Composite insulator housings have become a standard solution implemented on power grids worldwide and indeed are the preferred choice for application in heavily polluted service environments. To ensure performance under such conditions, a variety of design tests are performed, including: IEC 60587 – Inclined plane test; IEC 61462 (§7.3.2) / IEC 62217 (§9.3.2) – Accelerated weathering test; IEC 61462 (§7.3.3) / IEC 62217 (§9.3.3) – Tracking and erosion test at different shed profiles and dimensions; and IEC 62217 (2005-10) First edition – 5000 hours multiple stress test. Eric Moal and Volker Bergmann of Reinhausen Power Composites and Anna Sörgel of Siemens AG in Germany discuss pollution testing on insulators and specifically artificial pollution tests under DC stress as well as long-term outdoor tests under AC stress.
Insulator design depends on pollution level as well as on diameter, shed profile, material and creepage distance. Other important considerations include avoiding rain bridging, preventing localized short circuit between sheds, promoting self-cleaning, avoiding pollution ‘traps’ and controlling local electric field stress. Recommendations for shed profile of AC composite insulators are summarized in IEC TS 60815-3 and in IEC TS 60815-4 for DC composite insulators in uncontrolled environments. Compared to porcelain, the hydrophobic properties of silicone housings result in superior pollution performance, meaning lower creepage distance required. Moreover, in the case of large dimensioned units, composite insulators can save up to 40% of total weight versus porcelain. Also, the manufacturing process for composite insulators allows for bigger dimensions with a single piece, whereas porcelain insulator sections are usually limited in size in order to fit into most high-temperature kilns. Indeed, in severe pollution areas and for many UHV applications, composite insulators may be the only practical solution with regard to extremely high creepage distance. In regard to AC, there are IEC standards for design and testing and practical experience with composite insulators covers decades. In regard to DC applications, pollution behaviour is generally a more significant issue because of polarization and charge accumulation, with resulting possible higher pollution accumulation. In addition, a missing voltage zero crossing leads to different discharge behaviour.
Standards for pollution testing of composite insulators under AC and especially for DC applications are still not available. This has lead to special customer tests with increased requirements that are not necessarily widely accepted but rather based on individual experience. Additional longterm ageing tests under severe service conditions are therefore needed to provide a more complete overview of pollution behaviour.
Laboratory Tests: Artificial Pollution Test with DC
Suitability of insulator design can be verified by using recommendations in IEC TS 60815, by relying on operating experience or through laboratory testing. Customers typically ask for artificial pollution tests, especially for DC conditions, and several such tests have been performed according to IEC TS 61245 with pollution type A, i.e. that associated with inland, desert or industrially polluted areas. The test has to verify expected behaviour of different insulator designs at different pollution levels. Results are compared with requirements of IEC TS 60815-4.
Artificial pollution tests were performed in Germany on composite insulators using the solid layer method based on IEC TS 61245 for DC porcelain insulators. Additionally, preconditioning with kaolin according to CIGRE TB 555 was necessary to apply the pollution layer. The pollution suspension contained different quantities of salt and kaolin, described by ESDD (Equivalent Salt Deposit Density) and NSDD (Non Soluble Deposit Density). The following procedure was considered for these tests:
• Washing and drying insulator;
• Pre-conditioning by application of dry kaolin powder to suppress hydrophobicity;
• Application of uniform pollution layer (mixture of kaolin powder, sodium chloride and water);
• Measurement of ESDD and NSDD values according to IEC TS 61245 clause 6.5 and IEC TS 60815-1, Annex C to verify correct degree of pollution. If required, further adjustments done;
• Natural drying (around 24 h).
Fig. 1 shows the pollution layers tested with their respective ESDD and NSDD values. These correlate with site pollution severity classes according to IEC 60815-1. The 7 different pollution layers tested were labelled A to G.
Test samples with different alternating shed profile, diameter and pollution layer were chosen for the artificial pollution test (see Table 1). All samples were designed with a creepage factor of 4, i.e. in line with IEC TS 60815-4.
Verification of insulator pollution performance is usually done according to IEC TS 61245 as a withstand test and there are three individual test periods of 100 min each. The specified characteristics of the insulator being tested are confirmed, according to the standard, if no flashover occurs. To gather more information about limit values, modified up & down tests were also used. A starting test voltage was applied for the first 15 minutes. This voltage was calculated using equation 1, based on IEC TS 60815-4, where RSDD (reference DC site severity) and Cd (diameter correctionare introduced in equations 5 and 6. Then, voltage was increased every minute until the first flashover occurred. After flashover, the test voltage was raised as fast as possible to 90% of flashover voltage. Subsequently, voltage was increased again, as fast as possible, by 5% every 3 minutes until flashover. The procedure was continued for a minimum duration of 100 minutes, if no decrease in flashover voltage was observed at the last three flashovers.
Normalizing Test Results
To compare test results on insulators of different dimension, minimum flashover voltages are converted into specified creepage distances (SCD0, FO) according to equation 2 of IEC TS 60815-4. This also considers that, for average diameters Da > 250 mm, a correction has to be performed. Consequently, index ‘0’ for diameter correction and index ‘FO’ for normalized flashover values are given:
Decreasing flashover voltages result in increased values of SCD0, FO.