Investigation of water droplet‐initiated discharges on laser textured silicone nano‐micro composites using UHF and fluorescent fibre techniques

Science and Engineering Research Board, Grant/ Award Number: CRG/2019/005748/EEC Abstract Laser texturing has been carried out on the surface of the silicone nano‐micro composites to achieve super hydrophobic properties, and water droplet‐initiated Corona discharge studies were carried out. The Corona inception voltage (CIV) exhibits considerable enhancement with increase in the nano filler content under DC voltage compared with AC voltage. The corona inception voltage is high with the textured surface and is found to have direct correlation with contact angle of the composite specimen. The Corona inception voltage was measured using Ultra‐high frequency (UHF) and fluorescent fibre techniques. It is observed that the fluorescent fibre technique is more sensitive in identifying discharges. Frequency domain analysis of UHF signal shows a dominant frequency at 1 GHz and for fluorescent signal, the spectral content is in the range of DC to 10 MHz. The rise time and pulse width of the UHF signal increases with the increase in the nano filler in composite material. The energy content of UHF/fluorescent signal due to discharges shows similar trend with its increase in energy with variation in its magnitude of the signal formed. The pulse width of fluorescent signal formed due to water droplet‐initiated discharges under AC and DC voltage is almost the same, and with the textured specimen it is quite low than the non‐textured material.


| INTRODUCTION
Silicone rubber (SIR) is widely used as an outdoor insulation structure because of its good hydrophobic property with high pollution performance, temperature resistance, and ozone resistance with high electrical resistivity [1]. Under certain conditions, the hydrophobicity of SIR reduces due to surface roughness, increasing drops volume, ambient humidity and so on [2].
In power system network, the water droplet formed due to fog condensation or due to rain on the surface of the outdoor insulation structure may initiate corona discharge/surface discharge to occur leading to surface erosion of material [3]. Imai et al. have clearly indicated that the micro nanoparticles enhance the electrical, thermal and mechanical properties of the polymer composites [4]. For outdoor insulation structures, it is essential to have the characteristics of the insulating material to be hydrophobic, and possess high thermal conductivity with high discharge resistance. Earlier works have indicated that the addition of aluminium tri-hydrate (ATH)/nano alumina to SIR material could enhance the thermal conductivity of the material, thereby enhancing the erosion resistance to discharges [5][6][7]. Nazir et al. have clearly indicated that on addition of ATH to SIR, the nanoparticle fills the gap between the ATH and SIR thus forming a clear bond enhancing the hydrophobic property of the material with high discharge resistance [8]. Venkatesulu et al. have indicated that inclusion of optimum weight% of nano alumina have shown improved thermal conductivity with increase in resistance against corona discharges [9]. Such issues can also be avoided by having polymeric surfaces with super hydrophobic properties [10]. Super hydrophobic properties could be achieved by having low surface energy coatings [11,12], soft moulding [13], hot embossing [14], nano imprint lithography (NIL) [15] or by textured surfaces [16,17]. Jin et al. have indicated that flashover voltage is much higher with superhydrophobic surface than that of conventional SIR [18]. Recently Patil et al. have adopted nanosecond laser texturing and it showed increased hydrophobic properties of SIR [19]. Thus, the real existing problem of making a self-cleaning SIR is solved using the nanosecond laser texturing. Boulanouar et al. have indicated that textured surface shows a better behaviour of the electric field distribution than the traditional non-textured surface [16]. Arshad et al. have observed that the electric field near the water droplet formed gets enhanced and allows water droplet to roll allowing selfcleaning properties [20]. Sarathi et al. have indicated earlier that the water droplet-initiated discharges inject current with its rate of rise in nanosecond thereby radiating electromagnetic waves and also indicated that UHF technique is more sensitive for identifying such discharges [21]. Recently, Mahidhar et al. have studied the corona activity in liquid nitrogen-adopting UHF technique as well as in fluorescent fibre technique and reported that the fluorescent fibre technique is more sensitive technique to identify incipient discharges than the UHF technique [22]. Thus, it is essential to understand the characteristics of UHF signal formed due to water droplet-initiated discharges on super hydrophobic surfaces and to compare it with the signal obtained using fluorescent fibre technique.
Having known all these aspects, the following experimental studies are carried out with the SIR nano alumina micro ATH nanocomposites, by having laser texturing its surface to understand (i) Variation in hydrophobicity of the nanocomposites by contact angle measurement (ii) Variation in corona inception voltage measured adopting UHF and by fluorescence technique (iii) time and frequency domain analysis of UHF and fluorescent signal formed due to water droplet initiated discharges. Figure 1 shows the experimental setup for generating water droplet-initiated corona discharge. The experimental setup includes high voltage source, test electrode, fluorescent optical fibre sensor and UHF sensor for corona discharge identification. The desired AC voltage is generated using HV AC transformer (100 kV, 5 kVA, 50 Hz) and the required DC voltage is generated using voltage doubler circuit. The supply voltage is measured using a high voltage probe (ratio = 1000:1) and the injected current during the discharge process were measured (by using Hall sensor) simultaneously by using 200 MHz Tektronics oscilloscope.

| EXPERIMENTAL STUDIES AND SAMPLE PREPARATION
The Room Temperature Vulcanized Silicone Rubber -RTV8112 (Part-A) and Hardener RTV9858 (Part-B) (obtained from Momentive, USA) were mixed in the ratio of 10:1. The micro size powder (aluminium trihydrate (Al(OH) 3 ), 99.9% purity, with an average size of 5-10 µm, Astrra chemicals, India), and alumina nano powder (of 100 nm average size with 99.9% purity, obtained from Hongwu Nanometre, China) were heated in a temperature controlled oven for 24 h at 150°C to remove the moisture. The required weight percentage of micro and nano powder dispersed in 100 ml of ethanol and the mixture is subjected to mechanical stirring for 30 min, and kept for ultra-sonication in the pulse mode-9 s ON and 9 s OFF for 30 min, which maintains the temperature of the solvent at room temperature during the sonification process . This sonicated mixture is added at first stage to the base SIR (Part-A) and shear mixed for 30 min. Then the mixture is kept in the hot oven maintained at 100°C for 15 min to vapourize the ethanol content in the mixture completely. The Part-B is added to Part-A and shear mixed for 3 min and degasified and then filled into the mould. It is allowed to cure for 24 h under compression moulding with a pressure of 200 kPa. Different percentage weight composition of fillers in SIR composites that were used in the present study were fabricated, and its composition is provided in Table 1.
A nanosecond pulsed Nd: YAG laser with a wavelength of 1064 nm, 10 Hz frequency, 30 mJ of power with a laser diameter of 100 μm was used to fabricate square structures on SIR at 50 mm/sec scan speed. Scanning Electron Microscope (Thermo Scientific Model-Apreo-S with an operating voltage of 3 kV and resolution of 0.7 nm) was used to characterise the structured samples ( Figure 2). Figure 2a shows the square structures at lower magnification having width and spacing of 200 ± 12 µm. The magnified image is taken between the square structures as it can be seen from Figure 2b. The contact angle was measured using sessile water droplet technique using 20 μL volume drop. The surface temperature of the SIR nanocomposites after discharges were measured using Thermal Imager (Fluke Ti125 Infrared Camera). The water droplet movement was captured using Canon Powershot SX 50 HS digital camera in the video mode.
The fluorescence-based corona discharge detection system includes fluorescent optical fibre, a highly sensitive photodetector connected to high bandwidth digital storage oscilloscope (DSO). The green fluorescent fibre (Saint Gobain Crystals BCF 91 A) core is made of polystyrene and cladding with poly methyl methacrylate (PMMA). The fluorescent fibre emission exhibits a maximum at 430 nm, which is captured using a SiPM module (SensL's MicroFC SMA 10050). F I G U R E 1 Experimental setup for water droplet-initiated discharge studies A 10 cm length fluorescent fibre is kept at a distance of 5 cm from the discharge test setup to acquire light emitted during water droplet initiated discharges. One end of the fluorescent fibre is placed against the SiPM active area, while 29 V DC bias voltage is provided. The output voltage of the SiPM was pre-amplified (Agilent 8447D amplifier). Finally, the signal generated by the UHF sensor and the fluorescence sensor were measured by using a high bandwidth digital storage oscilloscope (DSO) with input impedance of 50 Ω. Figure 3 shows variation in the contact angle of textured and non-textured SIR nano-micro composite material. It is observed on texturing, irrespective of weight% filler content, a considerable increase in contact angle is observed when compared with virgin material.

| Contact angle measurement
The cause for the enhanced contact angle is attributed to the increase in surface roughness of the surface. The water droplet rests on the top surface of small pillars forming typical Cassie Baxter wetting state and finds air beneath it thereby enhancing the contact angle [23]. Figure 4 shows variation in water droplet-initiated corona discharge voltage with SIR nano-micro composites, under AC and DC voltages. In the present study, the water droplet caused corona inception voltage to be measured adopting UHF technique and by the fluorescent fibre technique. In the present study, the corona inception voltage by UHF/fluorescent technique is defined based on the first signal generated by the UHF sensor/fluorescent optical fibre generated signal. The results of the study indicate that the fluorescent fibre technique is much more sensitive and is able to identify corona discharge at a much lower inception voltage than the UHF technique. The CIV due to water droplet is high under negative DC voltage followed by positive DC and AC voltage. It is also observed that the water droplet caused corona inception voltage is high with the textured surface than in the smooth surface, under AC and DC voltages. The cause for it could be due to the increase in the creepage path formed due to textured surface reducing the tangential electric field and due to the variation in contact angle of the water droplet. In general, comparing Figures 3 and 4, it is realized that the water droplet-initiated corona inception voltage and contact angle variation have direct correlation. In addition, increase in filler content has showed an increase in corona inception voltage under DC voltage but only marginal increase is observed under AC voltage. The typical voltage and current trace observed at the point of inception and on arcing between the electrodes through water droplet is shown in Figure 5. It is observed that initial corona activity is observed to be occurring near the peak of the AC voltage. During discharge between the electrodes, the discharges appear to occur in the range of AC cycle. In addition to the corona activity, surface discharge activity occurs and its point of inception as well as propagation varies depending on its local condition. Thus, the discharge was observed in the entire phase of the AC voltage. Figure 6 shows variation in water droplet movement with textured and non-textured SIR nano-micro composites under AC and DC voltages. The droplet of 20 µl quantity was maintained constant throughout the experiment. The water droplet in the electrode gap tends to vary with time on application and the characteristic change in shape of the water droplet is different under AC and DC voltages. Nazemi et al. have indicated four modes of water droplet shape variation, varying with supply voltage magnitude and frequency [24].

| Corona discharge inception voltage
Under DC voltage, the water droplet will not vibrate on surface of the insulating material due to applied voltage, but it elongates towards one of the electrodes and suddenly deforms at a certain time [25]. Li et al. have indicated that on application of voltage to the electrode gap with water droplet on hydrophobic surfaces, it gets disrupts once the electrostatic stress exceeds the limiting value of the elasticity maintaining water droplet integrity. With super hydrophobic surfaces, relatively small frictional drag and adhesive drag occurs allowing the movement of the water droplet driven by the electric field possible, and the water droplet rolls in the electrode gap than vibrating or to form elongation of the water droplet [26]. Thus, the position of the water droplet gets  [27]. It is observed (from Figure 6) that the water droplet shape has not changed much with the droplet sitting on top of textured surface, even after corona discharge/arcing inception. In the smooth surface, the water droplet elongates and on quenching of arc, discontinuous water droplets are observed on the surface. Such water droplet left over is not observed with textured surface, under AC/DC voltages. Also, the level of carbonization observed with textured surface is much lower than that of the normal nanocomposite surface. Figure 7 shows variation in the surface temperature on arc quenching between the two electrodes. It is observed that the local temperature is always less with textured surface than nontextured surface, immediately on arc quenching. It is observed that the temperature decays exponentially and the rate of reduction is less with the textured surface. Kumagai et al. have clearly indicated that the addition of ATH can reduce decomposition products formed due to discharges and forms alumina on surface of the insulating material; thus, reduces the rise in temperatures at the surface of SIR because of its endothermic dehydration [28]. Ndoumbe et al. have clearly indicated that the water droplet initiated corona occurs when the local electric field exceeds 6.7 kV/cm [29]. Bruce et al. have carried out tracking test with SIR under DC voltage and have indicated that positive DC voltage has the highest peak and average leakage current causing severe damage [30]. Xie et al. carried out studies with water droplet-initiated discharges and have indicated that at the point of discharge inception, no variation in advancing the contact angle is observed. But on discharge inception, the local temperature of water droplet increases, which reduces the surface tension of the liquid and the water droplet elongates [31]. Nazemi et al. have indicated that peak magnitude of current is high under positive DC and number of discharges is high under negative DC voltage of low [24]. This discharge causes local temperature rise and is transferred to the insulating material causing damage to it. Similar characteristics is observed with SIR nano-micro composites.

| Frequency domain analysis of UHF signal and fluorescent signal
During the discharge process, it is essential to understand the characteristics of the total luminescence formed at the end of the fluorescent fibre (F ), which can be written as where I is the radiant flux caused due to corona/discharge/arc radiated light, T is the transmissivity on the surface of the fluorescent fibre. E is the effective conversion efficiency from the radiant flux caused due to the corona discharge due to water droplet/arc formation between the water droplet and the electrodes formed light refracted into the fluorescent fibre, K is the transmission loss and L is the length of the fluorescent optical fibre [32]. Similarly, Judd et al. have clearly indicated signal excitation of the UHF signals by partial discharges in GIS [33]. Both UHF and fluorescent techniques have high signal to noise ratio. Thus, in the present study, the typical fluorescent signal and the UHF signal formed at the time of water droplet initiated discharges in time domain and its corresponding FFT is shown in Figure 8. The UHF signal formed have dominant frequency near 1 GHz. The frequency response of Fluorescence optical fibre is observed to have energy content in the ranges from DC to 10 MHz. The acquired sensor signals were generally characterized by its rate of rise of signal, peak/peak to peak voltage, energy content and by duration of the signal. Earlier studies have indicated that optical technique of PD detection have direct correlation with the PD signal formed [34][35][36]. Based on UHF signal acquired due to PD activity, correlation between the UHF signal and the PD injected current signal, is not straight forward [33].
The energy content of the signal E has direct correlation with charge as q 2 . It can be written as where E i , E i-1 are the energy content of UHF signal at instant i and previous occurrence, respectively, ΔT = 0.1 ns is the time interval between sample points, V i is the magnitude of the UHF signal at instant i, and R L is the measurement impedance (50 Ω). One hundred signals obtained in sequence mode were analysed. The trend in increase in energy with increase in magnitude of the UHF/fluorescent signal is the same (Figure 9). Figures 10  and 11 show the rise time and the pulse width of the UHF and fluorescent signal, respectively, formed due to water dropletinitiated discharge under AC and DC voltage. Rise time is the time taken by the signal to rise from 10% to 90% of its peak magnitude. Pulse width is the time taken by the signal to fall to 50% of its peak magnitude. The rise time of UHF signal and the fluorescent signal formed due to water droplet initiated discharges is low under AC compared with that formed under DC voltage. It is observed that the rise time of the UHF signal formed is high under negative DC than positive DC. Also, not much variation is observed in the rise time of the UHF signal formed with the textured SIR nanocomposites. It is also observed that increase in weight% of the nano filler in SIR shows increase in rise time and pulse width of the UHF signal, under AC and DC voltage. Similarly, the rise time of fluorescent signal is high under DC voltage compared with AC voltage. In general, the pulse width of fluorescent signal formed under AC and DC voltages is nearly same. In addition, the pulse width of fluorescent signal formed due to water droplet-initiated discharges is lower with textured specimen than non-textured material. It is also observed that increase in the nano filler content in SIR is observed to have increase in the rise time and pulse width of UHF signal, under AC and DC voltages. With fluorescent signal formed due to water droplet-initiated discharges, the rise time and width of the signal appear to reduce with increase in nano filler content especially with AC voltage and not much variation is observed with the fluorescent signal observe under DC voltage. Sarathi et al. carried out studies to identify corona discharge activity in liquid nitrogen and have indicated that the sensitivity to discharges is high with the fluorescence fibre technique compared with the UHF signal [37]. Admittedly, further work needs to be carried out to understand the impact of DC voltage on the generated current pulses due to water droplet initiated discharges on textured/non-textured specimens with the signal measured through fluorescence fibre technique and UHF sensor to correlate the rise in temperature of the specimen due to discharges, generated rise time and pulse width of the fluorescence/UHF signal.

| CONCLUSIONS
The important conclusions accrued based on the present study are the following: � Laser texturing of SIR nano-micro composites provides enhancement in contact angle and is nearly the same with increase in the nano filler content. � The water droplet-assisted corona inception voltage is high with a textured surface than a smooth surface, under AC and DC voltages. It is realized that water droplet-initiated corona inception voltage and contact angle variation have direct correlation.  � The water droplet sitting on the textured surface shows no elongation/variation in shape, even after arc inception, under AC and DC voltages. � Temperature of the SIR surface after arc quenching is high and the temperature decays exponentially. The initial temperature is high with non-textured sample, but the decay rate is less after certain time with textured specimen. � The fluorescent fibre technique is more sensitive than the UHF technique in identifying Corona inception formed due to water droplet under AC and DC voltage. � The UHF signal generated due to discharge has dominant frequency near 1 GHz. The spectral content of the fluorescence ranges from DC to 10 MHz. � It is also observed that increase in weight% of the nano filler in SIR shows increase in rise time and pulse width of the UHF signal, under AC and DC voltage. Similarly, the rise time of the fluorescent signal is high under DC voltage compared with that of the AC voltage. In general, the pulse width of fluorescent signal formed under AC and DC voltages is nearly the same, and with the textured specimen it is lower than the non-textured material.