature and background field conditions. These results are used for a computer simulation for estimating full size double pancake coil performance of the system, based on the non-steady state heat conduction analysis. The calculated result seems to be a good tool for predicting coil performance for the large capacity energy storage operation.

Superconducting magnetic energy storage (SMES) devices of several tens of kJ class are generally suitable for voltage compensation for microgrids, which produce and distribute electric power to restricted areas. MgB2 material has been developed with superconducting properties by decreasing the production cost. Since hydrogen energy would be widely utilized to realize society with low carbon emission and stored in liquid state for reducing its volume, the power distribution system consisting of MgB2 SMES for compensation of voltage fluctuations cooled by the liquid hydrogen would be effective by synergy effect. However, the MgB2 introduction to large-scale devices is still not enough and under investigation. Our group carried out the investigations to develop MgB<sub>2</sub>cable and pancake coil for the SMES device with specific capacity. The bending strain-sensitive characteristic of MgB2 material forces us to design the twisted conductors and pancake coils with various parameters properly within its tolerable bending strains of both before/after heat treatment.

In order to minimize the weight of support structures for superconducting magnetic energy storage with relatively large storage capacity, the coil for the storage device is designed based on the virial limit, in which the hoop stress is well optimized but flatwise (FW) and edgewise (EW) bending strains would be applied when we use YBCO thin tape to construct the coil. The complex bending strains supposed to be applied to the tape are experimentally investigated by using newly developed complex bending testing device and sets of strain gauges. Even in the range of the presumed bending radius and tensile stress for the coil winding, several combinations of FW, EW bending, and tensile stress cause specific deterioration of superconducting characteristics. The Rosette analysis by using tri-axial strain gauge revealed that the extremely high shear strain would be in relation to the deterioration. This work will help developing a winding technique without any degradation of superconducting characteristics such as critical current and n value of the fragile material.

The force-balanced coil, which has a helical winding configuration, enables the reduction of electromagnetic forces. However, in the use of high-temperature superconducting (HTS) tapes, the in-plane curvature of helical coils may cause a decrease in the critical current due to the edgewise bending strain. The objective of this work is to establish the winding technique of helical coils without plastic deformations of HTS tapes, particularly YBCO coated conductors. The authors developed a prototype winding machine for HTS tapes and discussed the feasibility of the geodesic winding pitch, which canminimize the in-plane curvature variations. The assembly of the prototype winding machine has been finished. From the results of the test operation, the torsion control schemes based on the simultaneous four-spindle angle control system are visually confirmed. In order to estimate the applied edgewise bending strain due to the winding errors, the authors investigated the validity of a development view method of helical coils by using a flexible scale attached to the torus surface.

High-temperature superconductor power applications are still expected to be the main players in reducing carbon footprints. In superconducting magnetic energy storage, increasing the stored energy also increases the electromagnetic force, which generally requires heavier support structure. In terms of the force applied to conductors, it causes the extra bending and/or torsional strain of the REBCO tapes. To minimize the applied stress and to reduce the amount of electromagnetic force support structures, a force-balanced coil (FBC) that has a helical-like structure has been proposed. In this type of magnets, the superconducting thin tape experiences complex strains that come from flatwise and edgewise bending and tensile stress before coil operation. It is necessary to assess the applied strain distribution on the YBCO layer to fabricate an FBC. A newly developed experimental device to apply various strains to the tape and evaluate the strain distributions is presented. The results show that localized large strain will degrade the overall superconducting property.

Due to global issue of energy shortage, alternative energy such as solar and wind energy has to be introduced to the commercial electricity grid. A hybrid energy storage system which is able to convert unstable alternative energy into constant electricity has been proposed. The hybrid energy storage system is composed of hydrogen system and SMES system. This research concerns on the issue of continuous operation of SMES system by introducing a threshold voltage range into the SMES control method. This control method was applied to an SMES coil in a 1-kW-class hybrid energy storage model system. Using the control method, we could restrain the temperature rise of the SMES coil. An experiment has been carried out to confirm the effectiveness of the control method in continuous operation of the hybrid energy storage system. The fluctuation output of a solar power generator was appropriately compensated by electrolyzer and SMES systems.

At the SULTAN test facility, the assessment of the performance of toroidal field coils has been progressing. Unpredictable strand buckling was observed during the destructive investigation of the conductor after significant degradation of the current sharing temperature. The buckling direction was perpendicular to the Lorentz force (LF), and the mechanism causing this was due to the thermal shrinkage occurring because of the difference in the thermal contraction between the strand material and the stainless steel conduit. Our previous research was based on a two-dimensional string model and demonstrated that the observed 2 mm strand bending could have led to strand buckling if the total amount of slide at the contact crossover was assumed to be 80 mu m. To verify this assumption, we fabricated a device for the measurement of the friction force (shear force) between strands under a constriction force comparable to the LF of several hundred of kilonewtons per meter. The results for a Cr-coated 0.89 mm diameter strand surrounded by bare Cu strands indicated that the thermal contraction force would be sufficient to overcome the static friction force when the contraction force is reduced to a tenth of the maximum LF. The mechanism of the slide motion could be divided into two processes: separation of the inner wall of the conduit and the gradual separation from other strands due to a gradual reduction of the LF. A tribological analysis revealed that the real contact area increases by 10% when the shear force is applied; this result would help us evaluate the contact resistance between strands of conductors.

A Cable-in-Conduit-Conductor (CICC) is a composite conductor consisting of many superconducting strands twisted in multiple steps. The CICC has the characteristics which make it suitable for fusion magnets, but lower critical current than expected have been observed in some experiments. The resistance distribution between the strand and the copper sleeve in a "Wrap Joint" is expected to be inhomogeneous and to affect the current distribution and the critical current in the CICC. We measured the DC resistance distribution between the strand and the copper sleeve in a simple model Wrap Join at liquid helium temperature and observed an inhomogeneous resistance distribution. A 3-D strand path calculated by considering the manufacturing process of CICC and an algorithm from previous work were used to evaluate the resistance distributions between the strand and the copper sleeve. The calculated and measured resistance distributions showed the same overall trends. The homogeneity of the resistance distribution between the strand and the copper sleeve is strongly dependent on whether the strand is in direct contact with the copper sleeve or not. Maximizing the number of direct contacts between the strand and the copper sleeve of the "Wrap Joints" was effective for reducing any possible current imbalances.

Thermal stability for a butt joint of a central solenoid (CS) is experimentally and numerically estimated. The butt joint is fabricated using a strand bundle for the CS conductor, and the quench current of the butt joint is measured by changing the temperature of supercritical helium (SHe). The results show that even if the SHe flow slows to 50% of the rated flow, the temperature margin of the joint is 4 K, the butt joint is sufficiently stable. We also calculate the thermal stability of the butt joint by changing certain operating conditions. According to the simulation data, when connection resistance becomes high, from 2 to 5 n Omega, there is little change in the temperature margin. The experimental and numerical results suggest that the butt joint does not quench and can be operated with stability.

The assessment of the performance of toroidal field (TF) coil of ITER has been progressing. Unpredictable strand buckling was observed by the destructive investigation of the conductor. The buckling direction was perpendicular to the Lorentz force (LF), and the mechanism of it was due to the thermal shrinkage caused by the difference of thermal contraction between strand material and conduit. Our previous work utilized a 2-dimensional string model and demonstrated that the observed 2mm stand bending could have led to strand bending if the total amount of slide at the contact cross over was assumed to be 53 μm. To verify this estimation, we fabricated a device for the measurement of friction force between strands under constriction force comparable to the LF several hundred kN/m. Our results for Cr-coated 0.89mm diameter strand surrounded by bare Cu strand indicate that thermal contraction stress applied to strand of 45N would be sufficient to overcome the static friction force when the contraction force reduced to tenth of maximum LF. The mechanism of slide motion could be divided into two processes: separation of the inner wall of the conduit and the separation from other strands due to a gradual reduction of LF.