Integrating IIoT-based sensors and communication networks has made it possible to continuously track the lightning current flowing through line surge arresters and transmission line towers, enabling early detection of abnormal conditions and improved system reliability.

The post Advancements in Monitoring Surge Arresters & Lightning (Video) appeared first on .

The corrosion mechanism was closely studied and tests as well as simulations were carried out. Results indicated that corrosion on caps could increase contamination accumulation on insulator surfaces and decrease flashover voltage. Pin corrosion could decrease mechanical strength. This edited past contribution to INMR by Prof. Wang Liming of China’s Tsinghua University, proposed solutions.

Hivolt Power System

China

Wish Power (Thailand) Co. Ltd

Thailand

See more suppliers of Insulators for Substation Equipment

The scale of China’s power grid has increased rapidly as has development of the country’s UHV DC technology. Numerous projects have already been put into operation and are now optimizing allocation of the country’s energy resources. Early such projects included the ±800 kV Chusui Line, from Yunnan to Guangdong, and the Xiangshang Line, from Yunnan to Shanghai, both commissioned in 2010. Starting October 2011, however, corrosion phenomena were reported on the hardware of many porcelain and glass insulators on these lines. The number of iron cap corroded insulators on Chusui Line was more than 24,000 according to inspection results while over 2000 corroded insulators were discovered on the Xiangshang Line by March 2012.

With increasing operating time, the number of insulators having hardware corrosion problem also increased and this was deemed to threaten security and stability of the power system. Subsequently, 82 porcelain and glass disc insulators suffering from pin corrosion were randomly retrieved from the positive polarity of the Chusui Line in March 2013.

Causes of Corrosion

Corroded porcelain and glass insulator samples were hung in a V-string with cap-corroded insulators located at the negative polarity and pin-corroded ones concentrated at the positive polarity. The corrosion area of iron caps was typically the lowest part and there was an obvious rust channel on the insulator surface (as shown in Figs. 1a and 1b). In the case of pin-corroded insulators, the corrosion area was mainly the annular region at the cement-zinc sleeve interface (as illustrated in Figs. 1c and 1d). Moreover, the lowest portion of the zinc sleeve corroded more seriously than other parts.

Electrolytic reactions play a major role in the corrosion process due to the polarity phenomenon affecting hardware. Insulators that corroded most seriously were locates in forested areas where humidity is high and there is often continuous heavy fog in late autumn, winter and early spring. For example, the environment surrounding tower #407 on the Chusui Line is shown in Fig. 2.

The principle behind electrolytic corrosion of iron caps is shown in Fig. 3. The electrolytic loop here consists of a DC power supply, hardware and electrolyte. The iron cap connected to the positive side (ground side) of the power supply is an anode whereas the pin connected to the negative side is the cathode. Ferrous ions are formed by oxidation reaction of the iron cap whenever the insulator surface becomes damp.

F_e→F_e²⁺+2e^–

H⁺ and OH^– ions exist in the electrolyte as a result of water ionization and a reduction reaction occurs on the cathode side. Cations migrate to the cathode and anions moves to anode under applied DC voltage. Precipitates are formed by Fe²⁺ and OH^–. Due to the existence of oxygen in the solution, a further oxidation reaction can take place. The constituents of rust are shown as:

mF_eO+nF_e2O₃+pH₂O

where the values of m, n and p vary under different temperatures, pH values and oxygen contents.

Simulation Test Method

The spray water method was used to simulate the corrosion process affecting insulator hardware. Before the test, the copper electrode was fixed onto the surface of insulators. The distances between electrode and insulation element (i.e. porcelain or glass) were about 5 cm and 1cm for iron cap & pin tests respectively (see Figs. 4a and 4c). The metal wire connected to the other end of the copper electrode was fixed on the pin and locking device for the iron cap & pin tests respectively (see Figs. 4b and 4d).

During the test, an NaCl solution was sprayed onto the insulator surface to form the electrolyte. For the iron cap test, the voltage applied to the pin was in the range -0.8~-1.5 kV, the iron cap was grounded and the conductivity of the NaCl solution and its spray velocity were 8~10 mS/cm and 8~10 L/h respectively. For the pin corrosion test, the voltage applied to the pin was in the range of +0.4~+0.8 kV and the iron cap was grounded. Conductivity and spray velocity were 2~3 mS/cm and 2~3L/h respectively. The experimental set-up is shown in Fig. 5.

Influence on Mechanical & Electrical Characteristics

Iron Cap Corrosion

Three pieces of iron cap corroded porcelain insulators retrieved from tower #407 of the ± 800 kV Chusui Line were used to carry out a contamination degree measurement. The surface of these insulators was divided into three parts: one was the region with accumulation of corrosion by-products on the upper surface (area A); another was the area with no corrosion by-products on the upper surface (area B); the third was the lower surface (area C), as shown in Fig. 6. Measurement results are presented in Table 1.

Results indicated that ESDD and NSDD of area A were much higher than those of areas B and C, i.e. the contamination degree of the rust channel area was higher than for other areas.

Pollution Flashover Testing

The negative and positive polarities of the ± 800 kV Chusui Transmission Project were put into operation in June 2009 and June 2010 respectively. While this transmission line had not experienced any flashover incident, there were reports of discharge phenomena in Aug. 2011 at the negative polarity on tower #407. For example, sparks could be seen and there was resulting noise. The XZP2-300 type porcelain insulators on this particular structure had experienced corrosion early on and were already seriously corroded. As such, each February or March during annual maintenance with the line out of service every suspension type insulator was replaced. Therefore, flashover tests conducted on samples removed from this tower reflected their condition after one-year of service.

Pollution flashover tests were carried out in a 26m × 26m × 30m fog chamber with maximum output voltage of DC power supply being 1000 kV. Four groups of XZP₂-300 type porcelain insulators removed from the Chusui Line were used to conduct these tests. Groups 1 and 2 were retrieved in March 2013, while Groups 3 and 4 in March 2014. Groups 1 and 3 consisted of 14 iron cap-corroded insulators, and Groups 2 and 4 were made up of 14 pin-corroded insulators. These were hung in a V-string (76°) with each side made up of 7 insulator discs. The boost voltage method was used to conduct the flashover test and the flashover time of each group was 6. Test results are presented in Table 2.

Findings showed that the flashover voltage of iron cap-corroded insulators was 20% lower than that of pin-corroded insulators. Moreover, in order to analyze the cause of the heavier pollution concentrated in the rust channel area, the conductivities of solutions with different masses of rust were measured and compared with the same mass of NaCl.

It can be seen from the results (see Fig. 7), that rust is not soluble in water and has little influence on conductivity of the solution. Rather, the high pollution degree at the rust channel area is caused by the rough surface, making cleaning of contamination more difficult. The micromorphology of iron cap corrosion by-products is shown in Fig. 8.

Influence of Pin Corrosion

Mechanical Stress Calculation

The XZP₂-300 type porcelain insulators on tower #407 of the Chusui Line were used for simulation by means of the FEM method. These porcelain insulators were hung in double V-strings (76°) with each string made up of 69 units. The transmission line employs six-bundle LGJ-630/45 type aluminum steel reinforced (ACSR) conductor. Referring to standard GB/T 1179-2008, the diameter of each wire is 33.8 mm and its mass 2079.2 kg/km. The span between adjacent towers is 500 m and maximum conductor sag is 19 m so that actual length of wire between two towers is 501.8 m. Each tower bears the wire weight of the span evenly, namely:

The static tension that insulators bear in each string is:

The weight of each XZP2-300 type insulator is 17.1 kg. The first insulator counted from the cross-arm bears the maximum static tension:

As such, the total static tension on this insulator is:

Calculation results of mechanical stress and strain on the first insulator from cross-arm are shown in Fig. 9.

Fig. 9 shows that, in the case of the pin, the part exposed to air and the cement-zinc sleeve interface experience great strain under the stress. If the cross-section of this pin is reduced due to corrosion, the portion exposed to air is easy to pull off. Similarly, if the adhesive strength between pin and cement decreases due to expansion from corrosion by-products, the entire pin could be pulled from the iron cap.

Mechanical Tensile Failure Test

Mechanical load test in accordance with Chinese National Standard GB/T19443 were carried out on XZP₂-300 type porcelain insulators, including new production units, cap-corroded units and pin-corroded samples. In terms of new XZP₂-300 type porcelain insulators where rated mechanical failure load is 300 kN, the failure load of these insulators fell mainly in the range of 380~410 kN. But for pin-corroded insulators with less than 4 years of service, the final failing load was in the range of 340~370 kN. Final failure mode is shown in Fig. 10.

Test results indicate that, even in the case of pin-corroded insulators where the zinc sleeves have not been penetrated and pin cross-section has not been reduced, mechanical strength decreases. This is because hoop stress at the cement-zinc sleeve interface weakens the strength of the bond between them (see Fig. 11).

Remedial Solutions

Solution for Corroded Caps Still In-Service

A U-shaped zinc ring has been designed to suppress electrolytic corrosion of the iron cap of those insulators in service. as shown in Fig. 12.

To verify the effectiveness of this zinc ring, XZP₂-300 type porcelain insulators were used to conduct accelerated corrosion test with the spray water method, intended to simulate electrolytic corrosion. The iron caps were grounded and the voltage applied to pins was during the accelerated corrosion test was -0.8~-1.5 kV. Test insulators were hung in a 76° V-string – the same as on the actual transmission line. During the corrosion simulation, the electric charge quantity was set as 81,000 C, based on the maximum average annual amount of corrosion charge (i.e. 2618 C/year) obtained from the Chusui Line and given the insulator’s expected 30-year service life. The respective insulator sections with the biggest depth of corrosion in their iron caps, both with and without the zinc ring, are illustrated in Figs. 13 and 14.

Corrosion depth of the iron cap without a zinc ring is quite large. However, after installation of the ring, the section with the deepest corrosion is almost the same as that of the non-corroded area. Therefore, installing such a U-shaped zinc ring is deemed effective in suppressing further electrolytic corrosion of iron caps.

Solution for Cap Corrosion on New Insulators

In the case of new insulators, a zinc ring attached to the iron cap has been designed to suppress electrolytic corrosion (see Fig. 15). The structure and size of these rings were optimized according to test results from sample insulators.

In the simulated corrosion, the electric charge quantity was set as 45,000 C. The respective sections with the highest depth of corrosion in caps, with and without a zinc ring, are illustrated in Fig. 17.

TECNALIA Electrical Labs

Spain

ABB and PEHLA Laboratories

Germany

See more Laboratories

Solution for Corroded Pins

Based on the research described earlier, mechanical strength of insulators in services decreases even if their pins are not corroded and become thinner since expansion caused by corrosion by-products reduces bonding strength between cement and zinc sleeve. Thickening the zinc sleeve is therefore not effective here and the recommendation is rather to install an organic material sleeve onto the zinc sleeve (as in Fig. 18). Such an organic sleeve can be manufactured with either of two types of material: one a high temperature vulcanized (HTV) silicone rubber; the other a semi-conductive rubber made by adding conductive carbon black to HTV silicone rubber. This sleeve alters the corrosion process, effectively shifting the corroded area from the cement-zinc interface to the exposed portion of the zinc sleeve that does not bear any hoop stress, and avoids any decrease in bond strength between cement and zinc sleeve. Improved suspension disc insulators having these organic sleeves are then used mainly in regions with high humidity and serious pollution. For XZP₂-300 type porcelain disc insulators, units having ‘half package’ type, ‘flush’ type and ‘whole package’ type organic sleeves installed are designed as shown in Fig. 19.

Tests using the water spray method were conducted on XZP₂-300 type porcelain insulators having different organic sleeves installed. Electric charge quantity was set as 45,000 C, based on maximum average annual amount of corrosion charge on the Chusui Line (1500 C/year) and the 30-year expected service life of insulators. A 1000-kg cement block was then hung on these insulators for 6 months after which they were exposed to a tensile load test in accordance with Chinese National Standard GB/T19443. Tensile failure load and pins are shown in Figs. 21 and 22.

Fig. 22b confirms that installing the half package type organic sleeve alters the corrosion area on the zinc sleeve. The original corroded part of the cement-zinc sleeve interface bears radial stress whereas the new area of corrosion is the exposed portion of the zinc sleeve that does not have to withstand hoop stress. As such, expansion from corrosion by-products cannot have an adverse impact on insulator mechanical strength. Test results (see Fig. 21) verify the effectiveness of a ‘half package’ type organic sleeve. However, corrosion on pins with flush type as well as whole package type sleeve installed can reduce pin cross-section and decrease mechanical strength. Moreover, test results indicate that the part of the zinc sleeve between the two lines is packed by the organic sleeve and cannot be used as a sacrificial electrode to protect the pin. In other words, installation of the half package type organic sleeve can reduce the effective size of the zinc sleeve and lead to pin corrosion, as evident from the blue circle in Fig. 22b. In order to overcome this problem, it is recommended to increase the height of the exposed part of the zinc sleeve by 1~1.5 cm, based on existing height.

Conclusions

1. Hardware corrosion phenomena have occurred on ± 800 kV UHV DC transmission lines due mainly to electrolytic corrosion.

2. Iron cap corrosion can lead to more contamination accumulating on the insulator surface and decrease flashover voltage of the entire string.

3. Pin corrosion can reduce the mechanical strength of the insulator.

4. Zinc rings and U-shaped zinc rings can be used to suppress cap corrosion on new as well as on those insulators already in service.

5. Installing an organic material sleeve can solve the problem of decreased mechanical strength caused by expansion of corrosion by-products.

The post Overcoming Corrosion of Disc Insulators on DC Lines appeared first on .

Direct comparison of the relative performance of hydrophobic, coated glass with the steady hydrophilic behavior of the uncoated glass is key to understanding the dynamics of hydrophobicity in service. Fortunately, advances in leakage current monitoring sensors have now made this possible by allowing large amounts of data in the field to be collected in real time and under normal operating conditions. Moreover, work done within different CIGRE Working Groups has provided new knowledge, including proposals to standardize pollution test procedures for coated glass insulators.

This edited past contribution to INMR by experts at La Granja Insulators presented a probabilistic approach for dimensioning coated glass insulators in polluted environments based on leakage current monitoring, quick-flashover tests and comparing relative performance. This methodology can help evaluate flashover risk and quantify the benefit of hydrophobicity when dimensioning coated glass insulators in polluted service areas.

Protektel sp. z o.o.

Poland

Izoelektro d.o.o.

Slovenia

See more suppliers of Arresters

Approaches to Dimensioning

The deterministic approach is used widely for insulator dimensioning and is based on a worst-case analysis with safety factors to cover unknowns, as per IEC-TS 60815. However this approach can result in over-dimensioned designs given its significant limitations in reflecting the dynamic properties of hydrophobicity transfer materials (HTM) such as silicone coatings.

By contrast, a probabilistic approach for dimensioning insulators in polluted areas has been shown to be more effective. Such an approach considers the stress and the strength as probabilistic variables to evaluate risk for flashover of potential designs. Those designs are then selected that yield adequate insulator performance with respect to pollution conditions. Unfortunately, characterization of pollution severity, as typically done by means of ESDD measurements, results in limited available data to fit well into statistical distributions. Moreover, as with the deterministic approach, the benefits of hydrophobicity are not considered or developed into this approach.

Probabilistic & Dynamic Approach

Based on the probabilistic approach, this approach aims to include the dynamic effects of hydrophobicity of insulators in service so as to quantify the pollution performance enhancement offered by the silicone coating:

• Pollution stress of the site is described by the probability density function f(γ) of an extreme value distribution expressed in terms of site severity “γ”;

• Insulation strength is described by the cumulative distribution function P’(γ,λ) of a three-parameter Weibull distribution as a function of the same measure of site severity “γ” and a new term, “λ”, which expresses the hydrophobicity condition of the surface of the insulators. Note that, in case of uncoated glass insulators with the same insulator type, string configuration, etc, the strength function is expressed only in terms of site severity P(γ) and represents the hydrophilic condition

• Multiplying the f(γ) x P’(γ,λ) functions yields the probability density of flashover. Flashover risk is given by the area below the curve.

Monitoring Pollution Severity

Leakage current (LC) across an insulator string is a suitable parameter to assess its pollution performance. This is because, unlike other methods such as measuring ESDD and NSDD, leakage current reflects the wetting effects of the environment on the pollution layer. This can be monitored in real time on energized insulator strings in service and used to monitor pollution severity on glass strings as well as the hydrophobicity condition of silicone coated glass. In addition, the large amount of LC data monitored in the field allows for a good fit of the stress distribution f(γ) as well as the hydrophobicity condition (λ).

Site Equivalent Salinity

The relationship between leakage current in the field and standardized laboratory pollution severities can be established through testing. Site equivalent salinity is the pollution level of salt fog tests, according to IEC 60507, that yields comparable leakage current levels on the same insulator string, at the same voltage as in service.

Note that the pollution severity (γ) for stress and strength functions must be expressed in the same terms, e.g. in case of pollution associated with coastal areas (type B) i.e., salinity (kg/m³).

Data from the Field

Example of data collected from the field involved two ‘mirror’ insulator strings composed of (1) glass and (2) silicone-coated glass insulators, energized at 220 kV and monitored for a period of more than 3 years. To select the statistical distribution that best fits the LC data, several techniques were employed, such as the log-likelihood, AIC (Akaike) and BIC (Bayesian information criterion). Generalized Extreme Value (GEV) Type II, or Fréchet, yielded the best metrics.

Loglogistic and lognormal distributions could also have been used but a worse fit was observed on the tail. By contrast, GEV fit well for both uncoated glass and silicone coated strings. As expected, LC values were lower for the silicone-coated strings. Particularly, when comparing the ‘tails’, the benefit of the silicone coating becomes clear.

The damping ratio (ζ) concept was developed to express statistically the ratios LC silicone-coated / LC glass:

ζ = 1: Leakage current is the same in both strings, coated and uncoated. That typically happens when LC=0 with a 100% probability of being exceeded.

ζ < 1: Leakage current of the silicone-coated is lower (damped) than for the uncoated string. It is interesting to observe that damping is most effective for the highest currents, with less probability of being exceeded (the peaks), thereby demonstrating the real effectiveness of coatings.

Insulator Strength

Quick-Flashover Tests

This procedure was initially established as a diagnostic method for naturally and artificially aged composite insulators. However, recent experience has indicated that the method can also provide valuable information when applied to new insulators.

Research has found that the recurrent application of flashovers to silicone coated glass insulators leads to progressive reduction in their hydrophobicity, reaching a quasi-stable hydrophilic state after around 10 flashovers. (Note that number of flashovers may depend on insulator type, string configuration, test arrangement, etc.). This makes it possible to establish a relationships between number of applied flashovers in the laboratory and hydrophobicity of silicone coated insulators.

The test procedure was based on a proposal for standardization as in Cigre TB 691 but using a different way to evaluate data and results:

• A conditioning period of 20 minutes is applied at the specified salinity level. Voltage at this stage is about 90% of the estimated flashover voltage;

• After this conditioning period, test voltage is raised in 5% increments and kept for one minute at each level until flashover occurs. After flashover, the insulator is immediately re-energized at its initial voltage (90%);

• Test voltage is then raised in steps of between 2.5% and 3.5% every 5 minutes until flashover.

• Test is continued with 90% of the previous value of flashover voltage until the required number of flashovers has been obtained i.e. 10 flashovers. Standard deviation of the last 5 flashovers is recorded and if it exceeds 5%, the test shall be carried out to obtain 5 additional flashovers.

When comparing results with those obtained from testing a glass insulator string with steady hydrophilic performance, it can be observed that flashover values tends to converge. Identical trends were found when analyzing the leakage current.

Defining hydrophobicity level (λi) as the hydrophobicity condition of the silicone-coated glass insulator string after i flashovers in a QF test. And the damping ratio as the relative performance between silicone-coated / glass strings. It is possible to fit equations linking the performance of silicone-coated insulators in terms of conductance (LC/U) and flashover values to the hydrophobicity levels.

Note how both start at ζ = 0, which is the fully hydrophobic condition, and then tend to converge to ζ = 1, which corresponds the worst-case hypothetical scenario of losing completely the hydrophobicity (non-coated).

Insulator Strength

The pollution curve of a glass insulators string can be expressed as follows:

• CD is the creepage distance of the string (mm);

• U₅₀ is the voltage level with a 50% flashover probability;

• γ is the pollution severity;

• B and α are experimental constants.

Combining the previous equation with the FO damping ratio, ζ_U, pollution curves for silicone-coated glass insulators can be expressed as a function of pollution severity (γ) and hydrophobicity level (λ):

Strength is described by the cumulative distribution function of the strength of a three-parameter Weibull distribution. The fully hydrophilic condition, corresponding to glass insulator strings, which can be taken as the base and does not depend on hydrophobicity condition, (λ), is expressed as follows:

where:

Combining flashover damping ratios, ζ_U, described in the previous equation, silicone coated glass insulator strength can be obtained as a function of pollution severity and hydrophobicity condition:

Comparison between natural and artificial pollution tests has shown that the deviation for natural tests is greater than for artificial tests. In this respect, it is possible to adjust the above parameters to spread the deviation.

The dashed lines represent the different hydrophobicity levels of the insulator string (λ=1,…,10) while the solid line the fully hydrophilic condition corresponding to uncoated glass. It is nonetheless important to note that field experience shows that hydrophobicity is never fully lost, thereby always maintaining the benefit of a silicone coating.

The hydrophobicity condition λ can be determined by equating and solving the equivalent damping ratios monitored in the field and obtained in laboratory:

ζ_{LC_field}  ζ_{G_lab} (λ)

In this case: λ = 4. Note that, a period of time long enough to reach a stable surface condition is needed to obtain consistent results. Possible ageing effects can be simulated by projecting a higher λ.

As mentioned, it is interesting to observe that damping is more effective for the highest leakage currents, with less probability of being exceeded (the peaks). In other words, silicone coated insulators better reduce the highest leakage currents, demonstrating the benefit of hydrophobicity as well as effectiveness of a coating.

Stress-Strength

The multiplication f(γ) x P’(γ,λ) function gives the probability density of flashover and risk of flashover is given by the area under the curve:

String designs with different numbers of insulator units can be compared and evaluated by computing string creepage distance. If multiple silicone coated insulator strings are installed on the same line and exposed to the same conditions, risk of flashover increases as follows:

Conclusions

• A probabilistic method for dimensioning silicone-coated glass insulators in polluted environments was developed based on leakage current monitoring in the field and quick-flashover laboratory tests.

• Pollution stress at site is well described by an extreme value distribution (Fréchet).

• Pollution strength of silicone coated glass insulators is defined by a three-parameter Weibull distribution that includes a novel concept of hydrophobicity level (λ), which covers performance from fully hydrophobic to the hydrophilic condition. The latter is obtained from testing an equivalent uncoated glass string.

• This methodology is useful to evaluate flashover risk and to quantify the benefits of hydrophobicity when dimensioning silicone coated insulators for service in polluted areas.

References  
[1] H. de Santos and M. Á. Sanz-Bobi, “Novel Approaches to Assess the Mechanical Reliability of Toughened Glass Insulators Used in Transmission Lines,” in IEEE Transactions on Power Delivery, vol. 37, no. 3, pp. 2083-2089, June 2022
[2]  CIGRE Working Group D1.44, “Pollution test of naturally and artificially contaminated insulators,” Technical Brochure 691. 2017
[3] CIGRE Working Group B2.69, “Coating for improvement of electrical performance of outdoor insulators under pollution conditions,” Technical Brochure 837. 2021
[4] H. de Santos and M. Á. Sanz-Bobi, “Research on the pollution performance and degradation of superhydrophobic nano-coatings for toughened glass insulators,” Electr. Power Syst. Res., vol. 191, p. 106863, 2021
[5] C.S. Engelbrecht, R. Hartings, J. Lundquist: “Statistical dimensioning of insulators with respect to polluted conditions”, IEE Proc.-Gener. Transm. Distrib, Vol. 151 No.3, May 2004, pp. 321-326.
[6] Pigini, A., Gutman, I.: Evaluation of the performance of polluted insulators: the IEC simplified approach against the statistical approach. INMR World Congress (2013)
[7] H. de Santos and M. Á. Sanz-Bobi, “A Cumulative Pollution Index for the estimation of the leakage current on insulator strings,” IEEE Trans. Power Deliv., vol. 35, no. 5, pp. 2438–2446, 2020

The post Dimensioning Silicone-Coated Glass Insulators for Polluted Environments appeared first on .

Recently, there has also been an increasing trend toward tailor-made solutions that is pushing further development of polymeric materials for composite insulator and string designs. This has been triggered by strategic decisions in the power supply sector such as the move toward ultra-high voltages (> 800 kV, UHV) and expansion of HVDC transmission. At the same time, new material requirements are being defined in some utility specifications and there is great interest in compact designs to promote public acceptance of new lines. Maintaining the quality of composite insulators for expected long-term performance in such applications is yet another trend, especially since these insulators have transitioned from exclusive and expensive to being regarded as cost-optimized commodities. These trends are also reflected in ongoing refinement of existing test methods to verify quality as well as in development of new test methods that aim to transfer service experience into the design stage of composite insulators.

This edited recent contribution to INMR by experts at Pfisterer in Switzerland looked at current trends and aspects for development of new high temperature vulcanizing (HTV) silicone rubber formulations.

PFISTERER

Germany

EMCO Industries

Pakistan

See more suppliers of Insulators: Overhead Line

New Market Requirements

The following case study offers an example of developing a tailor-made, nitric acid resistant formulation of HTV silicone rubber material. The utility specification required that a cut section of the insulator needed to be stored for 100h at 30°C in 1 molar nitric acid. The material would pass if no crack formation was observed on the housing surface after a further drying period of 12h at 80°C. This requirement had been included in the specification after an incident of insulator failure initiated by housing material degradation, crack formation and exposure of the rod after strong corona discharges due to improper field grading. Nitric acid was generated under the specific service conditions that saw only low precipitation over the year. Such acid attack relates mostly to decomposition of fillers such as aluminum trihydroxide (ATH) in HTV:

During the initial investigation, nitric acid (HNO₃) resistivity was considered for commercially available as well as experimental formulations of peroxide curing HTVs (see Table 1). Tests were performed by storing three plate specimens of each material weighing circa 20 g for time intervals of up to 1000h in about 1 liter of 1 molar acid at room temperature (i.e. 23°C ± 1°C).

After defined times, the test specimens were extracted from the acid bath, dried at 80°C for 12h and evaluated, including:
• mass loss;
• tensile strength and elongation at break;
• visual appearance of degradation.
Material formulations with ATH filler showed significant crack formation and depth of cracks was measured (see Figs. 1 & 2). As expected, formation and depth of cracks for HTV 4 (with non-silane treated ATH) was most critical, whereas the unfilled HTV (i.e. without ATH) showed no crack formation. Other HTV formulations fell somewhere in between.

Figs. 3 & 4 show mass loss and tensile strength of the HTV specimens being investigated. As expected, all ATH-filled HTV silicone formulations showed significant mass loss and decrease in tensile strength in about the same proportion as depth of crack formation. The unfilled formulation showed almost no mass loss and decrease in tensile strength (not shown). Interestingly, if effective cross-section reduction by cracking is considered in calculating tensile strength, the material still retains its initial tensile strength despite some deviation in measurement of depth (see Table 2). From this, it can be concluded that the bulk material is still fully intact. These initial investigations conclusively demonstrate the need to improve nitric acid resistivity of HTV formulations.

Given the mechanism of so-called ‘stress corrosion cracking’ that occurs during crack formation by nitric acid attack, three conditions must be fulfilled:

• Presence of a chemical substance that can corrode the material;
• A material susceptible to corrosion from this chemical substance; and
• Mechanical stress applied to the material.
From this, it becomes clear that the greatest potential lies in improving material formulation, either by replacing the ATH filler by an inert one, by optimizing silane treatment, particle size and distribution of ATH filler or by implementing an additive that serves as a buffer for the acid. Most challenging for development of an acid resistant HTV is achieving such optimization without compromising other material properties, especially a high level of tracking and erosion resistance, mechanical properties and overall ease of processing. Moreover, all required material and design tests according to applicable IEC standards must still be passed.

Given a limited choice of different HTV components such as base polymer, curing agent and ATH fillers (including particle size, distribution and surface treatment), further additives were therefore evaluated with respect to:

• electrical properties (e. g. volume resistivity, tracking and erosion resistance);
• mechanical properties;
• multi-stress (e.g. accelerated weathering test);
• processing variables, especially viscosity, de-moulding ability, shelf life, and behaviour during storage.
Finally, an appropriate HTV formulation was found that successfully passes the acid resistance test, while also maintaining a high content of ATH filler (see Fig. 5).

Multi-Stress Evaluation of Nitric Acid vs. Tracking & Erosion Resistance

Polymeric materials in service are exposed to a combination of environmental factors (e.g. UV, acids) as well as electrical stresses (e.g. water droplet corona and dry-band arcing) that can simultaneously affect them. IEC TR 62039, defines 13 key properties, including minimum requirements, that must be fulfilled by polymeric materials for high-voltage outdoor applications. One of these is tracking and erosion resistance according to the test method described in IEC 60587. In the case of HTV silicone formulations, a high level of tracking and erosion resistance is related to ATH filler content. In parallel, nitric acid attacks and decomposes the ATH filler in an HTV silicone material. Therefore, the question arises of how previous nitric acid attack can affect the tracking and erosion resistance of these formulations.

To obtain information on combined material stress by nitric acid attack and tracking and erosion resistance, plate shaped specimens of HTV 1 (as per Table 1) were pre-stressed by storage in nitric acid for different time intervals and subjected to the tracking and erosion test according to IEC 60587. A 0.1 mol/l nitric acid concentration in de-ionized water was used for pre-stressing. Five samples, sized 120 x 50 x 6 mm, were placed in a container with 1.3 l acidic solution and stored at 35°C for periods of 96h, 240h, 504h, 1008h and 2184h. Compared to the above mentioned acid concentration of 1 mol/l, a 10 times weaker nitric acid concentration was used to avoid severe crack formation on samples after the selected storage time, which could then have additional influence on tracking and erosion resistance. Moreover, ingress of nitric acid into test specimen might also result in an increase in surface conductivity, which could also have an undesirable effect on results of the tracking and erosion test. Weight of the samples was measured both before and after storage in nitric acid in order to detect any mass loss during storage.

A tracking and erosion test was performed according to IEC 60587, Method 1 (i.e. constant voltage application), Criterion A (i.e. leakage current criterion) at 4.5 kV AC stress. Given these parameters, the tracking and erosion test was deemed passed after a duration of 6h if none of the 5 specimens tested:

• exceeded a leakage current of 60 mA for 2 s;
• showed a hole (bulk erosion) due to intensive erosion;
• started burning.

Fig. 6 shows the test set-up at the start. Results were analyzed by evaluating weight loss of the specimen during the test as well as the length, width and depth of any observed erosion. Fig. 7 provides findings on mass loss after nitric acid storage. As expected, according to Fig. 3, increase in mass loss with longer storage time could also be observed for the reduced nitric acid concentration of 0.1 mol/l.

Test results of the tracking and erosion tests after nitric acid storage for defined intervals are shown in Fig. 8. Evaluation of erosion depth is considered the most characteristic parameter for silicone elastomers with this test.

All samples passed the tracking and erosion test, regardless of their storage time in the nitric acid. Up to a storage time of 1008h, no change was observed in erosion depth while a storage time of 2184h resulted in a slight increase in depth. Visual inspection of test samples revealed an increase in superficial erosion (see Fig. 9).

Moreover, area of superficial erosion increased with time in acid storage. While after 96h it only covered a small centre area between the electrodes, this type of erosion covered the entire area between electrodes after 1008h storage. After 2184h, a complete layer of silicone was eroded between electrodes. This effect is thought to be related to increasing intensity of nitric acid attack over the course of the storage period by locally dissolving ATH filler from the surface of the silicone rubber. Despite the acid attack, the standard HTV silicone formulation selected was still able to pass the tracking and erosion test. This was because the intact material volume was able to withstand the phenomenon of bulk erosion, i.e. the typical failure mode for such material formulations during a tracking and erosion test.

New Test Method to Evaluate Core-Housing Adhesion

An essential interface for composite insulators in service is that between the core rod and the housing. Only good adhesion can assure reliable long-term performance by avoiding risk of moisture accumulation and flashunder development at and along this interface.

Recent years have seen an increase in the number of reports of damage where the root cause related to such lack of adhesion and this has been observed for post, long rod and hollow core insulators. Given this negative experience, there have been extensive investigations into developing an up-to-now missing test method for quantifiable adhesion testing of the core-housing interface. This included an international round-robin test with the participation of different test laboratories. The so-called ‘pull-off test’ is based on the water diffusion test specified in IEC 62217 and includes a mechanical adhesion test. Initially, the water diffusion test was developed to evaluate hydrolysis resistance of an insulator core. However, due to the high quality of FRP rods these days, it further developed into a method to evaluate the quality of the interface. Here, the pull-off test was used for evaluating existing insulator designs to gain more experience with this new method. Moreover, it was decided to evaluate the suitability of new silicone formulations during development since proper chemical bonding as a key property also relates to housing material formulation.

A pull-off test was conducted:

• insulator segments were prepared, acc. to IEC 62217, cl 9.4.2, including 1 shed;
• the water diffusion test was conducted by boiling samples for 100h in salt water and performing a voltage test;
• the test criterion: leakage current at 12 kV < 0.1 mA;
• afterwards, tensile tests were conducted perpendicular to the interface (see Fig. 10),
• the criterion: σ > 1.5 N/mm²,
• an additional criterion was to evaluate type of fracture (see Table 3).

These tests involved two different types of long rod (A1, A2), one post (A3) and two types of hollow core insulators (B, C) from different manufacturers. In addition, HTV and RTV (room-temperature curing) silicone rubbers served as the housing materials (see Table 4). Insulator types A1-A3 passed the pull-off test, demonstrating the desirable cohesive interface fracture. While A3 showed higher leakage current, this was related to larger core diameter. Insulator type B failed the test by exceeding the leakage current criterion and having too low pull-off tensile strength. This was in line with the undesirable adhesive interface fracture observed. Insulator type C passed the test but showed a mixture of cohesive/adhesive interface fracture. Based on this initial investigation, the test method was seen to yield a correct and quantifiable evaluation of interface quality. Still, adapting the leakage current criterion to core diameter size is probably necessary.

In a next step, this new method was applied during selection of a housing material in respect to sufficient interface adhesion. A total of 18 different HTV formulations were investigated using the pull-off test. Sample insulators were manufactured by injection moulding different material formulations onto FRP cores with a diameter of about 18.6 mm. Using identical conditions and designs for production of these sample insulators allowed direct comparison of leakage current and pull-off tension. Fig. 11 shows examples of the leakage current measurements obtained.

Test results for samples 3 and 7 showed significantly higher leakage currents and a subsequent manual adhesion test confirmed lack of adhesion in these samples. However, unsatisfactory adhesion was also evident from the rod surface of sample 4 and to a lesser extent for sample 14, although not recognizable from the leakage currents (see Table 5). This was attributed to localized reduction in adhesion, which did not have an impact over the entire 30 mm length of the specimen and increased leakage current due to accumulation of moisture at the interface. In summary, the test method was deemed suitable to evaluate new housing material formulations during their development.

In another example, a different FRP rod supplier was evaluated and the pull-off test used to confirm proper adhesion between housing and core. Preliminary investigations confirmed suitable properties, according to IEC 62217. The FRP rods passed the water diffusion test as well as adhesion with the standard housing material in a manual adhesion test. Nonetheless, insulator samples showed unexpectedly high leakage currents (in the range of ~200 µA versus the expected ~60 µA). The pull-off tension of ~2.0 N/mm² was in the expected range above 1.5 N/mm².

Additional evaluation of type of fracture showed proper adhesion with good chemical bonding. Since the housing material was a standard formulation that also passed this test with the currently used rods, the only root cause could be some chemical interaction between the rod surface primer-housing material with the salty solution during boiling. The question then arose whether the new rod supplier would pass the pull-off test. Despite exceeding the leakage current criterion, proper adhesion was confirmed with the pull-off tests, which was the main goal of the investigation. To generate an overall evaluation of the suitability of this test method, all tested long-rod insulator samples with HTV housing were presented so as to correlate leakage current with pull-off tension, with distinction made between cohesive and adhesive fractures (see Fig. 12).

Overall, there was no clear correlation between pull-off tension and leakage current for the long-rod insulators investigated and no conclusions about type of interface fracture could be drawn. Correct interpretation of leakage current is regarded as challenging since this is an integral parameter influenced by the test arrangement in general and by sample dimensions (i.e. shank diameter) in particular. Any analysis of potentially excessive leakage currents, must identify and take into account the relative contribution by the FRP rod, the housing and their interface.

In conclusion, in order to assure proper evaluation of the results from the newly developed pull-off test, the combination of leakage current, pull-off tension and type of interface fracture are all meaningful parameters to consider. Further experience will need to be gained to determine suitable limits. Currently, this test is under consideration for standardization in IEC 61109 as a product standard for composite long rod insulators.

CATU Test Laboratory

FRANCE

EDP Labelec

Portugal

See more Laboratories

Summary & Conclusions

Current trends in composite insulator technology relate to growing demand from UHV and HVDC applications. Another factor are special requirements now found in some utility specifications that may require new product development or distinct material properties, e.g. a nitric acid resistant HTV formulation.

To maintain the required high level of quality of composite insulators and to assure expected reliable long-term performance, new test methods are now under development and being considered for standardization. These are demonstrated by the example of the newly developed pull-off test to evaluate proper core-housing interface bonding. Composite insulators have been used for decades yet remain technically exciting products with potential for continuous improvement.

References

[1] C. Baer, F. Schmuck, J. Strumbelj, E. Tinner, J. Lachman, S. Kornhuber, J. T. Loh: Technical Demands to Improve Today`s Composite Insulator Reliability, Paper B2-221, CIGRE Centennial Session 2021
[2] J. Strumbelj, C. Baer, J. Lachman, F. Schmuck: Application of Multi-stress Test Methods to evaluate To-day`s Composite Insulator Reliability, Paper B2-669, CIGRE Session 2022
[3] K. O. Papailiou, F. Schmuck: Silicone Composite Insulators, Materials, Design, Applications, Springer, Berlin, Heidelberg, 2013
[4] J. Lachman: Lessons from 25 Years’ Experience Testing Polymeric Insulators, INMR World Congress, Tucson USA, October 2019
[5] I. Gutman, J. Lundengard, C. Ahlrot: Need of standardized adhesion test for composite insulators: lesson learnt from service experience, 20th Symposium on High-Voltage Engineering (ISH), 2017
[6] C. Ahlrot, P. Aparicio, A. Berlin, T. Condon, J.-F. Goffinet, I. Gutman, K. Halsan, R. Radosavljevic, K. Varli, K. Välimaa: New test procedure intended to evaluate adhesion of core/housing interface of composite insulators”. CIGRE D1-303, CIGRE Session 2020
[7] I. Gutman, C. Ahlrot, P. Aparicio, A. Berlin, T. Condon, A. Dernfalk, J.-F. Goffinet, K. Halsan, K. Kleinekorte, J. Lundengård, M. Radosavljevic, P. Sidenvall, S. Steevens, K. Varli, K. Välimaa: Development of Innovative Test Procedure for Evaluation of Adhesion of Core-Housing of Composite Insulators: from Root Cause of Failures in Service to Reproducible Test Procedure CIGRE Science & Engineering, N°20 February 2021

The post Optimizing Silicone Rubber-Based Housing Materials for HV Composite Insulators in Outdoor Applications appeared first on .

Leer artículo en español

The failure of an arrester almost always results in a complete short circuit inside its housing. In most scenarios, failure occurs due to dielectric breakdown, whereby the internal structure has deteriorated to the point where the arrester is unable to withstand applied voltage, whether normal system voltage, temporary power frequency overvoltage (e.g. following external line faults or switching) or lightning or switching surge overvoltages. There are a variety of reasons why an arrester might reach such a state. This edited past contribution to INMR by industry expert, Michael Comber, reviewed the principal modes by which arresters fail.

Protektel sp. z o.o.

Poland

EMCO Industries

Pakistan

See more suppliers of Arresters

Moisture Ingress

Perhaps the most common cause of arrester failure is moisture entering its interior. This implies that the arrester was:

• not well designed, or
• not properly manufactured, or
• damaged by some external force resulting in a compromise to its sealing system.

While the underlying cause in each case may be the same, the way this progresses to eventual failure can vary significantly.

In the case of a hollow core arrester where there is gas space around the column of MOV blocks (typically dry air or nitrogen), even a tiny leak can result in what is referred to as ‘seal pumping’ due to pressure differentials. For example, during the day suns heats the arrester such that the internal pressure increases relative to ambient and outward gas leakage occurs. When the arrester cools at night, this process reverses, with internal pressure dropping below ambient and external air (with all its moisture content) being drawn into the arrester. Such a cycle can repeat itself over many days, months or even years before the moisture inside builds to the point where there is a problem with reduced dielectric integrity.

In a solid core arrester design (with little to no internal gas space) this process will not take place. However leakage can still occur through imperfect end seals. In this case, moisture ingress is more due to ‘wicking’ – a process whereby moisture gradually finds its way down through interfaces between the MOV blocks and the materials in contact with them.

The manner in which dielectric integrity is degraded due to moisture ingress can also vary. The mere presence of moisture, if concentrated only within the gas inside a hollow core arrester, will not have significant impact on dielectric strength. Rather, it is how this moisture interacts with internal surfaces and materials that becomes the issue. It has been noted, for example, that moisture related failures of porcelain-housed arresters tend to occur more in the evening than during the heat of day. This is attributed to accumulated moisture condensing on the inside walls of the porcelain when it cools after sunset. Electrical strength across the wall is then progressively reduced until internal flashover occurs from end to end.

Moisture will typically not condense on the MOV blocks of an energized arrester because these generate enough heat to keep their temperature slightly higher than that of the surrounding gas. However, if the material used to coat or collar the blocks is hygroscopic, it can absorb moisture thereby causing some blocks to become more conductive on their outer surfaces. This essentially shifts voltage to other blocks and leads to higher conducted currents, external to some blocks but internal to others. Ultimately, the complete stack can no longer withstand the applied voltage. (Note: this scenario can be avoided by ensuring only non-hygroscopic collaring materials, such as glass, are used.) In the case of solid core arresters, moisture that has wicked into internal interfaces along either a portion or the complete length of the arrester can result in dielectric breakdown and failure.

Temporary Overvoltage (TOV)

Under normal operating conditions, which see the arrester energized at its maximum continuous operating voltage, Uc, the temperature of the MOV blocks rises only slightly above ambient. A point is then reached where the incremental heat generated is in equilibrium with the heat the arrester dissipates into the surrounding air. This represents the arrester’s normal thermal stability operating condition, depicted by the green dot on curve 1 of Fig. 1. Here, the blue curves represent power losses of the MOV blocks as a function of block temperature and the dark tan line represents the heat that can be dissipated from the arrester assembly, also as a function of block temperature. If the power frequency voltage across the arrester increases (e.g. due to a system disturbance, fault or switching operation), the MOV blocks conduct more current and begin to heat up.

As long as the overvoltage is below some critical limit, a new thermally stable operating point will be reached, albeit at a higher MOV block operating temperature, as depicted by the green dot on curve 2. However, should the overvoltage be of sufficient magnitude, the heat generated by the blocks remains greater than the unit can dissipate. A potential thermal runaway situation will then occur, as depicted by curve 3.

Should voltage return to normal (i.e. to the MCOV) before critical block temperature is reached, the arrester will remain thermally stable and will eventually cool to its initial condition, as depicted in Fig. 2. Here, curve A represents the conditions for normal operating voltage (i.e. MCOV) and curve B the conditions for an elevated voltage that can potentially lead to thermal runaway – even though this is avoided because voltage returns to normal before the critical temperature is reached.

On the other hand, if the overvoltage continues beyond the point at which critical MOV block temperature is reached, the temperature of the blocks continues to rise even if voltage returns to MCOV, as depicted in Fig. 3. In such a case (i.e. thermal runaway), the blocks eventually become so conductive that they can no longer support even MCOV and will be short-circuited, resulting in arrester failure.

Ageing of MOV Blocks

In the early days of metal oxide arresters during the mid to late 1970s, MOV blocks from all manufacturers exhibited some degree of ageing, whereby their power dissipation at any given voltage increased slowly, but continuously, over time. The resulting impact on arrester performance would be similar to that described for TOV – namely after some time in service, the power (heat) generated by the blocks would be basically similar to that resulting from a TOV occurring when the blocks were new.

As time progressed, the heat generated would be equivalent to that from a higher TOV on blocks in their new condition. Ultimately, the heat generated reached a point where no stable operation could be maintained, as depicted by curve 3 in Fig. 1. The blocks would then experience thermal runaway, just as if exposed to a sufficiently high, sustained TOV when in their original new condition.

This ageing characteristic of blocks was recognized early on and was addressed in ANSI/IEEE as well as IEC test standards by means of accelerated ageing tests. In these tests, sample blocks were subjected to MCOV for 1000h while maintaining block temperature at 115°C and it was considered that this was equivalent to 40 years service at 40°C. If at the end of 1000h the power dissipation was higher than at the start of the test, parameters for other duty tests had to be adjusted to account for this increase. The clear implication was that arresters passing the tests would be good for at least 40 years of service (provided of course that they operated within the specifications). With subsequent major improvements in processing technology, MOV blocks produced these days exhibit a characteristic whereby power dissipation actually decreases over time at any given voltage. This implies that they become more rather than less thermally stable during service and therefore are unlikely to cause arrester failure due to ageing.

Thermal Runaway from Surge Duty

The surge duty referred to here is that resulting from relatively high current surges due to lightning, switching of long lines or capacitor banks. Some of these may have very high amplitudes but relatively short duration (e.g. lightning surges), while others have much longer duration but with significantly lower amplitude (e.g. switching surges). Still, all have a charge content that, when passed through an arrester, results in a certain amount of energy absorbed by the blocks. This absorbed energy causes almost immediate adiabatic heating. MOV blocks typically have a specific heat capacity of about 3.3 J/cm³/°C, meaning that they will sustain a temperature rise of about 10°C for every 33 J/cm3 of energy (assuming this energy is input rapidly).

If the input of energy is excessive, the resulting temperature rise of the blocks may be such that the arrester is pushed into a thermal runaway condition. For example, with the arrester operating according to curve 1 in Fig. 1, the stable operating temperature will be that depicted by the green dot. If the temperature of the blocks is raised quickly (as a result of energy absorption) so that it becomes higher than depicted by the black dot on the same curve, then the arrester will not recover from this event and go into thermal runaway, as described earlier for a situation of prolonged TOV.

Damage to MOV Blocks From Surge Duty

One manifestation of the energy absorbed by the MOV blocks is rise in their temperature, as discussed above. However, if the energy is of sufficient magnitude and deposited over a relatively short period of time, the blocks can become irreversibly damaged.

For example, the resulting thermo-mechanical shock could cause them to crack into two or more pieces. In other cases, varistor blocks can be punctured in localized areas, either partially or completely through their body. In yet other cases, a pinhole type failure can occur at the edge of the block, possibly causing material to be removed from its outside surface. Typically, each such type of damage is accompanied by degradation of the block’s electrical integrity, manifest either by its inability to sustain another energy event without electrical breakdown or by a reduction in its capacity to support normal operating voltage. Both situations can sooner or later result in complete arrester failure.

The post Failure Modes for Surge Arresters appeared first on .

Steady-state sheath voltages are kept within limits by proper selection of cable section lengths, grounding the sheaths at strategic locations and applying sheath cross bonding for field cancellation. If sheath grounding or application of SVLs are performed improperly, the SVL and the cable itself are at greater risk of damage.

This edited past contribution to INMR by experts at Power Engineers in the United States provided basic information about proper SVL operation and focused on consequences to SVLs in the event of failure to ground the cable sheath.

DALEKOVOD OSO

Croatia

PFISTERER

Germany

See more suppliers of Cable Accessories

Listen to Online Lecture on Sheath Voltage Limiter Failure from Improper Bonding of Cable Sheaths by Jon Leman

Sheath Induced Voltages

It is important to first understand the fundamentals of induced voltages on cable sheaths and Fig. 1 shows a simplified cable cross-section. While actual cables include additional layers, these prominent layers are sufficient for the purpose of this discussion. Any voltage on the sheath contributes to voltage difference between core and sheath and between sheath and ground. Typically, the insulation provided by the jacket is the controlling factor for maximum allowable sheath voltage. Typical jacket materials have a breakdown voltage in the range of 20-100 kV/mm. If the jacket is compromised, this can lead to further deterioration and cable failure.

Fig. 2 is a representation of a horizontal cable arrangement but with insulating components not shown. Each core conductor is surrounded by a concentric sheath, which is grounded at one end. From Faraday’s Law and Lenz’s Law, AC current flowing in the core induces a longitudinal voltage in the conductor sheath. Since one end of the sheath is grounded, the induced voltage appears between sheath and ground at the open end.

Distance between phases is often relatively small in the case of cable installations. Therefore, currents in adjacent phases and other nearby circuits should be considered when calculating the voltage induced in any sheath. If the phase currents are balanced (i.e. same magnitude with phase angles 120° apart), field cancellation effects can help reduce severity of induced voltage. Single-phase installations that do not benefit from field cancellation could see higher induced sheath voltages for the same core current.

IEEE Standard 575 and references cited at the end of this discussion provide methods for calculating sheath-induced voltages. Simulation can also be used but care must be taken to properly model return currents and electromagnetic coupling between sheaths, phase conductors and ground conductors. Table 1 offers an example showing results calculated using equations from IEEE 575 and verified with simulation using ATP-EMTP. A 230 kV XLPE insulated cable is arranged horizontally with 0.25 m spacing (0.8 ft). Section length is 305 m (1000 ft) and the sheaths are grounded at one end. Jacket thickness is 3.8 mm (0.15 in) and load and reported fault currents have a frequency of 60 Hz. Note that none of the induced voltages approach breakdown strength of the jacket (i.e. approx. 76 kV = 3.8 mm x 20 kV/mm).

Transients such as those from switching or lightning have higher frequency content and couple more effectively through cable capacitance. This can cause much higher voltage difference between sheath and external ground. Indirect lightning strikes to ground can also increase ground potential, causing a large transient voltage to appear across the jacket.

A simple ATP-EMTP simulation illustrates the impact of switching transients. A surge type source with 10 µsec rise time, 350 µsec tail time and amplitude of 2.0 per unit on 230 kV was applied to the core of one phase of the 230 kV XLPE cable described above. Transient voltage on the sheath of the same phase reached a peak value just above 110 kV while sheaths on adjacent phases saw peak transient voltages of about 80 kV. Depending on material and thickness of the jacket, these voltages could be high enough to result in puncture.

Basics of SVLs

More detail is available in Refs #6, 7, 8 & 9. A typical insulation coordination approach using SVLs to protect a cable jacket consists of the following:

1. Maximum continuous operating voltage (MCOV) of the SVL must be selected so that it will not conduct current for any sheath voltages induced by the maximum normal or emergency steady state load currents or fault currents flowing in core conductors;

2. The SVL is selected with a volt-current (V-I) characteristic that results in conduction for high voltages induced from transient events, such as switching and lightning. Since insulating properties of the jacket are not well defined and not assured by industry standards, ample protective margins are recommended;

3. The SVL must have an energy absorption capability sufficient for possible events that could result in SVL conduction. Total energy ratings and temporary overvoltage curves provided by manufacturers should be consulted.

SVLs are applied at locations corresponding to sheath-to-ground voltage peaks. This includes the open end of single point bonded cable sections or at section terminations where the sheaths of cross-bonded cable configurations are transposed (see Fig. 3).

Rigorous coverage of insulation coordination analysis for SVLs is beyond the scope of this paper and additional information about specification of SVLs is found in the literature. Fig. 4 shows the characteristics of an SVL selected for the 230 kV XLPE cable installation example. Applying this SVL and simulating the switching transient described above results in clamping of sheath voltages to about 11 kV (down from 110 kV). Transient SVL current peaks at about 5500 Amps and SVL energy absorption is approx. 615 Joules.

Impact of Ungrounded Sheath

Fig. 5 offers a circuit representation of one phase of a single point bonded system. One end of the sheath is connected directly to ground and an SVL is connected between sheath and ground at the other end. The capacitor circuit elements represent the sum of the distributed capacitance between core and sheath and sheath and ground. So long as one end of the sheath is grounded, sheath-to-ground capacitance has negligible effect. Steady-state induced sheath voltages in this case are due primarily to magnetic fields from core currents which couple to the sheath (i.e. mutual inductance, not shown).

If connection to ground is broken or not installed, electrostatic effects dominate and voltage at the sheath will be determined by a voltage divider using the respective capacitances. If resulting sheath voltage is higher than the SVL’s V-I curve, the SVL will conduct. Returning to the 230 kV cable discussed above, assume cables are direct-buried in conductive ground (i.e. ground potential reference at the surface of the cable jacket). Capacitance per unit length between concentric cylindrical conductors can be calculated using equation (1). Table 2 lists input data and results of the voltage divider calculation, overlooking presence of the SVL.

where:

ε₀ is the permittivity of free space;
ε_r is the relative permittivity of the dielectric;
r_s is the inner radius of the outer conductor; and
r_c is the inner radius of the inner conductor.

Comparing sheath voltage result to the V-I curve in Fig. 4, it can be seen that voltage is high enough to cause the SVL to conduct. However, the steady-state current cannot be found directly from the SVL’s V-I characteristic because it is limited by core-to-sheath capacitance. Manufacturer data typically does not include V-I characteristics for currents below 500 Amps. If nominal voltage of the cable system is much larger than the lowest published voltage of the V-I characteristic curve, current through the SVL can be approximated by assuming all the voltage drop is across the core-to-sheath capacitance. The larger the voltage across the SVL with respect to nominal system voltage, the more this approach will tend to overestimate SVL current. In this case, current will be limited to:

Even though this current is low, the heating effects are substantial and can result in rapid deterioration and failure of the SVL. (This will be explored below). An error that is occasionally made in conjunction with failure to ground the sheath is to install SVLs at both ends of the sheath. The result is similar but the current in each SVL will be approximately half the total capacitive current. The example above is for a direct-bury cable in which the reference ground surrounds the jacket. This configuration was selected for simplicity in calculating sheath-to-ground capacitance. But what about other cable configurations? Figs. 6 and 7 compare results of electrostatic finite element analysis for three cable configurations. .

In each case, distance between sheath and ground plane is successively larger and such larger distances result in reduction in capacitance between sheath and ground. The corresponding increase in capacitive reactance between sheath and ground causes an increase in sheath potential due to the voltage divider. Note that these are steady-state voltages the sheath would see without an SVL connecting it to ground. In all three cases, the example SVL would conduct. The further the sheath is from reference ground, the higher will be the floating potential. However, current in the SVL will be approximately the same in all three cases since it is the core-to-sheath capacitive reactance that limits the flow.

Thermal Impact of Induced Current

Estimated SVL current in an ungrounded sheath, as discussed above, was 2.47 Amps rms (about 3.5 Amps peak). In reality, this current will be a distorted waveform since the SVL only conducts for a portion of each cycle. Fig. 8 shows a simulated result of SVL currents for each phase of the 230 kV case installation having ungrounded sheaths. The simulation uses the SVL’s V-I characteristic from Fig. 4. The program estimates the V-I characteristic for operating points outside those specified by manufacturer data. Variations and asymmetry in current waveforms are due to the non-linear V-I characteristic and series capacitance as well as from variation in induced sheath voltages caused by asymmetry inherent in the horizontal cable arrangement (A to B spacing = B to C spacing ≠ A to C spacing).

Fig. 9 plots power absorbed by the SVLs. Peak power is about 26.5 kW and average power is about 13.3 kW. This level of power dissipation would cause rapid heating and failure of the SVL given that a typical SVL has a maximum operating temperature of 60°C (140°F). If temperature is too high, properties of the zinc oxide blocks change and a thermal runaway condition can occur.

To estimate current required to reach maximum operating temperature, a thermal finite element analysis was prepared for a representative SVL (see Fig. 10 which lists the thermal characteristics of the model). The zinc oxide block was set as a variable volumetric heat source with heat loss through conduction and radiation at the housing-to-air interface. The SVL model was placed in a region about the size of a link box enclosure. Boundaries of the link box region were set to a constant temperature of 10°C (50°F).

The simulation iterates until equilibrium is reached between heat sourced by the ZnO blocks and heat flow out of the SVL. Maximum temperature in the SVL was then recorded. This was performed for a range of ZnO heat generation levels. Post processing calculations determined conduction currents that would produce the ZnO heat energy assuming a 5 kV residual voltage across the SVL and assuming the instantaneous power is about twice the average power. Fig. 11 provides results of the analysis. These values are only an approximation but nonetheless demonstrate that even little steady-state current is sufficient to push an SVL to its thermal rating.

SVL currents in the milliamp range are possible for distribution level circuits with short section lengths. Under these circumstances, the SVLs could operate for some time before failing and would be sensitive to ambient temperatures and other thermal conditions.

Conclusions

Sheath voltage limiters are important in order to reduce cable losses and to protect cable jackets from lightning and switching transients but their use requires that the sheath be properly grounded. Failure to ground the sheath results in a floating potential condition. If the steady-state sheath voltage exceeds the V-I characteristics of the SVL, the unit will begin to conduct continuous current, limited by cable core-to-sheath capacitance. If an ungrounded condition exists on a transmission cable sheath, SVL currents can be in the range of a few Amps. This corresponds to an instantaneous power in the range of tens of kW and SVL failure will then likely occur quickly. The process leading to failure can take longer on short distribution circuits where capacitive currents might be in the milliamp range. Analysis of ungrounded conditions is not necessary for proper specification of SVLs. However, the conceptual analysis and calculations completed above are for training purposes and also provide information to help troubleshoot SVL failures.

References

[1] Lapp Tannehill, “Insulation/Jacket Materials: Physical Properties Chart,” [Online]. Available: https://www.lapptannehill.com/resources/technical-information/insulation-jacket-materials-physical-properties-chart. [Accessed 16 September 2019].
[2] IEEE Power and Energy Society, “IEEE Std 575-2014, IEEE Guide for Bonding Shields and Sheaths of Single-Conductor Power Cables Rated 5 kV through 500 kV,” IEEE, New York, NY, 2014.
[3] C. Adamson, E. Taha and L. Wedepohl, “Determination of the Open-Circuit Shath Voltages of Cable Systems,” IEEE Proceedings, vol. 115, no. 8, 1968.
[4] M. Shaban, M. A. Salam, M. A. B. Sidik, Z. Buntant and W. Voon, “Assessing Induced Sheath Voltage in Multi-Circuit Cables: Revising the Methodology,” in IEEE Conference on Energy Conversion (CENCON), Johor Bahru, Malaysia, 2015.
[5] V. K. Gouramanis, G. C. Kaloudas, T. A. Papadopoulos, K. G. Papagiannis and K. Stasinos, “Sheath Voltage Calculations in Long Medium Voltage Cables,” in IEEE PowerTech, Trondheim, 2011.
[6] J. Woodworth, “Sheath Voltage Limiters Protect HV Power Cables,” Zimmar Holdings Ltd./INMR, 9 February 2019. [Online]. Available: https://www.inmr.com/sheath-voltage-limiters-protect-power-cables/. [Accessed 18 September 2019].
[7] D. Cao, X. Liu and X. Deng, “The Suitability Analyses of Sheath Voltage Limiters for HV Power Cable Transmission Lines,” in 2nd International Conference on Electrical Materials and Power Equipment, Guangzhou, China, 2019.
[8] A. Heiss, G. Balzer, O. Schmitt and B. Richter, “Surge Arresters for Cable Sheath Preventing Power Loss in MV Networks,” in 16th International Conference and Exhibition on Electricity Distribution, IET, Amsterdam, Netherlands, 2001.
[9] Insulect, “Sheath Voltage Limiters,” 2019. [Online]. Available: https://insulect.com/products/sheath-voltage-limiters-svl. [Accessed 17 September 2019].
[10] C. R. Paul, Analysis of Multiconductor Transmission Lines, Second Edition, Hoboken, NJ: John Wiley & Sons, 2008.
[11] J. He, R. Zeng, S. Chen and Y. Tu, “Thermal Characteristics of High Voltage Whole-Solid-Insulated Polymeric ZnO Surge Arrester,” IEEE Transactions on Power Delivery, vol. 18, no. 4, 2003.
[12] S. B. Lee, S. J. Lee and B. H. Lee, “Analysis of Thermal and Electrical Properties of ZnO Arrester Block,” Current Applied Physics, vol. 10, no. 1, pp. 176-180, 2010.

The post Sheath Voltage Limiter Failure From Improper Bonding of Cable Sheaths appeared first on .

This edited contribution to INMR by Dr. Poorvi Patel of the Electric Power Research Institute (EPRI) in the United States reviews the content of a 2022 IEEE Guide for dielectric frequency response (DFR) on bushings (C57.12.200)*. It also provides testing guidelines for time-based, type-based or phase-based DFR analysis and presents a case study in which DFR helped a utility identify bushings with potential issues.

* REPRINTED WITH PERMISSION FROM IEEE. COPYRIGHT IEEE 2022. ALL RIGHTS RESERVED

Hitachi Energy Transformer Components and Service

Switzerland

PFISTERER

Germany

See more suppliers of Bushings

Dielectric Frequency Response (DFR), also called frequency domain spectroscopy (FDS), is a diagnostic technology that has been practiced for many years. In mid-1990, the first commercial portable field device was introduced to the market, and the application was mainly to access the condition of paper-oil insulated cables. A few years later, the technology was also used on paper-oil insulated power transformers to measure the average moisture content of transformer insulation. In 2004, CIGRÉ developed Technical Brochure 254, “Dielectric Response Methods for Diagnostics of Power Transformers”, which focused on three dielectric response techniques: return voltage measurements (RVM), dielectric spectroscopy in time domain (TDS), and dielectric frequency domain spectroscopy. As DFR became more frequently used in the field to assess transformer insulation moisture, a Guide specific to this technology was needed.

In 2010, CIGRÉ published Technical Brochure 414, “Dielectric Response Diagnoses for Transformer Windings”, which combined acquired practical experience with the mathematical modelling approach to determine testing best practices. The document covered the fundamentals of the insulation model, measurement configurations, test performance, and a comparative analysis using Karl Fisher titration to validate using dielectric response methods to detect moisture estimation in solid insulation. In 2018, IEEE published C57.161 “Guide for Dielectric Frequency Response (DFR) Test Methods of Liquid Immersed Transformers”. The Guide was focused on how DFR could be used to estimate insulation moisture. However, at this time, there was more knowledge of DFR cases from the field, in which this technology was showed to be sensitive to insulation contamination such as carbon contamination and corrosive sulphur (Cu2S). The Guide presented a few such cases in the Annex.

Moreover, bushings are a vital part of high voltage transformers and reactors. It is important that all maintenance programs need to include bushing condition assessment. A problem with a single bushing can prevent the transformer/reactor from returning to service. CIGRÉ’s 2015 Technical Brochure 642 – “Transformer Reliability Survey” found that about 20% of transformers fail due to bushing-related issues. A bushing failure can result in equipment damage, increased system disturbance, and loss of load. The failure of an oil-filled bushing can be more problematic than dry-type bushings and can cause substantial collateral damage. Like transformer insulation or paper-oil cables, bushing insulation of oil-filled bushings is mainly paper and oil. Thus, DFR could also be used for condition assessment of bushings. Not limited to oil-impregnated paper (OIP) bushings, DFR can also be applied to resin-impregnated paper (RIP), resin impregnated synthetic (RIS), and resin bonded paper (RBP) bushings. Bushings could also be wall-type bushings or roof-type bushings, there is limited knowledge how DFR could be applied and analysed on these types of bushings today.

Performing DFR on capacitance graded bushings in the field is not as prevalent as performing DFR on transformers/reactors; however, it is gaining popularity. Significant research, lessons learned, and knowledge sharing from the field are needed to understand the value of this technology. Today, primary diagnostic testing on bushings is 50/60 Hz power factor and capacitance testing. Reviewing 50/60 Hz power factor and capacitance may not reveal developing condition degradation of the bushing prior to failure. The goal of DFR testing is to identify problematic bushings in early stages, as this technology measures the insulation power factor and capacitance over a frequency band.

In 2018, the IEEE entity group in China, together with the IEEE Transformer Committee, established a Guide for DFR on bushings. IEEE C57.12.200 was published in 2022. This Guide provides guidelines for performing DFR measurement of various types of condenser bushings (capacitance graded bushings) either in the field or in the factory. The Guide does not cover gas-insulated bushings but covers measurement technique, measuring equipment requirement, and interpretation of results based on comparisons. The Guide also covers theory related to DFR, such as temperature correction using the Arrhenius equation, the influence of surface creep currents, and the influence of voltage. The Annex describes a few DFR bushing case studies.

IEEE C 57.12.200, while not complete, is a step forward to guide testers and enable them to begin measuring and analysing bushing DFR. Future updates to this Guide need to collect more data and case studies to improve understanding of the value of the technology as well to better understand the types of abnormalities in early stages this technology can diagnose that other technologies cannot diagnose.

Performing DFR Testing

Just as with the 50/60 Hz power factor capacitance test, DFR is a nonintrusive, non-destructive off-line testing technique. Measurements are obtained in the frequency domain. The results can be displayed as capacitance and power factor as a function of frequency. The power factor reflects loss characteristics determined by polarization and conduction phenomena of insulating materials. On power transformers, the DFR measurement could be used to determine average moisture content of solid insulation and conductivity of liquid insulation. If contamination of transformer or reactor insulation is suspected (i.e., carbon contamination of corrosive sulphur contamination), the DFR has been quite successful at detecting such anomalies.

The use case for DFR measurement on bushings remains unclear. In several cases, where the dielectric losses at lower frequencies are higher on one bushing compared to similar bushing installed on the same transformer with identical bushing type, rating, and manufacturer (i.e., sister bushing units). When inspecting the problematic bushing, an abnormality such as contamination has been found. Using DFR on the transformer, the analysis provides an average moisture value, while for bushings, the proper algorithms are not available. If a bushing has higher moisture content compared to its sister bushings, it may be detected by comparing the DFR traces. The bushing insulation system (capacitance graded bushing insulation) consists of a composite dielectric media, OIP, RIP, or RIS (see Fig. 1). Most DFR testing on problematic bushings has been performed on OIP-type bushings. Testing on OIP bushings has shown that DFR can detect bushings that are in poor condition. However, more testing is required for other types of bushings such as RIP, RIS, and RBP bushings.

Test Set-up
The set-up and procedure of the DFR measurement is like the power factor measurement at power frequency described in IEEE Std 62-1995. The most common testing performed is the C1 capacitance and power factor testing. Fig. 2 shows the capacitance and dielectric loss trace of a bushing in good condition. Losses are typically quite low for a new bushing or a service-aged bushing with no issues.

The measurement can be performed on all types of bushings with test or voltage taps. Comparison with the reference can reveal the change in overall insulation condition. The impact of contamination and degradation is more prominent in the low-frequency range for bushings with paper insulation.

Temperature is a key factor that influences DFR traces, hence, it is very important to record the ambient temperature before testing. If the bushing is mounted on a transformer or other oil-filled apparatus, the top oil temperature should be recorded as well. Furthermore, the surface condition can also affect DFR traces. The traces can exhibit a different appearance, depending on whether the surface is wet or dry. Thus, cleaning the bushings before testing is needed for both C1 and C2 testing (see definitions below). The environmental conditions such as humidity and weather (if the measurement is performed outside) are also important to record.

Before testing, all the windings on which the tested bushings are installed, including the neutral if available, need to be short-circuited. All other windings accessible through bushing terminals should also be shorted and grounded. As an example, if a high-voltage (HV) side bushing is being tested on a two-winding transformer, the HV side needs to be shorted, and the LV side needs to be shorted and grounded. For an autotransformer, windings that have internal electrical connections should be shorted together. Failure to short the windings properly could introduce interference or resonance effects that can lead to erroneous results.

In some cases, the DFR testing is performed before installation on a transformer or in a storage facility. In these cases, the bushings can be tested in an oil-filled tank (to minimize surface leakage current on the oil side) or mounted on a grounded metallic rack. The bushing flange should be grounded to a solid ground.

Bushings should not be tested in the shipping container. The reason is that the DFR could consider the shipping cage, especially a wooden container, to be part of the bushing insulation. In any case, contacting the bushing manufacturer for more guidance before performing testing is advisable.

The C1 capacitance is the main bushing insulation measured from the conductor to the bushing tap (see Fig. 3). The C2 capacitance is the insulation/oil from the outermost capacitance layer to ground. Bushings rated up to 69 kV, on the other hand, have a built-in C2 capacitance that is dependent upon outer layers of paper with glue and an oil gap (Fig. 3a). This C2 capacitance could be influenced by external factors such as insulator surface contamination, and air and oil surrounding the bushing. Bushings rated at 115 kV and above have C1 and C2 capacitance that are influenced mainly by on the bushing condenser paper, which leads to more reliable capacitance and power factor test results (Fig 3b). If time is limited for the testing, the C1 DFR capacitance and power factor testing should always be performed.

Test Lead Connections for C1 DFR Testing
To perform C1 capacitance and power factor DFR testing of a bushing that is not installed on a transformer, the HV supply lead should be connected to the conductor, and the current measuring lead should be connected to the test tap or voltage tap (see Fig. 3). The ground lead of the test instrument should be placed on the same ground as the bushing flange ground.

For a bushing installed on a transformer, the connections are similar to when it is not installed. The HV lead should be connected on the conductor, and the current measuring lead should be connected on the test or voltage tap (see Fig. 4). However, in this case, the windings on which the tested bushings are installed, including the neutral if available, should be shorted (Fig. 4). The Figure shows an example for a two-winding transformer. All other bushings that are not tested should be shorted and grounded.

Test Lead Connections for C2 DFR Testing
Performing C2 capacitance and power factor testing for bushings above 115 kV may be worthwhile, as these bushings have a defined insulation path (see Fig. 3b). To perform C2 capacitance and power factor DFR testing of a bushing that is not installed on a transformer, the HV supply lead should be connected to the test or voltage tap, Fig. 3 and the current measuring lead should be connected to the conductor. The guard lead should be connected to the flange.

For a bushing installed on a transformer, the connections are similar to when it is not installed. The HV lead should be connected on the test or voltage tap, the current measuring lead should be connected on the conductor, and the guard lead should be connected to the ground. Also, in this case, the windings on which the tested bushings are installed, including the neutral if available, should be shorted. All other bushings that are not tested should be shorted and grounded.

Testing Mode & Steps to Take Before Starting Test
Before starting the testing, the bushing surface should be cleaned with plain rags. If a cleaning solution is used, ensure it does not leave a slight film residue on the bushing surface that could influence the testing results. Consult with the bushing manufacturer for recommended cleaning solutions. In most cases, using a clean, dry rag is sufficient. Furthermore, the ambient temperature should be recorded, and if the bushing is installed on a transformer, the top oil temperature should be recorded. The bushing temperature can then be calculated as the average of the top oil temperature and ambient temperature, as shown in Eq.1;

where,
T_body is the bushing temperature;
T_amb is the ambient temperature;
T_topoil is the top oil temperature of the transformer in which the bushing is installed.

The instrument settings for C1 testing should be Ungrounded Specimen Test mode “UST” and for C2 testing, should be Grounded Specimen Test “GST” mode.

The voltage for C1 testing is 200 V or 140 V RMS. If the testing is conducted in an electrically noisy environment (e.g., high levels of EMF), such as near energized equipment or a high-voltage, direct current substation, a higher voltage (e.g., 2000 V or 1400 V RMS)—may be needed. To supply a higher voltage, a voltage amplifier needs to be connected to the DFR measuring equipment.

When measuring DFR on a transformer, the recommended frequency band is typically 1000 Hz to 1.0 mHz. For bushings, the optimal best practice for a stop frequency is uncertain. If sufficient time is available, conducting the testing from 1000 Hz to 1 mHz for bushings is recommended. The IEEE C57.12.200 states that the frequency band should be at least from 1000 Hz to 10 mHz.

<p=class_1>

DFR Bushing Analysis

In one sense, analysis of DFR on bushings is similar to sweep frequency response analysis (SFRA). Both diagnostic technologies are based on the comparable trace method. The major difference is that SFRA compares impedance traces between 10 Hz and 2.0 MHz, while DFR on bushings compares the power factor traces (dielectric losses) typically between 1000 Hz and 10 mHz.

The bushing DFR trace analysis can be conducted by comparing the following:
• Time-based comparison. Compare measurement traces with a factory baseline or with results from previous routine testing on the same bushing—per IEEE C57.12.200*.
• Type-based comparison. Compare measurement traces with a baseline from the same type of bushing measured in the factory—per IEEE C57.12.200*.
• Phase-based comparison. Compare measurement traces with sister bushings installed on the same transformers—per IEEE C57.12.200*.

Typical Bushing Abnormalities Detected by DFR Measurements
DFR measurements on transformer insulation has been very successful in detecting abnormalities such as high amounts of insulation moisture or contamination. However, abnormalities found by conducting DFR on a bushing, whether it is oil filled or non-oil filled, remains in an exploration phase. More use cases from the field need to be conducted to understand the types of early signs of bushing issues that could be detected. Several users have reported certain types of faults and issues, such as X-wax contamination or shorted capacitive layers.
Assuming a valid C1 DFR measurement, several abnormalities could lead to unusual DFR traces*:
• High moisture content in solid insulation;
• High oil conductivity due to aging or overheating;
• Insulation deterioration (e.g., X-wax and carbon tracking);
• Excessive surface leakage current or badly deteriorated C₂;
• Shorted capacitive layers.

Temperature Correction Using Arrhenius Equation
Since insulation properties are dependent on temperature, DFR results need to be normalized to a common temperature, which is recommended to be 20°C for comparison and analysis. If the trace shape does not change over time, the bushing is in good condition. For DFT traces, the equation used for temperature shifting is called the Arrhenius equation (see Eq. 2). Correction from temperature T₁ to temperature T₂ could be achieved by shifting the frequency on a logarithm scale by factor L , which can be expressed by:

Where:
E_a is the activation energy of the insulation material in eV;
k_B is the Boltzmann constant (8.617× 10^-5 eV/K);
T is the temperature (°K) of the object.
Temperature correction is necessary for all three types of analyses: time-based, type-based, or phase-based. If all three bushings on a transformer are tested at about the same time, the phase-based approach may not require a temperature correction for proper analysis.

<p=class_1>

Time-Based Comparison
To perform a time-based comparison analysis, proper baseline DFR capacitance and power factor traces are needed. Time-based comparisons are made by comparing the DFR traces with baseline DFR measurement results. The baseline measurement can be obtained from a factory test or previous routine testing. Before a time-based analysis is performed, the measurements need to be corrected to 20°C using the Arrhenius Equation (Eq. 2). Fig. 5 shows power factor (PF) measurements of a 138 kV OIP bushing. The first routine testing of the bushing was made in June of 2020. The average temperature (calculated according to Eq. 1) at the time of the testing was 22.5°C. This test may be considered the baseline measurement for all future DFR testing comparison. In March 2022, the testing was repeated on the same bushing, and the bushing temperature was 8.4°C. A comparison of the DFR power factor (PF) traces measured in June 2020 and March 2022 without temperature correction as shown in Fig. 5a, seems to indicate that the bushing has undergone a significant change in its condition. However, Fig. 5b shows the traces corrected to the reference temperature of 20°C, which indicates the bushing condition has not changed since June 2020.

A general rule is that if the bushing condition is stable and its mounting condition has not changed, the capacitance and power factor curves will remain stable over time, as shown in Fig. 5b.

Type-Based Comparison
If the baseline measurements for the bushing tested are not available, a comparison can be made using DFR measurements from a bushing of identical design and rating type by the same manufacturer. The mounting condition also needs to be similar. A discrepancy may occur if a bushing installed on a transformer is compared with a similar type, manufacturer, and design of bushing, but the bushing is not installed on a unit. Also, for the type-based comparison, the temperature compensation to reference temperature 20°C prior to performing the analysis is important. Fig. 6 shows DFR power factor traces of two 245 kV, 600 A oil-filled bushings. It is noted that even if the voltage and current ratings are the same for the bushings, but the design is different, the DFR traces are not comparable. Fig. 7 shows a case where twelve 34.5 kV RIS bushings were measured. All the bushings are identical bushings from the same vendor. In this case, all the DFR PF traces are aligned.

Fig. 8 shows three power factor traces of 115 kV OIP bushings of the same type and manufacturer. The figure shows that two of the traces are aligned, while the third trace exhibits much higher losses. This could be due to bushing abnormalities such as higher moisture or contamination. In this case, closer investigation or replacement of this bushing would be recommended.

Phase-Based Comparison
If all the bushings on a three-phase unit, or three single-phase units (transformers or shunt reactors), are measured, then a phase-to-phase comparison can be performed. If the bushings are measured in the same testing occasion, the analysis could be performed without the need of temperature correction, as the three bushings probably operate at a similar temperature. The phase-to-phase comparison method (also called the phase-based comparison) is a variation on the type-based comparison. The bushings of the transformer or shunt reactor need to have an identical design and similar mounting condition before the analyses are performed. Fig. 9 shows three RIP bushings with the same design and manufacturer installed on a three-phase transformer. All traces are aligned, indicating no deterioration of the bushings.

Impact of Stray Currents on Bushing Surface
In some cases when performing C1-DFR testing on a bushing installed in a transformer, the traces become negative after a certain frequency (see Fig. 11). This behavior could be due to the influence of surface creep currents on the DFR results. These creep currents or surface leakage currents on bushing surfaces can affect the measured capacitance and power factor at line frequency and/or at lower frequencies. The effect of creep current can be minimized by cleaning the bushing on the air side (porcelain surface), but this may not resolve the issue. In some cases, it may be beneficial to measure the bushing when the bushing temperature is higher than the ambient temperature (i.e., at a higher transformer oil temperature).

Creep currents can also be present on the liquid side of bushing, along the surfaces of the part immersed in the insulating liquid of the transformer (see Fig. 12).

Three different effects are possible*:

• The creep current on the air side top part (T) of the bushing surface, that flows through RT and CT, adds a loss current, leading to the addition of a positive loss peak to the PF trace (Fig. 13).
• The creep current on the air side bottom part (B) of the bushing surface, that flows through RB and CB, adds a loss current, leading to the addition a of a negative loss peak to the PF trace (Fig. 11).
• The creep current on the liquid(oil) side (O) of the bushing surface, that flows through RO and CO, adds a loss current, leading to addition of a negative loss peak to the PF trace (Fig. 11).

Case Study

This case study was discussed in Annex G of IEEE C57.12.200 guide [1,12]. This case involves 25 OIP U-type bushings that were tested for 50/60 Hz power factor, DFR, and dissolved gas analysis (DGA). The testing was performed to identify deteriorating bushings, as many bushings of this type had begun to fail. Before the analysis was conducted, all the data was corrected to a reference temperature of 20°C. Moreover, all the bushings are installed on single-phase shunt reactors. This utility used the following guideline: if the bushing capacitance compared with both nameplate and previous tests has increased by 5%, the bushing should be further investigated or replaced due to possible insulation deterioration. All the DFR measurements were performed at 2.0 kV (1.4 kV RMS).

Fig. 14 shows that 60 Hz power factor testing of four of the 25 tested bushings was normal (green in the figure). However, the DFR traces showed that one of the bushings (OIP 2-green line) was in poor condition, and two of the bushings (OIP 13-blue curve and OIP 14-red curve) are in an initial phase of deterioration.

Fig. 15 shows the DFR power factor traces for all 25 bushings. Three out of the 25 bushings are in very poor condition and DGA of the oil also indicates that these bushings have a higher level of hydrogen.

Fig. 16 shows the hydrogen and acetylene (C2H2) levels of the 25 bushings. The three bushings (OIP 2, OIP 19 and OIP 25) have a higher level of hydrogen as well as C2H2. Figs. 15 & 16 indicate that one of the bushings is in its initial stage of deterioration, i.e. OIP 17. This bushing seems also to exhibit higher-level acetylene generation.

Forensic analysis of the bushings with the highest hydrogen level indicated that the bushings had a deposit of X-wax contamination (a yellowish deposit, shown in Fig. 17). X-wax is a by-product caused by partial discharge activity, generating a higher level of hydrogen and reacting with mineral oil to form solid particles of carbon and hydrocarbon polymers.

TECNALIA Electrical Labs

Spain

STRI

Sweden

See more Laboratories

Summary

Dielectric Frequency Response (DFR) on bushings is here to stay. In 2022 an IEEE Guide C57.12.200 (2022) describes methods for the gathering of DFR measurements for bushings and various methods of performing the analysis.

The discussion above highlighted points from the Guide and described an example in which DFR testing helped a utility make decisions on what bushings to replace and helped the utility diagnose a bushing that was in good condition. As this testing is becoming more prevalent, more experience and case studies are needed in which different types of abnormalities have been detected and reported. While the average moisture content of insulation in transformers can be determined, this is not currently possible for DFR testing of bushings. The DFR analysis for bushings is more of a comparative method, which can be conducted via a time-based, type-based, or phase-based analysis. The time-based and type-based require temperature correction using the Arrhenius methodology before the analysis is performed. For a phase-based comparison, temperature correction is not necessary if all the bushings have been tested at the same testing occasion.

More experience is also needed to determine the appropriate stop frequency for the testing. In the guide, the stop frequency was set to 0.01 Hz, but this may not provide sufficient data for proper analysis. Several cases have been reported in which DFR detected X-wax type of contamination inside the bushing. In a few cases in which DFR was performed, the traces appeared abnormal, but forensic analysis of the bushings revealed a puncture on a capacitive layer. Most of the reported cases re OIP-type bushings. Additional information is needed on failure modes that can be detected for RIP/RIS bushings by performing DFR. The critical question that needs to be answered is whether DFR can detect bushing problems in an early stage compared to conventional 50/60 Hz capacitance and power factor testing. Based on today’s knowledge, in most cases, the answer is ‘yes’.

Abbreviations & Acronyms
CIGRÉ Conseil International des Grands Réseaux Electriques
(Council on Large Electric Systems)
DFR dielectric frequency response
DGA dissolved gas analysis
EPRI Electric Power Research Institute
IEEE Institute of Electrical and Electronics Engineers
OIP oil impregnated paper
PF power factor
RBP resin bonded paper
RIP resin impregnated paper
RIS resin impregnated synthetic
RMS root mean square
SFRA sweep frequency response analysis

REFERENCES
1. IEEE Guide for the Dielectric Frequency Response Measurement of Bushings, IEEE C57.12.200, 2023. https://ieeexplore.ieee.org/document/10026269.
2. CIGRÉ. Dielectric Response Methods for Diagnostics of Power Transformers. Technical Brochure 254. Paris: CIGRÉ, 2004. https://cigreindia.org/CIGRE%20Lib/Tech.%20Brochure/254%20Dieelctric%20Response%20Methods%20for%20diagnostic%20of%20power%20transformers.pdf.
3. CIGRÉ. Dielectric Response Diagnoses for Transformer Windings. Technical Brochure 414. Paris: CIGRÉ, 2010. https://cigreindia.org/CIGRE%20Lib/Tech.%20Brochure/414.Dielectric%20dianost%20ses%20for%20Transformer%20windings.pdf.
4. IEEE Guide for Dielectric Frequency Response Test, IEEE C57.161, 2018. https://ieeexplore.ieee.org/document/8571325.
5. CIGRÉ. Transformer reliability survey. Technical Brochure 642. Paris: CIGRÉ, 2015. https://e-cigre.org/publication/642-transformer-reliability-survey.
6. IEEE Guide for Diagnostic Field Testing of Electric Power Apparatus – Part 1: Oil Filled Power Transformers, Regulators, and Reactors, IEEE Std. 62-1995 (Revision of IEEE Std 62-1978), 1995. https://ieeexplore.ieee.org/document/467562.
7. IEEE Standard General Requirements and Test Procedure for Power Apparatus Bushings, IEEE Std C57.19.00-2004 (Revision of IEEE Std C57.19.00-1991), 2004. https://ieeexplore.ieee.org/document/1440990.
8. IEEE Standard Performance Characteristics and Dimensions for Outdoor Apparatus Bushings, IEEE Std C57.19.01-2000 (Reaffirmed 2005), 2000. https://ieeexplore.ieee.org/document/836386.
9. IEEE Standard Performance Characteristics and Dimensions for Power Transformer and Reactor Bushings, IEEE Std C57.19.01-2017 ((Revision of IEEE Std C57.19.01-2000), 2018. https://ieeexplore.ieee.org/document/8410922.
10. IEEE Guide for Application of Power Apparatus Bushings, IEEE Std C57.19.100-2012 ((Revision of IEEE Std C57.19.100-1995), 2013. https://ieeexplore.ieee.org/document/6469143.
11. Peter Werelius, Mats Ohlen, Jialu Cheng, and Diego M. Robalino, “Dielectric frequency response measurements and dissipation factor temperature dependence,” 2012 IEEE International Symposium on Electrical Insulation, San Juan, PR, 2012, pp. 296–300, doi: 10.1109/ELINSL.2012.6251476. https://ieeexplore.ieee.org/document/6251476.
12. Diego M. Robalino, Ismail Güner, and Peter Werelius, “Analysis of HV bushing insulation by dielectric frequency response,” 2016 IEEE Electrical Insulation Conference (EIC), Montreal, QC, 2016, pp. 571–575, doi: 10.1109/EIE.2016.7548667. https://ieeexplore.ieee.org/document/7548667.
13. Peter Werelius, Matz Ohlen, and Joacim Skoldin, “Dielectric Frequency Response Measurement Technology for Measurements in High Interference AC and HVDC Substations,” in Techcon Asia-Pacific, 2011.
14. Guidelines on Specification and Maintenance of Polymer Bushings–2022. EPRI, Palo Alto, CA: 2022. 3002024619. https://www.epri.com/research/products/000000003002024619.
15. E. Ermakov, L. Jonsson, L. Melzer, “Approach to DFR analysis for condition assessment of 400 kV transformer and shunt reactor OIP bushings,” International Colloquium on Power Transformers & Reactors, Overhead Lines; and Materials and Emerging Test Techniques, November 2019, New Delhi, India.

The post Dielectric Frequency Response to Diagnose Condition of Bushings appeared first on .

Given the above, a group of European power companies funded a project to objectively benchmark performance of composite insulators. The research included collecting service experience as well as comprehensive testing of insulators, both from storage depots and removed from service.

Collection of service experience within the project’s framework formed the basis for this edited past contribution to INMR by Dr. Igor Gutman of the Independent Insulation Group (I2G), who coordinated the research. Primary interest was on overhead line insulators but these and substation insulators were analyzed separately because the two arrived in service at different periods.

Izoelektro d.o.o.

Slovenia

Dalian Insulator Group T&D Co., Ltd

China

See more suppliers of Support Insulators for Substations

Information on service experience with composite insulators has been comparatively scarce. Five CIGRE-inspired reviews are known to have taken place (the last still not officially published), but the information collected was limited. The main outputs included:

• Number of insulators installed;
• Maximum service period;
• Reliability;
• Typical reason for installation;
• Typical failure modes.

This first CIGRE survey in 1990 summarized service experience of composite line insulators at voltages higher than 100 kV, including suspension, tension and line post insulators. The total number installed was estimated as 140,000, and the volume of service experience (number of insulators times number of years) was 831,000. Thus, average service period was 6 years. Vandalism and handling were the main reasons for their application followed by pollution performance.

The insulator component that failed most often was the housing, explained by degradation of the material as was typical for first generation composite insulators. This was followed by failure at the core/housing interface. Reliability of composite insulators was estimated at 11×10^-4 per year – an estimation acknowledged to be unrealistically low due to the definition of failure used in the questionnaire. Utilities were asked to define failure as “any condition that led to the removal of insulators from a line”. For example, one utility installed a batch of 350 insulators of the same type and after a few years in service three insulators broke while others showed signs of degradation. Although the remaining insulators still appeared sound, the utility nonetheless decided to remove all 350 and thus reported 350 failures even though only 3 had failed in reality. Considering this, actual reliability was likely 10 to 100 times higher, i.e., 10^-4 to 10^-5 per year.

The second CIGRE survey was conducted and published in 2000. Again, the survey looked at insulators for voltage levels higher than 100 kV, including suspension, tension, line post insulators and interphase spacers. Total number of insulators installed was estimated at 700,000 and volume of service experience was 4,679,000. Thus, average service period was 7 years.

Separate data from utilities in the United States were included in this survey and provided an average service period of from 14 to 15 years. Again, vandalism and ease of handling were the dominant reasons for their application, followed by use in upgrade/compaction projects, polluted service areas and price. The prevailing failed component now was the core, which might indicate brittle fracture, followed by the core/housing interface. This survey used a different definition of failure, i.e. “an insulator that could not sustain the system voltage or mechanical load”. This is the definition used now. Average reliability of composite insulators was estimated at 10^-4 per year.

The third CIGRE-driven estimation of service experience for line insulators was presented in 2011 in a chapter included in a Technical Brochure (TB). This data was based mainly on a CIGRE survey together with more recent data provided by EPRI. The prevailing failure modes identified were brittle fracture and flashunder. This TB estimated reliability in the range 10^-4 to 10^-5 per year and considered mechanical failures to be the dominant failure mode. It should be noted that whereas data for conventional ceramic insulators refer to a consolidated technology representing about 100 years of experience, data for composite insulators included failures associated mainly with the first and second generations of this technology. As such, expected failure rates of presently available composite insulators should be much lower.

The fourth CIGRE review of service experience of line insulators involved WG B2.57 and was finished in 2021. The first draft of the TB did not offer any new information on service experience.

This first and the last items mentioned above in the CIGRE-driven survey on composite hollow core apparatus insulators in 2011 were included in the TB. Application of polymeric housings started on an industrial scale only in the early 1980s. Since then, the total quantity reached about half a million in the range 145 kV and above, based on data from 2006. Estimated market volume at the time was more than 50,000 insulators/year. If directly-moulded apparatus housings (e.g. surge arresters, cable terminations and bushings) at voltage levels above 60 kV are also considered, there would probably be one million units in service. Experience collected by this WG was limited but positive, with only minor degradation reported in a few cases.

Questionnaire & Response

Analysis of existing data on service experience led to the conclusion that such information was relatively scarce. A specialized questionnaire was therefore created and distributed to reflect the specific interests of the utilities participating in this project.

Fig. 1 shows an example of the questionnaire. A slightly more detailed questionnaire was also created and distributed to those utilities willing to contribute more data. These questionnaires were created in collaboration with all participants and the key issues to consider included:

• Compressed questionnaires (maximum 10-15 questions). Otherwise, they would never be answered.
• To be sent only to utilities where the project participants had contacts. Otherwise, probability to get a response would be low.

The questionnaires were distributed to 99 different companies and 53 answers were obtained providing a response rate of 54%. Only one of these 53 companies did not use composite insulators.

Unfortunately, not all large users of composite insulators on a country by country basis responded. Transmission voltage level in this study was defined as ≥110 kV.

Overhead Line Insulators

Total reported overhead line insulators installed at transmission voltages was estimated as 1.9 million while the total number of insulators installed at distribution voltages was estimated at 6.7 million. Thus, a combined total of 8.6 million units. This rather large number represented about 25% of the worldwide population, estimated to be between 30 and 40 million units based on discussions with international experts. Also, the upper limit for total number of composite insulators installed worldwide was verified during interviews with manufacturers.

Application of composite insulators is clearly commonplace with 98% of responding utilities using them. Insulators are used in standard as well as special designs of OHL (i.e., I- and V-strings; interphase spacers, insulated cross-arms, jumper supports and line posts).

Substation Insulators

Experience with some 260,000 composite substation insulators was reported. The total population of such insulators installed worldwide is thought to be between 2 and 3 million. Selection of composite insulators in this application is also common, with 92% of responding utilities using them. These insulators are used in a range of different apparatus, e.g. arresters, instrument transformers, circuit breakers, bushings) and also as station posts.

Maximum Duration in Service

Overhead Line Insulators

Maximum time in service was found to be similar for transmission and distribution voltage classes. Fig. 2 presents the data for both. Average maximum time was 24 years while maximum time was 40 years (i.e. still in service). There was also anecdotal data that several composite line insulators still operating in the Netherlands and Germany have been installed for more than 40 years.

Substation Insulators

Average time in service for substation insulators was 22 years and maximum time was 45 years. Fig. 3 summarizes these results.

Service Experience

Power companies were asked how they evaluate the overall experience with composite insulators. Three levels of answers, created by respondent companies, i.e., positive, mixed or negative were obtained for overhead line insulators and were marked as “+1”, “0” and “-1” respectively. This information is summarized in Fig. 4.

In the case of substation insulators, only two levels of answers were obtained, i.e., positive and neutral, comparable with “mixed” for line insulators. “Neutral” was considered as “smoother”, because it was built mostly on expectations for additional service. These levels were marked as “+1” and “0” respectively (see Fig. 5).

Overhead Line Insulators

Fig. 4 summarizes combined assessed service experience by transmission and distribution companies. For the majority, experience was positive (86%) compared to 4% negative and 10% mixed.

It is interesting to note that those companies who responded as “mixed” or “negative” did not use any pre-qualification procedures in their selection of insulators. Such procedures can be significantly different, from choosing among only a few qualified supplies to analyzing test reports and comparing these to the technical specifications created by the power companies. For example, 72% of utilities that considered that they have had positive experience with composite line insulators regularly used pre-qualification procedures or tenders with follow-up analysis of test reports versus only 28% that did not. It therefore seems worthwhile to apply such pre-qualification procedures.

Substation Insulators

Experience is shown in Fig. 5. For the majority, experience has been positive (94%), with the rest (6%) defined as neutral. By contrast to experience with line insulators, only 25% of those utilities that considered their experience with composite substation insulators as positive were applying pre-qualification procedures.

Reasons for Application

Overhead Line Insulators

The reasons why utilities chose to install composite insulators were left open for them to define and are summarized in Fig. 6. This part of the analysis combined transmission and distribution voltage levels. For each responding utility, one point was given for every specific reason mentioned. Ease of handling was the dominant reason, whereas the traditional reason, i.e. improving pollution performance, came next, followed by price.

Substation Insulators

Fig. 7 summarizes why utilities choose to install composite insulators at substations. The reasons are again created by the power companies questioned and there are fewer than was the case for overhead line insulators. Not surprisingly, safety against explosive failure was by far dominant reason, while easier handling was the second.

Causes of Failure

Overhead Line Insulators

Fig. 8 summarizes causes of failure or failure modes of composite line insulators. Again, analysis of the replies included both transmission and distribution voltage levels. Five types of failure dominated: flashunder, flashover, surface degradation, brittle fracture and damage caused by bird pecking.

In regard to surface degradation (which combines erosion, cracking, etc.), it might be that a utility observing this type of degradation would decide to replace insulators even though the damage was not critical. Thus, it was proposed not to consider this failure mode in the analysis. It was interesting to note that flashover was mentioned among the most common type of failure, especially given that some “experts” have claimed that it is rare if not impossible for composite insulators to flashover in the classical pollution dry-band mode known for ceramic insulators.

It is therefore likely that some utility or service provider staff may not have been able to recognize the difference between ‘flashover’ (i.e. an external breakdown along the surface) and ‘flashunder’ (i.e. an internal breakdown typically in the interface between core and housing). If this is indeed true, the flashunder failure mode was dominant and part of the ‘flashovers’ needed to be added to the cases of ‘flashunder’. In spite of this, number of failures reported would still be low because, as mentioned above, 86% of responding utilities considered their service experience as positive. However, fewer than 20% of utilities were able to present their exact number of failures.

One difficult question was which generation of composite insulators were actually represented in the failures being reported. The questionnaire summarized all data collected, thus conservatively combining experience for all generations of these insulators, presently assumed to be four to five. Actual situations in service are illustrated by examples of comments received from utilities that had used composite insulators for many years:

1. “On one OHL the degradation of specific HTV silicone rubber led to a change of the mechanical properties (brittleness when mechanically stressed) due to a chemical reaction with acids and UV”;
2. “For the current generation issues are very rare and are basically flashovers due to bird-streaming”;
3. “Since we only have the third generation of composite insulators in our network, we have not experienced any failures”;
4. “Flashovers were experienced only for one supplier”;
5. “We observed issues with 20 – 25 years first generation EDPM insulators in the late 80s and early 90s with loss of hydrophobicity and tracking/erosion. But more recent supply of silicone insulators seems to be OK”;
6. “Expected asset life of composite insulators is 40 years; in the heavy contaminated environment near the coast, we replaced a particular brand of composite insulators after 13 years in service after two brittle fracture failures”.

Substation Insulators

Fig. 9 summarizes causes of failure of composite substation insulators. Mechanical issues were the dominant failure mode and it was assumed that these related mostly to station post insulators. Normally, only limited surface degradation was observed on apparatus insulators and it was often a subjective decision by the utility whether or not to replace such insulators.

Reliability

Attempting to calculate reliability in terms of number of failed units per year was complicated. This was because service experience, in terms of aggregate number of insulators installed multiplied by number of years in service should be known, as are the exact number of failures. Thus, in cases where no failures were reported, failure rate was nonetheless assumed to be one failure over the whole service period. Otherwise, aggregate volume of installed insulators was estimated by multiplying 50% of the maximum time in service by number of installed units, since only maximum service time was available.

Overhead Line Insulators

Table 1 presents such an estimation based on limited data from 15 utilities that were used to assess reliability. For visualization purposes, the answers from these utilities were arranged according to a traffic light principle, as GREEN (positive), YELLOW (mixed) and RED (negative). Based on this, the following conclusions were drawn:

• ‘Positive’ experience with composite insulators typically relates to an annual failure rate of 10^-5 or less;
• ‘Mixed’ experience typically relates to an annual failure rate of 10^-4;
• ‘Negative’ experience typically relates to an annual failure rate of 10^-3;
• Average annual failure rate for all data collected in Table 1 was 10^-5.

Substation Insulators

Data on failure rates of composite substation insulators is limited because of their relatively short service time compared with line insulators and also because most have operated without failures. Table 2 presents some information for 5 utilities that reported exact number of failures. Aggregate service time of installed insulators was (as for line insulators) taken as half the maximum time in service. Based on this, ‘positive’ experience expressed by a utility typically corresponded to a failure rate of between 10^-4 and 10^-5. It can be assumed that, with increasing service experience, failure rate will become lower, as demonstrated by the experience of utility D. Given a greater share of new generation composite insulators entering service, reliability would be higher, approaching 10^-5 per year.

Use of Grading/Corona Rings

Fig. 9 analyzes the response from utilities that answered the question about application of grading rings. No grading devices were being used for distribution classes (i.e. below 110 kV). For transmission class insulators, general practice has been to use one grading ring (or sometimes only arcing horns) at the HV fitting starting from 110-132 kV and two grading rings (one at each end) starting from 220-275 kV.

GCC Electrical Testing Laboratory

Kingdom of Saudi Arabia

EDP Labelec

Portugal

See more Laboratories

Summary

Overhead Line Insulators

This project collected service experience linked to the operation of about 9 million composite insulators installed at transmission and distribution voltage levels. Information was collected using a questionnaire that received responses from about 50 utilities. Average time in service was 24 years (with a maximum of 40 years). Thus, the technology is considered as mature. These insulators were estimated to represent about 25% of the total population worldwide.

Replies received to the main questions formulated by participants in this project, can be summarized as follows:

• Application of composite insulators has become commonplace (confirmed by 98% of answers);
• Insulators are used on standard OHL and in special designs (I- and V-strings; interphase spacers, insulated cross-arms, jumper supports and line posts);
• The dominant reason for application is ease of handling (35%);
• Present service experience is considered as positive, as confirmed by 86% of replies. A large part of those utilities with positive service experience used pre-qualification procedures while utilities with negative or mixed experience typically did not use such procedures. Thus, it seems worthwhile to apply pre-qualification procedures;
• Types of insulator failure experienced were defined by the utilities on their own and not standardized. Therefore, it would be more correct to define them as “observations”. Five dominant types of failure were mentioned: surface degradation, flashunder, flashover, brittle fracture and bird damage from pecking. First commenting surface degradation (combining erosion, cracking, etc.), it might be that a utility observing this type of degradation decided to replace insulators, although this damage was not critical. Thus, it was proposed not to consider this failure type in further evaluation. Flashover is an interesting response because many CIGRE/IEC experts do not believe that classic pollution-driven flashover can take place on composite insulators. Even if it is believed that some ‘flashovers’ were actually ‘flashunders’, some classical surface flashovers should also have occurred. Assuming that a part of ‘flashovers’ really belonged to ‘flashunders’, this would become the single most dominant reason for failure. Subsequently, the ranking might be:

1. flashunder,
2. brittle fracture,
3. flashover, and
4. bird damage (this type of damage may be overrepresented because of the 6 Australian and New Zealand utilities responding to the questionnaire).

•  When considering flashunders as the dominant reason for failure of these insulators, the root cause is poor adhesion in the core/housing interface, allowing for moisture penetration. The moisture penetrates through the rubber housing by diffusion and typically condenses in the core/housing interface or, more rarely, penetrates through improper sealing. Thus, core/housing adhesion and the quality of sealing become two important issues to investigate. It is important to stress that both adhesion and sealing tests are non-standardized tests, i.e. not yet included in IEC standards. Evaluation of quality of adhesion is already recognized and under consideration in the ongoing revisions to IEC 62217 and IEC 61109. Development of methods to evaluate the quality of sealing is also underway in a separate project. Too high electric field can accelerate degradation due to poor adhesion of the core and housing and must therefore be controlled. Limits for electric field are also well established and under consideration in revision of IEC 61109.
• For transmission class insulators, general practice is to use one grading ring, or sometimes only an arcing horn at the HV fitting, starting at 110-132 kV. Two grading rings at both ends are applied, starting at 220-275 kV.
• Total service experience and total number of failures summarized in this document provide an average annual failure rate of 10^-5, which is in line with the ‘positive’ service experience subjectively expressed by most power utilities.

Substation (Apparatus & Station Post) Insulators

Service experience with 260,000 composite substation insulators was reported by 27 utilities. Total average time in service was 22 years, with a maximum of 45 years. Thus, this technology is also mature. The total number of insulators installed worldwide is estimated at between 2 and 3 million. A summary of important findings is as follows:

• Application of substation composite insulators is commonplace (92% of answers received). These insulators are used both for apparatus and as station posts;
• The dominant reason for use of composite apparatus insulators is safety against risk of explosion (40% of replies);
• Present service experience is considered as positive, as confirmed by 94% of respondents;
• By contrast to experience with composite line insulators, only 25% of these utilities used pre-qualification procedures;
• Types of failure experienced were defined by the utilities themselves and not standardized. Mechanical issues were the dominant failure type (37% of replies). It is assumed that these relate mainly to issues with station post insulators;
• Based on the limited amount of data, it is assumed that positive service experience typically corresponds to an annual failure rate of between 10^-4 and 10^-5.

References
• Yu. Gutman, E.A. Solomonik, V.N. Solomatov, Yu.N. Yashin: “Operation and Field Tests of Overhead Line Composite Insulators with Silicone Rubber Cover”, ISH-1993, Yokohama, Japan, 23-27 August 1993, 47.13
• CIGRE SC 22 WG 03.01: “Worldwide experience with HV composite insulators”, ELECTRA 130, December 1990, p.p. 69-77
• CIGRE WG 22.03: “Worldwide Service Experience with Composite Insulators”, ELECTRA 191, August 2000, p.p. 27-43
• CIGRE WG B2.21: “Guide for the Assessment of Composite Insulators in the laboratory after their Removal from Service”, CIGRE Technical Brochure No. 481, December 2011
• CIGRE WG B2.57: “Survey of operational Composite Insulator Experience and Application Guide for Composite Insulators”, 2020, the first draft, not finalized yet
• CIGRE WG A3.21: “Aspects for the Application of Composite Insulators to High Voltage (≥72kV) Apparatus”, CIGRE Technical Brochure No. 455, April 2011
• Gutman, A. Dernfalk, P. Sidenvall, J. Lundengård, C. Ahlrot, P. Aparicio, A. Berlin, T. Condon, J.-F. Goffinet, K. Halsan, M. Radosavljevic, K. Varli, K. Välimaa: “Rod to Housing Adhesion in Composite Insulators: Practical Evaluation in Collaboration with Utilities”, 2019 INMR World Congress, Tucson, USA, 20-23 October 2019
• Ahlrot, P. Aparicio, A. Berlin, T. Condon, J.-F. Goffinet, I. Gutman, K. Halsan, M. Radosavljevic, K. Varli, K. Välimaa: “New test procedure intended to evaluate adhesion of core/housing interface of composite insulators”, CIGRE-2020, D1-303
• Gutman, C. Ahlrot, P. Aparicio, A. Berlin, T. Condon, A. Dernfalk, J.-F. Goffinet, K. Halsan, K. Kleinekorte, J. Lundengård, M. Radosavljevic, P. Sidenvall, S. Steevens, K. Varli, K. Välimaa: “Development of Innovative Test Procedure for Evaluation of Adhesion of Core-Housing of Composite Insulators: from Root Cause of Failures in Service to Reproducible Test Procedure”, Cigré Science & Engineering, N. 20, February 2021, p.p. 171-182
• J. Philips, A.J. Maxwell, C.S. Engelbrecht, I. Gutman: “Electric Field Limits for the Design of Grading Rings for Composite Line Insulators”, IEEE Transactions on Power Delivery, Vol. 30, No. 3, June 2015, p.p. 1110-1118
• Gutman, P. Sidenvall: “Optimal Dimensioning of Corona/Grading Rings for Composite Insulators: Calculations & Verification by Testing”, INMR World Congress & Exhibition on Insulators, Arresters & Bushings, Munich, Germany, 18-21 October 2015

The post Application of Composite Insulators: Perceptions vs. Actual Service Experience appeared first on .