Pressure-Dependent Breakdown Voltage in SF₆/Epoxy Resin Insulation Systems: Electric Field Enhancement Mechanisms and Interfacial Synergy

Abstract

In SF₆ gas-insulated equipment, solid dielectrics critically degrade insulation performance by reducing the electric field’s ability to withstand gas gaps. To investigate the critical role played by solid dielectric surfaces during the initial phase of gas–solid interface discharge phenomena, this paper experimentally measures the AC breakdown voltage (U_bd) of both dielectric surface-initiated breakdown (DIBD) and electrode surface-initiated breakdown (EIBD) across eight types of post insulator samples. Tests are conducted in 36 mm SF₆ gas gaps under pressures ranging from 0.1 to 0.4 MPa. Combined with electrostatic field simulations, the results reveal that DIBD requires substantially lower U_bd than EIBD under comparable maximum electric field (E_max) conditions. As gas pressure increases, this difference becomes more pronounced. This phenomenon can be explained by three key mechanisms: First, due to the regulatory effect of dielectric materials and shielding electrodes on the electric field distribution, the high-electric-field zone along the gas–solid interface exhibits a longer effective discharge path compared to that in a pure gas gap. This configuration creates more favorable conditions for discharge initiation and subsequent propagation toward the opposite electrode. Second, microscopic irregularities on the dielectric surface induce stronger local electric field enhancement than comparable features on metallic electrodes. Third, in high-electric-field regions adjacent to the dielectric surface, desorption processes significantly enhance electron multiplication during gas discharge, and this enhancement effect becomes more pronounced as gas pressure increases, further lowering the discharge inception threshold. As a result, discharge initiation at dielectric interfaces requires less stringent electric field conditions compared to breakdown in a gas gap, especially at high gas pressure. This conclusion not only accounts for the saturation behavior in the U_bd-p characteristic of SF₆ gas–solid interface discharges but also explains why surface contaminants/defects disproportionately degrade interfacial insulation performance relative to their impact on gas gaps.

1. Introduction

Due to its excellent insulating and arc-quenching properties, SF₆ gas-insulated metal-enclosed switchgears (GISs) and SF₆ gas-insulated transmission lines (GILs) are widely used in modern power systems [1,2,3,4]. Epoxy resin combines key advantages such as high insulation strength, high mechanical strength, ease of molding, high thermal endurance, and resistance to environmental aging. As a result, it has become an irreplaceable insulating and structural material in power equipment [4,5]. Specifically, in GIS/GIL systems, epoxy-based basin insulators and post insulators play crucial roles in providing both electrical insulation and mechanical support. However, the surface discharge along the gas–solid interface also introduces the risk of insulation failure. Surface flashover caused by overvoltage, metal particles, surface charge accumulation, surface cracks, and other foreign matters or defects is the main form of fault that threatens GIS/GIL insulation systems [6,7,8,9,10]. This fault type requires a long maintenance time, causes significant economic losses, and has a negative social impact. Therefore, investigating discharge mechanisms at gas–solid interfaces and optimizing insulator electric field configurations remain critical research priorities in high-voltage insulation engineering.

Previous studies have explored the effects of dielectric and metal surfaces on the initiation and development of gaseous discharge. In 1987, Verhaart et al. [11] highlighted the importance of photoemission effects for surface streamer criterion, which was drawn from their experimental finding that the critical electric field strength for surface streamer discharge at an SF₆-PTFE gas–solid interface is lower than that in a pure SF6 gas gap. In 2003, Jorgenson et al. [12] found that the photoemission process on solid surfaces promotes the development of surface streamers: the number of electrons generated by photoemissions on a solid surface is 6% greater than that captured in N₂ with a pressure lower than 0.1 MPa through simulation and comparative experiments. Chvyreva et al. [13,14] established that under AC conditions, the Raether–Meek criterion for streamer inception at electrodeless gas–solid interfaces exhibits gas-dependent behavior: gases with higher electron affinity gases require greater electron accumulation for avalanche-to-streamer transition. Our previous experimental observations revealed that in 0.06 MPa SF₆, the solid surface demonstrates pronounced adsorption effects on surface-creeping discharges, influenced by both discharge stochasticity and electrostatic forces. These studies indicate that the participation of solid dielectrics promotes initial surface discharge. However, none of the above studies could clarify the differences between the initial discharge on the solid surfaces versus electrode surfaces in GISs/GILs under high-pressure SF₆ gas discharge conditions.

In this paper, epoxy resin insulators with adjustable inner and outer electrodes for electric field regulation were employed to simulate the electrode surface-initiated breakdown (EIBD) and the dielectric surface-initiated breakdown (DIBD) under AC voltage in SF₆. The influence of the maximum system electric field strength and the gas pressure on the surface breakdown voltage were studied and discussed.

2. Test Platform and Samples

2.1. Test Platform

The test platform used in the experiments is depicted in Figure 1. The AC 50 Hz voltage supply was used to power the setup. The voltage was generated by a variable auto-transformer that is adjustable in the range 0–380 V. It was then directed to a high-voltage gas-insulated transformer (with an output voltage of up to 500 kV) and applied to the samples through water resistance. The function of water resistance is to limit the breakdown current and protect the surface of insulators from severe erosion. Referring to the contents of IEC 60296-2020 and IEC 60243-1:2013 standards [15,16], this paper adopts the fast boosting method, which involves raising the power frequency voltage to 50% of the estimated breakdown voltage within 10 s, and then using 2% of the estimated breakdown voltage as the boosting rate to make the insulator surface break down within 1 min. Since the duration of the applied voltage in the experiment did not exceed 1 min, the influence of thermal effects on the discharge process was not considered in this study [5]. The voltage waveform was measured by an oscilloscope (DPO4104, Tektronix China Co., Ltd., Shanghai, China) with a bandwidth of 1 GHz and a sampling rate of 5 GHz, through measuring winding with a voltage ratio of 1000:1. The amplitude of the applied voltage is expressed here as a root-mean-square (RMS).

Following breakdown testing, all insulator samples underwent surface preparation which consisted of mechanical polishing with 2000-grit sandpaper (3M™) and ultrasonic cleaning in absolute ethanol. The surface roughness of the polished samples, as measured by the 3D optical profilometer (Up S-Dual 3D, RTEC Instruments, San Jose, CA, USA), was determined to be R_a < 0.5 μm and R_z < 3 μm, with occasional peak values (R_t) reaching up to 20 μm. To prevent the influence of discharge ablation traces on surface flashover voltage, the insulator sample was replaced after each breakdown. The used samples were polished with sandpaper before reinstatement for reuse. Insulator samples with severe surface ablation or carbonization that could affect subsequent measurements were discarded. The breakdown voltage (U_bd) was defined as the average of six valid measurements.

The test chamber was evacuated to a vacuum (gas pressure < 60 Pa), then filled with commercial SF₆ (purity > 99.997%) at 0.1 MPa. The gas pressure was monitored with a digital pressure sensor (Panasonic^® DP-102, Osaka, Japan). All tests were conducted in a climate-controlled laboratory at 25 ± 2 °C and 40–60% RH.

2.2. Samples

The insulator samples were made of epoxy resin doped with Al₂O₃ fillers, and were manufactured using the vacuum casting process. Three different surface structures (concave, wave, and cylinder) as well as two shielding electrode structures (rod–plate outer shielding electrodes with small inner shielding electrodes and plate–plate outer shielding electrodes with big inner shielding electrodes) were employed to regulate the electric field distribution. The height of the epoxy resin body was 30 mm, and the height of the embedded electrode’s exposed outer section was 3 mm; that is, the height of the gas gap was 36 mm, and the arc length along the insulator surface was 44 mm. Figure 2 presents the structure of the insulator samples.

Taking the gas gap composed of two types of external shielding electrodes into consideration, the numbering convention and electric field distribution of eight samples are shown in Figure 3. The electric field distribution of samples was calculated by the finite element method. Metal electrodes were modeled as equipotential conductor boundaries, and the relative permittivity of SF₆ gas and epoxy resin was set as 1.002 and 5.0, respectively. The applied voltage was set to 1000 V. A refined triangular free-mesh division achieved convergent computational results at 14,825 elements. Due to the simplicity of the geometric structure, the default solver settings in the software were sufficient to meet the computational requirements, and no specific additional configurations were necessary. The electric field distribution of the cross-section and gas–solid interface of the insulator sample is shown in Figure 3. Obviously, the E_max of samples (a) to (e) is located at the electrode surface while that of sample (f) to (h) is located at the insulator surface.

Previous research has established that the breakdown process in compressed SF₆ follows a sequential progression: (1) initial electron generation triggering streamer inception and corona formation, (2) pre-breakdown partial discharge (PD) development, and (3) stepped leader propagation across the gap, ultimately culminating in arc breakdown [17,18,19,20,21,22,23]. As is well known, although SF₆ gas theoretically has excellent insulation strength, its insulation strength can drop sharply under extremely non-uniform electric field structures [22,23]. The electric field inhomogeneity sensitivity is an intrinsic characteristic of SF₆ gas insulation systems. In uniform electric fields, discharge propagation to the opposite electrode occurs immediately upon initiation. Conversely, under non-uniform electric field conditions, discharge propagation requires significantly higher applied voltages to satisfy the critical propagation electric field strength threshold. GIS/GIL insulation gas gaps are characterized by a weakly non-uniform electric field distribution, in which the ratio between the maximum and the minimum electric field in the gap is in the range between 2 and 3 [22]. In this paper, the uniformity factor of the electric field (f) of samples is calculated by Equation (1), and the results are shown in

Table 1. Sample ID with the corresponding uniformity factor of the surface electric field distribution.

Shielding Electrodes	Rod–Plate				Plate–Plate
Surface profile	Gas gap	Concave	Wave	Cylinder	Gas gap	Concave	Wave	Cylinder
Sample ID	(a)	(b)	(c)	(d)	(e)	(f)	(g)	(h)
Location of Emax	Electrode surface					Dielectric surface
f	5.29	3.08	3.18	3.52	1.61	2.82	2.60	1.71

. Clearly, the electric field distribution in samples with plate–plate shielding is more non-uniform than in samples with rod–plate shielding.

$$f={\displaystyle \frac{E_{\max }}{E_{\mathrm{av}}}}={\displaystyle \frac{{hE}_{\max }}{U}}$$

(1)

where f refers to the uniformity factor of the electric field distribution, 1; E_max refers to the maximum electric field strength, kV/mm; E_av refers to the average electric field strength, kV/mm; h refers to the height of the gas gap, mm; and U refers to the applied voltage, kV.

In our previous experiments [24], we employed a photomultiplier tube (PMT) and a high-speed framing camera to observe the spatiotemporal optoelectronic distribution of the discharge process in sample a (f = 5.29) and sample b (f = 2.82) under lightning impulse voltage. Analysis of the discharge images indicates that the discharge inception consistently occurred at the location of the maximum electric field strength. Furthermore, based on the time delay between gap breakdown and the initial detection of the discharge signal, we found that for sample b (f = 2.82), breakdown occurred within 100 ns after discharge inception. This suggests that in slightly non-uniform electric field configurations, the breakdown voltage can be considered representative of the discharge inception voltage. The captured surface flashover images confirm that discharge initiation consistently occurred at the location of the simulated maximum electric field (E_max). This observation strongly supports the reliability of the uniformity factor (f) used in this study.

3. Results

3.1. Breakdown Voltage

Figure 4 illustrates the dependence of AC breakdown voltage on gas pressure for the samples, covering a range from 0.1 to 0.4 MPa. In the gas gap with rod–plate shielding, after the introduction of post insulators, E_max is still at the electrode surface and its value decreases, as shown in Figure 3a–d, so the breakdown voltage (U_bd) of the insulator samples is higher than that of the pure gas gap. However, in the gas gap with plate electrodes, E_max shifts from the electrode surface to the insulator surface and its value increases, as shown in Figure 3e–h, so U_bd decreases after the introduction of post insulators.

As shown in Figure 5a, in the rod–plate structure, the breakdown voltage increment (ΔU_bd) decreases sharply with increasing gas pressure, while ΔU_bd only decreases slightly for the case shown in Figure 5b. The most significant distinction between these two sample sets is their electric field non-uniformity: the former exhibits strong electric field non-uniformity (f > 3), while the latter demonstrates moderate non-uniformity (f < 3).

3.2. Breakdown Electric Field Strength

Theoretically, the discharge inception condition of SF₆ gas (E_cr/p = 88.5 kV·mm⁻¹·MPa⁻¹) is constant [25,26], which means E_cr should increase linearly with gas pressure. Although discharge initiation distorts the background electric field distribution, the AC breakdown voltage in a quasi-uniform electric field gap remains primarily determined by E_max in the gas volume [22]. The maximum value of the electric field strength at breakdown voltage (E_bdmax) of all samples was calculated by Equation (2), and the results are presented in Figure 6.

$$E_{\mathrm{bdmax}}=U_{\mathrm{b}\mathrm{d}}{E}_{\max }$$

(2)

where E_bdmax refers to the maximum electric field strength along the insulator surface at breakdown voltage, kV/mm; E_max refers to the maximum electric field strength under 1 000 V, kV/mm; and U_bd refers to the breakdown voltage, kV.

As shown in Figure 6, the E_bdmax of samples under rod–plate shielding is lower than the theoretical value when the gas pressure is higher than 0.3 MPa, while the E_bdmax of samples under plate–plate shielding is lower than the theoretical value when the gas pressure is higher than 0.1 MPa. These results may be attributed to two factors: First, since electron swarm studies are typically conducted under low-pressure conditions with small electrode gaps (where the pd product is minimal), the established criteria may not be applicable to high-pressure, large-gap configurations. Second, although the samples were polished with sandpaper, protrusions may still remain on the electrode and insulator surface. The breakdown voltage of SF₆ gas gaps exhibits strong sensitivity to the electric field inhomogeneity, particularly at elevated gas pressures [22,27]. Consequently, the measured breakdown voltages are consistently lower than theoretical streamer inception values under such conditions.

3.3. Relationship Between Breakdown Voltage and Maximum Surface Electric Field Strength

Theoretical analysis predicts a significant inverse correlation between U_bd and E_max in SF₆ gas gaps under identical geometric conditions. However, the descending trend of the U_bd-E_max curve in this study exhibits a distinct hump; U_bd even increases rather than decreasing as E_max increases from 0.0718 kV/mm to 0.1015 kV/mm, as shown in Figure 7.

The samples were divided into two groups according to the discharge initiation position, and Figure 7 was redrawn to obtain the two U_bd-E_max curves shown in Figure 8. It can be seen that the descending trend of the U_bd-E_max curve for the DIBD group is faster than that for the EIBD group.

Although the mechanisms of gas discharge inception differ under AC and impulse voltages, experimental results obtained under impulse conditions can still serve as a valuable reference for trends observed in AC voltage scenarios. Here, we utilized our previously acquired dataset comprising breakdown voltages (U_bd) under standard lightning impulse and the electric field distribution profiles for 48 distinct post-type insulator configurations. These measurements were obtained in a 36 mm SF₆ gas gap at 0.2 MPa, for which the data points were more densely distributed across the varying range of E_max [28]. The breakdown voltage was also divided into two groups: EIBD and DIBD, and the separate phenomenon is more obvious in the U_bd-E_max curves, as shown in Figure 9. This implies that in experiments with large sample sizes and under standard lightning impulse, the trend of U_bd decreasing with the increase in E_max in DIBD is more significant.

4. Discussions

4.1. Influence of Solid Dielectric Surface on Electric Field Distributionn

As illustrated in Figure 3, E-distribution along the dielectric surface exhibits significantly greater complexity compared to that in a gas gap. In a gas medium, the electric field lines originate perpendicularly from the electrode surface and exhibit a monotonic decay with increasing distance from the electrode. Conversely, E-distribution along the surface of a solid dielectric is governed by both the material properties and the geometry of shielding electrodes, resulting in a composite electric field comprising a tangential component parallel to the solid dielectric surface (E_t) and a normal component perpendicular to the solid dielectric surface (E_n). This dual-component characteristic leads to non-monotonic variations in the surface electric field strength between electrodes.

The influence of surface micro-protrusions on the electric field distortion along dielectric interfaces cannot be neglected. To systematically evaluate the influence of microscopic surface protrusions on E-distributions, identical rounded-triangular defects (height: 0.10 mm, base width: 0.20 mm) at the E_max locations of both electrode and dielectric surfaces were introduced to samples e and h, respectively. The distorted E-distribution caused by such a micro-protrusion under the breakdown voltage is shown in Figure 10.

Obviously, the enhanced E*_max at the electrode surface is increased by 125%, which is much higher than the 36.9% recorded along the solid dielectric surface. However, E-distribution in a 3 mm long strip-shaped region near the location of E_max along the solid dielectric surface significantly increased from 70%E_cr to 80%E_cr, while only the electric field strength in a circular region with a radius of about 0.2 mm near the protrusion was significantly enhanced. This is also one of the reasons why E_max for dielectric surface-initiated discharge is significantly lower than the theoretical value. The electric field simulations are conducted under idealized assumptions, whereas in practice, both electrode and solid dielectric surfaces exhibit dimensional deviations from design specifications and surface roughness. Consequently, for samples with weakly non-uniform electric field configurations (f < 3), the experimentally determined E_bdmax systematically underestimates theoretical predictions derived from E-distribution. The numerical values obtained from simulation calculations cannot exactly match the actual E-distribution in the samples during breakdown experiments, but they can still accurately characterize the overall E-distribution.

4.2. Influence of the Solid Surface on the Critical Conditions for Gaseous Discharge

According to Raethe–Meek’s criterion [29,30,31], the transition from electron avalanche discharge to self-sustained streamer discharge occurs when the following quantitative conditions are met:

$${\int }_0^l{\alpha }_{\mathrm{e}\mathrm{f}\mathrm{f}}(E,p)\mathrm{d}l={10}^8$$

(3)

where α_eff is the effective ionization coefficient. Specifically, α_eff = α − η if α > η, and it is zero otherwise. α is the ionization coefficient, 1; η is the attachment coefficient, 1; E refers to the applied electric field strength, kV/mm; p refers to the gas pressure, MPa; and l is the distance that the avalanche can propagate in mm. The integration is normally performed along the electric field line.

The stochastic initiation of gas discharges stems from its nonlinear relationship to the local electric field configuration, where both the spatial extent and geometry of the electric field enhanced regions exhibit a greater determinative influence than E_max alone.

The charged particles produced by the discharge will further distort the background E-distribution. After the electron avalanche initiation, the modified spatial E-distribution can be characterized by Equation (3). Streamer propagation becomes sustainable when two critical conditions are met: (1) the electron density at the electron avalanche head exceeds ~10⁸, and (2) the avalanche head radius reaches approximately 30 μm [32]. Under these conditions, the space charge-induced electric field becomes sufficient to significantly modify the background E-distribution. Remarkably, streamer advancement can persist even when the enhanced local E_max remains slightly below the conventional breakdown threshold (E_a > 75 kV·cm⁻¹·MPa⁻¹). This clearly demonstrates the importance of space charge effects in discharge development [32].

$${E=E}_{\mathrm{a}}+E_{\mathrm{s}\mathrm{p}}=E_{\mathrm{a}}+{\displaystyle \frac{N_{\mathrm{s}}\mathrm{e}}{{4\mathrm{\pi }{\epsilon }_0(z_0+R_{\mathrm{s}})}^2}}$$

(4)

where E_a is the applied electric field strength, kV/mm; E_sp is the electric field strength produced by space charge, kV/mm; N_s is the number of charges at the head of the electron avalanche, 1; R_s is the radius of the head of the electron avalanche, μm; e is the charge of a single electron, C; and z₀ is the axial distance from the start of the secondary electron avalanche, mm.

As depicted in Figure 11, the discharge processes in a pure gas gap are primarily driven by the ionization and attachment of SF₆ gas molecules in the active region, which are highly dependent on the local electric field strength and gas pressure [31,32,33].

The traditional gas–solid interface discharge models, as reviewed in [34], generally employ a two-dimensional axisymmetric model with parallel plate electrodes and a cylindrical smooth solid dielectric. Thus, the direction of the applied electric field is parallel to the gas–solid interface.

When E_max occurs at the solid dielectric surface, the role of E_n in the discharge process becomes critical, as illustrated in Figure 12. First, the solid dielectric surface acts as a potential barrier, restricting the motion of free electrons and reducing their probability of escaping from the active discharge region. This confinement distorts the local E-distribution, enhancing ionization probability. Second, high-energy photons and electrons bombard the dielectric surface, liberating additional free electrons at the gas–solid interface [11,12,13,14]. Under the influence of E_n (perpendicular to the interface), these electrons are driven toward the surface, where they become trapped or bound, leading to surface charge accumulation. Extensive experimental evidence confirms that surface charge accumulation significantly alters the surface flashover voltage and discharge morphology [35,36,37]. Thus, E_n amplifies the solid surface’s positive feedback effect on discharge propagation.

The process of generating and trapping free electrons on the solid surface can be simplified to the surface ionization coefficient (α_d) and the surface attachment coefficient (η_d), which are mainly determined by the local electric field strength. Therefore, the effective ionization coefficient α_eff in the Raether–Meek criterion for solid dielectric surface-initiated discharge should be modified as follows:

$${{\alpha }_{\mathrm{e}\mathrm{f}\mathrm{f}}(E,p)=\alpha }_{\mathrm{g}\_\mathrm{e}\mathrm{f}\mathrm{f}}(E,p)+{\alpha }_{\mathrm{s}\_\mathrm{e}\mathrm{f}\mathrm{f}}={(\alpha }_{\mathrm{g}}(E,p)-{\mathrm{\eta }}_{\mathrm{g}}(E,p))+{(\mathrm{\alpha }}_{\mathrm{s}}(E)-{\mathrm{\eta }}_{\mathrm{s}}(E\left)\right)$$

(5)

where α_eff, α_{g_eff}, and α_{s_eff} represent the effective ionization coefficient of the gas–solid interface, gas gap, and solid dielectrics, 1; α_g and α_s represent the ionization coefficient of SF₆ gas gap and solid dielectrics, 1; and η_g and η_s represent the attachment coefficient of the SF₆ gas gap and solid dielectrics, 1. E refers to the applied electric field strength, kV/mm; p refers to the gas pressure, MPa.

α_{g_eff} varies with gas pressure, while α_{s_eff} is only related to the intrinsic parameters of the solid dielectric material. As the gas pressure increases, the magnitude of E_cr required for the initiation of gas–solid interface discharge also rises. As a result, the role of electrons emitted from the solid dielectric surface in the discharge process becomes more significant, potentially generating seed electrons that trigger the initiation of discharge.

In conclusion, both E_n and the gas pressure increment contribute to enhancing the effective ionization process on the gas–solid interface. E_n promotes the ionization of electrons on the solid surface. As gas pressure increases, the applied electric field strength required for discharge initiation rises correspondingly. This pressure-dependent enhancement effect is particularly pronounced for E_n. Combined with factors that distort the electric field, such as defects on the solid surface, the critical condition for gas discharge initiation at the solid dielectric surface is generally lower than that at the electrode surface.

It must be emphasized that the above analysis is qualitatively derived from macroscopic experimental results in conjunction with fundamental discharge theories, and is intended solely as a reference for researchers engaged in the structural design and experimental investigation of related insulation components. More accurate theoretical studies should be based on controlled-variable investigations of surface conditions and precise observations of optoelectronic signals during the discharge inception phase.

5. Conclusions

Although it is widely recognized that introducing a solid dielectric weakens gaseous insulation, the critical conditions governing surface discharge inception—often the immediate trigger for dielectric failure—remain largely unexplored. This paper experimentally measured the AC breakdown voltage of surface-initiated discharge and electrode-initiated discharge in SF₆ gas at pressures ranging from 0.1 MPa to 0.4 MPa. Combined with electrostatic field simulations, the study analyzed the influence of the solid dielectric on the initiation and development of discharge. The conclusions are as follows:

(1)

Under AC voltage conditions, the critical electric field for solid dielectric surface-initiated breakdown in SF₆ gas is significantly lower than that of metal electrode surface-initiated breakdown under identical electric field distribution. Although the discharge mechanisms differ, the same disparity phenomenon is also observed under standard lightning impulse voltage.

(2)

The discharge process on dielectric surfaces demonstrates higher sensitivity to microscopic irregularities. According to the electrostatic field simulation results, while micro-protrusions significantly amplify the maximum electric field strength on the electrode surface, their influence on the electric field distribution in the adjacent gas region remains limited. In contrast, similar protrusions on a solid dielectric surface cause more extensive distortion of the electric field along the direction of the electric field lines.

(3)

From the perspective of discharge mechanisms, a qualitative analysis indicates that as gas pressure rises, the critical electric field strength required to initiate discharge at the gas–solid interface increases, and more free electrons are released from the solid dielectric surface. This is further promoted by the normal component of the electric field, as well as local electric field distortions due to surface defects, resulting in significantly lower discharge initiation thresholds compared to a pure gas gap.

(4)

These conclusions underscore the qualitative role of localized electric field distortion and material surface properties in discharge behavior. Further research employing partial discharge imaging and surface functionalization techniques is essential to quantitatively elucidate the proposed mechanisms and generalize these findings.