Utilities, including the Bonneville Power Administration (BPA; Portland, Oregon, U.S.), are performing live-line maintenance safely and effectively. However, since the original live-line maintenance procedures were developed for BPA's first 500-kV lines, there have been significant changes in the design of the line as well as the type of tools. The proliferation of compact tower designs is especially significant because of their reduced electrical clearances. These changes have driven the need to determine the electrical safety margin of current live-line techniques.

Full-scale tower models based on BPA's compact 500-kV line designs were constructed and tested at the Carey High Voltage Lab in Vancouver, Washington, U.S. BPA's maintenance department used the test results to specify the conditions under which live-line insulator replacement could be performed.

Insulator replacement is a common task performed on high-voltage transmission lines. Porcelain and glass insulator sheds can break from environmental factors or vandalism. Because an insulator string with broken sheds is more likely to flashover, it must be replaced. If a weakened insulator flashes over while crews are replacing it, maintenance personnel could be injured. Therefore, when developing a live-line insulator replacement technique, it is important to understand how the number of broken insulators, the position of broken insulators and the presence of live-line tools influence the performance of the insulator string.

Laboratory Test Program

An insulator's critical flashover (CFO) value, the voltage at which the probability of flashover is 50%, is used as an indicator of the dielectric strength. The CFO value depends on the type of stress applied. Lightning impulse, switching impulse and power frequency are the common stresses considered on the power system. Because a switching impulse with positive polarity results in the lowest CFO values, this type of stress was used at the Carey Lab to evaluate the live-line techniques.

Two full-scale tower models were constructed in the laboratory and used as a platform for testing. One tower model represented the center phase of a single-circuit configuration. The other tower model represented the top phase of a double-circuit configuration. The critical portion was modeled using a metal framework having the same dimensions as the actual tower. The test line at the lab was configured to represent a typical 500-kV line using a triple-bundle conductor. The test line, insulators, tower model and live-line tools were arranged in a realistic configuration. A series of switching impulses were then applied using a 5.6-MV outdoor impulse generator. The resulting flashovers were analyzed to determine the CFO voltage for the configuration.

Test Results

How many broken insulators can be tolerated? If live-line replacement of broken insulators is to be performed safely, it is important to quantify the extent to which broken insulators weaken the string electrically. Using the tower models previously described, the CFO was measured on a V-string with no broken insulators. The insulators were then broken two at a time until a total of eight insulators were broken in the same string. The test results showed that, indeed, the broken insulators reduced the dielectric strength by an average of 4% per broken unit.

Is the location of the broken insulators important? Previous studies have shown that the location of the broken insulators is critical, with the greatest influence observed when the broken units are located at the energized end of the insulator string. In the case of the double-circuit tower model, this finding was confirmed by the current test results. As the broken insulators were shifted up the string away from the energized end, the dielectric strength improved by an average of 1% per insulator. However, the single-circuit tower window was weakest with two unbroken insulators at the energized end.

How do different insulator types compare? BPA uses three types of insulators on transmission towers: glass, ceramic and non-ceramic. Each of these insulators responds differently to damage. Glass insulators, because of the way they are made, shatter completely, leaving little insulation in place. In field experience and in previous laboratory experiments, porcelain insulators typically break in chunks and retain significant dielectric strength, even when broken. However, in this study, the porcelain insulators lost most of their dielectric strength when broken. As a result, the performance of broken ceramic insulators was almost identical to the performance of broken glass insulators. Non-ceramic insulators (NCI) were not considered in this test program because BPA does not replace them while the line is energized.

How does the tower configuration affect the results? As previously mentioned, tests were performed on a single-circuit tower window and on the top phase of a double-circuit tower. Because the tower window is a higher-stress area than an outside phase, BPA's compact tower design had greater clearance in the tower window (305 cm) than on outside phases (254 cm). This increased clearance was reflected in the test results, which showed that the tower window was 9% stronger than the outside phase.

Do live-line tools reduce the flashover value? Perhaps the most interesting question of all is whether the live-line tools reduce the flashover value of a damaged insulators string. Surprisingly, in most cases, the dielectric strength improved with the live-line tools in place. The live-line insulator replacement procedure was divided into the following discrete steps with the assumption that movement of the tools between steps does not affect the dielectric strength:

1. Base case, no tools.

2. Add strain poles and yoke.

3. Add cradle, bale stick and other tools.

4. Swing insulators up to crossarm.

5. Swing insulators down.

6. Add ladder and mannequin.

7. Same as Step 3, but use dielectric rope in place of strain poles.

In some cases, it is easy to understand why there would be an improvement with the tools in place. In Step 5, for example, the damaged insulator string is lowered out of the way and the strain poles support the line. The strain poles have a higher strike distance than the insulator string, and a higher dielectric strength is expected. On the other hand, it is not known why adding the strain poles and yoke to the damaged insulator string results in an improvement. Nevertheless, most tools either improve the dielectric strength or do not change it significantly.

The test results provide a better understanding of the influence of each test variable; however, they are not complete by themselves. A methodology must be established for applying the results in a practical sense. The CFO values for each scenario or configuration must be categorized as acceptable or unacceptable in terms of risk to personnel.

Electrical breakdown is a random phenomenon and is best described by probability of occurrence. The laboratory tests measured the CFO level. Recall that CFO is the level at which a 50% probability of flashover exists. The lowest measured CFO in this study was 783 kV, which is considerably greater than the maximum operating voltage of BPA's system. However, a surge could raise the voltage to an abnormal level during the live-line maintenance. Lightning, switching and faults cause the most common surges on the power system. Live-line maintenance is not allowed when there is a risk of thunderstorms; therefore, lightning does not require consideration. Switching surges also can be ruled out, because reclosing is disabled on any circuit that is undergoing live-line maintenance. Hence, a fault is the most likely cause of a surge during live-line maintenance.

When a conductor is faulted, a voltage is induced on the unfaulted phases. Phases on the same circuit can experience as much as a 30% increase in voltage due to induction. There also can be a 30% increase in the conductor-to-structure voltage due to neutral shift. How these voltages relate to each other varies due to the time difference between the initial voltage change and the actual current flow. An increase in voltage on one of the unfaulted phases of 1.6 per unit has been assumed to represent the worst case. With a system voltage of 550 kV and a surge magnitude of 1.6 per unit, 720-kV peak is the worst-case voltage magnitude at the work site.

Another important consideration when applying these test results is the maximum allowable probability of flashover. Ideally, a flashover probability of zero would be used. However, this would rule out live-line maintenance altogether. Some level of risk must be tolerated, but how much? BPA has chosen a 3σ probability based on NESC guidelines. Given a normal probability distribution and a known standard deviation, the acceptable CFO level is then computed by adding three standard deviations to the worst-case surge. In this case, the result is a CFO of 805 kV. Any scenario that results in a lower CFO is unacceptable.

These tests have shown that on BPA's compact 500-kV towers, an insulator string with as many as four broken units can be safely replaced using current live-line tools and techniques. This knowledge will allow live-line maintenance to proceed on the compact lines, thereby increasing their availability. Future testing is planned to study double V-string configurations, different conductor bundles and other possibilities. These projects demonstrate that high-voltage testing will continue to play an important role in the safe and economical operation of the nation's power grid.

Jeff Hildreth works at BPA's Carey High Voltage Lab in Vancouver, Washington, where he is concentrating on data acquisition, computer control of test apparatus and high-voltage testing. Hildreth, who received a BSEE degree from Georgia Institute of Technology in 1995 and an MSEE degree in 2002, has worked on the design of high-speed digital circuits at Intel in Oregon and on the testing of power system components at the NEETRAC High Voltage Lab in Forest Park, Georgia.
jghildreth@bpa.gov

Don Gillies received a BSEE degree from Washington State University and worked for BPA from 1949 to 1979 when he retired. Since then, he has been a consultant for utilities and institutes on safety and maintenance problems. Gillies is a fellow of the IEEE and CIGR… and is active in IEEE committees concerned with transformers and T&D engineering. He is U.S. technical advisor to IEC/TC78.
dagillies@bpa.gov