Field failures of composite insulators in Sweden have been linked to insufficient adhesion between silicone rubber housings and the fiberglass core and, in some cases, to high electric fields at the high-voltage end fittings, leading to partial discharges at the interface.

The post Experience at Swedish DSOs with Composite Line Insulators (Video) appeared first on .

This edited contribution to INMR by S. Gobi Kannan of TNB’s Grid Div., in cooperation with K. Hafizzudin & other experts at TNB Labs, presents findings from a forensic analysis conducted on these bushings. The conclusions at the end should be noted carefully by both manufacturers and users of high voltage components produced using liquid silicone rubber insulation.

Nanjing Electric Insulator Co.,Ltd

CHINA

Hivolt Power System

China

See more suppliers of Bushings

Site Inspection

An inspection was carried out in the area where the deteriorated bushings were stored. This assessment revealed visible surface degradation, including extension cracking and discoloration, particularly in areas exposed to sun and moisture.

The cracks appeared have formed due to prolonged exposure to ultraviolet radiation, temperature fluctuations, and humidity. Signs of surface contamination were also observed – an indication of potential interaction with airborne pollutants. These observations suggested that the material had undergone chemical and physical changes over time, potentially affecting structural integrity and performance of the bushings.

Visual Examination

The polymeric bushings were visually examined, which revealed generally poor condition. There were signs of heavy contamination with dirt and dark deposits, suggesting prolonged exposure to environmental contaminations or operational wear. In addition, the edges of the silicone rubber insulator had chipped, which could compromise insulation properties and mechanical integrity of the bushing. The outer surface exhibited significant deterioration, categorized as ‘alligatoring’ cracking where the material forms a rough surface (see Figs. 3,4 & 5).

By contrast, a new bushing inside the original crate was selected to serve as reference for the original condition. A close-up examination was conducted on a new silicone rubber bushing stored in a wooden crate under a roofed storage area (see Fig. 7). The insulator surface appeared clean and shiny, with no evidence of cracking or deterioration. Proper storage conditions included protection from direct sunlight, moisture, and airborne contaminants – all of which contributed to preserving the bushing’s physical integrity.

Visual inspection was also carried out on a reference bushing installed at the substation and coming from the same manufacturer. The silicone rubber insulator in this case showed a mildly contaminated surface, likely due to exposure to the outdoor environment. Despite this, the material remained intact, with no signs of cracking or structural damage. Its condition suggested that, while some surface dirt accumulation had occurred, the bushing retained its mechanical and insulation properties.

Material & Elemental Analysis

To evaluate the extent and nature of deterioration in the polymeric bushing material, a comprehensive series of laboratory-based material analyses were conducted. These tests were intended to identify physical, chemical, thermal, and structural changes to the silicone rubber insulator with focus on correlating environmental exposure to observed failure modes. The scope of the material analysis covered both qualitative and quantitative methods to ensure a thorough assessment of degradation mechanisms. The following techniques were employed:

Together, these different analyses provided a detailed understanding of chemical and mechanical degradation processes that may have affected the bushing’s housing material.

Fourier-Transform Infrared Spectroscopy (FTIR) Analysis

FTIR analysis was carried out on the silicone rubber insulator to identify any sign of degradation. Two specimens, labelled as ‘reference’ and ‘deteriorated’ samples were selected from the silicone rubber insulator of the substation and warehouse, respectively. The analysis was conducted in accordance with ASTM E1252:1998 (2021) – Standard Practice for General Techniques for Obtaining Infrared Spectra for Qualitative Analysis.

For the Part 1 Analysis, the top surface of both Sample A and Sample B was placed directly onto a Golden Gate Diamond Attenuated Total Reflectance (ATR) accessory and scanned in reflectance mode for 16 times from 4000 cm-1 to 600cm-1 using FTIR spectrometer to obtain their individual infra-red spectrum. For the Part 2 Analysis, at 0 hour and after each exposure time during accelerated weathering test (UV exposure test), the exposed side of each sample was placed directly onto the ATR accessory and scanned using the test parameters described above.
Table 1 shows results for Part 1. Except for the functional groups highlighted in bold, the test results indicate characteristics absorption peaks of a silicone-based material. Results also infer that the deteriorated sample had undergone stages of degradation due to presence of OH, C=O and C=C functional groups.

Differential Scanning Calorimetry (DSC) Analysis

Differential Scanning Calorimetry (DSC) Analysis was carried out on the silicone rubber insulator. DSC is a method used to determine oxidation onset temperature (OOT) at which a material begins to oxidize in the presence of air or oxygen. Oxidation is a chemical reaction in which a substance combines with oxygen, leading to formation of oxides.

Two specimens, labelled as ‘reference’ and ‘deteriorated’ samples were selected from the silicone rubber insulator of bushings at the substation and warehouse, respectively. Oxidation onset temperature (OOT) was determined using a DSC analyzer according to ASTM E2009:2023 – Standard Test Methods for Oxidation Onset Temperature of Hydrocarbons by Differential Scanning Calorimetry. Approximately 10 to 20 mg of test specimen taken from the deteriorated and reference samples were heated from 30°C to 350°C at intervals of 10°C/minute in the presence of oxygen.
For this test, the oxygen gas flow rate was maintained at 50 ml/minute throughout the testing period. Results are shown in Table 2.

TGA Analysis

Thermogravimetric (TGA) Analysis was also carried out on the silicone rubber insulator. This test determines changes in the mass of a sample as a function of temperature under a controlled atmosphere. Two specimens, labelled as ‘reference’ and ‘deteriorated’ samples were selected from the silicone rubber insulator of the bushings at the substation and warehouse, respectively.

Approximately 10 to 13 mg of test specimens was analyzed in accordance with ASTM E1131:2020 – Standard Test Method for Compositional Analysis by Thermogravimetry.

Sample A and Sample B were analyzed using the following test parameters in a TGA analyzer equipped with an auto sampler:

• Temperature Program:
i. Heat from 25°C to 600°C at 10°C/minute in nitrogen
ii. Heat from 600°C to 800°C at 10°C/minute in nitrogen
• Gas flow rate: 50 ml per minute
Results are shown in Table 3.

From Table 3, it is evident that although the composition of both samples is similar, i.e. about 22 wt.% total organic content and 78 wt.% ash/inorganic filler content, onset of decomposition (ignition) temperature of the deteriorated sample is higher than for the reference sample. Factors that contributed to the higher onset of decomposition (ignition) temperature value in the deteriorated sample are most probably the same as in the DSC-OOT results.

In summary, while specific mechanisms may vary between oxidation and decomposition, factors influencing onset temperature for both processes in degraded samples often overlap. Factors such as chemical changes, thermal history, mechanical stress, and environmental exposure all play an important role in determining the thermal stability and decomposition behavior of polymers such as silicone rubber.

Water Immersion Test

A Water Immersion Test was carried out on the silicone rubber insulator to determine the amount of water that the material can absorb when exposed to moisture or liquid and to assess the material’s porosity, durability and suitability. Two specimens, labelled as ‘reference’ and ‘deteriorated’ samples were selected from the silicone rubber insulator of the substation and warehouse bushings, respectively, and a water immersion test was conducted according to ASTM D570:2022 – Standard Test Method for Water Absorption of Plastics.

Three pieces of test specimens, each cut from reference sample and deteriorated sample, were immersed in distilled water at room temperature for 24h. After this period, the rate of water absorption was calculated based on percentage change in mass. Results are shown in Table 4 and indicate that, despite the higher OOT and onset of decomposition (ignition) temperature values, the average water absorption value of deteriorated sample is significantly higher than that of reference sample. This is probably due to material degradation in the deteriorated sample.

Accelerated Weathering Test

An accelerated weathering test was carried out on both samples to simulate ultraviolet exposure to the silicone bushing over time. This test was to identify any sign of degradation on the silicone bushing when exposed to UV rays for a period of 7 years (from the year of manufacture until the year the bushing was discovered to have deteriorated).

The analysis was conducted in accordance with Cycle 1 of ASTM G155:2021 – Standard Practice for Operating Xenon Arc Lamp Apparatus for Exposure of Materials.

The pump assembly with the cover fully closed was placed inside a xenon chamber and subjected to the following test parameters:
i. Light source: Xenon Arc
ii. UV irradiance: 0.35 W/m2 at 340 nm
iii. Black panel temperature: 63°C
iv. Relative humidity: 50%
v. Filter: Daylight BB
vi. Exposure duration: 1400 hours (to simulate 7 years of usage. Each 200 exposure hours represent to 1 year of usage)

In this analysis, the presence of OH, C=C, and C=O groups is taken as indicators of material degradation. The results are shown in Table 4.5.
Table 5 illustrates that reference sample exhibited greater resistance to the accelerated weathering test compared to the deteriorated sample. For the deteriorated sample, initiation of degradation was detected after 1200 hours of exposure.

Energy Dispersive X-Ray (EDX) Analysis

Energy Dispersive X-Ray (EDX) Analysis was carried out on the silicone rubber insulator. This test was to identify any foreign material which might contribute to deterioration of the silicone rubber insulator.

Two specimens, labelled as ‘reference’ and ‘deteriorated’ samples were selected from the silicone rubber insulator of the substation and warehouse, respectively. Results are shown in Table 6.

Scanning Electron Microscopy (SEM) Examination
Scanning Electron Microscopy (SEM) examination was carried out on the cross-section of the silicone rubber insulator to examine the cross-sectional characteristic of the deteriorated sample.

Figs. 27 & 28 show the SEM images.

EDP Labelec

Portugal

PSW – Siemens Energy Testing Laboratories Berlin

Germany

See more Laboratories

Nuclear Magnetic Resonance (NMR) Spectroscopy Analysis

Nuclear magnetic resonance (NMR) spectroscopy was carried out on the silicone rubber insulator to identify any UV sensitive components, including:

i. Ethyl vinyl acetate (EVA)
ii. Ethyl propylene rubber (EPR)
iii. Ethylene propylene diene monomer (EPDM)

Two specimens, labelled as ‘reference’ and ‘deteriorated’ samples were selected from the silicone rubber insulator of the substation and warehouse, respectively.

Each sample (10 g) was soaked in a flask containing 50 mL of dichloromethane for 90 min. The solvent was evaporated under reduced pressure by a rotary evaporator and the residue (liquid at room temperature) was dissolved in 0.5 mL of deuterated chloroform (CDCl3) for NMR analysis.
The NMR analysis was performed on an NMR Avance NEO 300 MHz spectrometer (Bruker, Germany) at 25 °C. Chemical shift of the proton NMR spectrum was calibrated according to chloroform residual peak at 7.26 ppm.

A strong signal around 0.07 ppm was observed in the proton NMR of the good sample, which is attributed to Si-Me. The spectrum of deteriorated sample showed a distinct sharp signal around 4.9 ppm, which cannot be attributed to any of the proposed UV sensitive components (EVA, EPR and EPDM).

Analysis & Discussion

Visual examination of the deteriorated sample from the warehouse showed ‘alligatoring’ cracking appearance on the silicone rubber insulator when pinched, while the reference sample from the substation did not show this when the same action was applied.

FTIR analysis on as-received samples showed that the deteriorated sample had experienced degradation whereas the reference sample had not.
The oxidation onset temperature through Differential Scanning Calorimetry (DSC) Analysis showed that the decomposition temperature for the deteriorated sample was higher (320°C) compared to the reference sample (309°C).

As in the TGA analysis, the decomposition temperature for deteriorated sample was higher (400°C) compared to the reference sample (388°C).
It can also be seen that the OOT value from analysis of the deteriorated sample is higher than for the reference sample. Several factors could contribute to the degraded sample having a higher oxidation onset temperature, as shown in Table 7. The higher oxidation onset temperature of the deteriorated sample compared to the reference sample could be attributed to a combination of these factors, reflecting the complex interplay between material properties, environmental conditions, and degradation mechanisms.

Overall, as silicone rubber degrades, it can undergo cross-linking or chain scission processes, leading to formation of new chemical structures. These changes often result in a material with a higher degradation temperature, whereas new structures are more stable and require higher temperatures to break down.

The water immersion test showed that the deteriorated silicone rubber sample absorbed more water than the reference sample, meaning that its hydrophobicity had been reduced.

Accelerated weathering test (UV exposure test) revealed that both reference and deteriorated sample do not show any sign of degradation up until 1000h. However, evidence of material degradation on the deteriorated sample started to show up at 1200h and on. No sign of material degradation was detected on the reference sample after 1400h.

EDX analysis on deteriorated samples show that contamination on the surface is composed mainly of Si (silicon), Al (aluminium), Ti (titanium) and Mg (magnesium). Except for Ti, these elements are common elements for soil dirt. No corrosive elements were found on the insulator. The source of the Ti was not known.

SEM images on the deteriorated sample cross-section close to outer surface showed that the degraded layer had an average thickness of 114-130 µm.

Nuclear magnetic resonance (NMR) spectroscopy analysis showed no signs of UV sensitive components in the silicone rubber, including Ethyl vinyl acetate (EVA), Ethyl propylene rubber (EPR) and Ethylene propylene diene monomer (EPDM).

Conclusions

Comparison between the reference sample (at the substation) and the deteriorated sample (in the warehouse) is summarized below:

Despite the same year of manufacture for both sets of bushings, laboratory examination clearly showed that the deteriorated bushing from the warehouse had undergone degradation. By contrast, the reference sample (from the substation) remained in good condition.
Reduced hydrophobicity of the deteriorated silicone bushing can be a contributing factor to the degradation. Degradation could not be attributed to presence of UV sensitive components since these were not detected on the silicone material. Degradation caused by corrosive deposits was not possible since none were found.

The variation in oxidation onset temperature (OOT) between two batches of liquid silicone rubber (LSR) observed through thermogravimetric analysis (TGA) provides indication of inconsistencies in the manufacturing process, which may have contributed to the field-observed degradation in service. Specifically, presence of alligator cracking—a pattern of surface fissures resembling reptile skin—is commonly associated with premature ageing, embrittlement, or loss of elasticity in silicone materials.

Such cracking is often a manifestation of poor polymer chain integrity or loss of filler-matrix bonding, both of which are consistent with a lower OOT and reduced thermal stability. In LSR systems, rapid curing cycles, coupled with low-viscosity formulations, make the process highly sensitive to variations in temperature, catalyst efficiency, and filler dispersion. If these parameters are not tightly controlled, the result could be an incomplete cross-link network, chain scission, or uneven additive distribution—all of which can compromise the long-term resistance of the material to UV, corona discharges, or thermal cycling.

The alligator cracking observed could thus be a surface expression of internal chemical degradation triggered by inadequate vulcanization or unstable filler bonding. These defects may not be visually evident at the time of manufacture but can propagate over time under environmental and electrical stresses, leading to irreversible surface damage and eventual failure.

In addition, ageing test results demonstrated that significant molecular degradation, as detected by FTIR, only became evident after 1400h of UV exposure, highlighting that standard ageing durations may not be sufficient to reveal true long-term behavior. Therefore, OEMs should consider extending the UV ageing test duration to a minimum of 1400 hours, or beyond, to fully capture the onset of degradation mechanisms relevant to field performance.

Therefore, the combination of lower oxidation onset temperature and observed alligator cracking reinforces the conclusion that LSR materials are more vulnerable to property deviation from process inconsistency, and that critical applications such as HV bushings demand batch-level quality verification, including TGA and surface inspection, to ensure long-term reliability.

The post Deterioration of Polymeric Transformer Bushings: Experience in Malaysia appeared first on .

This edited past contribution to INMR by Rizally Priatmadja and other engineers at PT PLN (Persero) discussed the preparatory and design phases of the project and reviewed the installation process. There is also an evaluation of initial performance post-operation.

Wenzhou Yikun Electric

China

Guangzhou MPC Power International Co. Ltd.

China

See more suppliers of Arresters

PLN UIT JBT manages a transmission network of nearly 13,000 kmc and 19,035 towers divided into 619 circuits covering Central and West Java. Most disturbances are dominated by lightning. Figs. 1 and 2, for example, show transmission line outage records between 2017 and 2022, accounting for an average of 43% of total cumulative disturbances and an average of 90% of total natural disturbances. Based on PLN’s Forced Outage Information System (FOIS) application record, 90% of lightning disturbances occur on 70 kV transmission lines and among those with the highest total cumulative outages is the Parakan-Kadipaten OHL.

The 70 kV Parakan-Kadipaten OHL is in a region with isokeraunic levels ranging from 25 to 50 thunderstorm days per year. Table 1 provides FOIS fault records and for the 70 kV Parakan-Kadipaten line, which between 2016 and June 2023 recorded 20 fault occurrences due to lightning. During the 7.5 year period, 7 towers were impacted by lightning strikes, making this the most prevalent incident. Various measures were undertaken to decrease frequency of outages, including implementing a multi-rod grounding (MRG) system and reducing tower inductance. Nevertheless, there was a notable rise in fault incidents during the following several years.

In response to these issues, at the end of 2022, PT PLN (Persero) started co-operation with an EGLA supplier to evaluate and improve the lightning performance of transmission lines. Starting Dec. 9, 2022, approximately 79 EGLAs were installed in specific configrations along the transmission line with poor lightning performance. Subsequently, from Dec. 9, 2022 on, only 3 outages due to lightning were recorded on this line, confirming the effectiveness of this protection system.

Below are details the transmission line lightning performance studies and evaluation carried out in this partnership project. Methods are presented to select arrester characteristics and define the quantity and the optimized arrester locations along the line. Field experience obtained during the first 6 months and line performance/reliability of the system after EGLA application are compared with performance before arrester installation.

Literature Review

1. Shielding Failure Analysis

In order to investigate preventive measures such as shield wires, it is important to analyze the parameters of striking distance and exposure radius. To ensure protection of the OHL, it is imperative that exposure radius does not surpass the striking distance. Fig. 1 depicts the conventional electromagnetic model (EGM) of conductors and shield wires. The variable r_g, denotes radius of the lightning strike from the point of contact with the Earth, while D_c represents the distance at which shielding failure occurs. In addition, r_c represents the radius of a lightning strike of a specific magnitude on the phase conductor and shield wire. A number of striking distance equations were identified and present study utilizes Brown-Whitehead constants, outlined below.

Ng is ground flash density whereas yearly flash incidence, NL, is a measure of flash occurrences on shield wires per 100 km.

The variable, b, represents distance between shield wires and outermost phase conductors whereas the variable, h, represents height of the tower. In order to approximate annual occurrence of flash incidents on each phase conductor, the shielding failure rate (SFR) is utilized..

In this context, the variable “l” represents the length of the line, “IMSF” is maximum lightning current that can occur before shielding failure and “f(I)” represents probability of corrence of peak lightning current. Moreover, considering that Ic represents crucial stroke current in kiloamperes (kA), total rate of shielding failure flashover (SFFOR) to the phase conductor can be determined as follows.

When a flashover is triggered by a lightning strike, the voltage at the highest point of the transmission tower will reach its maximum value at the peak of the lightning current. Therefore, the voltage at the top of the tower (VTT) can be expressed in terms of the lightning current, I(t).

(dI(t))/dt is a measure of lightning steepness, L_T represents inductance of the tower and R_E signifies tower footing resistance. To compute L_T, it is necessary to obtain the values of the tower surge impedance, Z_T, tower height, H_T, and the speed of light, c.

Moreover, the tower, functioning as a variable-impedance transmission line, partitions the tower section into a series of parts with distinct surge impedance values that correspond to the height and average radius. Hence, overall surge impedance, ZT, of the tower can be determined by summing the surge impedance ZT of each individual part of the tower.

2. Back-Flashover Analysis
Implementation of effective shield wire protection measures has the potential to result in an inadvertent discharge towards the phase conductor. Occurrence of a lightning strike on the shield wire or tower structure results in substantial potential differences across the line insulator. Consequently, when this potential surpasses insulation strength, backflashover occurs. Anticipated annual backflashover rate (BFOR) per 100 kilometres is calculated as follows:

Probability, denoted as I_c, of the initial stroke current surpassing the critical current, I_c, can be approximated as:

SUTT 70 kV Parakan-Kadipaten Profile

The 70 kV Parakan-Kadipaten OHL line has operated for 50 years, delivering energy from the Parakan Hydropower Plant with a total capacity of 10 MW. Initially, the conductor was operated with a voltage level of 30 kV but in 1973, the line’s voltage level increased to 70 kV. Average distributed load is 113.25 Amperes, with a maximum capacity of 380 amperes. To monitor the line’s performance against potential lightning disturbances, PLN has four LDS sensors installed at several locations on the island of Java.

One of the reasons for choosing this transmission line in this proof of concept project was ease and flexibility of line shutdown during the installation process. Since a special placement study process did not accompany initial installation, possible future placement optimization could be carried out flexibly.

Lightning Characteristics & Statistics

Lightning characteristics were studied across the sub-transmission line for the period from 2017 to 2022, including peak current data, flash density and periodic flash distribution. Over the past 6 years, there were a total of 2691 lightning occurrences along the 70 kV Parakan-Kadipaten OHL, with 92.87% being negative in polarity. Fig. 5 depicts the distribution of peak current frequency for both negative and positive polarities of lightning flashes along the line from 2017 to 2022. As can be seen, most currents fell between 10 and 22 kA.

In general, cumulative data over the years 2017 to 2022 showed that lightning strikes to the Parakan-Kadipaten line peaked in March. The time these strikes occurred between 1 pm and 12 pm with the highest number of strikes falling from 2 pm to 4 pm.

Lightning Performance Review

Two approaches were used to assess the impact of frequent disturbances in the Parakan-Kadipaten 70 kV SUTT area. The first was the value of tower leg grounding in that section. Grounding resistance of tower legs plays a significant role in providing lightning protection when a backflashover occurs.

Results of tower foot grounding measurements in this area are shown below:

Additionally, the current methodology relies on historical strike data from PLN’s Lightning Detection System (LDS) program. Lightning strike data utilized in this study covered the period from 2017 to 2022. The next step involved acquiring disturbance history data and disseminating post-disturbance inquiry findings to illustrate the areas impacted. Utilizing the data provided, it became possible to analyze features of lightning disturbances, specifically on the 70 kV OHL of Parakan-Kadipaten segment.

The relationship between altitude of transmission towers and occurrence of disturbances was found not to be directly influenced by geographic conditions nor by the intensity of lightning strikes in a line section. Disruptions experienced from 2018 to 2023 had direct correlation with elevated frequency of lightning strikes observed on each individual tower, i.e. the higher the frequency of lightning strikes, the greater the likelihood of disturbances. Rice fields, for example, are typically characterized by a greater incidence of lightning strikes and a higher prevalence of lightning disturbances in terms of geographic distribution.

Pilot Project of EGLA Application on 70 kV Parakan-Kadipaten OHL

a. Transmission Line Lightning Arresters

One of the reasons for choosing EGLAs in this project was cost savings. In addition, EGLAs have a longer lifetime than NGLAs because there is no ageing from exposure to continuous leakage currents. The spark gap is designed to function below the insulator basic insulation level.

Installation on the line is carried out by removing the arcing horn to ensure the EGLA performs properly and fulfil the requirement, EGLAfo < LIWLinsul. The gap on the EGLA must be appropriately designed only to work against lightning overvoltage. When the gap discharge requirements are met at switching overvoltage, should there be damage to the SVU, the line will potentially trip during the energizing process. As such, EGLAs are not recommended to be installed stand-alone on EHV systems because there is the need to prevent overvoltage switching. Thus, EGLAs are most suitable and most effectively installed in systems up to 150 kV. Table 3 provides the specifications of the EGLAs installed on the 70 kV Parakan-Kadipaten OHL.

A total of 80 EGLA sets were planned to be installed alternately on 28 towers of the Parakan-Kadipaten line with the configuration shown in Table 4. Installation considerations were based on past experience by the supplier on 20 kV systems and also on tower construction in the field.

Lightning surge arresters (LSAs) were prioritized for installation in locations with the following criteria:
• high value of footing resistance;
• shielding angle of earth wire;
• towers often struck by lightning.

In cases where there is no budget constraint and in order to realize the best possible performance, LSAs are applied on every phase.

Review of EGLA Implementation
Starting Dec. 2022, when EGLAs were first installed, until Nov. 2023, there have been a total of 3 disturbances caused by lightning. All occured on the Parakan-Kadipaten line and were triggered by short-circuit of insulators. As seen in Table 5, most occurred on insulators without an EGLA installed. CIGRE TB 855 explains that in cases where the lightning surge travels along the line from the stroke point, the lightning fault can be transferred to adjacent towers having no LSAs – either by the OHGW given successful shielding or via power line conductors if there has been shielding failure. This suggests the advisability of installing LSAs on adjacent towers beyond the high lightning risk target section. Note that LSAs may not protect an insulator assembly on the next tower, given the traveling time of a lightning surge across the span between towers.

Case 1

Disturbance that occurred on the transmission line Parakan-Kadipaten #1-2 caused reclosing at both the Parakan Substation and the Kadipaten Substation, indicating that the relay worked. The distance relay functioned in all three phases, R-S-T, Zone 1. The chart from the Disturbance Fault Recorder at both side substations described a significant increase in current in phases R-S-T-N. There were voltage defects in all three phases, which means disturbance caused by active system from a lightning surge. Climbing inspection revealed that several sets of insulators had flashed over, as shown in Table 6.

Case 2

The second case was a disturbance that occurred on transmission line Parakan-Kadipaten #1, which caused reclosure at the Parakan Substation and the Kadipaten Substation, indicating that the relay worked. It was a distance relay that worked in phase R-N, Zone 1. Based on the Disturbance Fault Recorder at Kadipaten Substation, there was a voltage drop in phase R from 38 kV into 5.6 kV and increasing current from 29A into 248A. Morteover, the Disturbance Fault Recorder at Parakan Substation recorded that there was a voltage drop in phase R from 39 kV into 10.1 kV and increasing current from 28A to 592A. The DFR chart showed that disturbance was caused by a lightning surg.

Case 3

A disturbance that occurred on transmission line Parakan-Kadipaten #2 caused reclosing at the Parakan Substation and the Kadipaten Substation, again indicating that the relay worked. This was a distance relay working in phase R-N, Zone 1. Based on the Disturbance Fault Recorder at the Kadipaten Substation, there was a voltage drop in phase R from 38.6 kV into 4 kV and current increasing from 30A to 273A. A Disturbance Fault Recorder at the Parakan Substation recorded that there was a voltage drop in phase R from 39 kV to 12.7 kV and increasing current from 27.6A to 596A. The chart from the DFR showed that the disturbance was caused by a lightning surge.

ICMET Craiova – National Institute For Research, Development And Testing In Electrical Engineering

Romania

EDP Labelec

Portugal

See more Laboratories

Discussion

There has been a history of disturbances caused by lightning surges on the Parakan-Kadipaten transmission line between 2018 and 2023, i.e. 20 cases of disturbances caused by lightning surges. Of these, 17 occurred before installation of EGLAs and only 3 after installation. As shown in Fig. 10, the Parakan-Kadipaten Line can be divided into 2 categories: Cluster A (green rectangle) and cluster B (blue rectangle). Cluster A defines towers sited near rice fields while cluster B defines those in hilly areas.

As many as 18 disturbances with 2 times happened in sequence or equal to 90% incident, the disturbance occured in the cluster A area. Towers in rice field have higher vulnerability to being struck by lightning since there are no other structures which can be struck. Disturbances that occurred after installation of EGLAs occurred on Towers #17, 15,16 and 11. According to further investigation, the discharge that occurred was marked by flashover of the insulator. Most flashovers occurred where EGLAs were not installed. Apart from EGLA installation, ageing of insulators has become one of factors in the protection system of a transmission line that is not performing optimally. This is confirmed by the test result of the dry impulse test. The test result in Table 8 stated that dry impulse test failed because maximum flashover that allowed is 2 times.

To obtain optimal lightning protection performance on transmission line, towers that have high critical level need EGLA installation using top-bottom configuration.

This can be achieved through options such as:

1. Adding installations of EGLAs to achieve top-bottom. For example, these can be installed in Cluster A which has high criticality;
2. Relocating EGLAs installed in Cluster B and maintaining towers in Cluster B with minimum configuration while looking to maintain low footing resistance (i.e. <10 ohm);
3. Replacing ageing insulators to obtain optimal results from the protection system against lightning surges.

Conclusions

The 70 kV Parakan-Kadipaten Line was selected for application of EGLAs to test their effectiveness with a particular installation design. Over the period from Dec. 9, 2022, to Aug. 2023, there have been 3 disturbances that resulted in short circuits to insulators that were not equipped with EGLAs.

Based on observations during Qtr 1, 2023, after installation of the EGLAs there was a 33% decrease compared to Qtr 1, 2022 under conditions of all lightning strikes, which also decreased from 625 to 435.

Lightning performance calculations using Sigma SLP software show that the existing configurations do not significantly reduce disturbances on the 70 kV Parakan-Kadipaten Line. By retaining existing numbers, installing EGLAs in a new configuration may end up providing more significant improvements.

References
[1] CIGRE WG C4, Procedures for Estimating the Lightning Performance of Transmission Lines – New Aspects, no. June. Paris: CIGRE, 2021.
[2] PT.PLN (Persero), “70 kV OHL of Parakan-Kadipaten Lightning Detection System Report,” Cirebon, 2023.
[3] CIGRE 855, Effectiveness of line surge arrester for lightning protection of overhead transmission line, no. December. 2021.
[4] IEEE, Guide for Improving the lightning performance of Transmission Lines. 1997.
[5] R. Bhattarai, R. Rashedin, S. Venkatesan, A. Haddad, H. Griffiths, and N. Harid, “Lightning performance of 275 kV transmission lines,” Proc. Univ. Power Eng. Conf., pp. 2–6, 2008, doi: 10.1109/UPEC.2008.4651622.

The post Applying EGLAs on 70 kV Overhead Line: Design & Lightning Performance appeared first on .

Testing of MV cables is needed to ensure safety for people, to verify proper workmanship and to increase reliability and availability of the network.

The post Statistics & Experience with VLF Withstand Testing & PD Diagnostics of MV Cable Networks (Video) appeared first on .

Service experience has shown that, from the electrical point of view, application of NGLAs typically improves line performance with relatively few failures. Nonetheless, utilities have reported that installations are sometimes compromised by mechanical issues such as failure of connection leads or disconnectors. While arrester standards do not include mechanical tests or requirements for the connections, IEC 60099-5 suggests an arrester life expectancy of at least 25 years. Given these considerations, EPRI began a research project to better understand causes of lead breakages so that necessary functional improvements could be included in future user specifications.

This edited past contribution to INMR by experts at EPRI reviewed findings and offered insight into requirements for mechanical testing, with focus on NGLAs since these are the type most commonly applied in the United States.

DALEKOVOD OSO

Croatia

EB Rebosio SRL, A Gruppo Bonomi Company

Italy

See more suppliers of Laboratory & Field Testing Equipment

• Connections should allow free movement of phase conductors, which can take the form of conductor swing, Aeolian vibrations, galloping and sub-span oscillations;
• Connections and associated hardware should withstand all mechanical forces to which they are subjected;
• Connections must be durable enough to withstand fatigue due to movement;
• Installations and connections should not place excessive mechanical loading on the arrester or on disconnector attachments;
• Arrester disconnection should occur in a controlled manner, without consequential damage to the unit or to other equipment.

Lead wires are used to connect an arrester unit to the phase conductor or to a grounded part of the line support structure and there are two types of lead configurations:

1. Flexible rope type wires with crimped lug connectors

These configurations are mostly used in applications where the lead is not subjected to high mechanical loads (as in Fig. 1).

2. Chain intertwined with lead wire

These configurations have been successfully applied in cases where significant mechanical loads are expected.

Since the primary function of the lead is to establish an electrical connection, it should be designed to carry (a) low magnitudes of continuous current flowing through the arrester during normal operation; (b) high impulse discharge currents when the arrester conducts a surge; and (c) power frequency fault current if the arrester fails. Since the lead also implicitly establishes a mechanical connection, it is also necessary to consider these mechanical forces when designing the lead attachment.

Fig. 2 shows examples of common lead failures, which include:

• Lead wire pulling out of the lug, a common problem and possibly the result of mechanical overload or poor crimping of the lead connection;
• Edge of the crimped lug ‘sawing’ into the lead conductor thereby severing strands and mechanically weakening the lead, such as when a mechanically loaded lead is not aligned with the lug;
• Chain connections where the electrical lead connection woven through the chain wears down due to continuous lead movement.

Load Characterization Testing

Current standards do not provide guidance specific to applying line arresters nor do they prescribe mechanical tests for leads. Therefore, with the goal of developing suitable laboratory test methods for leads and disconnectors, EPRI constructed a simulated field test to quantify typical in-service mechanical lead loads under different environmental conditions. The simulated field test measures mechanical loading of the arrester and leads under actual service conditions while also monitoring and recording any mechanical degradation on the components. This information then provides a good basis for subsequent laboratory testing.

Instrumenting an energized installation was not cost justified in this case since only the mechanical forces on the lead are of interest. Three typical arrester configurations were therefore installed on a de-energized test line commissioned at EPRI’s High Voltage Test Facility in Lenox, MA in Jan. 2015 (see Fig. 3). Mechanical loads in the leads are monitored continuously but only maximum load over a 2-minute interval is logged. Special precautions were taken to ensure load cells and their added weight did not influence the loads in the leads. A dedicated weather station was used to continuously reconcile impact of the outdoor environment during testing and backed up by another logging weather station attached to a mast only a short distance away. The most important parameter, wind speed, is monitored with a 2-D ultrasonic type anemometer, which has no moving parts to wear out or to introduce errors from drag. Ultrasonic anemometers have the further advantage of more accurately measuring lower wind speeds.

Load measurements taken over an 18-month period for each of the three arresters were then compared and data from the position experiencing the highest loads was reduced to a simplified test cycle. An acceleration factor was achieved by applying loads at much shorter time intervals (16 hours of testing equivalent to one year of service life). The original intent was to develop a test that enabled accelerated ageing of just the arrester leads. However this was later extended to mechanical testing of the arrester body as well.

Accelerated Ageing Testing

To simulate tensile forces in arrester leads, a test rig was built enabling a series of load cycles to be applied precisely to the leads and hence the arrester body (see Fig. 4).

The intent of the test rig was to closely simulate the types of failures experienced by arrester leads in the field and to achieve this in an accelerated timeframe. Once a repeatable and realistic failure mode has been achieved, the device can then evaluate and compare different lead configurations and materials. Method of load application (i.e. actuator type and location) changed during the project based on types of physical degradation experienced by leads during the course of testing. For example, the initial lead loading method was using a pneumatic actuator, as seen in Fig. 4. However, it was discovered that the arrester’s relatively low inertia resulted in an unrealistically low load limit in the lead, no matter the size of actuator or pressure used.

The next generation of load application method was by swinging the arrester to fixed angles, which when released resulted in desired loads in the leads (see Fig. 5). This more closely resembles how an arrester lead is actually loaded, i.e. via conductor and arrester wind motion, but required the top link of the arrester to be replaced with a simple pivot. The loads in leads were easily controllable with this method of actuation and the higher test loads required could be obtained. Nevertheless, this test configuration resulted in frequent failure of the arrester’s top attachment stud. Although such failures have been reported in the field, the high rate experienced during testing indicted that this loading method was not realistic. The final and still used generation of load application is via an underslung lever mechanism that contacts and pushes the bottom of the arrester (see Fig. 6). An advantage of this approach is that the arrester with its original top link assembly can be tested without modification.

The leads of arresters tested using this rig configuration have exhibited closest comparison to types of failures actually experienced in service (see Fig. 7). Subsequent testing in the current rig will evaluate the effect on life expectancy of undesirable lead lug orientations (e.g. tightened so they are not in-line with the load) as well as different lead materials and configurations.

In addition to accelerated ageing of arrester leads, the impact of repetitive shock loading on arrester functionality will be monitored. A selection of arresters has been electrically tested prior to being used for lead ageing. After an equivalent 30 years of mechanical testing, these will be electrically tested once more to allow comparison.

Conclusions

Non-gapped line arresters are increasingly being applied to improve lightning performance of transmission lines. While these arresters perform well electrically, their installation is often compromised by failures of the connection leads or the disconnector. Service experience shows that many of these failures are due to installation related issues. For example, connections between arrester and energized conductor or grounded structure are often subjected to static and dynamic loads which could lead to fatigue or overloading, resulting in either broken connections or damage to the arrester. The fact that present test standards do not address requirements for arrester connecting leads prompted EPRI to develop a mechanical measurement and testing approach to better understand these issues. However, method of testing must be carefully selected so that failure modes most closely replicate those actually experienced in service. Even minor changes to test rig configuration can result in significant variations in types of failure observed.

References
1. Cigré WG 33.11 Task Force 03: “Application of metal oxide arrester to overhead lines”, Electra No. 186, October 1999, pp 83-112.
2. Williamson J., “Lightning Protection and Surge Arrester Application on NB Power Transmission Lines”, IEEE PES Transmission and Distribution Conference and Exposition, 2008.
3. Overhead Transmission Line Lightning and Grounding Reference Book 2011. EPRI, Palo Alto, CA: 2011. 1023429.
4. Application of Transmission Line Surge Arresters. EPRI, Palo Alto, CA: 2010. 1019954.
5. Guide for the Application of Transmission Line Surge Arresters. EPRI, Palo Alto, CA: 2009. 1017709. 

The post Testing Connection Leads for Transmission Line Arresters appeared first on .

Different techniques have been used to mitigate such risk including periodic insulator cleaning as well as measures to enhance insulation strength such as by applying RTV coatings or booster sheds. However, these tend to be labor-intensive and often limited by operational constraints of an energized transmission line. Performance of insulators under pollution will therefore depend on the intrinsic properties of their housing material.

Compared to ceramic insulators, composite polymeric insulators—particularly those made with silicone rubber—offer superior resistance to pollution flashover. Silicone insulators combine excellent hydrophobicity, resistance to chemical degradation, and excellent insulation properties, which makes them ideally suited for contaminated environments.

But a key element in the long-term performance of silicone insulators relates to their ability to regain their initial water-repellency should this property be temporarily suppressed or lost due to repeated exposure to contaminants and moisture. The recovery mechanism is essential to maintaining long-term performance in outdoor applications.

This edited contribution to INMR by experts at K-Line Insulators in Canada reports on findings of a study undertaken to compare hydrophobic recovery of silicone insulators supplied by different manufacturers after exposure to repetitive contamination cycles.

Guangzhou MPC Power International Co. Ltd.

China

Newell (A PGC Company)

United States

See more suppliers of Silicones & RTV Silicone Coatings

One of the most widely used and well-established methods to assess water repellency of polymeric materials is contact angle (CA) measurement between water droplet and polymer surface. Contact angle is defined as the angle at the intersection of the liquid–solid and liquid–vapor interfaces. The point where the solid, liquid, and vapor phases meet is known as the “three-phase contact line.” Generally, a small CA (<90°) indicates high wettability (hydrophilicity), while a large CA (>90°) indicates low wettability (hydrophobicity).

Interfacial tension between a solid and liquid (γsl) is governed by the surface tension of the liquid (γlv) and the surface energy of the solid (γsv). The surface tension of a liquid reflects the cohesive forces between its molecules, whereas the surface energy of a solid represents the strength of intermolecular attractions at its surface.

Assuming water is the liquid of interest, the CA between a water droplet and a solid surface is primarily determined by the solid’s surface energy, which is influenced by its chemical composition. Consequently, factors such as surface treatments, cleanliness, and contamination can significantly alter the solid’s surface energy and, thus, the CA.

Contaminants such as soluble dust and salts tend to exhibit high surface energy and hydrophilic behavior, leading to a shift in surface wettability from hydrophobic to hydrophilic once these accumulate on insulator surfaces. Silicone rubber is intrinsically hydrophobic, a critical property, given that presence of water can degrade electrical performance by reducing surface resistance. Hydrophobicity helps maintain insulation integrity by repelling moisture.

However, loss of hydrophobicity can diminish electrical insulation and pollution withstand as well as accelerate ageing of silicone insulators. Notably, rate of hydrophobicity recovery, which mitigates ageing effects, depends largely on presence and migration of low molecular weight siloxane (LMWS) chains within the material.

Low Molecular Weight Silicone Chains

Surface hydrophobicity of high-temperature vulcanized (HTV) silicone rubber insulators plays a critical role in their superior performance. Compared to other organic polymeric insulation materials such as ethylene propylene diene monomer (EPDM), silicone insulators exhibit better suppression of leakage current and partial discharges under heavily contaminated conditions. This hydrophobicity arises from the material’s low surface energy (16 – 21 mN.m^-1), which is attributed to the presence of methyl groups (-CH₃) in its chemical structure.

Fig. 2 illustrates the chemical structure of polydimethylsiloxane (PDMS). PDMS chains are structured with siloxane bonds (Si-O) as the backbone and methyl groups (-CH3), as functional substituents. Among common organic substituents, the methyl group has the second-lowest critical surface tension—surpassed only by the fluoroethylene group (-CF3). As a result, hydrophobicity of silicone rubber is comparable to that of fluorinated polymers such as polytetrafluoroethylene (PTFE).

Surface hydrophobicity of HTV silicone rubber insulators can degrade over time due to factors such as corona discharge, dry band arcing, and the accumulation of dust or contaminants. The hydrophobic recovery of PDMS is attributed primarily to 3 mechanisms:
1. reorientation of methyl groups through conformational changes;
2. condensation of silanol groups during the crosslinking process; and
3. diffusion of LMWS to the surface.

Among these, presence and migration of LMWS from the bulk to the surface play a critical role in hydrophobic recovery of HTV silicone insulators (see Fig. 3). Notably, LMWS can migrate through the bulk material and form a hydrophobic film over surface contaminants—even under heavily polluted conditions—significantly enhancing long-term performance and stability of silicone insulators.

A deeper examination of silicone rubber’s molecular structure reveals that its broad molecular weight distribution of siloxane components underpins this unique property. Siloxane molecular weights can range from less than 25,000 g/mol for oil-like silicone fluids to over 500,000 g/mol for fully polymerized silicone chains. This range allows the smaller LMWS molecules to migrate more easily through the bulk material, provided they are evenly distributed. Migration to the surface reduces surface free energy, thereby restoring hydrophobicity.

Therefore, evaluating mobility of LMWS and hydrophobic recovery capability is essential to the long-term reliability of high voltage silicone insulators. To avoid need for costly maintenance procedures such as manual cleaning, the hydrophobicity transfer property becomes a critical operational characteristic for housing materials designed to maintain surface hydrophobicity under polluted conditions.

Efficiency of this migration depends on several factors:

1. LMWS concentration & molecular structure
More mobile, low-viscosity LMWS tend to migrate more quickly and uniformly.

2. Crosslink density & network architecture
Tightly cross-linked systems hinder LMWS mobility, thereby reducing recovery rates.

3. Filler content & type (e.g. ATH)
This can facilitate or obstruct LMWS diffusion, depending on dispersion and interfacial interactions.

Past research by K-Line Insulators used the method outlined in IEC 62217 ED3 to evaluate and compare the hydrophobic recovery of silicone insulators from different material manufacturers. In the present study, this work was extended to investigate the long-term hydrophobic recovery performance of the same 3 silicone rubber insulator manufacturer materials.

Several factors influence rate and effectiveness of hydrophobic recovery of silicone rubber insulators:

1. Temperature and Humidity
Elevated temperature and humidity levels can accelerate recovery by enhancing the mobility of LMWS.

2. Exposure Time & Contaminant Type
Duration of exposure and the nature of the contaminant impact the degree of recovery. Certain contaminants, such as salts or oils, may require prolonged exposure or higher temperatures for effective recovery.

3. Surface Roughness
Surface texture of the silicone rubber can influence hydrophobic recovery. Smoother surfaces generally show improved recovery characteristics.

4. Material Composition
Specific formulation of the silicone rubber inclu

Here, the focus was on exposure time and material composition as primary factors influencing long-term hydrophobic recovery of silicone rubber. Exposure time was evaluated through multiple contamination /recovery cycles. Each sample underwent repeated surface contamination until it could no longer regain its hydrophobic properties.

Methodology

The Hydrophobicity Transfer Test was conducted in accordance with IEC 62217 ED 3 (Clause 9.3.5). Test samples were extracted from the sheds of the silicone insulator housings. Prior to testing, the samples were thoroughly cleaned using isopropyl alcohol to remove any mold release residues, followed by rinsing with deionized water. After cleaning, the samples were allowed to dry for at least 24h under standard laboratory conditions. To define the test area, the samples were covered with adhesive aluminum foil, leaving a 30 mm × 30 mm window exposed (see Fig. 4).

The pollution layer thickness for this test was determined by thickness of the foil. To achieve the required thickness of 0.36 mm, three layers of 0.12 mm aluminum foil were stacked on top of one another. A contamination slurry was then prepared and applied to the exposed window to simulate polluted conditions and evaluate hydrophobicity transfer behavior.

The slurry was prepared by mixing 7.5 g of untreated medium-grain silica powder (average particle size of 3 µm; CABOT, USA) with 3.5 mL of a water-isopropanol solution (65 vol% water, 35 vol% isopropanol). The mixture was stirred to ensure homogeneity. To minimize isopropanol evaporation, the slurry was applied to the surface immediately after preparation. Excess slurry was removed using a plastic stick to produce a smooth and uniform coating (see Fig. 5).

Contact angle (CA) measurements were performed using freely available software developed by the Biomedical Imaging Group, based at EPFL (École Polytechnique Fédérale de Lausanne). This research group has published two peer-reviewed studies presenting advanced methods for high-accuracy CA measurements. In this investigation, the Low Bond Axisymmetric Drop Shape Analysis (LB-ADSA) method was employed, which is based on fitting the Young–Laplace equation to droplet shape profiles extracted from image data. To ensure accuracy and reproducibility, water CA measurements were conducted at multiple locations on each sample. Average CA values and corresponding standard deviations were reported for all samples in this study.

Fourier Transform Infrared Spectroscopy (FTIR) was used to evaluate surface chemical composition during the hydrophobic recovery process. The analyses were carried out using a Cary 630 FTIR Spectrometer (Agilent, USA) in Attenuated Total Reflection (ATR) mode. Spectra were recorded across the infrared range of 400 to 4000 cm⁻¹ to capture relevant chemical functional groups present on the silicone rubber surfaces.

Results & Discussions

The surface wettability of the contaminated silicone surfaces of insulator manufacturers #1, #2, and #3 was monitored at specific intervals by measuring CA values (see Table 1). While the objective of the previous study was to determine rate of hydrophobic recovery after a single contamination cycle, the focus now was to evaluate the hydrophobic recovery rate over 5 consecutive contamination cycles.

All samples were initially hydrophobic, as expected, since the housing material was silicone rubber, which possesses intrinsic hydrophobicity. Following contamination, all samples became superhydrophilic (i.e., CA = 0°) due to the slurry covering the surface and masking the methyl groups responsible for the hydrophobic properties of the silicone rubber.

However, each sample exhibited different rates of hydrophobic recovery. On the first day after applying the contaminant slurry, Insulator Manufacturer #1 reached a CA of 38°, while Manufacturer #2 and Manufacturer #3 showed CAs of 22° and 18°, respectively. By the second day, the CAs had increased to 68° for Manufacturer #1, 38° for Manufacturer #2, and 22° for Manufacturer #3, indicating a faster recovery rate for Manufacturer #1. By the fourth day, Manufacturer #1 had regained its defined hydrophobic property level, whereas the other two samples had not yet recovered to the same extent.

To further evaluate the hydrophobic recovery of the other manufacturers, the samples were monitored over an extended period until their CAs approached a near-plateau. Manufacturer #2 eventually reached a nearly hydrophobic level, with a CA of 88 ± 2°. By contrast, Manufacturer #3 showed no significant change in CA, even after 32 days.

At this stage (i.e. 32 days after applying the first layer), the second layer of contamination was applied using the same procedure as the first, on top of the first layer. Overall, the hydrophobic recovery rate was slower compared to the first contamination cycle. After four days, only Manufacturer #1 approached a near-hydrophobic CA (85 ± 3°). By day eight, Manufacturer #1 had reached a fully hydrophobic CA of 94 ± 2°, while Manufacturer #2 had only reached 80 ± 4°. Manufacturer #2 eventually reached a near-hydrophobic CA of 88 ± 1° after 16 days. Manufacturer #3 recovered to approximately the same CA level it had reached after the first contamination cycle, showing limited improvement.

The experiments proceeded with the third layer of contamination (i.e. 16 days after applying the second layer). At this stage, there was a longer delay in the hydrophobic recovery across all manufacturers’ materials. It took 10 days for Manufacturer #1 to reach a hydrophobic CA level. However, even after 16 days, neither of the other two manufacturers’ materials were able to reach full recovery.

The decline in hydrophobic recovery continued with the application of the fourth contamination layer. It took 20 days for Manufacturer #1 to regain its hydrophobicity, while Manufacturer #2 and #3 reached a CA of 84 ± 2° and 19 ± 3° by that time, respectively. In the fifth contamination cycle, it took 22 days for Manufacturer #1 to regain its hydrophobicity. Manufacturer #2 recovered to nearly the same level as in the fourth cycle, reaching a CA of 83 ± 2°. Manufacturer #3, however, completely lost its ability to recover hydrophobicity.

Fig. 6 summarizes time-dependent hydrophobic recovery of the 3 silicone rubber insulator materials from different manufacturers over 5 contamination cycles.

Manufacturer #1:
• Initial Recovery (Cycle 1): Fastest recovery among all; reached full recovery by Day 4.
• Cycle 2: Recovery slowed slightly; reached near-hydrophobic CA (85 ± 3°) by Day 4 and full hydrophobicity (94 ± 2°) by Day 8.
• Cycle 3: Recovery deteriorated further; took 10 days to regain hydrophobicity.
• Cycle 4: Required 20 days to recover.
• Cycle 5: Took 22 days to regain hydrophobicity.
• Trend: Progressive degradation in recovery rate, but still able to fully recover in each cycle.

Manufacturer #2:
• Initial Recovery (Cycle 1): Moderate recovery; CA reached 22° on Day 1, 38° on Day 2, and near-hydrophobic level (88 ± 2°) after extended time.
• Cycle 2: Reached 80 ± 4° by Day 8 and 88 ± 1° by Day 16.
• Cycle 3: Recovery slowed further.
• Cycle 4: Achieved CA of 84 ± 2° by Day 20.
• Cycle 5: Recovered to a similar level as Cycle 4 (83 ± 2°) after 22 days.
• Trend: Slower and less consistent recovery with each cycle; did not reach full hydrophobicity in later cycles.

Manufacturer #3:
• Initial Recovery (Cycle 1): Weak recovery; CA reached only 18° on Day 1 and showed minimal improvement after 32 days.
• Cycle 2: Reached similar CA as in Cycle 1; no significant improvement.
• Cycle 3: Continued limited recovery.
• Cycle 4: No specific improvement observed.
• Cycle 5: Completely lost ability to recover hydrophobicity.
• Trend: Consistently poor hydrophobicity recovery, with performance degrading further in later cycles.

Fig. 7 shows the ATR FTIR absorption spectra of silicone rubber insulators from the three manufacturers before contaminant slurry application. The absorbance peaks observed at approximately 1000-1110 cm-1, 805-855 cm-1, and 1245-1275 cm-1, correspond to the molecular vibrations of the Si-O-Si symmetric stretch, Si(CH3)2 symmetric stretch, and Si(CH3) groups, respectively. Additionally, the broad peak between 3000-3500 cm-1 is attributed to the molecular vibrations of aluminum trihydrate (ATH) particles present in the samples.

Fig. 8 illustrates the evolution of surface chemical bonds for the three silicone rubber insulator manufacturers on the 2nd and 4th days after the first layer of contaminant slurry application. The significantly higher CA observed for Manufacturer #1 after 2 days is attributed to the emergence of Si(CH₃) peaks in its FTIR spectrum. By day 4, the FTIR spectrum of Manufacturer #1 closely resembles its pre-test profile, which explains the high CA value of approximately 90°. Manufacturer #2 began showing Si(CH₃) peaks by day 4, corresponding to an increased CA of 79°. In contrast, Manufacturer #3 exhibited no significant spectral change over the 4-day period, aside from a slight increase in the Si(CH₃)2 peak, resulting in a modest increase its CA (34°).

The observed FTIR spectral changes provide insight into the hydrophobic recovery mechanisms of the silicone rubber surfaces. The presence and intensity of Si–CH₃ related peaks serve as indicators of the surface’s hydrophobic methyl groups, which are responsible for water repellency. For Manufacturer #1, the rapid reappearance of these Si(CH₃) peaks within four days suggests efficient migration or reorientation of low molecular weight siloxane chains to the surface, restoring the hydrophobic character.

This molecular-level recovery aligns with the significant increase in contact angle observed. Manufacturer #2’s slower and less pronounced Si(CH₃) peak development reflects a more gradual or limited re-establishment of hydrophobic methyl groups at the surface, consistent with its intermediate contact angle recovery. Conversely, Manufacturer #3 showed minimal spectral changes, indicating a poor recovery of hydrophobic groups on the surface. This lack of significant molecular reorganization likely results in its persistently low contact angle and weak hydrophobic recovery.

Overall, FTIR data confirm that the recovery of surface hydrophobicity is closely linked to the regeneration of methyl group functionality on the silicone rubber surface. This process is influenced by the availability and mobility of low molecular weight species within the bulk material, which differs among manufacturers’ different formulations and contamination cycles.

Fig. 9 illustrates the evolution of surface chemical bonds in silicone rubber samples from three manufacturers, measured 22 days after the 5th application of contaminant slurry. The FTIR spectra align with the CA measurements. Manufacturer #1 exhibited stronger Si(CH₃) and Si(CH₃)2 peaks, correlating with higher CA values. In contrast, Manufacturer #2 displayed the same peaks but at much lower intensity, resulting in reduced hydrophobicity. This observation suggests that Manufacturer #2 had a weaker recovery performance compared to Manufacturer #1, potentially due to differences in crosslink structures, crosslink density, the availability and type of LMWS, and the migration rate of LMWS to the surface. The chemical bonds in Manufacturer #3 closely resembled those of silica powder, indicating a failure to recover hydrophobicity. This suggests that Manufacturer #3 lacked sufficient LMWS in the bulk material to migrate to the surface and mask the contaminant layer.

Summary & Outlook

This research evaluated the hydrophobic recovery behavior of silicone rubber insulators from 3 manufacturers subjected to 5 repeated contamination cycles. Findings, derived from time-resolved contact angle (CA) measurements and FTIR spectroscopy, offer insight into the degradation and re-generation mechanisms of surface hydrophobicity under environmental stress.

Key Observations:

1. Manufacturer #1 material consistently demonstrated superior hydrophobic recovery, with rapid restoration of surface methyl groups, as indicated by the early and strong reappearance of Si(CH₃) and Si(CH₃)₂ peaks in FTIR spectra. Even though the recovery rate gradually declined with successive contamination cycles—requiring up to 22 days in the 5th cycle—this material was still able to fully restore its initial hydrophobicity in all cases.

2. Manufacturer #2 material showed moderate performance, with increasingly delayed and incomplete recovery in later cycles. The FTIR spectra revealed delayed development of methyl group peaks and lower peak intensity overall, suggesting that either the quantity or mobility of low molecular weight siloxanes was limited. By the 5th cycle, although a near-hydrophobic state was reached (CA ≈ 83°), the surface never fully regained its pristine hydrophobic condition, even after 22 days.

3. Manufacturer #3 material failed to recover hydrophobicity after repeated contamination, with contact angles remaining near zero from the 3rd cycle onward. FTIR data supported this finding by showing little to no regeneration of methyl-functional groups, implying an absence or exhaustion of LMW chains in the bulk. This suggests that the material either lacked sufficient LMW species from the outset or that its cross-linked network restricted migration of these molecules to the surface.

These results emphasize that not all silicone rubber insulator materials offer equivalent long-term performance in polluted environments. They also illustrate that some materials that perform well after a single contamination cycle may still degrade significantly under repeated exposure.

KERI High Power & High Voltage Laboratories

South Korea

PSW – Siemens Energy Testing Laboratories Berlin

Germany

See more Laboratories

REFERENCES
1. Hu, Y.Y., X. Yu, G.G. Wang, and M. Lu, Surface Wettability of the Contaminative Silicone Rubber. Applied Mechanics and Materials, 2014. 675: p. 31-37.
2. Yu, J., et al., Rapid hydrophobicity recovery of contaminated silicone rubber using low-power microwave plasma in ambient air. Chemical Engineering Journal, 2023. 465: p. 142921.
3. Papailiou, K. and F. Schmuck, Silicone composite insulators. 2013: Springer.
4. Vasudev, N., et al. Long term ageing performance of Silicone rubber insulators under different conditions. in 2009 IEEE 9th International Conference on the Properties and Applications of Dielectric Materials. 2009. IEEE.
5. Hillborg, H., M. Sandelin, and U.W. Gedde, Hydrophobic recovery of polydimethylsiloxane after exposure to partial discharges as a function of crosslink density. Polymer, 2001. 42(17): p. 7349-7362.
6. Homma, H., et al., Diffusion of low molecular weight siloxane from bulk to surface. IEEE Transactions on Dielectrics and Electrical Insulation, 1999. 6(3): p. 370-375.
7. Yuan, Y. and T.R. Lee, Contact angle and wetting properties, in Surface science techniques. 2013, Springer. p. 3-34.
8. Jothi Prakash, C. and R. Prasanth, Approaches to design a surface with tunable wettability: a review on surface properties. Journal of Materials Science, 2021. 56: p. 108-135.
9. Amin, M., M. Akbar, and S. Amin, Hydrophobicity of silicone rubber used for outdoor insulation (an overview). Rev. Adv. Mater. Sci, 2007. 16(1-2): p. 10-26.
10. Zhu, Y., M. Otsubo, C. Honda, and S. Tanaka, Loss and recovery in hydrophobicity of silicone rubber exposed to corona discharge. Polymer degradation and stability, 2006. 91(7): p. 1448-1454.
11. Maghsoudi, K., Carreira, A.J., Hydrophobic Recovery of HTV Silicone Rubber Insulators. INMR Woorld Congress, 2023.
12. IEC 62217 ED3: Polymeric HV insulators for indoor and outdoor use – General definitions, test methods, and acceptance criteria. 2019.
13. http://bigwww.epfl.ch/demo/dropanalysis/.
14. Stalder, A.F., et al., Low-bond axisymmetric drop shape analysis for surface tension and contact angle measurements of sessile drops. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 2010. 364(1-3): p. 72-81.
15. Stalder, A.F., et al., A snake-based approach to accurate determination of both contact points and contact angles. Colloids and surfaces A: physicochemical and engineering aspects, 2006. 286(1-3): p. 92-103.
16. Maghsoudi, K., G. Momen, R. Jafari, and M. Farzaneh, Direct replication of micro-nanostructures in the fabrication of superhydrophobic silicone rubber surfaces by compression molding. Applied Surface Science, 2018. 458: p. 619-628.

The post Hydrophobic Recovery of HTV Silicone Insulators Under Repeated Pollution Exposure appeared first on .

Leer artículo en español

Power system engineers all face a similar dilemma when trying to identify which components offer the quality they require – especially in today’s global market with so many competitors. The choice becomes even more important in the case of surge arresters, whose function is to prevent costly damage to assets due to lightning or switching surges.

How quality is defined will of course depend on the specific type of arrester under consideration. For a distribution arrester, its ability to disconnect from the system if overloaded is critical. This is also high on the list when assessing the quality of a transmission line arrester. By contrast, much different factors are relevant when evaluating the quality of a station arrester.

Designing and manufacturing quality surge arresters often requires testing and characterization beyond what is indicated in the standards. Additional checks and tests may prove necessary to guarantee the quality of any arrester that is put into service.

This edited past contribution to INMR by arrester expert Jonathan Woodworth, presented his views on key factors to consider when purchasing high voltage surge protective devices.

Hivolt Power System

China

Wish Power (Thailand) Co. Ltd

Thailand

See more suppliers of Arresters

Station class arresters protect the highest value assets at a substation and are part of a grid that serves millions. It is therefore vital they are produced to the highest level of quality possible. The latest design to join this family is the composite hollow core arrester, which combines the benefits of polymer-housed and porcelain-housed designs. The various elements that comprise such a product are shown in Fig. 1 and critical characteristics start with a rubber housing material that can withstand years of stress, no matter the environment.

The distribution type arrester differs from the station arrester in that it is installed only for lightning protection. This small yet key component of a power system protects lines that directly serve end customers. Similarly, the principal function of a transmission line arrester is to protect the HV system from lightning damage. A vital component on both – but not on station class arresters – is the disconnector. If not for the disconnector, an arrester that fails would short out the system and lead to a sustained outage. The disconnector removes the failed unit from the system should it become a short circuit, causing no more than a system ‘blink’.

Common Features of Quality Arresters

Robust Materials with Proven Service Record

A quality housing is perhaps first on the list if an arrester is to function for decades without problem. Different polymeric housing materials are not all equal in performance, so it is vital to select one that has demonstrated good service history, even under severe environmental conditions. Arrester suppliers must be able to provide technical support to demonstrate, through testing and field experience, that the materials they use will perform over the long term.

Protection Levels

Effective lightning and switching protection are what is required most when purchasing an arrester. Discharge voltage and sparkover values are measures of this protection and the lower these are, the better. Consideration of this critical parameter should be paramount when selecting an arrester. Additionally, a quality supplier will always have an audit trail to verify all the characteristics published for its arresters.

Failure Mode

If an arrester is overloaded beyond its design capability and fails, it is important that it does so in a manner that does not jeopardize nearby high value assets. To ensure that the failure mode claimed for an arrester is as published, a certified test report should be made available, ideally from an independent and reputable laboratory. These tests are complex and running them improperly can result in arresters that pass yet are not capable of a safe mode of failure.

Moisture Resistance
Resistance to moisture ingress is among the most significant features of a quality arrester. Products designed and manufactured without proper testing and verification of this capability will probably not survive long under demanding environmental stresses. A common misconception is that polymer-housed arresters are intrinsically moisture resistant; in fact they are not. The arrester sealing function should therefore be primary in design and also during the manufacturing process. To assure good quality, it is advisable to request proof that the arrester will withstand all expected environmental conditions and also how this has been verified. IEEE standards offer a minimum set of tests that must be passed, but high quality arresters are usually tested beyond normal standards.

Quality Disks

The MOV disk is the ‘heart’ of an arrester. Without quality disks that perform as specified, an arrester will simply not provide the desired level of system protection.

Durable Terminals

‘Durable’ here means not easily damaged by mishandling during installation. These parts must therefore be made of materials that will last the full service life of the unit and not cause installation problems such as cross threading, partial discharge or breakage. A second aspect of robust terminals is long-term resistance to weathering. Rust that forms on a connector, for example, can eventually wash down the housing and reduce its wet flashover capability.

Dependable Disconnector

The disconnector can be equal in value to the arrester itself if its successful operation helps avoid the cost of a long-term outage.

For this reason, it is important to know the service history of the disconnector. Questions to ask in this regard include: how long has the design been in use; what is the principle of operation; how low and how high in current will it operate; and how has performance been verified during production. A quality arrester will also have a disconnector that is approved as non-hazardous for transport by public regulatory agencies. This means that, if involved in a fire, it must not explode and endanger first responders.

Reliable Venting

In the event an arrester becomes overloaded, it is essential that it vent internal gases rapidly and efficiently to avoid complex fracture and potential collateral damage to nearby assets. Polymer and porcelain-housed arresters must both perform flawlessly under such a scenario.

Proven Dielectric Materials

The materials comprising an arrester need good dielectric properties, depending on where they are used. For example, good dielectric is required wherever the material experiences voltage stress during the arrester’s long service life. Dielectric materials with a proven history in high voltage applications should be well known to manufacturers and only these are to be used in a quality arrester design.

Clear & Complete Certified Test Reports

The quality and design of any arrester can be determined by looking at the certified test report and arrester manufacturers should always have this data available to customers. Reports need to be clear, concise and accurate, with tests that follow IEEE C62.11 or IEC 60099-4. Data presented should be self-explanatory and easily understood.

High Power Frequency Withstand Capability

High power frequency withstand (TOV withstand) capability is not necessarily a given in any arrester. In fact, some arresters may offer only poor temporary overvoltage withstand capability and fail during a single line to ground fault. It is advisable to only install arresters with high TOV withstand capabilities, as verified by the TOV curve in either catalogue literature or certified test reports. It is also vital that this characteristic be tested routinely during production.

Safety Label

Differently rated distribution arresters can appear similar at first glance and the practice of using a safety label was therefore adopted in the 1990s. Such a label is essential to line personnel installing arresters should a lower rated arrester accidentally be packaged in a carton containing higher rated units. If a 10 kV arrester was packed in an 18 kV carton, for example, its installation could pose a hazard. A large label makes it easier for the installer to see the rating and reduces risk. This is the 3^rd arrester label found on quality arresters and should ideally be a mandatory requirement by the user.

Dependable & Reliable Routine Tests

Routine tests on arresters and their components are critical during production. Discharge voltage, TOV, watts loss, leakage current and partial discharge are all critical parameters that should be monitored to be within manufacturer specifications.

The post Identifying Quality Surge Arresters appeared first on .

At the same time, there have been reports of negative service experience and corresponding calls to update and further improve existing standards. For example, one of the most important design aspects for these insulators is the longitudinal interface between the core rod and the polymeric housing. Indeed, failures at this interface have been one of most critical aspects leading to end-of-life situations. Based on this, questions have been raised whether present standards and the procedures used to test interface design are sufficiently sensitive and whether any design test result can ever be transferred to routine test behavior.

This edited past contribution to INMR by experts at EGU HV Laboratory, Pfisterer Switzerland and ČEPS presented findings from laboratory investigation into failures of polymer insulators at the critical core-housing interface.

LAPP Insulators GmbH

Germany

Wish Power (Thailand) Co. Ltd

Thailand

See more suppliers of Insulators: Overhead Line

Polymeric insulators have been used worldwide on overhead lines and for other applications for decades, predominantly for AC but also in several DC applications. Choice and development of suitable material compositions and methods of manufacture have improved significantly since the first generation. Nonetheless, the basic design principle has not changed in the terms of how the technical requirements are shared among the different elements and materials used in this construction. Apart from avoiding premature ageing of the housing, it has also become evident that it is critical to protect the fiberglass reinforced (FRP) rod and its interface with housing against moisture. This will prevent failure by insulator puncture (also called ‘flashunder’), which can result in complete mechanical disintegration and dropped conductor.

Based on this, it is now clear that reliable long-term performance can only be achieved if a polymeric insulator is designed to have:

• a consistently reliable and high-quality manufacturing process;

• adequate design that reflects its service conditions;

• appropriate choice of housing material to match service and environmental conditions; and

• functional, reproducible quality of interfaces in terms of long-term adhesion.

Service Experience & Laboratory Testing

Service experience has confirmed that poor adhesion along the rod-housing interface is a key trigger for development of internal tracking. This will eventually result in insulator puncture should the tracking path achieve sufficient length that the remaining unaffected length can no longer withstand an overvoltage or even line-to-ground voltage. This conclusion has only been reinforced by years of experience with failures during design testing of polymeric insulators. For example, years of laboratory testing have revealed that about 10% of polymeric insulators fail during design and type testing. Fig. 1 shows how these failures have been distributed among the different design tests.

As evident, insulators fail mainly during interface testing either by a steep-front impulse test (41%) or by the 30-minute power frequency withstand test (14%), i.e. 55% of all types of insulator failure under laboratory testing conditions relate to this interface. Moreover, insulator samples that failed during the water diffusion test also indicate poor housing-to-rod bonding rather than low rod quality. That means up to 60% of the polymeric insulators being tested under laboratory conditions fail due to issues with interface quality. Clearly, based on service and testing experience, bonding quality represents the major determinant with respect to failure of polymeric type insulators.

Special Water Absorption Test Sequence

Laboratory studies were conducted with the goal of setting up a special test sequence. This sequence included pre-stressing to simulate water absorption in an accelerated manner through the bulk of a polymeric housing material as well as moisture propagation in the interfaces, i.e. along an FRP rod with areas of poor or no housing-to-rod bonding.

The special thermal cycling water absorption test sequence consisted of the following:

• Tap water was used and heated in a vessel to approx. 54°C;

• Temperature was kept at that level for one hour;

• Water was cooled to room temperature (approx. 24°C), i.e., one thermal cycle consisted of a heating phase and a cooling phase with total duration 8 to 9 hours;

• Cycles were continuously repeated for a specific period;

• Finally, a 30-minute AC withstand voltage test acc. to IEC 62217 was conducted with temperature of the insulator shank of each sample monitored using an infrared camera. Maximum temperature change was recorded.

The main goal was to compare behavior of insulator samples under specific pre-stress conditions under different numbers of thermal cycles in water. This would simulate moisture absorption phenomena through the housing materials and along FRP rod-to-housing interfaces. The different test objects included special test samples, insulator samples from among those provided for regular design testing and samples removed from service. All insulators tested were manufactured with silicone rubber housing material, either high-temperature vulcanizing (HTV) silicone or liquid silicone rubber (LSR).

1. Insulator Samples with HTV & LSR Housing Materials: Case Study 1

Special polymeric long-rod insulator samples with HTV silicone housings were made for testing purposes to simulate different qualities of housing-to-rod bonding. The samples tested either had no chemical bonding or 100% chemical bonding.

Next, two LSR long-rod insulator samples made for the purpose of design testing were also included in the test scenarios. Those samples were expected to have little to no housing-to-rod adhesion based on experience with previous samples from that same production batch. None of these samples had earlier been subjected to any design or type testing.

The following test procedure was applied to all samples:

• Initial testing done acc. to IEC 62217: Test samples were boiled for 42 hrs. Then, the steep-front impulse test was conducted. The 30-minute AC withstand test was also performed on samples with temperature measurements before and after the test;

• A ‘rest’ period of one month to allow moisture to evaporate from the silicone housings;

• A special thermal cycling water absorption test sequence, i.e. after application of different numbers of thermal cycles (e.g. 65, 96, 121) samples were removed from the vessel and subjected to steep-front impulse testing as well as the AC withstand test with measurement of temperature;

• Peel testing was done on selected samples.

Table 1 summarizes test results.

All samples passed initial testing. Those samples with LSR housings (expected to have little or only poor bonding) showed higher temperature rises after the initial 30-minute AC withstand test (i.e. 16.6°C and 19.5°C). Nonetheless, all samples still passed criteria given in IEC 62217 in regard to interface testing.

The temperature rise in the LSR samples also increased during application of the water absorption test procedure. Then, sample #LSR2 failed the AC withstand test due to internal puncture. Temperature rise was above 37°C.

A peel test was performed on samples #LSR1 and #LSR2 after 65 thermal cycles of water absorption and after the AC withstand test. As expected, no bonding was observed on either sample. Moreover, after internal flashover, sample #LSR1 showed carbonized tracks along the rod-to-housing interface due to lack of bonding. Sample #LSR2 also showed carbonized tracks due to internal electrical discharges during the AC withstand voltage test (see Fig. 2).

Samples #2B0 (i.e. with no chemical bonding) and #5B1 (with 100% chemical bonding) both passed all tests. But sample #2B0 showed a slightly higher temperature rise after application of several thermal cycles of the water absorption procedure. Clearly, more time was needed for moisture ingress to reach the FRP rod through the bulk of the HTV housing material.

2. Insulator Samples with HTV & LSR Housing Materials: Case Study 2 

Special samples with HTV housing and without any chemical bonding between rod and housing were used. Moreover, these samples were manufactured either with or without chemical sealing at the triple point, i.e. sample #92B0 was made with chemical sealing at the triple point while sample #93B0 was made without chemical sealing at that point (see Fig. 3).

The samples with LSR housing were from among those supplied by a customer for standard design testing for product qualification according to IEC 61109 and IEC 62217. Table 2 summarizes test results.

Fig. 3 shows special insulator samples with HTV housings after 250 of thermal cycles.

Different behavior was observed for insulator samples with LSR housing and HTV housing materials with respect to rate of moisture ingress and propagation along their rod-to-housing interfaces. For example, the test sample with HTV housing, no bonding but with sealing at the triple point (#92B0) showed no notable temperature increase at the shank and passed all criteria given in IEC 62217 at each stage of testing (see Fig. 4). Obviously, the mechanical bond between FRP core rod and housing created by the housing’s shrinkage during the curing process was sufficiently strong to prevent moisture propagation along the FRP rod-to-housing interface

The sample with HTV housing, no bonding and with no sealing at the triple point (#93B0) recorded a high temperature rise at the shank and failed the criteria in IEC 62217 at each stage of testing (see Fig. 5). This was due to lack of sealing and likely weaker mechanical adhesion of the housing to the FRP rod.

Sample #LSR4 (with LSR housing) recorded a temperature increase of more than 10°C after 35 thermal cycles (312 hours) and 55 thermal cycles (492 hours) but still passed the criteria in IEC 62217, meaning Δt < 20 K. Subsequently, it punctured after 155 thermal cycles (1395 hours) and a critical hot spot was localized at its middle (see Fig. 6).

Peel testing was subsequently conducted on intact sample #LSR3 and showed good chemical bonding between rod and housing (see Fig. 7).

Peel testing on sample #LSR4 showed virtually no chemical or mechanical bonding of the housing to the FRP rod (see Fig. 8).

Summary of Case Studies 1 & 2

A special water absorption test method was used to investigate quality of the FRP rod-to-housing chemical bond in polymeric insulators. Different behavior was observed for samples with LSR housing and HTV housing materials with respect to rate of moisture ingress and propagation along the rod-housing interface.

Samples with LSR housing and poor or no bonding failed by puncture after 155 thermal cycles (1395 hours). Most significant was that samples with no bonding still passed the criteria in IEC 62217 (and also IEEE C29.11) which specify that maximum temperature rise of the housing surface must be below 20°C (K). That value is clearly too high and does not guarantee samples without sufficient chemical or mechanical bonding will be detected during testing. By contrast, samples with good bonding showed only minor temperature rise at the housing surface i.e. no moisture propagation along the interface could happen.

Test samples with an HTV housing, no bonding but with sealing at the triple point showed no notable temperature increase of the shank and passed criteria given in IEC 62217 at each stage of testing. But samples with HTV housing, no bonding and with no sealing at the triple point showed a high temperature rise of the shank and failed criteria in IEC 62217 at each stage of testing. This was due to lack of sealing and probably weaker mechanical adhesion of the housing to the FRP rod.

It is noteworthy that samples with HTV housings and no chemical bonding (i.e. no primer applied) showed much better adhesion to the rod than samples with LSR housing and no bonding. Apart from different water absorption behavior, this may also relate to different processing and curing conditions. Whereas HTV silicone rubbers are usually cured at temperatures of about 160 to 180°C, they experience higher thermal shrinkage than LSR rubbers which are usually cured at 100 to 120°C. As such, higher mechanical pressure to the FRP rod can be achieved for HTV samples, which helps reduce risk of water accumulation at the interface. Fig. 9 shows HTV samples with and without chemical bonding after peel testing. Even without applied primer, test samples still showed high mechanical adhesion to the rod, which would result in high tear strength and low leakage currents during the water diffusion test.

3. Performance of Samples with HTV Housing from Service: Case Study 3

Several long-rod polymeric insulators exhibiting some deterioration in their housing surface were removed from a 400 kV line after 8 to 10 years in service and two full-scale insulator samples were taken for testing. Three test specimens were then cut from each sample representing a top part (i.e. grounded side in service), a middle part and a bottom part (i.e. energized in service).

Test specimens were subjected to:

• Modified test on interfaces according to IEC 62217 with specimens boiled for 100 hours, followed by the 30-minute AC withstand test with surface temperature monitored by an IR camera;

• Rest period of one month to allow moisture evaporation from the silicone housing;

• Special water absorption test sequence (8 cycles / 72 hours), followed by the 30-minute AC withstand test with surface temperature monitored using an IR camera.

The same test specimens were used for these test procedures. Table 3 summarizes findings.

Figs. 10, 11, 12, 13 show thermal images of the samples after the boiling procedure and special water absorption pre-stressing (8 cycles) at 15 minutes and 30 minutes of AC withstand testing.

Peel testing at the top parts of samples #I1A & #I1B was conducted to check level of adhesion of the silicone housing to the fiberglass core (see Fig. 14).

Summary of Case Study 3

Investigation of insulators removed from service revealed the following:

• Maximum temperature increase of shanks was below 10°C (K) with regard to the modified test on interfaces according to IEC 62217 with boiling procedure applied. i.e. samples passed testing;

• Maximum temperature increase does not vary, comparing measurements at 15 and 30 minutes of AC withstand testing;

• Maximum temperature increase of the shanks was below 10 K in regard to the modified test on interfaces according to IEC 62217 with boiling procedure applied. i.e., samples passed testing.

• Maximum temperature increase of the shank of sample #I1B was higher than 25°C (K) when applying the special water absorption procedure (8 cycles). i.e., samples failed evaluation criteria according to IEC 62217;

• Maximum temperature increase varies if comparing measurements done at 15 minutes (45°C) and 30 min (35°C) of AC withstand testing;

• Peel testing showed reduced level of adhesion of housing to core for sample #I1A. For sample #I1B, almost no adhesion between housing and core was detected.

Long-Term Water Absorption Stressing of Insulator Samples with HTV Material

Long-term behavior was also studied of samples with HTV housings under specific pre-stressing conditions. Samples #1B0 and #2B0, without chemical bonding between FRP rod and housing, were subjected to long-term thermal cycling water absorption testing. Table 4 summarizes maximum temperatures recorded during 30-minute withstand voltage testing after different numbers of cycles.

Fig. 15 charts maximum temperature rise of the housing surfaces of samples #1B0 and #2B0 during 30-minute withstand voltage test after particular numbers of thermal cycles.

Samples showed a temperature increase in their housing surface over time (i.e. with increasing number of applied thermal cycles) that was proportional to the amount of moisture that had penetrated through the silicone material.

Sample #1B0 showed much slower moisture propagation along the housing-to-rod interface than did sample #2B0, probably due to a different level of mechanical adhesion of the housing to the rod. Peel testing was not carried out since samples were to be used for further water absorption testing.

TECNALIA Electrical Labs

Spain

ICMET Craiova – National Institute For Research, Development And Testing In Electrical Engineering

Romania

See more Laboratories

Conclusions

A special thermal cycling water absorption test sequence was investigated to verify the quality of housing-to-rod bonding in polymeric insulators. Different behavior was observed for insulator samples with LSR and HTV housing materials with respect to rate of moisture ingress and propagation along the rod-housing interface.

Samples with LSR housings and poor to no bonding failed by puncture after a certain number of thermal cycles applied. Given this, criteria specifying a maximum temperature rise of housing surface to be below 20° K, as per IEC 62217 and other standards, is too high and does not guarantee that samples with insufficient chemical and mechanical bonding will be detected.

HTV housed samples with no bonding but with sealing at the triple point showed no notable temperature increase of the shank and passed criteria given in IEC 62217 at each stage of testing. This was because the mechanical bond between FRP core rod and housing due to housing shrinkage during curing was strong enough to prevent moisture propagation along the rod-housing interface. By contrast, HTV housed samples with no bonding and with no sealing at the triple point showed a high temperature rise of the shank and failed the criteria in IEC 62217 at each stage of testing. This was likely due to lack of sealing and probably weaker mechanical adhesion of the housing to the rod.

Polymeric long-rod insulator samples with HTV housing taken from service passed the modified interface testing with the boiling procedure involved. But samples with poor bonding failed AC withstand testing after the special thermal cycling water absorption procedure.

Differences in quality of bonding of samples with HTV and LSR housing must also be taken into account in view of the different silicone material formulations and manufacturing processes used in industry. HTV materials are usually cured at temperatures of approximately 160 to 180°C and therefore a higher thermal shrinkage to the rod.

A 30-minute AC withstand test with temperature monitoring according to IEC 62217 is a highly useful test method. While leakage current measurements did not show high sensitivity, an improved test circuit with separate resistive and capacitive parts of leakage current would help.
Simulating the failure mechanism of in-service polymeric insulators having no or very poor bonding is difficult since this takes place over a long-term when insulators are energized. Regarding insulators with HTV silicone housing materials, longer water absorption pre-stressing might be needed to reach an adequate moisture ingress level such as to seriously affect electrical performance.

The special thermal cycling water absorption procedure using tap water is more severe than the boiling procedure given in IEC 62217. This is because this procedure enhances capability for water to penetrate through the bulk of the silicone material due to its lower specific conductivity and relative permittivity. In the case of a high temperature increase on the insulator shank, i.e., high real current flowing through the housing-rod interface, moisture evaporation starts about 15 minutes after AC withstand testing. This might result in reduction of the shank’s temperature in the testing interval between 15 and 30 minutes.

As such, when evaluating quality of bonding of a housing to the FRP rod, it is important to consider polymeric material type and properties, chemical bonding and also mechanical bonding. Future work will concentrate on further studies of moisture ingress phenomena through silicone housing materials as well as through the triple point when there is improper sealing. Quantification of thermal changes, related to real current flow through an insulator, will also be studied in more detail.

Better understanding different possible levels of bonding quality for polymeric insulator interfaces will help improve evaluation of present condition as well as predicting future service performance. IEC 62217 is currently under revision.

References
[1] IEC 62217:2012 Polymeric HV insulators for indoor and outdoor use – General definitions, test methods and acceptance criteria.
[2] Lachman, J.: “Lessons from 25 Years’ Experience Testing Polymeric Insulators” INMR World Congress 2019, Tucson, USA.
[3] Baer, C., Schmuck F., Strumbelj, J. Tinner, E., Lachman, J., Kornhuber S., LOH, J. T.: “Technical Demands to Improve Today`s Polymeric Insulator Reliability” Cigre General Session 2020, 2021 and 2022, Paris, France.

The post Testing Sheath-to-Rod Bonding to Assure Long-Term Performance of Polymer Insulators appeared first on .