If one looks back some 30 years or so, large national power companies such as EDF in France, CEGB in the UK or ENEL in Italy were important participants in the development process for electrical insulators. In co-operation with local manufacturers, these utilities often set the applicable standards and testing requirements and sometimes even provided test sites (such as Martigues in France) to assess field performance. But that was the utility model of the past.
In more recent years, with de-regulation and the re-focusing of the utility industry on its core business of electricity supply, such close co-operation with insulator suppliers has diminished and even disappeared … except in a handful of countries. One of these exceptions is South Africa, where the national utility, Eskom, continues to play a high profile role in the field of insulator development. Indeed, a proportion of turnover has been allocated each year to R&D activities that, among other things include monitoring the service environment and assessing performance of different insulator types and designs. Eskom has also helped pioneer new insulator test methods and related monitoring equipment.
INMR traveled to South Africa in 2013 to meet Eskom engineers and visit substations and lines scattered across the vast Western and Northern Cape.
With a sprawling network comprising some 300,000 km of transmission and distribution lines, Eskom has had a long history of problems due to insulator failures. When the country’s 400 kV ‘backbone’ transmission grid was established between 1966 and 1969, the first such lines employed glass cap & pin strings with the relatively short creepage of 15 mm/kV. In coastal areas, however, industrial and marine pollution combined with local weather patterns caused a high incidence of flashovers, especially during late winter and early spring. In addition, vandalism at various locations resulted in numerous broken glass discs. Locally produced porcelain cap & pin insulators were tried as an alternative but soon abandoned because of frequent punctures in areas with lightning.
Given these types of problems, by the late 1970s some of the glass cap & pin strings on Eskom’s transmission system began to be replaced by porcelain long rods having a specific creepage distance of 25 mm/kV. Although superior performance was expected due to their single unit construction and aerodynamic profile, the number of flashovers actually increased. Various mitigation techniques, such as hand washing, were then considered but such maintenance proved costly and mostly ineffective. By 1987, a program of washing from helicopters was introduced in its place, partly because of the speed with which the work could be done and also because it was often difficult for trucks to reach affected lines.
According to Eskom maintenance personnel, porcelain long rods required frequent washing in some areas and it was difficult to predict when was correct time. To help, a device was built based on evaluating the radio interference generated when pollution reached a critical level. Nevertheless, the cost of helicopter washing also proved high and, during extreme pollution events, even periodic washing did little to prevent flashovers. Nor was increasing specific creepage distance of strings an option because of limitations associated with minimum clearances.
During the late 1980s, mass pollution flashovers occurred in the eastern province of KwaZulu Natal and it was at this stage that Eskom became among the first utilities in the world to consider widespread application of composite insulators. Actually, several hundred silicone insulators had already been installed on lines located in heavily-polluted areas and, except for a handful of mechanical failures, these performed well electrically – even with a specific creepage distance of only 23 mm/kV.
A decision was therefore made in 1990 to specify silicone insulators with 30 mm/kV to progressively re-insulate all 275 kV and 400 kV lines located in areas exposed to severe pollution. As a result, by 1995 Eskom had over 15,000 composite insulators in service on its transmission network and overall performance was judged very good. This selective re-insulation reportedly resulted in customers experiencing from 100 to 200 fewer voltage depressions each year. In addition, there were cost savings as well from no longer having to wash lines. Excluding customer benefits, the US $ 7 million reportedly invested in the project had a payback of less than 10 years while also solving persistent past problems such as vandalism and pin corrosion on glass strings located near the seacoast.
Over the following years successive generations of composite insulators found their way onto Eskom’s network and the utility began developing expertise in assessing which designs were good and which were not. The basic strategy was to become ‘an informed buyer’ rather than simply depending on external sources. Part of this philosophy also involved fully understanding the actual service environment for insulators by monitoring pollution levels on a regular basis.
Eskom insulator expert, Wallace Vosloo, who helped implement this policy for years and published an influential text covering the field, remarked, “many utility people thought that IEC 61109 Annex C was all that was needed to guide their specification of composite insulators. But, in our case, we soon found out that it was simply not applicable to our environment.” As proof of this, Vosloo noted that in 2003, when Eskom instituted its program of qualifying potential insulator suppliers at the Koeberg Insulator Pollution Test Station (KIPTS), only 60% of the products tested over the subsequent 3-year period passed. Moreover, many of those that failed the tests also failed in service, even though these often fulfilled all the requirements set down in IEC 61109 Annex C. This, he concluded, demonstrated that meeting this standard alone – at least as far as South Africa was concerned – was not sufficient proof of expected good performance in service.
These days, even with a fast-growing population of composite insulators selectively installed throughout the country, widespread flashovers still occur in parts of South Africa due to different underlying causes. Sanjay Narain, one of Eskom’s Chief Engineers in transmission line engineering and insulation design, explained that in KwaZulu Natal, for example, these problems have usually been linked to high local humidity combined with the frequent burning of sugar cane. The Western Cape, by contrast, tends to be dry and widespread bush fires occur in cycles of about every 7 years, due to the time needed for re-growth of vegetation. These fires create high loading of airborne particulate matter that drifts out to sea and mixes with salt, before being blown back to land and settling onto lines. Clean insulators can become critically polluted in a very short time. The last major flashover problem in the area occurred in 2007 and the one before that, in 2000, apparently blacked out almost the entire Cape Town region.
According to Narain, at the time many local lines were still insulated with glass and it was only these that experienced flashovers. Moreover, some substation porcelain with as much as 31 mm/kV specific creepage also flashed over. “Because this problem was absent on lines insulated with polymers,” he reported, “we decided to carry out a massive project where nearly a dozen of our 400 kV lines were re-insulated with silicone rubber in place of glass. Hopefully, we solved cyclical pollution flashover problems due to brush fires.”
By 2013, Eskom’s transmission network was about to undergo a fundamental shift, with a new backbone of 765 kV. Said Nishal Mahatho, an insulator specialist at the utility’s Research & Innovation Department, “much of our system has been based on 400 kV and 275 kV, although we have two 765 kV lines insulated with glass that operate in the middle of the country. But we have changed philosophy and are moving in a phased approach toward a 765 kV grid and by 2020 to ±600 kV DC system as well. This means that there is even more work than normal going on when it comes to insulator research.” One example was a project at KEPCO’s Gochang Test Station in Korea where five different designs of composite insulators, already approved by Eskom, were compared to one another and also to the standard glass strings used on existing 765 kV lines in South Africa. The environment is coastal and every effort has been made so that the comparison is as realistic as possible, even using the same planned conductor bundle configuration.
Mahatho explained that leakage current is being measured on each insulator design and correlated against weather conditions. The insulators are installed in double I-strings and are apparently among the first to be monitored using sophisticated new leakage current sensors from EPRI in the U.S. Periodic UV and IR testing with special cameras is also being performed on each. Said Mahatho, “the idea was to pit the possible suppliers against each other and see how their insulator designs will perform in an energized test in that type of service environment. For example, because of the high voltage, electric field will be one of the issues that might contribute most to premature ageing.” He went on to state that the test lasted 2 years with at least 90% of this time under voltage and that parameters examined for each insulator included hydrophobicity and electric field distribution, in addition to leakage current. Insulators were also removed after a year to allow close-up visual examination before putting them back up.
Narain referred to the generation of insulators for Eskom’s 765 kV network and noted, “we are dealing in this case with composite insulators where corona ring design and placement may prove even more vital to performance than the silicone rubber housing. Narrain also pointed out that Eskom philosophy has always been that insulator corona rings must take care only of the insulator and not necessarily aim to also protect associated hardware. “For this reason,” he said, “you will find separate grading rings for hardware on a lot of our 400 kV and 765 kV lines.
Narain compared the South African topography to an upside down saucer in that there is a sharp rise from most coastal areas to plateaus that typically reach 1500 m or more. This means that a high proportion of Eskom lines have their insulation affected by altitude and therefore local test laboratories, such as the National Electric Test Facility have been intentionally sited to operate under reduced air density (in the case of NETFA at 1625 m).
Narain emphasized that the entire process of insulator selection at Eskom has been systematically refined so as to remove decisions based mostly on past experience. Instead, a large body of reliable information has been compiled on pollution across the country to record and track environmental stresses at every line and substation.
“In the past, he said, “selecting insulation was driven mainly by experience and perceptions. Instead, we have moved to eliminate subjectivity from decision-making and base our line insulation specifications not on guessing ESDD levels but rather on actual measurements that provide us with a statistical representation of the real pollution situation. This reduces the band of uncertainty and means we can fine-tune our selection of ideal insulator dimensions and geometry. For example, instead of specifying from 30-40 mm/kV, we can be precise and state we want e.g. 35-36 mm/kV.”
Narain explained that pinpointing pollution levels across the country and correlating these against the flashover behavior of actual insulator profiles from a test laboratory makes it possible to plot stress versus strength. “We put all this type of information together to arrive at the optimal insulator choice,” he said. “In other words, once we know the flashover characteristics of any particular insulator’s profile (i.e. the flashover voltages under different pollution severities) we can be certain of how it will respond in our service environment. It also means that different insulator creepage values can be specified in the case of a single very long line.”
One example of this type of thinking put into practice was the extension to Cape Town of the 765 kV Gamma-Kappa Line. The first 200 km of this 350 km line were insulated with glass strings having 20 mm/kV. However, once the line passes the mountains near the heavily agricultural Ceres Valley, it enters a zone of high pollution and insulation will be therefore changed to composite types with 31 mm/kV. “Of course”, noted Narain, “if there is not much difference in acquisition cost between low creepage and high creepage insulation, we might decide to go with the higher level along the whole route. Environments are dynamic, with possible increases in pollution levels, and we could therefore gain extra security.”
At the same time, he and colleague Vosloo warned that line design engineers must be cautious in specifying high creepage in every situation. Said Vosloo, “there is an optimized creepage which balances the risks of flashover and ageing. Too low a creepage increases risk of flashover but too high may increase risk of premature ageing.”
With the planned re-structuring of the transmission grid to one based on 765 kV AC and ±600 kV DC, there was a drive to establish a new KIPTS with greatly expanded facilities to handle testing at much higher voltages and also under DC. According to Vosloo, who was based there from 1993 and who oversaw much of its expansion from the initial 22 kV and 66 kV rigs to 11 kV, 33 kV and 132 kV, this work was expected to last about five years. The enlarged site will also include a high voltage hall as well as pollution chamber where flashover performance will be measured of test objects taken down from the station. The largest hurdle was seen as installing sophisticated power source equipment.
Given the amount of time needed for the expanded Koeberg site to be finished and commissioned, insulator testing at Eskom was planned to continue using a modularized scheme based on facilities enclosed inside large shipping containers placed inside a substation. Among these, a salt fog chamber was planned, with office, test bay and associated power sources and support equipment. There were plans as well for an inclined plane test set-up for AC and DC as well as a tracking wheel. Said Vosloo, “we feel confident in saying that our formalized testing requirements at KIPTS, imposed on every insulator supplier to Eskom since 2002, have helped advance insulator technology in general. The proof is that by 2013 some 90 percent of products tested passed – a 50% increase over the pass level at the start. More important to Eskom, we also eliminated most of the problems we once had on transmission lines.”