São Paulo, one of the most densely populated metropolitan regions in the world, presented extreme challenges for deployment of new high-voltage infrastructure due to saturation of urban space, complexity of existing underground utilities, and unavailability of rights-of-way for overhead transmission.

The post Challenges Designing Underground HVAC Transmission Line in Brazil’s Largest City (Video) appeared first on .

Traditional inspection methods are periodic, labor-intensive, expensive, and often fail to detect insulators at imminent risk of failure. Strategies to detect incipient failure on transmission line infrastructure need to improve.

The post Applying IoT Devices, AI & Machine Learning to Predict Failures on Remote Transmission Lines (Video) appeared first on .

As discussed in this edited contribution to INMR by R. F. Dickson, L. Porras, D. Aranguren and S. Mejía, an expert system was developed, integrating modules for information, gathering real-time monitoring and prediction. The monitoring and information modules generate near-term failure predictions using lightning characteristics such as density, current, polarity, and multiplicity.

Wish Power (Thailand) Co. Ltd

Thailand

Hivolt Power System

China

See more suppliers of Cutouts

This lightning monitoring and prediction tool enhances the supervision and control of ISA Intercolombia’s transmission assets. It boosts situational awareness at the Control Center, helping engineers respond more effectively to lightning-caused outages. Faster response times can reduce service interruptions, improve system reliability, and lower annual asset unavailability—ultimately benefiting the company’s operational performance and profitability. When a risk alert is issued, the Control Center can proactively assess system conditions, implement mitigation strategies, and coordinate restoration efforts.

The tool also aids decision-making by providing real-time diagnostics of failure causes and weather conditions during outages. This allows for immediate action to restore service or conduct deeper analysis using oscillograph data, especially for failures not caused by lightning. Since lightning is a leading cause of transmission line outages, the system indirectly helps monitor other risks like wildfires or vegetation interference. Moreover, the system supports right-of-way maintenance by warning personnel about lightning threats, helping them decide whether to proceed with or to postpone work. The insights generated also support engineering and project teams in designing new transmission lines.

Transmission System

Colombia’s power transmission system is composed of various power utilities, including ISA Intercolombia, which operates approximately 12,000 km of transmission lines nationwide. These lines are segmented by different voltage levels, as shown in Table 1.

Considering that one of the most important duties of power utilities is to ensure the reliability and availability of assets, it is essential to evaluate all possible causes that could affect their operation. Based on historical data, Fig. 1 describes the four most recurrent causes of transmission line failures over a 10-year period (2014–2024).

According to the previous figure, failures caused by lightning account for 79% of all transmission line failures and with respect to lightning activity, the Table 2 shows the number of lightning strokes detected in the total Colombian territory (1,142.000 km2) varying from 23.9 million in 2021 to 46.7 million in 2023. Lightning incidence presented an increase during the four-year period with a very high value reported in 2023 mostly influenced by the La Niña Phenomenon that marked the prevailing raining conditions in that period. As a reference parameter, maximum peak currents are also given in Table 2. In particular, the number of strokes within a 10 km buffer around the ISA´s power lines vary from 2.99 million in 2021 to 8.5 million in 2023.

As described, the power line influence area´s strokes vary from 12.5 to 18.2% of the total lightning activity detected in the entire territory during the last four years. Regarding the magnitude of lightning currents within the 10 km buffer around transmission lines (Table 2), the maximum discharge current has been increasing over the past four years, reaching a peak of 369 kA in 2024. It suggests not only a rise in the number of lightning strokes, but also an apparent increase in their maximum current, meaning that each year this phenomenon could pose a greater threat to transmission systems.

Fig. 2 provides the number of failures caused by lightning over the same four-year period (2021–2024). Despite the variations observed in the number of lightning strokes over the years, lightning-related faults tend to remain stable each year except in 2024, which showed a decrease. This reduction may be due to improved classification of fault causes in transmission lines, resulting from better data interpretation enabled by analytics and enhanced accuracy of the lightning information system.

Given the annual behavior of both lightning and the resulting failures, a deeper analysis per month is illustrated in the Figs. 3 and 4 to identify stronger correlation between these variables.

Although, as previously mentioned, 2023 experienced an abnormally high number of lightning strokes, it can be observed that the behavior of failures tends to follow the peaks and valleys of lightning activity over months. This indicates that lightning-related faults are, in some measure, influenced by lightning density. In addition, Fig. 4 provides a detailed overview of the months exhibiting elevated faults, primarily driven by lightning strikes. The most impacted months are April, May, and October.

Figs. 5 and 6 present complementary perspectives on lightning activity and its impact on electrical infrastructure. Both figures cover the same period and circuits, yet they emphasize different trends. Fig. 5 illustrates the distribution of lightning strikes across the electrical circuits with the highest failure rates, while Fig. 6 presents the failure rates of these same circuits, normalized per 100 km per year.

A comparison of the two figures reveals that a high density of lightning strikes does not necessarily correlate with higher fault rates. This observation suggests that other factors—such as the quality of grounding systems, the presence of surge arresters, tower and insulation design, and the effectiveness of lightning detection and monitoring systems—play a significant role in mitigating lightning-related faults.

Fig. 6 further indicates that the transmission line “500_Circuit1,” which is the most susceptible to lightning-related faults, operates at 500 kV and is situated in a region with a high lightning strike density—reaching a peak of 55 flashes/km²/year, as shown in Fig. 7. Additionally, it is observed that some 230 kV lines, despite experiencing fewer lightning strikes, exhibit failure rates comparable to those of the 500 kV lines. This suggests that these 230 kV lines may possess specific vulnerabilities—such as insulation deficiencies—that increase their susceptibility to faults. As a result, these lines should be closely monitored and evaluated for potential improvements.

Lightning Detection System

The analysis based on lightning data from the Colombian Total Lightning Detection System (CTLDS), illustrated in Fig. 1, has improved the precision in determining the true causes of failures. The CTLDS comprises 24 sensors, and its effectiveness has been reported in multiple studies, in accordance with the evaluation procedures described in IEC62858.

The system achieved an average location accuracy of 0.182 km and a detection efficiency of 97% for lightning-caused failures in power lines. These results meet the IEC 62858 standards, which recommend a detection efficiency above 80% and a location accuracy within 500m. Additionally, Betz [6–7] has presented performance evaluations of LINET (LIghtning NETwork) technology and comparisons with other lightning detection systems worldwide.

Efficiency of Lightning Detection Systems

The effectiveness of the Lightning Location Systems (LLS) largely depends on factors such as the number of sensors, the distance between them (baseline), and their geographic distribution across a region. Detection Efficiency (DE) is generally expected to be high when a lightning flash is captured by at least six sensors positioned in different directions. However, in real-world scenarios, DE can be influenced by additional factors such as complex terrain, ground conductivity, local noise, and physical obstructions (CIGRE TB376 WG c4.404.

To assess variations in DE across the coverage area, Relative Detection Efficiency (RDE) methods are used. These methods help identifying regional deviations in detection performance. DE maps, whether theoretical (based on network layout) or empirical (based on peak current data), reveal actual performance differences that may impact failure detection (Figs 9 & 10). As a result, RDE calculations are integrated into failure correlation processes and are crucial for timely and accurate failure analysis and reporting.

Detection Efficiency is of crucial importance on the correlation criteria since when the detection efficiency is higher than 95%, the correlation is precise in both time and location. Nevertheless, when the detection efficiency is lower than 80%, a miscorrelation may be obtained. To address this limitation, satellite-based lightning mapping, such as that provided by the Geostationary Lightning Mapper (GLM), is employed. Due to its measurement principle based on optical sensing is highly affected by atmospheric conditions, a systematic combination of both ground-based and satellite lightning detection systems is proposed to obtain reliable failure analytics even when DE is reduced.

Therefore, the correlation module integrates two advanced lightning detection systems to enhance fault cause analysis. The first one is based on LINET (Lightning Detection Network) technology, a high-precision ground-based system. The second system is the Geostationary Lightning Mapper (GLM), an optical sensor onboard NOAA’s GOES satellites. GLM operates from geostationary orbit and continuously monitors total lightning activity.

This correlation method, which consists of two lightning detection systems, sometimes supports the correlation of failures that are not caused by lightning, such as those resulting from vegetation interactions with overhead conductors. A real example will be illustrated later.

Failure Cause Estimation Tools for Transmission Lines
With the ongoing energy transition and rapid expansion of transmission networks, current power reliability standards require faster fault detection, reporting, and restoration. This means that when a transmission line goes out of service, it’s essential to accurately determine the root cause of the failure. Lightning Detection Systems support real-time decision-making by enabling a quick online assessment of whether lightning was responsible and pinpointing the strike location at the time of the outage, allowing for immediate corrective action. Given that around 79% of transmission line failures in Colombia are due to lightning, high-precision total lightning detection systems are crucial. They make it possible to identify the specific tower affected and confirm the true cause of the fault.

At ISA INTERCOLOMBIA, a Substation Automated Equipment Management System (SAGES) plays a critical role in supporting real-time operational activities (see Fig. 11). This system automatically retrieves fault records from both substations associated with a transmission line following a disturbance. Utilizing this data, SAGES executes fault location algorithms and computes the corresponding fault impedance. These impedance values, along with other critical fault characteristics and data from the lightning correlation module, are then processed through a decision matrix. This integrated analysis enables the system to accurately identify the root cause of the fault, improving diagnostic accuracy and supporting timely restoration efforts.

Lightning Correlation Module

The correlation method (as described in Fig. 12) consists in processing time of failure with a precision of milliseconds and using a correlation area expanded to the complete line extension from SUBSTA to SUBSTB. With this method, the failure location from one line end is not used, which minimizes the correlation processing time and avoids the errors involved in the location calculation.

The correlation assessment is based on an index denoted as SCORE, given by:
SCORE=Y₁ (t)×Y₂ (R)                   (1)
Y₁ (t) and Y₂ (R) are time and distance functions, given as follows:

where,
Y₁ (t) is the SCORE time function with a range of [0, 1].
Y₂ (R) is the SCORE distance function with a range of [0, 1]
t is the absolute time difference between the failure and the stroke detection time in seconds.
R is the radial distance from the stroke location to the transmission line in km.
σ_t y σ_R are smoothing parameters for both time and distance differences, which are defined based on experimental errors from real correlations over the Colombian power transmission system (Time error of 600 ms and radial distance error of 3.5 km for exact correlations – Table 4), when comparing lightning data with protection relay data.

SCORE assessment is performed as given in Table 3:

According to Table 3, SCORE values in the range [1.00, 0.951] indicate an exact correlation (highlighted in red). Values between [0.95, 0.80] suggest a highly probable correlation (highlighted in orange), while those in the range [0.78, 0.60] represent a probable correlation (highlighted in yellow). Values below 0.59 indicate that correlation is not possible (highlighted in green).

In this method, the correlation determining variable is the time absolute difference t, since it takes advantage of the GPS synchronization that exists between the lightning detection network and the protection relays together with the Supervisory Control and Data Acquisition (SCADA) system, which provide the control center personnel with the sequence of events (SOE) that occur in the power system, with a time stamp in milliseconds.

In this way, during a transmission line failure, the control center performs correlations using the time stamp of the first signaling of the SOE protection operation, which is instantaneous during the occurrence of a failure. In most of the times, the time errors between the SCADA systems, the electrical protections and the lightning detections are less than 100 ms. Radial distance errors, which are the sum of the tower geographic error and the stroke location errors (mean stroke location error less or equal to 500 m) is approximately 3.5 km for exact correlation (see Table 4). This correlation method allowed improving the failure-cause classification for transmission lines. Given its accuracy, it is a very important input for the root cause analysis applied to substation equipment and transmission lines.

Processing & Reporting

A reporting module allows querying the lightning activity over a transmission line (within a perimeter of 5 km, along the entire length of the circuit), a substation or a specific geographic coordinate. Failure date and time (with a resolution in milliseconds) are the input data, with the selection of a given affected asset (line or substation), with several output modes such as e-mail or visual interface. Reports are generated in standard formats (such as PDF) and are automatically delivered to operators and decision makers. A correlation example is given in Fig. 13.

In Fig. 13,
• Date: failure timestamp with precision in milliseconds
• Current: Stroke peak current, in kA (ND: Not determined, due to this stroke was detected by GLM).
• Dist (R): Perpendicular distance from the stroke location to the affected transmission line, in km.
• Dist (SUBST A): Distance from Substation A to the failure location, in km.
• Dist (SUBST B): Distance from Substation B to the failure location, in km.
• Tower: Closest transmission tower to the failure stroke
• Score: correlation index used to evaluate the correlation quality

As described in the example of Fig. 13, a strong correlation was found between the failure time and location with a flash detection composed by a first stroke with a peak current of 145.8 kA followed by a series of subsequent strokes within a flash duration of 236 ms detected in a distance range from 370 to 3360 m. The SCORE in this case provided a result of 1 corresponding to a “Exact” correlation according to the SCORE sensitivity evaluation of Table 4.

Using Geostationary Lightning Mapper (GLM) Data to Identify Vegetation-Induced Faults
A real case in the Colombian transmission system is illustrated below, where a 500 kV circuit failed on Jan. 6th at 13:30 due to arcing through vegetation. This failure was uniquely correlated using GLM, with the following characteristics:

• No rainfall or thunderstorms were observed in the area surrounding the circuit within a one-hour time window (Fig. 14).
• No correlation was detected using LINET (Lightning Detection Network) technology, a high-precision ground-based system.
• The correlation using GLM was exact (Figure 15).
• The fault location algorithm, based on measurements from both ends of the transmission line, yielded 47.23 km from Substation A—similar to the 47.35 km reported by the GLM system.
• Failure analysis based on fault records indicated a high-impedance fault (10 Ohms) with characteristics consistent with vegetation interference.
• After a transmission line inspection, a Guadua—native bamboo species—was found very close to the faulty conductor, being identified as the cause of the failure (Fig. 16).

Laboratory Center of Koncar Electrical Engineering Institute

Croatia

Powertech Labs Inc.

Canada

See more Laboratories

Real-Time Supervision
Real-time supervision module is a graphic interface that monitors lightning activity over the transmission lines (see Fig. 17). Lightning data from the CTLDS feeds the systems to perform a variety of analytic tasks based on the Dynamic Stroke Density Method introduced by Pérez et al. as a new dynamic risk function for power transmission lines.

The Dynamic Stroke Density DSD during an active thunderstorm with respect to a specific position x of the power line is given by eq. (1) where N is the number of strokes in a given period of time, K_i (s) is a gaussian kernel function for distance and F_i (I) is a sigmoid function for the stroke peak current.

The Gaussian kernel function K_i (s) is given by eq. (2) where σ_s is a “smoothing factor” of distance and s is the distance from a given stroke to the position x.

The sigmoid function for the stroke peak current F_i (I) is given by eq. (3) where I is the stroke peak current, I_BFO is the critical peak current for back flashover of the line position x and σ_c is a “smoothing factor” of current. The critical peak current for back flash over can be estimated by its simplified form I_BFO=3CFO/((1/3)Z_T+(2/3) R_T ) where CFO is the critical flash over voltage of the insulation at the position x, Z_T is the surge impedance of the tower and R_T is the ground resistance.

Note that the previous formulation supposes that the higher the lightning peak currents detected in an approaching thunderstorm, the higher the risk, since typically most transmission line failures are due to back flashovers caused by direct lightning strikes overhead ground wires with very high peak currents; further considering that once a high current stroke has been detected, there is a higher probability that another high current stroke will occur again. In addition, previous assumptions result more applicable for high voltage systems, i.e. nominal voltage higher than 220 kV, and needs to be re-formulated for lower voltage ones. In any case, the influence of the detected stroke peak currents will be reflected in the values taken by σ_c.

The DSD for the complete path of a transmission line can be defined as the integral of the function given in (1) along the total length of the transmission line. In discrete form and considering ∆l as a constant span between subsequent transmission towers, DSD is given by (4) where N is the number of strokes in a period of time, M is the number of line segments, ∆l is the length of the line segment (span between subsequent towers), s_ij is the distance between the stroke i and the segment j, I_i is the peak current of the stroke i and I_BFOj is the critical current for back flashover in the segment j. For simplicity, eq. (4) is written in terms of the gauss error function erf.

DSD behavior during a real lightning disturbance event in a transmission line is illustrated in Fig. 18. In this case, a 500kV transmission line with 415 towers and 212 km in central Colombia reported a failure at 21:57 local time on 21 July 2018.

The DSD risk function is used to estimate the power failure risk by considering both the lightning hazard time-space dynamic evolution and the line vulnerability, where a simplified thumb rule for situational identification is given by: i) DSD < 0.5 (Low failure probability), ii) 0.5 < DSD < 1 (Medium failure probability), iii) DSD > 1 (High failure probability).

The main functionalities for the real-time monitoring are: i. graphical identification of lightning activity on transmission lines, ii. efficient calculation of failure probability on lines and fast identification of affected assets, iii GIS data structure for precise information deployment about paths, characteristic, physical and electric data relevant for the power line performance, iv. GIS data structure for lightning and meteorological data analyses.

Fig. 19 illustrates the process diagram of the lightning-caused failure forecast for power lines. Detailed tower physical characteristics are assimilated and loaded in a dynamic database (DB-Towers) which is used to compute the vulnerability indexes at every tower site. As previously mentioned, the DSD risk function is based on a gaussian kernel computation supported on the real time lightning detection and the lightning forecast obtained from adapted weather forecast models. Vulnerability parameters and DSD modeling are combined to compute the failure risk at a given tower site. The failure history and continuous monitoring support a learning loop to continuously improve forecast skills.

Fig. 20 describes the method used to evaluate the prediction skill of the previously described model. It uses a ROC curve based on the True Positive Rate – TPR and False Positive Rate FPR according to different DSD thresholds. The prediction skill is given by the Area Under the Curve – AUC as illustrated in Figure 20. As can be noted in the example, AUC of 0.95 for a 230 kV line is obtained whereas an AUC of 0.79 is found for a 500 kV line, both in mountainous conditions.

This method is currently used to evaluate the prediction performance for different power lines under different conditions and to improve the model itself by considering additional predictive variables related to the thunderstorm physical characteristics, orographic conditions, seasonal patterns, constructive characteristics and protection elements of the power line, among many others.

Safety Application for Right of Way Activities
Safety issues related to lightning strikes during right-of-way (ROW) activities in transmission lines are critical due to the high energy and unpredictability of lightning events. Workers during routine patrols are provided with a mobile application, which alerts them about lightning activity nearby.

As depicted in Fig. 21, three warning areas are used to trigger alerts; the smaller one is the Area Of Concern AOC, where a lightning flash has the potential to affect people safety, the medium one is the Warning Area WA, where lightning detections are used to trigger alerts to prevent people in risky activities in AOC, in this case, personnel should suspend activities and begin seeking shelter and, the Coverage Area CA, which is used to monitor close thunderstorms without risk. Red, orange and yellow colors are used to indicate the warning level related to AOC, WA and CA.

Conclusions

This study presents an Expert System designed to analyze data and predict lightning-induced failures in transmission lines, with the objective of enhancing the reliability and operational efficiency of the electric service provided by ISA INTERCOLOMBIA. A key contribution of the system lies in its ability to identify root causes of failures, enabling proactive and informed decision-making to support preventive and predictive maintenance strategies that meet or surpass regulatory quality standards. Ensuring personnel safety and optimizing operation and maintenance costs are also central goals.

The system’s centralized architecture is highly scalable and adaptable, allowing it to be configured for individual transmission lines or other electrical assets based on their specific technical parameters. To support more integrated, secure, and reliable transmission system operations, the platform incorporates advanced correlation techniques and analytical modules for real-time monitoring and short-term prediction (nowcasting within 1 hour). The predictive algorithms are capable of processing large volumes of complex data autonomously, delivering effective performance that is expected to improve further with the development of specialized models tailored to meteorological and electrical dynamics.

References
[1] Aranguren D., Inampues J., López J., Torres H., Betz H. Cloud to ground lightning activity in Colombia and the influence of Topography. Journal of Atmospheric and Solar-Terrestrial Physics.
https://dx.doi.org/10.1016/j.jastp.2016.08.010. 2016
[2] IEC62858. Lightning densities based on lightning location systems (LLS) – general principles. International Electrotechnical Commission, 2015.
[3] Aranguren D., González J., Cruz A., Inampués J., Torres H., Sarmiento P. Lightning strikes on power transmission lines and lightning detection in Colombia. 2017 International Symposium on Lightning Protection (XIV SIPDA), Natal, Brazil, 2017.
[4] Gonzalez J., Aranguren D., Inampués J., Cruz S., Torres H., Gomez J., Arango C., Asenjo F. Lightning Information Management Systems, a Useful Tool to Make Accurate Decisions. Study Case: Chivor – Rubiales Transmission Line. International Conference on Lightning Protection 2018, Rzeszow, Poland.
[5] Pérez E., Espinosa J., Aranguren D. On the Development of Dynamic Stroke Density for Transmission Line for Power System Operational Applications. International Journal of Electrical Power and Energy Systems, Volume 116, March 2020, 105527.
https://doi.org/10.1016/j.ijepes.2019.105527
[6] H. Höller, H.-D. Betz, K. Schmidt, R. Calheiros, P. May, E. Houngninou, G. Scialom. “Lightning characteristics observed by a VLF/LF lightning detection network (LINET) in Brazil, Australia, Africa and Germany”. Atmos. Chem. Phys., 9, 7795-7824, 2009.
[7] Betz, H.-D., “Utilization of Lightning Data for Recognition and Nowcasting of Severe Thunderstorms”, EGU General Assembly, May 02-07, Vienna, 2010.
[8] L. Porras, R. Dickson, G. Fonseca, and D. Aranguren, “Supervision and Forecast of Lightning Threat on Transmission Lines,” presented at the CIGRE Session, Paris, France, Paper 11736, 2024.
[9] R. Dickson and D. Aranguren, “Performance of Lightning Detection Systems for the Identification of Failure Causes in Transmission Lines,” presented at the CIGRE Session, Paris, France, Study Committee C4 – Power System Technical Performance, PS3, Q3.4, Oral contribution, 2024.

The post Analyzing Impact of Lightning on Overhead Transmission: Experience in Colombia appeared first on .

High temperature vulcanized silicone high voltage insulators are known for exceptional hydrophobic properties such that water droplets stand separately over the surface and do not form a continuous film.

This edited past contribution to INMR by the Dept. of Applied Sciences at the University of Quebec in co-operation with K-Line Insulators reviewed opportunities to further improve pollution flashover behaviour.

Listen to Online Lecture on Super-Hydrophobic Silicone for High Voltage Insulators by A.J. Carreira

Super-hydrophobic surfaces offer a water contact angle (WCA) >150˚ and a contact angle hysteresis (CAH) or sliding angle (SA) <10˚. This feature has attracted significant attention in a wide range of applications requiring anti-corrosive, ice-phobic, anti-biofouling, non-wetting and self-cleaning surfaces. Inspired by the unique self-cleaning of lotus leaves, also known as the “lotus effect”, this property is being widely used in industrial applications. Since pollution particles are generally hydrophilic, they tend to adhere to water droplets rather than to a super-hydrophobic, self-cleaning surface.

Two main approaches are used to produce a super-hydrophobic surface: 1. roughening the surface of a material with low surface energy and 2. depositing a low surface energy material on an already rough surface. Silicone rubber is a low surface energy material and it can be roughened sufficiently to produce a super-hydrophobic surface. A self-cleaning silicone rubber surface is capable of repelling water droplets and thereby removing contaminant particles that adhere to it, i.e. any dust and pollution is easily removed as water droplets roll off its super-hydrophobic surface. Therefore, fabrication of a super-hydrophobic surface having self-cleaning properties can effectively address issues associated with contaminated insulator surfaces.

Although several methods are available to produce super-hydrophobic surfaces, direct replication was selected since it is simple, efficient and inexpensive and can also easily be industrialized. Using this method, the surface does not undergo chemical changes. Rather, an appropriate micro-nanostructure is created on the surface of a negative replica. A positive replica is then produced by means of injection molding (see Fig. 1). The super-hydrophobic silicone rubber surfaces thus created demonstrate ultra water repellence as well as self-cleaning properties under different contamination conditions.

Assessment of Super-Hydrophobic Properties

The smooth sample of silicone rubber originally had a WCA ~116º ± 2º and CAH >30º (see Fig. 2a & c) whereas the new replicated silicone surface showed a WCA ~161º ± 1.3º and CAH ~7º (see Fig. 2b & d). Due to the low CAH, water droplets on the super-hydrophobic surface easily rolled off, demonstrating a self-cleaning capability.

Micro-nanostructures on the silicone rubber surface allow achieving such high WCA and low CAH. Creation of such micro-nanostructures after the direct replication process on the silicone surface has been studied using SEM analysis (see Fig. 3).

Ultra Water Repellent Silicone Rubber Surfaces

To demonstrate super-hydrophobic properties, a sample of micro-nanostructured silicone rubber surface was immersed in a water-filled Petri dish. The super-hydrophobic silicone surface remained dry while the height of the water reached up to ~3 mm (the so-called Moses Effect). The blue water droplets demonstrate dryness of the surface.

Water jet tests serve to confirm ultra water repellent properties of a surface. Here, the water jet was applied by a syringe equipped with a needle, under normal force, on both hydrophobic and super-hydrophobic samples placed on a flat surface. When the jet was sprayed on the smooth silicone rubber, water accumulated on its surface (see Fig. 5a).

But when a water jet was applied to the super-hydrophobic silicone rubber surface, it rebounded (see Fig. 5b). This demonstrates that a current of high-pressure water cannot remain on the surface of micro-nanostructures, i.e. the super-hydrophobicity of the manufactured surface is stable.

The behaviour of a surface exposed to continuous impact of water droplets is representative of water repellence and this behaviour can be perceived when surfaces are exposed to rainfall. In such a test, water droplets of certain diameter are released from a height and impact silicone rubber surfaces placed on 10˚ tilted stand. As seen in Fig. 6a, two consecutive water droplets adhered to the smooth silicone rubber surface upon impact.

However, water droplets rebounded off the super-hydrophobic silicone rubber surface without leaving a residue (Fig. 6b). It is worth noting that formation of a secondary or satellite droplet was also observed in the case of the manufactured super-hydrophobic surface and further emphasizes its ultra-water-repellent property.

Self-Cleaning Silicone Rubber Surfaces

A super-hydrophobic silicone rubber surface with ultra water repellent properties and low CAH is self-cleaning. This self-cleaning property was compared to that of the smooth normal silicone rubber surface by adding the same amount of fine kaolin particles as contaminants to each surface.

As seen in Fig. 7a, a water film forms on the smooth surface with kaolin. By contrast, water droplets readily clean the surface of the super-hydrophobic silicone rubber along their flow path. As seen in Fig. 7b, only several water droplets are sufficient to clean the surface.

A dirty multi-component suspension of SiO2 particles, salt, carbon black and glycerine was added to water to evaluate functionality of a super-hydrophobic surface under polluted wet conditions. The hydrophobic and super-hydrophobic samples were immersed in equal amounts of suspension and then maintained at ambient temperature of 70˚C for 2 h. Figs. 8a and b show the surface appearance of each sample after water evaporation. While the surface of the smooth silicone sample is completely covered by the dirty sediment, most of the super-hydrophobic silicone rubber surface remains clean.

After cleaning each surface using the same amount of de-ionized water, it was observed that water accumulated on the smooth surface, forming a continuous film (see Fig. 8c). However, the surface of the super-hydrophobic surface readily cleaned off remaining dirt sediments. Such remarkable self-cleaning behaviour of super-hydrophobic surfaces is highly desirable when insulation materials are being used in highly polluted service areas.

Conclusions

Micro-nanostructured silicone rubber surfaces were manufactured by direct replication using an injection molding process with the goal of helping overcome insulation pollution flashovers. These silicone rubber surfaces showed super-hydrophobic properties, i.e. a WCA >150˚ and CAH <10˚, and also demonstrated extraordinary water repellence, confirmed by immersion in water testing as well as water jet impact and water impact tests. These experiments were beneficial in investigating functionality of super-hydrophobic silicone rubber surfaces exposed to rainfall. Manufactured surfaces also demonstrated self-cleaning properties due to their low CAH. Kaolin powders were easily removed from super-hydrophobic surfaces using water droplets and super-hydrophobic surfaces showed excellent self-cleaning properties when exposed to dirty, wet conditions. Although they accumulated only negligible residue on their surface, the dirt residue was easily cleaned using several water drops. This confirms that super-hydrophobic surfaces can be considered a highly effective option to reduce risk of pollution flashover of insulation in highly polluted service areas. References
[1] http://www.inmr.com/pollution-flashover-insulators/
[2] Maghsoudi, K., Momen, G., Jafari, R., Farzaneh, M., & Carreira, T. (2018, October). Micro-Nanostructured Silicone Surfaces for High-voltage Application. In 2018 IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP) (pp. 179-182). IEEE.
[3] Gençoğlu, M. T., & Cebeci, M. (2008). The pollution flashover on high voltage insulators. Electric Power Systems Research, 78(11), 1914-1921.

The post Self-Cleaning Properties of Super-Hydrophobic Silicone for High Voltage Insulators appeared first on .

This edited past contribution to INMR by Bas Verhoeven of Kema Labs in the Netherlands reviewed findings based on 30 years testing cables and accessories.

Klöckner Desma Elastomertechnik GmbH

Germany

DALEKOVOD OSO

Croatia

See more suppliers of Laboratory & Field Testing Equipment

Cables: Type, Pre-qualification & Commission Testing

International standards for testing at medium, high and ultra-high voltages include 3 distinct groups of tests. The first is type testing intended to verify design and manufacturing. Testing a cable system comprised of the cable and accessories such as joints and terminations (open-air, SF6 or oil plug-in) takes place in a laboratory with various electrical tests followed by heat cycle tests that simulate the loading pattern under normal use. Testing non-electrical properties of cable materials is also part of a type test.

The second group – pre-qualification testing – is only applicable to high and ultra-high voltage cables. After successful completion of type testing, the cable is installed under realistic outdoor conditions that mimic normal installation methods, including being buried directly in the ground and with some sections installed in tunnels. A voltage of 1.7 U_o is applied to the cable and 180 heating cycles are applied, one every 2 days. A pre-qualification test lasts one year.

The third group is commissioning testing of a newly installed cable system. After installation in the field, the cable is energized to e.g. 1.7 U_o for one hour to detect possible errors in workmanship.

The relevant IEC standards for AC cables are:

• IEC 60502-2 for Cable 6kV–30kV
• IEC 60502-4 for Cable Accessories 6kV–30kV
• IEC 60840 for Cable & Accessories 30kV–150kV
• IEC 62067 for Cable & Accessories 150kV–500kV

STL Guides in regard to IEC 60840 and 62067 were issued in 2019 with the goal of harmonizing interpretation and execution of these standards, with focus on pre-qualification tests.

Although international standardization for DC cables is less developed than for AC cables, these types of cable have specific test requirements for the temperature gradient over the XPLE-insulation material and a superimposed DC and LI test. Current standards for DC cables include:

• CIGRE 496
• IEC 62895:2017

Failure Statistics from Type Testing Cables

Manufacturers come to test laboratories such as Kema Labs for a formal type test once they believe their products meet all the requirements set out in the standards. However data on intial failure rate when entering into a type test program, based on more than 1000 cables and accessories tested, indicates a failure rate of about 25%.

Fig. 1 shows failure rate for medium voltage cables and accessories as well as these in the case of high and ultra-high voltages combined. Medium voltage terminations and joints show a high failure rate, due mainly to improper materials used in the heat shrink technology. Moreover, high voltage cables tend to show a higher failure rate compared to the medium voltages. This is because MV cable manufacturers try to enter the HV market due to potentially higher profit margins but often face manufacturing quality issues at higher voltages.

Fig. 2 shows year-on-year initial failure rate for MV and HV cables and accessories. Straight trend lines all have a positive inclination indicating that initial failure rate is increasing over time instead of the normal expectation of decreasing due to improved technology, experience and skill. This increase is likely the outcome of huge market pressure on lowering cost, which risks affecting quality of cables and accessories and eventually resulting in more network blackouts.

Failure modes of cables and accessories vary depending on voltage class and Table 1 provides an overview for both MV and HV.

Newly installed XLPE cables require a commissioning test to verify correct installation before a cable system is connected to the main network. The best way to conduct such a test is application of voltage above nominal system voltage that is maintained for a certain period of time.

For example, the IEC standard recommends 1.7 Uo for one hour. KEMA Labs has tested many newly installed cable systems across Central Europe and Fig. 3 shows the total circuit length of cable tested per year. The red line represents number of breakdowns normalized per 100 km of installed cable circuit. While the long-term average should be only 2 failures per 100 km installed circuit length, there is a peak of nearly 8 breakdowns per 100 km circuit length. This was found to have been caused by improper cable jointing in projects where there was great pressure to complete and energize the circuit. A secondary reason was hiring of insufficiently qualified jointers.

Statistics on Installed Cable Failure Analysis

Kema Labs also conduct independent power failure investigations (PFI) of in-service cable systems. Fig. 4 shows that the root cause of failures in over 225 of these PFIs is almost equally distributed between the cable, the joints and the terminations. Root cause failure modes for cables were:

• 33% production related;
• 19% installation related;
• 17% due to external damage; and
• 11% due to design issues.

For joints and terminations, root causes of failures were:

• 57% installation related;
• 13% design related; and
• 8% production related.

Design issues as the root cause of failures in any PFI of in-service cables and accessories are surprisingly high and lead to the question whether these components have been certified. Since the reply is almost always “yes”, it becomes clear that the certification process may not always be sufficiently transparent. By contrast, data also shows that components tested in an independent laboratory do not show such a high proportion of design related root cause of failures.

Pressure Relief Test for Surge Arresters

The pressure relief test is important for surge arresters. In case the energy dissipation is too high, due to ageing related or other defects, the arrester can explode and eject debris into the surroundings. Debris can damage nearby equipment resulting in potential loss of a substation bay. Standards include a pressure relief test to determine if debris is not ejected beyond a radius of 3m.

Data from pressure relief tests conducted as part of type testing arresters indicate that approximately 25% do not meet the standard. Fig. 5 for example shows debris one second after the pressure relief test. Many large hot particles can been seen well beyond the maximum permitted 3m radius.

It has become common practice in many countries for surge arresters to remain in service until failure. But such failure of an arrester basically can result in an explosion within the substation. It is therefore important to ensure that only surge arresters that have been properly certified to meet this standard are kept in service.

The post Failure Statistics from Testing Cables & Surge Arresters appeared first on .

This edited contribution to INMR by Prof. Dr.-Ing. Klaus-Dieter Haim at the University of Applied Sciences in Zittau/Görlitz in Germany presents a survey of CIRED requirements for the test of cable screen connection for 66 kV offshore cables and normal 20 kV land cables. This will result in new proposals on how best to assess results and acceptance criteria as a first step toward a test standard for cable screen connections.

PFISTERER

Germany

DALEKOVOD OSO

Croatia

See more suppliers of Cable Accessories

Up to now, there have not been major problems with the screen of medium voltage cables as well as with their connections to joints and terminations.

However, the issue of cable screen connections is becoming increasingly important for reliability of medium and high voltage cable systems due to increases in

• cross-section of cables;
• maximum conductor current and therefore increased screen current;
• failure rates of joints and terminations.

Use of full ampacity in cables in combination with increased cable cross-sections leads to higher conductor and screen currents. This is the reason for upcoming new problems, such as illustrated in Figs. 1 and 2.

To avoid a rising number of joint failures of cable screen connections, CIRED Working Group WG 2017-1 has proposed procedures for testing and evaluating medium voltage power cables. But this recommendation is close to the connector standard (IEC 61238 -1-3) and only applies for cables with copper wire screen and direct connection of screen wires (as shown in Fig. 3) where such application is possible.

More than 95% of cable screen connections are not as presented in Fig. 3. Even for those cables with copper wire screen, use of modern all-in-one joints with incorporated screen contact systems (see Fig 4) requires a different screen contact solution for the joint.

For all cables without a copper wire screen (i.e. more than 50% of MV cables), a special screen connection device is required. There are growing numbers of medium voltage cables with a laminated aluminum screen and, for this type, screen connection is more complex (see Fig 5).

Below is a discussion of the test procedure proposed as well as test results for 4 different cable screen connection designs of 66 kV separable plug-in terminations for offshore array cables. Also, results of tests on the 5 most often used screen connection designs for standard 20 kV cables. Based on these results, recommendations are given for modifying CIRED acceptance criteria.

Not all screen contact systems currently in use offer satisfactory long-term performance. The need to evaluate all the hundreds of various possible screen contact solutions requires a test procedure as well as assessment and evaluation criteria that are different from the connector standard.

Comparison of Conductor & Screen Connections

The test standard for MV cable conductor connectors (IEC 61238-1-3) is an internationally accepted test procedure and, with certain modifications, used worldwide. But for cable screen connections there is still no international standard available. The final report of the CIRED WG has been a first step in this direction.

The description of the test sample and test procedure is comparable with IEC 61238-1-3: 2018 and a good, practical approach. Assessment of test results and the acceptance criteria in the WG’s final report is also comparable to IEC 61238-1-3: 2018. This is a problem since the different physical behavior of cable conductor contacts and cable screen contacts requires a different evaluation of test results.

Fig. 6 shows the test set-up of a mechanical connector and the reference conductor without connector. A compression or shear bold connector generates high and permanent contact pressure. Measurements of contact resistance in a long-time test and load cycling over 1000 cycles show a stable contact resistance (see Fig. 7).

A cable screen contact, where the contact pressure is typically made by a constant force spring (see Fig 8) cannot generate high contact pressure. The results of contact resistance measurement are therefore not stable and that is why it is important to use different acceptance criteria for both contact systems.

Progress by CIRED WG 2017-1

This part presents laboratory test recommendations for MV cable screen connections according to the final report of the CIRED WG. The recommendation has its basis in the electrical part of IEC 61238-1-3.

This presentation is divided into 5 parts. The first describes how to determine the test current used in the heat cycles and short-circuit test. The second defines the test samples while the third covers resistance and temperature measurements. The fourth deals with the test program and finally, there is assessment of results and acceptance criteria.

Flowchart Diagram for Testing
Fig. 9 shows a flowchart diagram for the test of screen connection devices.

Test Screen Currents
The screen current (INS) applied during the heat cycles shall be at least 20% higher than the maximum allowed screen current in the screen connection device during operation (IN), (INS ≥ 1.2 ∙ IN). The current shall be alternating at power frequency (50 or 60 Hz). If the test is performed for a specific application, IN can be determined prior to the test through simulations and/or trials. Minimum screen test current is 30A and the maximum can reach 360A, depending on operating current.

The screen current applied during short-circuit shall be equal to or higher than the maximum short circuit current the screen connection device will be subjected to during service. This can be chosen according to the CENELEC HD620 standard by which the cable screen has been tested.

Test Set-up
The screen connection devices shall be installed as specified in their respective assembly procedures. The assembly procedure used shall precisely indicate all assembly steps and tools, materials, tightening torques etc. Note that only the actual screen connection device needs to be installed in the test loop and not the complete joint or termination body. If the screen connection device is an integrated part of the joint or termination body these devices need to be included.

Test Objects
The test objects are defined between two voltage measurement points. The suitable measurement points will depend on how the metallic screen(s) and screen connection device(s) are constructed.
One of the three alternatives below shall be used:

1. Test object Alternative 1: Between the midpoint of two adjacent cables. This alternative can be used if the screen at the midpoint of the cables has an equalized contact surface, or if this can be made. The contact surface must be made accessible, such as by removing cable jacket.

2. Test object Alternative 2: At the midpoint of two screen connection device links. This alternative can be used if the screen connection device link has an equalized contact surface, or if this can be made.

3. Test object Alternative 3: Between the middle of two cable screen device links and midpoint of the cable. The number cables in the circuit may be reduced to ½ compared to the other two alternatives. This alternative can only be used for cable screens consisting of one metallic element, such as aluminium laminate, and where the screen connection device link has an equalized contact surface.

Fig. 10 presents design of a test cable with one screen, lengths and voltage measurement points while Fig. 11 shows this for cables with two or more screens.

Test Circuits
The test circuit (see Fig. 12) shall consist of six test objects, one reference cable and two different voltage sources. One for the screen current and one for the conductor current (nominal current of the cable).

The resulting current through the screen is the combination of applied current by the screen current source and the induced current by the magnetic field generated by the current in the cable conductor. A cancellation of the induced current in the screen is possible by application of the test setup presented in Fig. 13. Figs. 14 and 15 shows the test set-up for cables with two or more screens.

Heat Cycle Test Program
A current, INS, is applied to the screen and a current, INC, to the conductor. The purpose of INC is to add heat to the cable such that thermal equilibrium is reached at a higher temperature of the screen than INS would allow by itself. Heating duration is then set as the time required to reach thermal equilibrium, i.e., where none of the temperatures measured varies by more than 2°C during a 10 min period (see Fig. 16).

Short-Circuit Test Program
Two short-circuits shall be applied after the 150th heat cycle. The short-circuit shall be applied to the cable screen only. Screen connection resistance measurements shall be made before and after the two short-circuits (not between).
In the test report, the short-circuits shall be defined either by:

• Maximum temperature, time and approximate current; or
• Actual current and time and approximate maximum temperature.

Before each short circuit, the test loop shall have a temperature between 15 and 35°C.

Assessment & Acceptance Criteria

The design of test samples and the test procedure is a good start and can be a basic of a future test standard for cable screen connections.

The evaluation of the results is much more complex compared with the conductor connections and need more investigations and test experiences to fix final acceptance criteria.

Tests were started in 2014 at the University of Applied Sciences in Zittau involving 20 kV cable screen connections with much different contact designs, as supplied by a major cable accessory supplier. Additional tests were then done in 2021 on more samples (i.e. ALMA, ALINA,…), this time according to recommendations in the final report of CIRED WG 2017-1 and continued with more load cycling and resistance measurements. However, it has still not been possible to offer a definition of what constitutes a good versus a bad contact system for cable screen connections.

TECNALIA Electrical Labs

Spain

PSW – Siemens Energy Testing Laboratories Berlin

Germany

See more Laboratories

Conclusions & Recommendations

The different physical behaviour of cable conductor contacts and cable screen contacts makes it problematic to set up the same acceptance criteria for conductor connections and screen connections.

Considering the long-time measuring results of the screen contact resistance, a proposal for the resistance acceptance can be:

• Initial scatter δ should be maximum 0.5;
• Resistance factor change YA should be maximum 0.2;
• Resistance factor change YB should be maximum;
• Resistance factor ratio λ should be maximum 6.0

Results for two much different screen contact designs for 20 kV (see Fig. 17) and for 66 kV (see Fig. 18) make it clear that evaluation of a suitable test procedure for cable screen connections is far more difficult than for cable conductor connections.

In Fig. 17 the yellow graph (screen connection with the name ALMA, see Fig 19) represents the most frequently used screen connection solution. This contact system cannot fulfil even the moderate requirement for the resistance factor ratio of λ = 6.0. This illustrates the challenges when it comes to setting evaluation and acceptance criteria for cable screen connection devices.

In the contact system, ALMA, the copper screen wires fold back on the cable sheath for a length of 5 cm. This area is covered by the copper mesh tube and the contact pressure is performed using a constant force spring.

References
[1] CIRED WG 2017-1 Final Report of the Working Group
[2] IEC, “61238-1-3: 2018,” in Compression and mechanical connectors for power cables – Part 1-3: Test methods and requirements for compression and mechanical connectors for power cables for rated voltages above 1 kV (Um = 1,2 kV) up to 30 kV (Um = 36 kV) tested on non-insulated conductors, ed, 2018.
[3] CENELEC, “HD620: Distribution cables with extruded insulation for rated voltages from 3,6/6 (7,2) kV up to and including 20,8/36 (42) kV,” ed

The post New Test Recommendations for Metallic Screen Connections of MV & HV Cables appeared first on .

Transient Recovery Voltage (TRV) of a circuit breaker is voltage difference measured between each side of the circuit breaker to ground. The most severe TRV from an amplitude point of view follows interruption of the first phase to clear an ungrounded three-phase fault. The shift in system neutral results in high amplitude TRV. While the probability of this fault is low, it is a basis for rating a circuit breaker’s TRV capability.

TRV should be measured across the terminals of all circuit breakers in the substation under study and three-phase ungrounded (LLL) fault, three-phase grounded (LLLG), and short line faults (SLF) should be applied near the breakers. There should be several fault locations to find out the worst system TRV on the breakers in order to test breaker TRV capabilities. Short line faults should be applied on transmission lines near the substation under study, with the basis for these being a single line-to-ground fault (SLG). N, N-1 and N-2 conditions should be applied in these studies.

EMCO Industries

Pakistan

Guangzhou MPC Power International Co. Ltd.

China

See more suppliers of Arresters

A circuit breaker clearing a capacitor bank is also prone to restrike due to TRV. Transmission lines and equipment connected to the station busbar impact the TRV for the three-phase ungrounded fault test. When studying transient recovery voltage, it is necessary to ensure that all suitable lines, transformers and bus equipment are represented in the transient simulation. It is a good idea to include a lumped capacitor to ground on each phase of the station busbar to encompass all bushing, winding and stray capacitances that might exist at the station. The value used will be an approximation, depending on number of transformers and other equipment connected to the busbar.

Simulation Studies

Selecting System Parameters

The system study simulator should estimate capacitances in the substation. Transformer capacitance (HG, LG and HL) and lump remaining capacitances on the bus as a capacitance to ground should be represented. Capacitor voltage transformers (CVTs) and buswork are the dominant capacitance to ground. Apply different types of faults on the line side of the breaker. Open the breaker. Observe the voltage built up across the breaker terminals (on the first phase to open). Compute the dV/dt and compare against the breaker TRV ratings. Plot the voltage across the breaker and ensure it is below the breaker TRV withstand voltage. IEEE or IEC standards should be used as reference.

Opening a breaker on the transmission line coming into a station, the worst case is if there are only transformers connected to the bus. If there are any other lines, consider opening them and performing the study under contingency conditions to obtain the worst case. Also, check if there is a grading capacitor across the breaker terminals and, if so, include it in the model. Select a system time step such that it would be possible to get multiple points in the initial rate of rise of breaker TRV voltage.

A good portion of this work consists of developing system models (frequency dependent line models, transformers with saturation and proper transformer connections and grounding).

Time Step & Simulation Time

Accuracy of the digital simulation can be affected by use of a time step that is too large or too small. Time step selection depends on steepness of the transient recovery voltage.

Range of time step = 0.5 to 10μs; typical = 0.5μs

Range of simulation length = .20 to .50 seconds; typical = 0.25 seconds.

Guidelines for TRV Studies

A. Transmission Lines

Transmission lines and cables in electric power systems are non-linear in nature due to frequency dependency in conductors (skin effect) and the ground and earth return path. A frequency dependent model should be used in simulation studies. Dimensions and data are required for each transmission line represented in the network diagram. This can be given at the tower and should include conductor sag. Shield wire dimensions and resistance should also be provided. For some short lines where the electromagnetic wave traveling times are small compared to the simulation time step, a lumped pi-circuits representation is to be used. For longer lines with larger wave traveling times, the distributed parameter line model is used.

B. Power Transformers

Simulation of transformers requires an understanding of some of their basic properties involving both core and winding configurations. This is complicated by the fact the core of the transformer is prone to saturation leading to the phenomena of inrush current, remanence, geomagnetic current effects and ferroresonance. Saturation representation of the transformer should be included in the simulation. The transformer model should have provision for a tap changer. Transformer coupling capacitances (HG, LG, and HL) should be represented.

Each transformer connected to the station busbar under study has strong influence on TRV due to the dominant bushing and winding capacitances. The magnetic coupling and winding inductance should be included in TRV studies. The overvoltage transfer through a transformer is a function of transformer stray capacitances. Thus, knowledge of transformer capacitances is required. By combining the bushing capacitance into the winding capacitances, the transformer model with two windings is represented by a winding inductance plus the equivalent circuit of three coupling capacitances: CH – capacitance of the high voltage side to ground; CL – capacitance of the low voltage to ground; and CHL – capacitance of the high voltage side to the low voltage side.

C.  Circuit Breakers

Breakers are modelled as simple time controlled switches in electromagnetic transient studies. Circuit breaker coupling capacitance should be represented.

D. Substation Power System Modeling

The first step in performing a TRV study in digital simulation is to build up the power system model of the study subject. This requires detailed representation of the substation and nearby power system. The time varying voltages at the breakers in question will be computed and examined. An onsite visit to collect data of substation layout is recommended. Load flow raw data file in PSS/E data format should also be obtained. The raw file is useful to determine the equivalent circuit of the external power network. A special tool called E-TRAN is used to convert PSS/E power flow raw data file to PSCAD case file. It also does the system reduction and finds the equivalent circuit for a specified part of the power system.

E. Power System Description

The substation consists of two 138 kV transmission lines through two breakers to a common bus. Two transformers convert the 138 kV to 13.8 kV of distribution level and provide power supply for a residential area. The substation is connected to the power network through the 138 kV transmission systems. Networks further away from the substation are represented by equivalent circuits. The following sections give the details of data calculation and set-up of the study model.

Post Insulators & Air Switches Model 

This equipment is represented by stray capacitance of 100 pF.

Capacitor Voltage Transformer (CVT) Model

The CVT is represented by CN = 8000 pF based on IEEE C37.011 standard.

25 kV Underground Cables Model   

Cable Bergeron Model representation is used in the simulation. The Bergeron model is a simple, constant frequency model based on traveling waves and is useful for studies where it is important to obtain the correct steady state impedance/admittance of the line or cable at fundamental frequency.

Station Layout Model

A detailed station layout diagram is represented in the PSCAD/EMTDC, which contains the physical length of each busbar section between station equipment, circuit breakers, transformers, capacitive voltage transformer, lead length, and underground cables. The TRV waveform is measured with a voltage meter across the terminals of the breaker.

Surge Arrester Model

The voltage-current characteristics of metal-oxide surge arresters are a function of incoming surge steepness. Protective characteristics for surge arresters showing crest discharge voltage versus time-to-crest of discharge voltages are available from manufacturers. The arresters can be modeled as nonlinear resistors with 8 x 20μs maximum voltage-current characteristics.

When TRV is to be investigated, the MOV surge arrester representation is significantly different from switching and temporary overvoltage studies. The surge arrester model recommended uses two sections of non-linear resistance designated as A0 and A1, separated by an R-L filter. The non-linear resistors, A0 and A1, can be modeled in PSCAD as a piecewise linear V-I curve with characteristics defined point-by-point. The number of points selected to represent the non-linear resistance depends on smoothness desired. The surge arrester model parameters are then further calibrated by comparing to voltage versus current (VxI) characteristic data from the arrester supplier. This method of arrester modeling is documented in IEEE and in the literature. Fig. 1 shows the representation of the model.

Equivalent Network

When transients are being studied, the whole network usually does not need to be modeled as it does in transient stability studies (where transients are much slower and can reach further into the network). Care is required in selecting the sources representing the short circuit impedance and where the source is located. Since electromagnetic transient programs are generally unable to solve very large networks (<3000 busses), the studies should be performed by introducing network equivalents in the client model system in order to represent the network outside of the study boundaries. The power system simulator should identify a portion of the network for a direct translation into the electromagnetic transient program model. The equivalent is a multi-port, which means that the representation will be correct for steady state as well as for open circuit and short circuit conditions, and contains voltage sources, to match PQ flow (net real and reactive power) and represent the generation in the equivalent network.

Case Studies – TRV Results

A. TRV Simulation Tests

TRV is measured across the terminals of the 145 kV and 15 kV circuit breakers. Three-phase ungrounded (LLL) faults and three-phase grounded (LLLG) faults are applied near the 145 kV and 15 kV breakers. Several fault locations are applied to find out the worst system TRV on the 145 kV and 15 kV breakers in order to test breakers TRV capabilities.

Plot the voltage across the breaker and ensure it is below the breaker TRV withstand voltage (envelope curve) as per IEEE and IEC standards.

Short line faults (SLF) are applied on the transmission lines with location a short distance out from the 145 kV breaker terminals. These SLFs are single line-to-ground faults (SLG).

The breaker is considered ‘PASSED’ if the TRV measured in the simulation does not exceed the TRV envelope curve. Otherwise. the breaker is considered ‘FAILED’ in the simulation. If the breaker does not pass, mitigation actions have to be taken.

Different TRV envelope curves (100%, 60%, 30%) of the circuit breaker are selected based on short-circuit current measured during the fault.

B. Fault Locations

Fig. 2 shows fault locations of the TRV simulation studies.

C. TRV Envelope Curves

IEC 62271-100 envelope curve is used in the simulation and compared with the TRV measured values across the terminals of the pole of the circuit breakers. Fig. 3 shows the IEC envelope curve and Table 1 provides IEEE and IEC TRV envelope data for 15 kV circuit breakers.

D. Mitigating TRV

According to IEEE C37.011, when the inherent TRV of the system exceeds standards, the user has three major alternatives outside of re-configuring the system:

a) Using a circuit breaker with higher voltage rating or a modified circuit breaker;
b) Adding capacitance to the circuit-breaker terminal(s) to reduce rate of rise of TRV;
c) Consulting the manufacturer concerning the application.

As long as a circuit breaker is applied within its symmetrical current and voltage ratings, one of these methods should result in a satisfactory application.

E. Alternative Mitigation Solution

Placement of a zinc oxide surge arrester across the circuit breaker terminals is an alternative solution to reduce TRV, as shown in Fig. 4.

F. TRV Results

Several mitigation solutions are applied in order to reduce the TRV measured across the terminals of a pole of the circuit breakers under the N-1 condition. These include surge capacitors to ground, grading capacitors across the breaker contacts, surge arresters and smooth reactors. Simulations are applied with and without the surge arrester at the high voltage power transformers side. When the measured TRV exceeds the TRV envelope curve, a PSCAD/EMTDC component is modeled to show a ‘FAIL’ flag sign (no fail=0/fail=1). Figs. 5 and 6 show the result of a TRV simulation where a three-phase ungrounded fault (LLL), usually the most severe TRV condition, is applied near breaker CB 138-2 at the substation side, with CB 138-1 open – both without an arrester and with a 132 kV (106 MCOV) voltage rating surge arrester across the terminals respectively. Fig. 7 shows that the arrester absorbed energy curve of the TRV simulation (total of 20.5 kJ), which is small compared to its rated energy handling capability (i.e. 3.4 kJ/kV MCOV).

T60 envelope curve is used in the simulation when only one of the 138 kV transmission lines is connected to the substation (N-1 condition). The highest short circuit current in the simulations is lower than 15 kA_rms, which is lower than 60% of the breaker rated short circuit current (31.5 kA)

Conclusions from TRV Simulations

Critical factors in simulation of TRV include:

• capacitance on the bus; other transmission lines connected to the bus;

• selection and location of faults being cleared; and

• transformer inductances and stray capacitances.

These studies have represented the substation under study in detail – specifically a model generally used for lightning studies has been used (where individual buswork components, stray capacitance are represented) which is more accurate than simply adding up the total capacitance of all elements on the bus and lumping it into a single capacitance in the simulation. Data for critical elements (such as transformer capacitances and CVTs) is available and used in the studies. Also, a sufficiently small-time step is used so the rapid build-up of voltage can be observed. The data and simulation methods used in these studies are consistent with generally accepted practice.

TRV simulation results are directly compared to IEC/IEEE and breaker TRV curves. If the voltage across the breaker terminal exceeds the IEC envelope, a fail indicator is generated. TRV simulation failure is reported for several cases (clearing LLLG and LLL faults) when breaker voltage is compared to the allowable curve for clearing a fault with a short circuit current.

KERI High Power & High Voltage Laboratories

South Korea

Powertech Labs Inc.

Canada

See more Laboratories

References
IEEE Std C37.011, IEEE application guide for transient recovery voltage for AC high-voltage circuit breakers rated on a symmetrical current basis.
IEC 62271-100, High-voltage switchgear and controlgear – Part 100: Alternating-current circuit-breakers
PSCAD Electromagnetic Transients User’s Guide.
IEEE WG 3.4.1 l, “Modeling of Metal Oxide Surge Arresters,” IEEE Transactions of Power Delivery, Vol. 7, No. 1, January 1992, pp 302-307.
Verdolin, R., “Modeling of ZnO Surge Arresters”, The University of Manitoba – Power Systems Transient Simulation, May 1992.
M. Nobre, W. L. A. Neve, B. A. de Souza, “An Alternative to Reduce Medium-Voltage Transient Recovery Voltage Peaks”.

The post Applying Surge Arresters to Mitigate Breaker Transient Recovery Voltage appeared first on .

PFISTERER

Germany

EB Rebosio SRL, A Gruppo Bonomi Company

Italy

See more suppliers of Cable Accessories

Types & Tasks for Cable Accessories

The majority of all power cables are extruded XLPE insulated cables. These consist of a conductor, an inner conductive layer, the main insulation layer, an outer conductive layer, an outer conductor and an outer insulation/jacket.

At places where two cable sections are connected, theses layers have to be re-built in the same way by applying a cable joint or splice. At the cable termination, the layers have to be treated such that they can withstand all resulting electrical and environmental stresses. This is done by mounting a cable termination.

Besides these main cable accessories, there are several combinations and modifications such as connectors and pluggable systems. The main resulting tasks include:

1. Electrical connection;
2. Field grading;
3. Sealing;
4. Withstanding weathering.

Medium voltage plug-in systems are a good example for the first task. A rather soft silicone elastomer cone is pressed by a metallic spring into the male resin part and provides for a perfect electrically insulated interface. Apart from the low hardness of the silicone elastomer, its high gas permeability is another advantage since any gas trapped in the interface diffuses out shortly after mounting. As mentioned, one of the main goals during early development of silicone accessories was to pre-fabricate a ready-to-use part consisting of all insulating and field grading parts. Fig. 5 shows an example of an older resulting design of joint. All grading parts in this example were made from conductive silicone elastomers. Other solutions applying silicone elastomers with high permittivity are also possible. Silicone elastomers, especially liquid silicone elastomers, have been found to be an ideal material for all necessary step-molding processes since they are low in viscosity and allow for good flow into the mold and towards the triple-points located between mold and pre-inserted conductive parts.

Proper sealing is another important task of a silicone cable accessory. The outstanding mechanical properties of soft silicone elastomers facilitate slip-on mounting while the perfect and lasting ability to shrink to original dimensions, even after long expansion time, allows for cold-shrink technology. Both mounting technologies also benefit from the high gas permeability of silicones. Proper mounting and sealing are supported as well by special silicone pastes available for T&D applications. Pastes can be modified so as to allow easy dismantling even after long time or of a migrating type that disappears from the interface in a short time. Most important in the case of outdoor applications is the general capability of silicone elastomers to withstand weathering. Nevertheless, silicones for these applications are specially modified to ensure lasting stability against weathering stresses as well as the possibility of resulting electrical and erosive stresses during service. An example, shown in Fig. 7, is a termination that provides perfect hydrophobicity. Experience and formulating know-how allow for silicones that can withstand the rather short-term stresses encountered during type tests (e.g. salt fog) as well as long-term stresses during service. It is important to consider pollution affecting outdoor cable accessories. In this regard, the ability of suitable silicone elastomers not only to remain hydrophobic but also to show good hydrophobicity transfer to the pollution layer is important.

Types of Silicone Materials

The principal materials used in silicone cable accessories are silicone elastomers. The early years of their production were characterized by 2-component, low viscosity elastomers designed to cure at low temperatures and parts were mainly cast or vacuum cast. Production output at the time was low compared to the sophisticated technologies available today. Nevertheless, these materials are still available, e.g. for prototyping or for manufacturing of small series of products. State-of-the-art materials for cable accessories are either low hardness, high-consistency silicone rubber (HCR), available as peroxide curing types or as 1- or 2-component addition curing materials for fast and very fast injection molding applications or extrusion respectively. Press molding is not recommended since state-of the-art accessories require a high accuracy of molds. The other option is using liquid silicone rubbers (LSR) that usually come as 1 to 1 two component systems. These materials are now used to manufacture all types of cable accessories. LSRs with 30 … 50 Shore are typically applied for high-throughput injection molding. Modified versions of HCR and LSR with low electrical volume resistivity or high permittivity are also available.

Silicone pastes are also available in a variety of modifications. Extremely soft silicone elastomers, i.e. silicone gels, first saw application in the low voltage area below 1000 V. So-called ‘gel-joints’ allowed proper electrical installation and sealing just by closing shell-like parts pre-filled with soft silicone gel. Thanks to special fillers, new potential applications for silicone gels are becoming evident to manufacturers.

Selected Properties of Silicones for Cable Accessories

There is as yet no standard for definition of general requirements for materials used in cable accessories. IEC 62039TR:2007, Selection guide for polymeric materials for outdoor use under HV stress, covers materials for outdoor cable accessories and defines the following key properties:

• Resistance to tracking and erosion;
• Resistance to corona and ozone;
• Resistance to chemical and physical degradation by water;
• Tear strength;
• Volume resistivity;
• Breakdown field strength;
• Resistance to chemical attack;
• Resistance to weathering and UV;
• Resistance to flammability;
• Arc resistance;
• Glass transition temperature;
• Hydrophobicity

Resistance to tracking and erosion is generally understood to be an important property of all insulating materials for outdoor application and is tested according to IEC 60587 (see Fig. 12). The general minimum requirement according to IEC 62039TR:2007 is a classification of 1A 3.5 but quality silicone elastomers usually withstand the higher classification of 1A 4.5 (Fig. 13) and provide for higher network security.

Breakdown field strength is understood to be another important property for such insulating materials. IEC 62039TR:2007 sets the minimum value of 10 kV/mm when measured acc. to IEC 602431-1. It is important to take into consideration, however, that this minimum does not reflect the properties of silicone elastomers. Moreover, the old version of IEC 60243-1 has since been replaced by a new edition, i.e. IEC 60243-1:2013, Electric strength of insulating materials – Test methods – Part 1: Tests at power frequencies which contains a new sub-clause ‘Elastomers’ with the following recommendations:

• Use test specimens of (1.0 ± 0.1) mm thickness;

• Unequal electrodes … shall be used (see Fig. 14);

• In the case of elastomers of low hardness, e.g. silicone rubbers, a suitable casting material shall be used as embedding material or surrounding medium, respectively (see Fig. 15);

Application of the new edition of IEC 60243 leads to measurements that are generally higher compared to the values collected using electrodes in air/fluids (see Fig. 16). It could be claimed that only by applying the proposed electrode arrangement in combination with embedding, can differences in specific breakdown voltage of silicone elastomers be sorted out. Unfortunately, the pure number is of limited interest since it usually neither reflects real electrical field condition in molded parts nor gives information about long-term behavior of materials under stress. Nevertheless, variations in the values show that there are clear differences between different silicone elastomers

The post Silicone Elastomers, Gels & Auxiliary Materials for Cable Accessories appeared first on .