**1. Introduction**

Silicone rubber has been widely used as an electrical insulating material of polymer insulators of power transmission and distribution lines, because of advantages like light weight, easy handling, and water repellency. Electrical and mechanical degradation of silicone rubber is caused by discharges on its surface, exposure to ultraviolet ray, surface contamination, and heavy rain.

With respect to degradation of silicone rubber caused by discharge, arc discharge gives serious and irreversible damage to silicone rubber when compared with partial discharge. To improve resistance to tracking and erosion of silicone rubber caused by arc discharge, alumina trihydrate (ATH) is usually added to silicone rubber [1]. Water of crystallization contained in ATH evaporates by absorbing energy of arc discharge, resulting in good performance. Arc discharge usually occurs on silicone rubber surface under sever condition such as high electric field, heavy contamination and wetting [2]. However, occurrence probability of such conditions may be low.

On the contrary, partial discharge occurs more easily under less sever condition. Partial discharge is a localized breakdown observed in a small portion of an insulation system subjected to high electric field. It does not cause short-circuit of the insulation system. In the case of polymer insulators, partial discharge may be observed at the triple junction point-like interface of silicone rubber/electrode/air and silicone rubber/water droplet/air [3,4]. Partial discharge generated in air attacks the surface of silicone rubber. It is difficult for partial discharge to breakdown silicone rubber by one attack because its energy is not sufficient enough. However, silicone rubber subjected to partial discharge over a long period of time will be eroded gradually from the surface to the bulk and finally breakdown. Knowledge of partial discharge degradation of silicone rubber is considered essential to electrical insulation design of composite insulator. However, researches focused on partial discharge degradation of silicone rubber have not been performed much.

Analysis of contact angle and SEM images of surface after 48 and 96 h partial discharge treatment show that the resistance to partial discharge is improved by adding nano-sized silica to silicone rubber [5]. Effect of partial discharge on rheological and chemical properties of silicone rubber has been discussed in terms of oxidization and localized temperature rise, moisture, and depolymerization [6]. Another paper reports that deposited charge on silicone rubber surface by partial discharge may have an impact on its flashover voltage, but the material degradation caused by partial discharge is limited [7]. To understand partial discharge degradation phenomena of silicone rubber, further fundamental long-term experiments under various conditions in accordance with appropriate procedures seem necessary.

The authors have carried out experiments in rather actual conditions, where silicone rubber was subjected to partial discharge in salt or clean fog [8,9]. To simplify the experimental condition by avoiding the effect of water droplet on silicone rubber, we started experiments in the atmosphere without fog [10]. The present paper reports a series of results obtained in accordance with proposed experimental procedures, which are suitable to study degradation of hydrophobic silicone rubber caused only by partial discharge. The repeatability of results is good and the acceleration rate can be changed in a certain range by adjusting applied voltage and spacing of the gap where partial discharge occurs. The effect of alumina trihydrate (ATH)—a typical additive for mitigating tracking and erosion of silicone rubber—on partial discharge degradation and breakdown is studied by using five kinds of silicone rubber samples including different amount of additives. Results of chemical analysis show that ATH is decomposed by partial discharge as reported in the case of arc discharge. However, it is suggested addition of ATH lowers permittivity and resistivity of silicone rubber, which results in active partial discharge in air gap and shorter lifetime to breakdown.

## **2. Experimental Procedures**

#### *2.1. Sample and Electrode*

Five kinds of silicone rubber samples of the same size (45 × 45 × 2 mm) were used as samples. The matrix of each sample was identical. Respective samples had different amount and surface treatment of additives: ATH and silica. The physical properties of samples [11] are summarized in Table 1.


**Table 1.** Physical properties of samples.

1 Part by weight. 2 Treated stands for rounded surface of particle.

A bundle of four steel round bars of 6 mm diameter was used as high voltage electrode; each bar had a hemispherical tip of 3 mm radius of curvature. The ground electrode was a stainless steel disc of 30 mm in diameter. A silicone rubber sample was placed on the ground electrode. An air gap of 1 or 2 mm spacing was formed between the tip of the round bars (high voltage electrode) and a sample surface. The electrode system is shown in Figure 1. Five electrode systems in total, each sample was placed in one electrode system, were set in an acrylic chamber of 1 m3.

The level of applied ac voltage was determined so that no arc discharge occurred throughout the experiment and surface degradation progressed as fast as possible only by partial discharge. In the present study, 6.6 and 8.5 kVrms (60 Hz) were selected for air gap spacing of 1 and 2 mm, respectively. Voltage was applied simultaneously to five electrode systems for 8 h and then interrupted for 16 h for recovery of hydrophobicity of sample surface. The 24-h cycle was repeated 50 or 100 times, during these cycles degradation performance of five samples is categorized into 3 groups. Applied voltage and current of each sample were monitored and recorded. A schematic diagram of the experimental setup is shown in Figure 2.

In the present study, samples were not tested under the same partial discharge conditions (for example, amount of charge, and number of pulses). The magnitude of applied voltage was fixed instead, considering the actual usage condition of polymer insulators at the site. It result in di fferent partial discharge activities among samples because localized electric field in the air gap depends on permittivity and conductivity of the sample.

**Figure 1.** Electrode system.

**Figure 2.** Schematic diagram of experimental setup.

#### *2.2. Analyses of Surface Erosion*

Contact angle was measured with a Drop Master DM500 (Kyowa Interface Science Co., LTD.) just before voltage application and just after voltage interruption of every cycle. A water droplet of 1 μl (conductivity: ~50 μS/cm) was dropped on a horizontally placed sample. Measurement was performed at five spots on the eroded area of a sample by partial discharge and the average contact angle was calculated.

Surface erosion of a sample was evaluated after 8-h voltage application with a surface roughness meter (TOKYO SEIMITSU CO., LTD., Surfcom 1400D) every three cycles. Measurement was carried out by 3 mm-long linear motion of a sensor on a sample surface.

Confocal microscope (Lasertec Corporation, OPTELICS HYBRID C3) enabled close observation of a sample surface of the area 1.5 mm × 1.5 mm. Images of samples were obtained every ten cycles. The arithmetic average roughness Ra of the surface in the observed area was also available. Ra gives the average of the absolute values of the roughness profile ordinates [12].

Infrared absorption spectrum was obtained every three cycles with a Fourier transformation infrared spectrophotometer (JASCO Corporation, FT/IR 6300).
