2.1 Experimental Set-Up

A superconducting cable consists of mainly three parts: termination, core, and joint box. A brief structure of a superconducting cable is shown in Fig. 1. Fig. 2 shows the schematic view of a termination part. This paper deals with the design of a termination part in Fig. 2.

In this study, experiments on the electrical breakdown characteristics are carried out in order to establish the criterion of creepage discharge along the surface of a spacer in insulation gas as well as sparkover characteristics of insulation gas, GN$_2$. It is known that dielectric characteristics of SF$_6$ is superior to those of GN$_2$.[ref.] Therefore, dielectric experiments on GN$_2$ is conducted and design of a termination part considering the dielectric characteristics of GN$_2$ is derived in order to design in consideration of inferior case. Schematic drawings of an electrode system for sparkover and creepage discharge experiments are shown in Fig. 3. In this figure, ‘HV’ denotes high voltage applied from a power supply to a sphere electrode. Dielectric experiments on sparkover and creepage discharge characteristics are conducted under AC and lightning impulse (Imp.) voltages at various pressure conditions. Also a schematic drawing of dielectric experiment is shown in Fig. 4. Sphere-to-plane electrode systems for sparkover and creepage discharge experiments with various diameters of a sphere electrode and gaps between a sphere and a plane electrode are used and specifications are described in Table 1 and Table 2, respectively. High voltage is applied to a sphere electrode, whereas a plane electrode is grounded.

Every electrical breakdown experiment is conducted under AC and Imp. voltages. An AC voltage generator has a maximum output of 100kV with a frequency of 60Hz and the ramping up rate is set to 1kV/s. The maximum capacity of an Imp. voltage generator is 600kV. Also, it has a wave front time of 1.2μs and a wave tail time of 50μs. All experiments are repeated seven times under the same conditions. Weibull distribution analysis for experimental results is conducted by using a statistical analysis software, Minitab.⁽⁴⁾

2.2 Experimental Results

Various experimental conditions are applied to experiments in order to verify the influence of a field utilization factor to dielectric characteristics. A field utilization factor indicates the field uniformity of an electrode system and is computed by the ratio of mean electric field intensity (E$_{MEAN}$) and maximum electric field intensity (E$_{MAX}$). Fig. 5 shows the electrical breakdown voltage (V$_{BD}$) of GN$_2$ according to gap under the absolute pressure of 0.4 and 0.5MPa. Based on the experimental results, it is known that V$_{BD}$ of GN$_2$ increases as pressure increases. V$_{BD}$ proportionally increases as pressure increases due to the probability of ionization decreases as the electron velocity decreases. The decreased velocity of the electron causes reduced collision energy for electrical breakdown.⁽⁴⁾

2.3 Criterion of Electric Field Intensity at Sparkover

It is reported that the electrical insulation design could be conducted by using the derived functional relationship according to a field utilization factor.⁽⁵⁾ It is revealed that the electric field intensity at sparkover voltage is dependent on a field utilization factor dependent on the configuration of an electrode system.⁽⁴⁾

As shown in Fig. 6, maximum electric field intensity at electrical breakdown voltage (E$_{BD,MAX}$) shows a saturated tendency as a field utilization factor increases. Therefore, the saturated minimum electric field intensity at sparkover is selected in this study to secure maximum electrical reliability of a termination part. In Fig. 6, it is found that E$_{BD,MAX}$ under AC voltage is smaller than that under Imp. voltage. The detailed experimental results on E$_{BD,MAX}$ along the surface of a solid material in GN$_2$ are dealt in the previous study, ⁽²⁾ and .⁽⁴⁾ The saturated E$_{BD,MAX}$ of GN$_2$ under various pressures and applied voltages is shown in Table 3. The saturated E$_{CD,MAX}$ of creepage discharge along the surface of solid materials such as GFRP and epoxy resin in GN$_2$ under various pressures and applied voltages is presented in Table 4 and 5, respectively. It is found that the saturated E$_{CD,MAX}$ of creepage discharge is lower than those of penetration sparkover.

3.1 Electrical Design Method

The criteria of E$_{BD,MAX}$ and E$_{CD,MAX}$ shown in Table 3, 4, and 5 are used to design an electrically reliable termination part of a 23kV superconducting cable. The standard for electrical insulation design of a termination part conforms to the withstanding voltage shown in Table 6.⁽⁹⁾ The relative permittivity of materials used for a finite element method (FEM) simulation is shown in Table 7.⁽²⁾

The structure of a termination part to be designed is shown in Fig. 7. A brief design considering E$_{BD,DESIGN}$ of a termination part is conducted by using the FEM simulation software, COMSOL. E$_{BD,DESIGN}$ denotes E$_{BD,MAX}$ of the designed model calculated by FEM simulation. An electrical SFGAS, a safety factor at the gaseous part can be calculated by formulae (1). EBD,CRITERION represents a criterion of E$_{BD,MAX}$ obtained by the creepage discharge experiment.

The design of an edge ring is proposed to mitigate E$_{BD,MAX}$ as shown in Fig. 8. The SF of an edge ring could be calculated by (2). In this case, the radius of an edge ring is set to 30mm to mitigate E$_{BD,MAX}$ at the edge part.

The surface length of a spacer made with glass fiber reinforced plastic (GFRP) and epoxy resin is calculated considering an SF which could be calculated by using (3).

3.2 Electrical Design Analysis

As shown in Fig. 8, E$_{BD,DESIGN}$ is reduced from 3.78kV/mm to 1.87kV/mm by adopting an edge ring with the radius of 30mm. An SF which indicates the stability of a designed system at rated voltage is calculated as 2.99 for AC and 2.16 for Imp. voltage. An SF for electrical design could be calculated by (1). It is verified that an SF at the edge part where a vertical part and a horizontal part meet could be reduced by applying a shield ring called an edge ring. Also, a shield ring design is suggested to reduce the triple point effect where the spacer and current lead meet.⁽⁸⁾ In Fig. 9 (a), it is found that E$_{BD,MAX}$ generated in GN$_2$ around a spacer without a shield ring is higher than that at an edge part in Fig. 8 (a). However, E$_{BD,MAX}$ at the intersection part at the end of a spacer is reduced from 5.12kV/mm to 1.25kV/mm by installing a shield ring with the diameter of 20mm. The calculated SF according to the surface length of a spacer made with epoxy resin is shown in Fig. 10. As shown in Fig. 10, it is found that ECD,DESIGN decreases when the surface length of a spacer increases by installing a shield ring around at the end of a spacer. Therefore, the triple point effects could be reduced and the value of SF could increase. Fig. 11 shows the calculated electrical SF of a spacer made with epoxy resin according to the surface length.

By referring to the data in Table 4 and 5, it is found that the SF of a spacer made with epoxy resin could be more than 2 when the surface length exceeds 40mm and the SF of a spacer made with GFRP could be more than 2 when the surface length exceeds 60mm.

3.3 Mechanical Design Analysis

As shown in Fig. 12, a spacer should mechanically withstand at the pressure exceeds 0.5MPa from an SF$_6$ part and 1.5MPa from a GN$_2$ part. For the mechanical design of a spacer, the side angle and the thickness of a spacer are considered. The SF for a mechanical design could be calculated as (4).^(2,9) A structural analysis is performed by the FEM simulation tool, COMSOL Multiphysics. The main parameters of a solid insulation material are shown in Table 8.

von Mises stress to a spacer is calculated by FEM simulation to estimate the yield stress of a material under complex loading, tensile strength indicates the maximum tensile load of a material can withstand prior to fracture.⁽⁹⁾

As shown in Fig. 13, it is revealed that von Mises stress decreases as the side angle increases. The tendency of von Mises according to the thickness of a spacer is calculated and shown in Fig. 14. It is known that von Mises stress decreases as the thickness of a spacer increases. Inversely proportional to the value of von Mises stress, the value of SF increases as the thickness of a spacer increases. In terms of mechanical properties, the thickness of a spacer for both materials (epoxy resin and GFRP) should be larger than 10mm to get the SF more than 2.