The Ultimate Guide to Choosing Aerodynamic Glass Insulator

Past technical recommendations such as IEC/TR as well as newer specifications such as IEC/TS are both used by power utilities for insulator selection in polluted outdoor environments.

You will get efficient and thoughtful service from Yipeng.

This edited contribution to INMR by Dr. Wallace Vosloo, retired from Eskom, Richardo Davey of Eskom Research Testing and Johannes Bekker at the University of Stellenbosch in South Africa reviewed different possible approaches in this regard.

Using IEC/TR 

Selecting porcelain and glass insulators for three-phase a.c. systems up to 525 kV phase-to-phase using IEC/TR (Guide for the selection of insulators in respect of polluted conditions) has been based on service experience as well as laboratory testing under naturally and artificially polluted conditions. As stated in this document: Simple general rules should assist choosing the insulator, which should give satisfactory performance under polluted conditions.

Although it is advised to determine pollution type and severity and then use laboratory tests to help select insulators, many utilities have instead simply chosen to standardize on one or more values of minimum Specific Creepage Distance. Here, SCD (the ratio of the leakage distance measured between phase and earth over the r.m.s. phase to phase value of the highest voltage for the equipment) is recommended for various pollution levels (I to IV), namely 16, 20, 25 and 31 mm/kV(Um) respectively for Light, Medium, Heavy and Very Heavy pollution environments. Pollution levels I to IV are determined mainly from IEC/TR Tables (shown in Fig. 1), service experience or by doing site pollution severity (SPS) measurements using Equivalent Salt Deposit Density (ESDD) or Directional Dust Deposit Gauges (DDDG) methods. Table 1 shows typical ESDD and DDDG values used to classify site pollution level and minimum required SCD.

Fig. 2 and Table 2 summarize basic profiling rules recommended in IEC/TR , which are normally followed.

Influence of insulator diameter on pollution performance is also considered for insulators with average diameters of between 300 and 500 mm. In the case of diameters greater than 500 mm, it is recommended that specific creepage distances be increased by 10% and 20%. Laboratory tests, in accordance with IEC 507, are also recommended to evaluate an insulator’s pollution performance but are rarely used by most power utilities. Greasing or washing is recommended for areas with severe pollution and/or low natural washing.

For a.c. porcelain and glass insulators, a utility would typically specify maximum connection length taking into account live line work, minimum dry arcing distance, and minimum creepage distance. Then, it would be stated that insulator profile must comply with IEC/TR , the focus being on simplicity and ease of use.

Case Study Using IEC/TR for Selecting Insulators

It is proposed to erect 132 kV lines in areas with pollution levels ranging from Light to Very Heavy using standard glass cap & pin disc insulators (U120B, F12/146). Assuming the parameters given below, how many discs (n) would be required per string?

System highest voltage (Um) = 145 kV

Minimum required Dry Arcing Distance (DAD) of the insulator string is  mm (i.e. for a high lightning area)

Disc spacing (s) = 146 mm

Arcing distance per disc (a) = 210 mm

Creepage distance (CD) per disc = 320 mm

The arcing distance of a disc insulator string = a + (n – 1) s

Thus, = 210 + (n – 1) x 146 and n = ( – 210) / 146 + 1 = 9.84 (in other words, 10 discs are required)

In terms of pollution level, as shown in Table 1, the specific creepage distances needed for Light, Medium, Heavy and Very Heavy pollution areas are 16, 20, 25 and 31 mm/kV (Um) respectively.

Number of discs required per string = (CD)/320 where CD = (Um x SCD)

Table 3 gives values calculated using (145 x SCD)/320.

Utilities have typically been using 7 to 14 standard glass discs per string for 132 kV lines. More recently, with the advent of new ‘active’ polymeric insulation materials that interact with the environment, utilities have continued to use IEC/TR since nothing else was available in IEC.

Typically, for a.c. polymer insulators, utilities such as Eskom would specify maximum connecting length (taking into account live line work), minimum dry arcing distance and minimum creepage distance. They would also state that insulator profile must:

1. comply with IEC/TR based on in-service and test station experience;
2. have open aerodynamic alternating shed profile with S/P ratio ≥ 1; and
3. that the material must be hydrophobic, with good hydrophobicity transfer capabilities.

Then, for e.g. ease of stock, minimum SCDs of 20 mm/kV for Light to Medium and 31 mm/kV for Heavy to Very Heavy pollution areas would be specified.

Insulation requirements for both UHV a.c. and d.c. insulators are more complex and normally determined along with technical experts from manufacturers (mostly members of Cigré WGs).

Using Latest Specification IEC/TS “Selection & Dimensioning of HV Insulators for Polluted Conditions”

The following major changes have been made with respect to IEC/TR :

• Encouraging use of site pollution severity measurements, preferably over at least a year, in order to classify a site instead of the previous qualitative assessment (see below).
• Recognition that ‘solid’ pollution on insulators has two components: one soluble and quantified by ESDD; the other insoluble and quantified by NSDD.
• Recognition that, in some cases, measuring layer conductivity should be used for SPS determination.
• Using results of natural and artificial pollution tests to help with dimensioning and to gain more experience in order to promote future studies to establish a correlation between site and laboratory severities.
• Recognition that creepage length is not always the sole determining parameter.
• Recognizing the influence of other geometry parameters and of the varying importance of parameters according to the size, type and material of insulators.
• Recognition of the varying importance of parameters according to type of pollution.
• Adoption of correction factors to attempt to take into account influence of the above pollution and insulator parameters.

Fig. 3 shows three approaches proposed for selection and dimensioning of insulators.

Fig. 4 provides Eskom’s specifications for determining site pollution severity and pollution performance curves.

Practical Example Using Three Approaches as per IEC/TS for Insulator Selection

It has been proposed to erect a 132 kV substation and interconnecting lines at Koeberg Nuclear Power Station (KNPS) along the South African west coast. The area is approximately 600 m from the coastal high-water mark and is exposed to strong coastal winds, low rainfall and regular salt fog events. The insulators (substation: 189 – posts (4 kN), 294 – hollow cores (average diameter <400 mm), 387 – long rods/strings (120 kN, ball and socket) and lines: 768 – long rods/strings (120 kN, ball and socket)) should have a maximum connecting length of 1.48 ± 0.02 m, minimum insulation length of 1.2 m, minimum dry arcing distance of 1.5 m and minimum specific creepage distance of 31 mm/kV (CD ≥ mm). In addition, insulators with open aerodynamic alternating shed profiles, S/P ratio ≥1, hydrophobic material and good hydrophobicity transfer capabilities are preferred.

The question is what insulation is required to ensure risk of insulator flashover is minimal, with mean time between flashover (MTBF) of at least 50 years?

Approach 1: Use Past Experience

Duinefontein 132 kV Substation (≈600 m from the coastal high-water mark) is situated only ≈1 km from the new proposed 132 kV substation and interconnecting lines area at KNPS.

In , the 60 porcelain station post insulators installed at the Duinefontein Substation that have a specific creepage distance of 32.4 mm/kV were upgraded (along with all other insulation) using room temperature vulcanized silicone rubber coating (RTV SR A). Prior to this upgrade (since washing did not work and greasing had only a 6 to 12 month effective lifespan) the substation experienced flashovers on an annual basis, including the Type B instantaneous event of February . Since the upgrade, no flashovers occurred to date (18 years later), which includes the catastrophic Type B instantaneous pollution event experienced in February . Indeed, the RTV SR A coating still has excellent hydrophobic properties.

Personnel from the KNPS Weather Station (≈250 m from Duinefontein Substation) classify the area as follows:

• Average ambient temperatures between 14° and 20°C (minimum 4°C and maximum 36°C);
• Exposed to strong coastal winds (gusts up to 35 m/s);
• Low rainfall area (≈320 mm per year) with only 3 to 5 rainy days in summer (≈80 mm)
• High humidity levels at night/early morning and regular salt fog events (≈40 per year).

ESDD, NSDD and DDDG pollution measurements at the Duinefontein Substation from March to date, were used to calculate ESDD2% = 0.165 mg/cm2 (STDEV = 0.57), average ratio of NSDD/ESDD = 1.1 and monthly average DDDGave = 382 µS/cm.

The flashover of a bare porcelain 132 kV breaker support insulator (having specific creepage distance 32.4 mm/kV) during the catastrophic Type B instantaneous pollution event experienced in the Cape in Feb was used to estimate minimum uniform pollution present on insulation during this event, namely ESDD = 0.4 mg/cm2 and NSDD = 0.182 mg/cm2. Field experience has shown that, while instantaneous pollution events will not occur annually, it can be conservatively assumed that one event occurs each year.

Pollution levels at Duinefontein Substation are as follows:

Type A:  40 natural pre-deposited pollution events per year with critical wetting (ESDD2% = 0.165 mg/cm2 with STD deviation of 0.57 and ESDD/NSDD ratio = 1.1).

Type B:  One instantaneous conductive fog pollution event per year (ESDD2% = 0.4 mg/cm2 and NSDD = 0.182 mg/cm2).

The SPS levels obtained from ESDD and NSDD measurements at Duinefontein Substation (see Fig 5) classify the area for Type A pollution (E6/7) d – Heavy and Type B (E7) e – Very Heavy. DDDG measurements, after climatic factor correction, classify the area as Very Heavy.

The 60 RTV SR A coated porcelain station post insulators with SCD of 32.4 mm/kV at Duinefontein Substation (≈1 km from the new proposed 132 kV substation and lines area) have provided excellent performance for 18 years. Note: these same 60 bare porcelain station post insulators had flashed over more than once annually.  No washing or greasing was recommended.

Koeberg Insulator Pollution Test Station (KIPTS) (≈50 m from the coastal high-water mark) lies about 2 km from the new proposed 132 kV substation and lines area.  Monthly average DDDGave =  µS/cm measured at KIPTS is Extreme (≈6.6 times higher compared to Duinefontein Substation). Findings of some insulator research and tests done at KIPTS over 15 years are presented in “Power Utility Perspective on Natural Ageing and Pollution Performance Insulator Test Stations”.

From natural insulator pollution performance experience gained at KIPTS the following:

Posts: Porcelain post insulators with SCD = 38 mm/kV will flashover more than 3 times per year at KIPTS;

RTV SR A coated porcelain post insulators with SCD ≥ 24 mm/kV will have no flashovers per year at KIPTS. RTV SR A coated porcelain transformer bushings at KIPTS with SCD = 30 mm/kV will perform well for 15 years;

Open aerodynamic alternating shed profile with S/P ratio ≥ 1 and a hydrophobic material with good hydrophobicity transfer capability will work best.

Hollows: SR hollow core insulators with SCD ≥ 28 mm/kV will have no flashovers per year at KIPTS;

Open aerodynamic alternating shed profile with S/P ratio ≥ 1 and hydrophobic material with a good hydrophobicity transfer capability will work best.

Longrods:  SR longrod insulators with SCD ≥ 22 mm/kV will have no flashovers per year at KIPTS;

Open aerodynamic alternating shed profile with S/P ratio ≥ 1 and hydrophobic material with a good hydrophobicity transfer capability will work best.

Note: Corona rings should be installed on 132 kV longrod insulators.

Strings: Standard glass cap & pin disc insulator (F12/146) strings with SCD of 27 mm/kV will flashover more than 3 times per year at KIPTS. The same insulator string with RTV SR A coating applied will have no flashovers per year;

Fog-type glass cap & pin disc insulator (F120P/146) strings with SCD of 37 mm/kV will have no flashovers per year at KIPTS. The same insulator string with RTV SR C coating applied will experience similar leakage currents to SR longrod insulators.

Note: Expect poor natural washing/cleaning and pin erosion problems.

Approach 2: Measure & Test

Approach 2 (as per Fig. 3) is used along with Section 12 of IEC/TS -2 for porcelain and glass, and Section 12 of IEC/TS -3 for polymeric insulators.

The pollution and climate at Duinefontein Substation (as in Approach 1) give the expected pollution severity levels and climate in the area of the proposed 132 kV substation and interconnecting 132 kV lines as follows:

Type A: (E6/7) d – Heavy as 40 natural pre-deposited pollution events per year with critical wetting (ESDD2% = 0.165 mg/cm2 with STD deviation of 0.57 and ESDD/NSDD ratio = 1.1).

Type B: (E7) e – Very Heavy as one instantaneous conductive fog pollution event per year (ESDD2% = 0.4 mg/cm2 and NSDD = 0.182 mg/cm2).

Candidate insulators were selected as shown in Table 4, where possible with a maximum connecting length of 1.48 ± 0.02 m, minimum insulation length of 1.2 m, minimum dry arcing distance of 1.5 m, minimum specific creepage distance of 31 mm/kV (CD ≥ mm). Insulators with open aerodynamic alternating shed profile with S/P ratio ≥ 1, and hydrophobic material with good hydrophobicity transfer capabilities. Porcelain post, standard glass and fog type glass insulators were included as reference.

Laboratory pollution U50% flashover voltage (using the rapid flashover test method) curves at three pollution levels (as in Fig. 4) – SDD of 0.06; 0.12 and 0.48 mg/cm2 with NSDD ≥ 0.1 mg/cm2 was obtained using:

• for porcelain and glass insulators the Solid Layer Test Method (see Table 5) according to IEC using Procedure B and spray gun for applying the Kaolin composition and Annex B.3.2. The degree of pollution on the test insulator was determined using the SDD method. Recommendations as given in Annex D and E were followed.

• and, for polymeric insulators according to modified Solid Layer Test Method (see Table 5) with pre-conditioning procedure, and with/without recovery according to Cigre TB 555 and Cigre TB 691.

The rapid flashover laboratory solid layer pollution test done on SR Longrod A insulator to determine U50% = 225 kV at SDD = 0.12 mg/cm2 and NSDD = 0.1 mg/cm2 with 48-hour hydrophobicity recovery is shown in Figure 6 as example. The rapid flashover laboratory solid layer pollution test results for all the candidate insulators are shown in Table 4.

The candidate insulators’ pollution U50% flashover voltage results in Table 4 were then converted into flashover stress along the test insulation length HT = 1.2 m as  in kV/m and is presented as a three-point approximated inverse power law curves against pollution level SDD in mg/cm2 in Fig. 7.

The candidate insulator pollution flashover performance curve constants A in kV/m and α was determined for equation U50%/Ht = A · SDD-α and the values are shown in Table 4.

Um-ph/H was calculated as 70 kV/m using the specified insulation length H = 1.2 m, and Um-ph = 83.7 kV (the highest system r.m.s. phase to ground voltage that the insulator to be supplied will be subjected to).

For more Aerodynamic Glass Insulatorinformation, please contact us. We will provide professional answers.

The candidate insulator was accepted for further calculation if  U50%/Ht> 70 kV/m in the SDD range of 0.12 to 0.48 mg/cm2.

Insulator pollution flashover performance curve constants A in kV/m and α, of the candidate insulators was used along with pollution severity levels and climate in the area of the proposed 132 kV substation and interconnecting 132 kV lines in the statistical approach as per Annex G of IEC/TS -1 in order to optimize insulation selection. Fig. 8 shows the Insulator Selection Tool, a commercially available statistical software.

As example, using the IST for MTBF of 50 years 768 – 132 kV SR Longrod A insulators require:

• SCD of 20.2 mm/kV when exposed to 40 natural Type A pre-deposited pollution events per year with critical wetting (ESDD2%= 0.165 mg/cm2 with STD deviation of 0.57 and ESDD/NSDD ratio = 1.1) (see Fig. 9);

• and SCD of 22 mm/kV when exposed to one Type B instantaneous conductive fog pollution event per year (ESDD2%= 0.4 mg/cm2 and NSDD = 0.182 mg/cm2) (see Fig. 10).

The MTBF obtainable within the required connecting length of 1.48 ± 0.02 m and SCD needed for MTBF of 50 years was calculated for the reference and candidate insulators using the IST. Results are shown in Table 6.

Approach 3: Measure & Design

As per Fig. 5, the area proposed for the 132 kV substation and interconnecting lines can be classified for Type A pollution as Class d – Heavy and for Type B pollution as Class e – Very Heavy. Thus, in the worst case, a minimum SCD of 31 mm/kV is needed for the reference glass disc insulator.

As per IEC/TS -2, the following is recommended for porcelain and glass insulators:

• Aerodynamic, alternating sheds on long rod insulators, post insulators, hollow core insulators;

• Anti-fog profile for disc insulators;

• p1 – p2 ≥ 15 mm, s/p ≥ 0.65, c ≥ 25 mm, l/d ≤ 5, 5˚ ≤ α ≤ 25˚ and CF ≤ 4 (lowest risk options used);

• No altitude correction required;

• 10% increase in SCD for hollow core insulators with average diameter > 300 mm and < 400 mm.

As per IEC/TS -3, the following is recommended for hydrophobicity transfer polymeric insulators:

• SCD could be reduced or increased depending on the environment or pollution level (no clear advice given);

• Aerodynamic alternating sheds;

• p1 – p2 ≥ 18 mm, s/p ≥ 0.75, c ≥ 40 mm, l/d ≤ 4.5, 5˚ ≤ α ≤ 25˚ and CF ≤ 4 (lowest risk options used);

• No altitude correction required;

• 10% increase in SCD for hollow core insulators with average diameter > 300 mm and < 400 mm.

Insulators must have a minimum SCD of 31 mm/kV (CD = mm) and hollow core insulators with minimum SCD of 31 x 1.1 = 34.1 mm/kV (CD = mm) are recommended. Table 7 provides a summary and comparison of results obtained for these three Approaches.

Discussion

As per Table 7, the findings of Approach 1 and 2 correlate with one another, showing that both approaches would work if correct data required as per Fig. 3 is available. The practical examples demonstrate that both approaches lead to a selection having good accuracy. Moreover, as per Table 7, Approach 3 in general would result in over-dimensioning of required insulation.

Table 8 offers a comparison of the recommendations as per IEC/TR and specifications as per Approach 3 in IEC/TS -1 for the practical examples. Findings are similar in regard to specific creepage distance. However, Approach 3 takes into account use of hydrophobicity transfer materials, which could result in a one class lower SCD, different profile parameters and anti-fog instead of open aerodynamic disc insulators.

Conclusions

In summary it has been demonstrated with practical examples, how the old technical recommendation, i.e., IEC/TR and the new technical specification i.e., IEC/TS are being successfully applied by a utility, in this case, Eskom, to select and dimension outdoor insulators for polluted environments (Note: Ageing and failure modes are not taken into consideration in this discussion).

The importance is also shown of site pollution severity measurements, climatic conditions, identifying pollution type and practical use of data collected from natural pollution test sites, in-service insulators, laboratory pollution flashover tests and statistical evaluation.

Past service experience and laboratory testing revealed risk of great variance in the quality of glass insulators from different manufacturers, depending on raw materials, process and know-how. This finding applied to all components – from the glass shell to metal fittings to cement. Moreover, experts report that it has not always been possible to distinguish insulators of inferior quality based solely on past standard tests. This is because insulators sent for type and sample testing can be specially selected for this purpose. Moreover, requirements for routine testing during production are not always stringent enough to screen out insulators of inferior quality.

Combined with this development, demand for uninterrupted transmission has increased worldwide and especially in Europe. Improved modelling and specialized software have also made it possible to optimize insulator selection and dimensioning, resulting in more highly stressed insulators. All these factors have meant greater need for uniform high quality since the failure of any disk in a string becomes that much more serious.

A decade ago, large power supply companies in Europe co-operated to establish test procedures and criteria that would better reveal the true quality of toughened glass insulators. This edited past contribution to INMR by Kjell Halsan (Statnett), Igor Gutman and Johan Lundengård (then at STRI) provided the impetus for changes to standards, now implemented, and also discussed findings from laboratory testing of glass insulators.

Norwegian transmission system operator, Statnett, has had a policy of closely monitoring the quality of its glass insulators and representatives have attended sample testing for all major deliveries. In addition, operating data on shatter rate per 100,000 glass insulators per year has been collected and analyzed for long transmission lines. These statistics showed that shatter rate of discs can vary significantly among suppliers and sometimes even for different batches delivered by the same supplier. Given this, Statnett conducted a project to ensure purchase of only high quality glass insulators from both existing and new suppliers – with emphasis more on life cycle cost than on acquisition cost.

In Sweden, past refurbishment of a 220 kV line by TSO, Svenska Kraftnät, employed glass cap & pin insulators selected on the basis of usual tendering procedures including compliance with IEC requirements. Insulators were installed in a mostly clean but humid area, typical of Scandinavia. Abnormal discharge activity and audible noise were reported on these insulators after only a short time in service, most often when humidity was highest. Field inspection of the line with a daylight UV-camera confirmed the activity and suggested that the insulators could be suffering from defects in their cement.

These observations became a concern since such unexpected initial behaviour for new insulators could influence long-term performance. Detailed investigations were performed, including voltage tests at operating voltage, and revealed discharge activity on the glass surface around the pin, just outside the cement. This was especially evident if the electric field distribution in this part of the insulator was affected by water sprayed onto it, to simulate humid service conditions. Resistance measurements using a hand-probe showed that the weakness observed was due to unusually high conductivity as well as unfavourable geometry of the cement. RIV levels, for example, had a wide spread among insulators from different manufacturers and this was interpreted as an indicator of poor quality control.

Given this, it was proposed to expand test requirements for tenders to include RIV measurements at 20 kV and 24 kV to reduce risk of similar problems in the future. Such a requirement is easy to implement and would help ensure that all insulators purchased are more robust with respect to discharge activity under humid conditions.

Preliminary Investigation

The aim of the first stage in the investigation was to identify effective yet practical test methods that could be used to qualify manufacturers in terms of the quality of toughened glass insulators they supply. This included a workshop with experts, discussions with leading test laboratories and manufacturers and a review of the literature. The production process for glass insulators was also examined closely to identify critical issues that can affect final quality. This was then used to identify tests that could be applied to differentiate between good and inferior quality insulators. The following points were highlighted:

• Overview of the components in a glass insulator disc;

• Walk-through a glass disc fabrication process;

• Quality control aspects during manufacture and assembly;

• Relationship between key properties and test methods to verify these.

Subsequently, a combination of the following tests was identified as suitable for more detailed investigation:

1. Steep-front puncture test in air to test the dielectric strength of the glass shell. This test method might need refinement with respect to steepness of the voltage wave applied.

2. RIV testing to verify pin cavity design and quality of cement used. This test was seen as maybe needing modification to verify the longevity of the RIV control measures that manufacturers use, e.g. coating with bitumen.

3. Ultimate tensile test to verify mechanical design. This is a standard test for insulators but could be modified for larger required sample size to better understand failure modes as well as how a manufacturer can control mechanical failure.

4. A test verifying residual mechanical strength can be an important tool to verify that, even if an insulator breaks, the conductor will not be dropped.

5. Thermal shock test to check the quality of the glass shell.

The first two, i.e. impulse puncture (called further steep-front) test and RIV test were subsequently chosen for the preliminary verification.

From discussions during this phase of the project, it was also concluded that testing alone might not always guarantee quality. Buyers would also need to implement a process for them to evaluate the quality control procedures used by the manufacturer. This would include regular factory inspections to ensure that these procedures are always implemented during production. In addition, a library of photos was created to assist the inspector to visually identify potentially problematic units for subsequent sample testing.

In this project phase, an overview of the entire manufacturing process was created to highlight important process parameters to be scrutinized during any factory visit. This guide (or questionnaire) would make it easier for a representative from the power supply company to go through the production and quality control processes of any manufacturer.

Limited Laboratory Testing

Limited testing was performed on 3 types of glass insulators selected from inventory at Statnett and Svenska Kraftnät. The aim was to conduct comparative testing on insulators assumed to have different levels of quality, i.e. ‘high’, ‘acceptable’ and ‘low’. These insulators were subjected to only two types of tests, selected as the most promising based on the literature, opinion of experts and previous tests performed for these power companies. The two tests were: RIV testing in dry/wet conditions followed by resistivity measurements of the cement; and steep-front testing (i.e. impulse puncture withstand test in air).

The RIV test was performed according to Swedish standard SS 447 10 11 (). A pre-stress voltage of 30 kV was applied for 5 minutes after which the RIV level was read at test voltage reduced to 24 kV. Then, test voltage was decreased to 20 kV and the RIV level was read once more. According to this standard, the RIV level must not exceed 60 dB (µV) at 20 kV.

Steep-front testing was performed according to IEC cl. 5.6 and Canadian standard CSA C411.1-10 cl. 6.6 and witnessed by representatives from STRI and Statnett. The results (Table 1) demonstrated that both tests were able to reveal insulators with ‘low’ quality in two of their elements, i.e. cement and glass. As such, it was recommended that these tests be part of an extended test matrix.

As one useful finding, results of initial laboratory testing showed that test insulators should be randomly selected since it seemed that manufacturers could produce a smaller batch with better quality insulators. Another practical output was that 8 out of 10 insulators randomly selected from one batch in storage failed the steep-front test by shattering. This confirmed not only the need for further work to prevent such insulators from ever reaching the stocks of power utilities but also the importance of carrying out steep-front as a sample test on randomly selected unit.

Development of New Test Matrix & Criteria

The aim of this phase of the project was to evaluate whether or not the test methods selected would be able to reveal insulators of variable quality. This would be accomplished by collecting a selection of insulators of differing quality such that some would be expected fail (thereby confirming that the test matrix can catch these) while some would be expected to pass (verifying that the test matrix was not unreasonably stringent). It was decided not to request the test insulators from manufacturers. Instead, these were selected at random from batches already purchased by participating power companies for line construction and refurbishment. In total, batches from 5 different suppliers with factories all over the world were chosen. Moreover, based on the specific request of participating utilities, the mechanical strength classes were limited to: 120 kN, 210 kN and 300 kN.

The results of 3 types of tests, i.e. RIV test (performed at STRI), steep-front test and thermal-mechanical test (both performed at EGU) were presented in a paper at the CIGRE General Session in . In order to verify how different tests can affect and reveal insulators of poor quality, a test program according to the scheme proposed by the CSA (outlined in Fig. 2) was performed using 3 different types of glass insulators, with 30 units of each type.

When comparing test results from the mechanical failing load test on the same type of insulators, no significant variation was found between the 3 groups of insulators going through different numbers of tests (batch 1, 2 or 3). Problems with quality of glass discs were revealed during the thermal-mechanical test, steep-front test and mechanical failing load test. Thus, it was proposed not to use the thermal-mechanical test to pre-stress the glass before the steep-front test since it was felt this would be overly complicated.

The test program concentrated on RIV and steep-front tests and details on both test methodologies are as follows:

RIV Test

A pre-stress voltage of 30 kV was applied first for 5 minutes (before the RIV level was read). Test voltage was then decreased to 24 kV and RIV level was read again at this voltage. Then, test voltage was decreased to 20 kV and RIV level read once more. The same procedure was applied for dry and for wet insulators. Wetting the insulators was done by placing them upside-down and pouring de-ionized water onto the cement area close to the pin. The insulators were moistened this way for 15-20 hours before the test. The acceptance criterion to check if the lot passes the RIV test will be according to the formula: Average RIV level ≤60 dB-Risk factor x Standard deviation. Risk factor for 20 kV is considered as 1.6 and 1.2 for 24 kV. Application of the risk factor formula will make it more difficult for insulators having a large spread to pass. As such, there is already additional quality control.

Steep-Front Test

Tests were performed according to IEC cl. 5.6 with a series of 5 positive, 5 negative, 5 positive and 5 negative impulses. Based on preliminary results, it was decided that no impulses with reduced amplitude should be applied between polarity changes. Test voltage for the steep-front was according to IEC . At relevant steep-front test voltages, each impulse resulted in flashover. The acceptance criterion was according to C411.10.1, i.e. the insulator passes if every steep-front impulse voltage application results in external flashover and the test record or peak voltage indicator does not show a large reduction in voltage between impulse applications. To evaluate results from the complete batch tested, it was decided to apply acceptance criteria in accordance with IEC , cl. 5.8, i.e. requiring that no punctures are accepted. If only a single unit is punctured, a re-test procedure shall be applied. Also, punctured units shall be included in the mechanical failing load test.

After the steep-front test, all insulators were subjected to the mechanical failing load test performed according to IEC -1 cl. 19.2. The acceptance criterion according to CSA C411.1-10, cl. 6.13.2 was used. Similar to the risk factor formula applied for results of RIV testing, the CSA has a formula to evaluate results of the mechanical failing load test: Quality factor Q=4≤ (Average failing load-Rated failing load)/(Standard deviation). A summary of these tests is presented in Table 2 that outlines results for 16 insulator types. It is clear the tests revealed different qualities of insulators.

Discussion on RIV Tests

Conductivity measurements provide some indication of possible issues with cement under humid service conditions, i.e. high RIV levels. However, there is no clear correlation to RIV testing. For example, some Type ‘A’ insulators that failed the RIV test showed little or no resistivity of their cement and partial discharge activity came from the bottom of the cap. During the latest series of tests, performed in , it was found that two cameras should ideally be used to allow simultaneous observation of discharges from the top or bottom as this could occur with different insulator designs.

There is information in the literature that glass insulators experiencing partial discharges between the cap and disc (i.e. at the top) will also have a higher shatter rate due to deterioration of the surface. Indeed, in this project those insulators with high RIV activity between the cap and disc had an abnormally high shatter rate in service.

It therefore seems logical that when evaluating an insulator for purchase, a power supply company should carry out an RIV test under dry conditions. This should be done at the maximum possible level of voltage for the shortest typical string length for that voltage class. For example, in the case of the TSO in Sweden, this can be 24 kV for 400 kV overhead lines, whereas it might be 20 kV for Statnett or CEPS, the TSO in the Czech Republic. It would also seem reasonable to apply the same criterion for the RIV test under humid conditions, although results in this case suggest that this might prove too strict a test criterion. Therefore, further analysis of all results will be made, taking into account which insulators were considered as acceptable and then changing the risk factors of 1.2 and 1.6.

From all test results presented, it can be seen that the batches where glass shells shatter during the mechanical failing load test also have a higher probability of failing the steep-front test. However, at present both tests are still felt necessary. This is because of the distinct difference between a steep-front test and a mechanical failing load test. In the case of the former the glass shell is put under very high electrical stress, whereas in the latter mechanical load is increased to the maximum tension stress that results in mechanical separation of the insulator.

Proposed Test Matrix & Criteria

The proposed test matrix should consist of 3 tests: RIV, steep-front and residual strength. The guidelines for the requirements and criteria have been developed in the framework of this project and are shown in Fig. 3.

Summary

The two key issues under focus, i.e. high RIV levels and varying shatter rate, can be revealed by additional tests. Based on a comprehensive research program performed between and , the following tests have been verified and are recommended to verify the quality of glass cap & pin insulators supplied by different manufacturers: modified RIV measurements under dry and wet conditions; modified steep-front test; and modified residual strength test. Actual test methods have been discussed and an important point is to ensure that insulators for these tests are selected in a random manner.

For more information, please visit Composite Insulator.