The surface of polymeric insulators is typically made of a hydrophobicity transfer material (HTM) and this differs from both porcelain and glass. But hydrophobicity level on such insulators can change over time, depending on voltage and environmental stresses. That means pollution withstand can vary as well. As such, while both laboratory and field pollution withstand for porcelain and glass insulators are a function only of their geometry, in the case of polymeric insulators this will also vary with surface condition. For ceramic insulators, pollution withstand is the design value under polluted conditions. In the case of composite insulators, however, ageing can occur at voltages significantly below withstand if leakage current is present – even if of low magnitude. Different design margins therefore have to be taken into account versus porcelain and glass insulators.
Many attempts have been made over the years to set up suitable pollution tests for composite insulators. These include tests based on salt fog, with particular attention to the quick flashover method and also solid layer tests where insulators are pre-conditioned with kaolin. The insulator is then contaminated with a slurry of kaolin and salt with different recovery times (as described in CIGRE TB 555). Similarly, there have been tests with solid layer pre-conditioning of insulators with kieselguhr followed by contamination with a slurry made up of kaolin, kieselguhr and salt to simulate the range of wettability levels from 1 to 7.
Such experience is presently available mainly for line insulators and for AC voltages. By contrast, experience with large diameter station insulators and also under DC has been more limited even though these are the cases where design is determined largely by pollution. This article by industry expert, Alberto Pigini, analyses available test experience and resulting indications for insulator design.
Pollution Tests with Standardized Procedure
Extension of the standard procedure based on three ‘one hour’ withstand tests (as per IEC 60507 or 61245) to composite insulators is problematic. This is due to the fact that pollution performance depends greatly on surface condition of the insulators and this will vary with degree of conditioning. For example, Fig. 1 presents results obtained under AC voltage on one polymeric insulator type by pre-treating the surface with a slurry of water and kaolin or with only kaolin (clay). Different surface rubbing procedures have been used with the goal of investigating insulator pollution performance in terms of unified specific creepage distance (USCD) versus surface wettability (W), the latter determined according to IEC-TS 62073. Dependence of pollution performance on surface hydrophobicity is evident. Consequently, adopting the standardized procedure would need, a priori, to define the pre-conditioning methodology in order to obtain repeatable, reproducible as well as representative results.
Salt Fog Method
Experience with Quick Flashover Method
This method is performed using a variable voltage procedure. Initially, a voltage lower than the expected flashover voltage is applied and maintained for 20 minutes. Then, voltage is raised in successive 5% increments until flashover, maintaining voltage for 5 minutes at each step or up to flashover. In the event of flashover, a voltage equal to 90% of the previous level is applied. Again, voltage is increased in steps of 5% until the next flashover. The test continues until a stable flashover voltage is obtained, i.e. the difference among the last 5 flashover voltage values is less than 5%. The average of these last 5 flashovers is then assumed to be an estimate of insulator performance.
An example of such a procedure is shown in Fig. 2. Here, tests with AC voltage aimed to demonstrate the value of this methodology as a diagnostic tool to compare performance of new insulators, insulators aged in the laboratory and insulators removed after a few years’ service in the field. Testing on a new (i.e. virgin) insulator highlights a very high flashover value. In particular, the first flashover value is close to that obtained on an air gap equal to the insulator’s arcing distance. Then, flashover voltage rapidly decreases due to the conditioning caused by subsequent flashovers and reaches some stable value.
The test was repeated after a 1000 h artificial ageing test by exposing the insulator to repeated weekly stress cycles as in Fig. 3. In this case, there is no significant decrease in flashover voltage with number of flashovers (i.e. the insulator already appears conditioned by ageing). Furthermore, the insulator showed a significant decrease in flashover voltage at stabilization compared to the virgin insulator.
A third insulator of the same type, this time sampled from a 150 kV line after three years of operation, was also tested with the quick flashover method. The results, reported in Fig. 2, are very similar to those obtained on the insulator after 1000 h laboratory ageing. This confirms the value of this method as a diagnostic tool.
The quick flashover method was applied to evaluate insulators of different types and materials (see Table 1). The insulators were made of different polymeric materials (ethylene propylene rubber, EPR, ethylene propylene diene monomer rubber, EPDM and silicone rubber, SR) and also had different geometric characteristics. Quick flashover tests were performed both before and after the 3000 h ageing test carried out by repeating the weekly cycle shown in Fig. 3 (see test arrangement in Fig. 4).
Comparison of these insulators in terms of the USDC required under DC voltage is shown in Fig. 5. With the extreme ageing conditions being simulated, performance of the EPDM and EPR insulators was actually better than that of the SR insulators – probably because under those conditions a material’s resistance to tracking and erosion is more important than its HTM characteristics. The results of the quick flashover test, together with all the other evaluations (e.g. visual inspection of degradation, leakage current, etc.), were used to obtain an order of merit of the different insulators and also as a diagnostic to obtain indications of relative performance following the laboratory ageing test. Results of the tests at a salinity of 80 kg/m3 are plotted in Fig. 5 in terms of USCD.
It is interesting to compare these USCD values with those of a proposed USCD-SES model where SES (site equivalent salinity) is assumed equal to Sa. Such a comparison is reported in Fig. 6 for both unaged (top) and aged insulators after the 3000 h test (bottom). The same Fig. also reports on the USCD-SES model for ceramic cap & pin insulators. This comparison highlights that, while unaged insulators require lower USCD values than suggested by the USCD-SES model for composite insulators, insulators after the 3000 h test require USCD values that are in good agreement with the model.
Quick flashover tests are also useful to obtain indications about optimal insulator profile. For example, Fig. 7 plots test results of USCD vs. shed spacing and highlights that the smaller the inter-shed distance, the higher the required USCD, as expected.
Results confirm that salt fog and in particular tests using the quick flashover method are useful to obtain an indication of pollution performance of composite insulators. Incidentally, the quick flashover method is much faster and also less costly than the standardized method for ceramic insulators. However information on performance of composite insulator under the quick flashover methods yields only preliminary indications for design, as may also be derived from ageing tests as well as service experience.
In particular, referring to the ageing tests, these were carried out by energizing the insulators shown in Table 1 at a test voltage of 70 kV in DC corresponding to an average USCD of 49 mm/kV (i.e. 44% higher than the average value suggested by Fig. 5 for new insulators).
In spite of the above margin between voltage from the quick flashover method and test voltage selected for ageing, insulator degradation occurred following the ageing tests, as shown in Fig. 8. Quantification of erosion in each case is shown in Fig. 9, which reports on degree of erosion of the coating expressed in % of the local thickness, after 3000 hours DC ageing.
Results confirm that in the case of composite insulators it is not sufficient only to select an appropriate USCD to avoid flashovers in service (as for ceramic insulators). Instead, it is also necessary to allow additional safety margins to limit degradation as much as possible (e.g. by selecting USCD values that guarantee very low leakage currents in service).
Solid Layer Test Experience
Fig. 10 reports results obtained on composite insulators with partially or fully suppressed hydrophobicity. Comparison is also made with a curve proposed for design. The wide dispersion of data is evident and this is likely explained by the varying degrees of hydrophobicity transfer and recovery during testing. Another possible explanation could be difference in methodologies, given lack of a standardized procedure. Finally, these differences may derive from differences in shed geometry, as per Fig. 11 that reports USCD required as a function of creepage factor (i.e. the ratio between creepage and arcing distances).
Attempts have been made to establish an agreed procedure that basically follows IEC 61245 after having first pre-conditioned the insulator using dry kaolin. This would also take into account the HTM characteristics of test insulators by prescribing different times between coating with the pollution slurry and the time of testing to allow for recovery. Variations in flashover voltage allowing for such recovery time, however proved limited (i.e. 10 to 20%).
Investigations mainly covered AC voltage. The few tests made under DC voltage gave a USCD50 of about 27 mm/kV at SDD=0.3 mg/cm2 (i.e. in the lower range of Fig.10).
Attempts were also made to arrive at some procedure to simulate different hydrophobicity levels by contaminating the insulators with a different type of slurry. Here, a range of wettabilities were obtained by using different non-soluble materials in the slurry, e.g. kaolin to simulate limited recovery (i.e. W close to 7), kieselguhr with a small percentage of kaolin to obtain a wettability of 5-6 without any recovery time after contamination and 1 to 3 by allowing a recovery period of several days. While these tests were made in AC, it can be assumed that results could also apply to DC as far as p.u. is concerned. Dependence of USCD on wettability level is shown in Fig.12. It is interesting to note that polymeric insulators offer good performance even when they have lost part of their hydro-repellence (namely up to a wettability level to about 5).
This could be explained by considering that, for high hydrophobicity levels, the flashover process along the surface becomes more difficult and the preferential path is across air. There is therefore far more limited influence of pollution severity and insulator geometry (including diameter).
Design of Composite Insulators for Contaminated Service Environments
As discussed above, assessing the pollution performance of polymeric insulators is made more complex because this depends on evolving surface conditions. Moreover, even if representative short-term pollution performance could be established, applicability to design would be questionable since the problem for composite insulators is usually not to prevent flashover but rather to avoid deterioration during service. As such, prevailing opinion is that design of polymeric insulators for now should be based mainly on field experience. It is interesting in this regard to consider the opinion of WG members, while preparing the recent version of IEC 60815 Part 4:
“Although there is some positive experience with validation by testing of traditional glass and porcelain insulators, this experience is mainly with pollution withstand tests where the degree of over-design is unknown. Any such experience is mainly lacking for composite insulators and the problem is accentuated by the continuing rise in system voltages where over-design may result in unrealistic insulator lengths or heights. Hence, for this first edition, the verification of a chosen insulator solution by testing is entirely subject to agreement.”
In spite of the still limited information on performance of composite insulators under DC voltage, attempts have been made to give preliminary design indications based on available information. Specific software was set up to obtain indications for both ceramic (porcelain and glass) and polymeric insulator design. The software covers both solid layer and salt fog conditions but only the former is examined below:
First of all, one has to establish desired insulation performance to be taken into account during design. In the case of composite insulators, this starts with the limited laboratory and field experience available. Schematization is to be optimized in the future based on systematic tests. Fig. 13 shows the assumed trend of USCD with SDD for insulators having an average diameter of 250 mm and a creepage factor of 2. Influence of insulator geometry in terms of average diameter and creepage factor is then shown in Figs. 14 and 15 respectively (SDD=0.1 mg/cm2).
Once assumptions are made on insulator performance and expected pollution severity, required specific creepage distance compatible with required line performance can be evaluated. For example, calculations were performed assuming:
• ceramic insulators with average diameter 250 mm and CF=3.2;
• composite insulators with average diameter 100 mm and CF= 3.2 after service with a wettability level W=6;
• maximum pollution severity (2% probability) SDD=0.1 mg/cm2;
• 10 critical wetting events and only 100 insulators at assumed maximum contamination for entire line.
The number of flashovers that can be expected for ceramic (black line) and for composite insulators (red line) as a function of selected USCD is shown in Fig. 16.
Here, assuming 1 flashover per year would be acceptable for the line, 50 mm/kV would be necessary for ceramic insulators while only 35 mm/kV would be sufficient for composite insulators. It should be taken into account, however, that composite insulators should operate far from flashover conditions in order to limit leakage currents that could lead to deterioration. This could be achieved by requiring higher margins for composite insulators, e.g. allowing a number of flashovers every 10 years for composite types versus 1 flashover per year for ceramic insulators, the USCD for composite insulators would be 45 mm/kV – a value close to that for ceramic insulators.
• Performance of polymeric insulators at a given pollution severity is not given by a unique value since this depends on surface condition, which varies during insulator service life.
• In spite of different possible pollution test methods, there is not yet agreement on methodology to be adopted that meets the criteria of always being representative, reproducible and repeatable.
• Investigations up to now have mostly covered line insulators. Information about influence of insulator geometry on required USCD at different hydrophobicity levels is still limited, especially for DC.
• Attempts have been made to schematize the different influences based on the limited quantitative/qualitative information already available and this will be optimized based on results of subsequent investigation.
• Application of ad hoc software yields indications on required creepage distances for composite insulators.
• Design of composite insulators should take into account not only the goal of limiting flashovers but also ensuring leakage currents are contained to very low values, even when most of the initial hydrophobicity has been lost. Calculations indicate that, given this requirement, USCD of composite insulators can end up being close to that of ceramic insulators