One of the important design differences between silicone composite insulators and porcelain cap & pin or long rod types is their smaller shank diameter, made possible by the high strength glass-fibre reinforced core rod. This contributes to superior pollution performance, such as in those situations where hydrophobicity is temporarily lost. However, this slimmer profile also means smaller metal end fittings and the issue of corona then has to be considered, even at lower voltage levels. For example, as far back as 1992 CIGRE WG B2.03 provided a general recommendation that, starting from 220 kV, the application of a suitable corona ring on the HV side is necessary mainly because of the potential impact on RIV. Subsequently, research using specialized software became available to more accurately simulate how corona affected critical areas of a composite insulator (e. g. the triple point) as well as the effect of water droplet corona. Yet in spite of all this progress, the following are just a sampling of examples, which demonstrate that the threat posed by corona is still not always being dealt with properly:
• inappropriate use of hardware on a composite insulator (i.e. one designed for a cap & pin insulator string);
• corona ring incorrectly installed;
• no simulation or testing done for a dead end/tension installation, even though this is considered to be potentially more severe for corona damage than on suspension strings;
• manufacturing defects;
• unwise cost savings on composite insulator end fittings (e.g. too small or without a collar);
• no corona rings used at 245 kV.
The IEEE Taskforce on Electric Fields and Composite Insulators published a paper that provided an excellent summary of the state of the art in this critical area. The paper not only highlighted the proposed corona protection designs offered by different manufacturers depending on voltage rating but also shows possible installation faults and provides assistance on modelling electric field. It also differentiated between electric field intensities that influence insulator string behaviour under dry as well as wet conditions (water droplet corona phenomenon).
Three important thresholds (shown as rms values) are mentioned in this document. These thresholds must not be exceeded when the insulator string is being simulated under dry conditions with no pollution (see Table 1). The impact of air density also has to be considered in cases of installations at altitudes significantly higher than sea level. Local rules typically provide the necessary correction factors in such cases.
CIGRE WG B2.03 published a Technical Brochure (TB 284) titled Use of Corona Rings to Control the Electrical Field along Transmission Line Composite Insulators. Various 2-D and 3-D simulation programs were investigated, which use finite element or boundary element methods, and comparison of results has shown that, if correctly used, these programs provide similar results. An interesting approach was adopted during this work involving 3-D simulations on two different types of 145 kV insulators. Laboratory tests were then conducted to measure the corona inception and extinction voltages on their HV end fittings in order to verify the simulations. The tests were performed at two different laboratories with the test configuration kept constant for each, i.e. same conductor, conductor clamp, height above ground, distance to steel cross-arm. This experimental configuration was also modelled using 3-D simulations of E-field. The voltage at which the electric field at the most exposed portion of the HV end fitting reached 22 kVrms/cm was then referred to as ‘calculated corona threshold voltage’ (see Table 2).
Comparing calculated corona threshold with corona extinction voltages measured at the two laboratories, it was observed that the difference between measured and simulated values was less than 10% of the minimum value for Insulator Type 2 and even lower for Insulator Type 1. Moreover, the differences were such that, in all cases, the simulation tended toward lower values. Hence, if anything, the simulated results were always on the ‘safe’ side when it comes to design margins.
TenneT’s Wintrack in the Netherlands is a project that uses a compact 420 kV double circuit line. But this highly innovative line design had to meet stringent new values of acceptable electromagnetic field stress as specified by the government. This meant that for the composite insulators selected, permitted threshold of electric field was even lower than that specified from common practice in the U.S. (summarized in Table 1).
To verify this, comprehensive simulations were performed using finite element and boundary element methods and included modelling the neighbouring phases, insulators, hardware and conductors (see the FEM simulation illustrated in Fig. 3). These simulation models showed that neighbouring phases caused electric field to increase by about 10%, which then had to be taken into account when designing combined corona and power arc protection devices. Finally, after a successful demonstration that all these required values were fulfilled and that the protection devices also met aesthetic demands, full testing was completed.
The phenomenon of corona on composite insulators has now been well documented. With the progress in computer speed and performance as well as the availability of suitable software programs, associated electric field stress on the insulator can be calculated, even after taking the overall structure near the string into account. Analysis of field data has also confirmed that empirically determined thresholds can be deduced from long-term experience. Still, care must be taken whenever new designs are developed or when there are innovative packages, such as Wintrack, that combine new string hardware and insulators. Today, since cost is a driving force in much of line construction, design engineers must be aware that too great an emphasis on economics alone may lead to the risk of under-dimensioned components with only limited service lives.