Competitive pressures combined with regulatory constraints limited new line construction at Public Service Electric and Gas (PSE&G), while demand growth resulted in increased loading on an aging electric transmission infrastructure. Live-line techniques enabled PSE&G, Newark, New Jersey, U.S., to perform routine maintenance functions without taking customers out of service. Live-line work also provided alternatives as PSE&G worked to meet contractual obligations without incurring penalties.

PSE&G experienced a series of “blue sky” trip-outs on its 138-kV lines — the backbone of its 1100-mile (1770-km) overhead transmission system. The cause of these trip-outs was flashovers of electrically degraded insulator strings. The insulator strings, 60 to 70 years old, are comprised of nine individual porcelain suspension units located between large arc rings.

PSE&G evaluated the effects of broken and shorted insulators on the operating characteristics of the line, while investigating the live-line tools and techniques to replace defective insulators. Because no industry methodology existed, a team of PSE&G tower workers assembled to tackle the problem. The team developed mechanical work procedures. However, a need existed to confirm safe live-line replacement techniques for damaged insulator I strings on 138-kV double-circuit steel lattice structures.

PSE&G engaged the EPRI Engineering and Test Center — Lenox, Massachusetts, U.S. (operated by EPRIsolutions, Palo Alto, California, U.S.) to measure the phase-to-ground and phase-to-worker electrical flashover withstand for various broken insulator configurations. PSE&G shipped field structures, insulators and hardware to the Lenox Center. EPRI personnel prepared a full-scale mock-up of the 1920s vintage 138-kV Type “K” double-circuit steel lattice tower for the Lenox test chamber according to PSE&G drawings. Switching impulse sparkover tests were performed with a varying number of broken and shorted insulator units removed from the field and tested in the I-string configuration. Tests were performed on the top, middle and bottom phases.

Typical live-line working tools were installed in the Lenox test chamber with workers simulated by mannequins dressed in conductive suits (Fig.1). Various worksite scenarios were tested to determine the 50% probability sparkover voltage, U_50, which is calculated using the up-and-down method as referenced in IEEE STD 4-1995.

The Locke porcelain suspension insulators that were subjected to lab testing had been in service since the 1920s when the line was constructed.

Assuring Adequate Protection

The number of damaged units, their distribution within the string, the presence of the worker and the presence of live working tools influenced the disruptive discharge/withstand performance of insulator strings with defective units. Therefore, it is crucial to identify the worst-case scenarios for evaluation that result in the lowest withstand voltage at the worksite under realistic live working conditions. The safety of the worker is assured if this withstand voltage exceeds the stresses by a suitable margin. If sufficient margin is not available, live work cannot be performed or additional means must be employed to reduce worksite voltage stresses.

Calculate First

Estimating the remaining dielectric strength of an insulator string with defective units is critical in developing live-line working procedures. CIGRÉ Electra document TF 33.07.02 suggests the following formula for insulator strings with defective units in the worst location:

U_b/U₀ = 1 - 0.8 k (n_b/n₀)

Where U_b is the 50% disruptive discharge voltage of an insulator string with nb defective units in string with n₀ total units and U₀ is the 50% disruptive discharge voltage of the same string with all healthy units. The coefficient k is an empirical coefficient that characterizes the average state of defective units. For cases where less than 50% of the string contains shorted units, the equation gives reasonable results.

After performing switching impulse flashover tests, the test conditions revealed a k value of 1.0. High-speed photography provides a view of a flashover as shown in Fig. 2.

U₅₀ test data is plotted in Fig. 3 after being corrected for atmospheric conditions. Data is presented for top, middle, and bottom phases with broken insulators and arc rings installed. The test insulator strings contained anywhere from one to four broken units.

The U₅₀ values of each of the three phases exhibit similar behavior. U₅₀ is nearly independent of the number of broken units up to three, because all sparkovers in these cases occurred between the large arc rings. For cases of I-strings with four or more units broken, the U₅₀ value decreases as nb increases.

As shown in Fig. 3, a comparison of U₅₀ values with and without tools indicates the effects of tools range from insignificant to a reduction in U₅₀ of about 11%, depending on the details of worksite conditions.

Analysis of test data indicates that little difference exists between the sparkover values obtained from strings containing broken insulators and those strings containing shorted insulators. Test data also indicates that the CIGRÉ formula for the calculation of the effects of defective insulators is conservative, which gives confidence that the procedures developed by the PSE&G tower team will provide safe working conditions.

Effects of Tools and Workers

Several tests explored the effects of tools and workers in various hotsticking scenarios. Workers represented by electrically conductive mannequins were placed in typical work locations (Fig. 4). In accordance with normal work procedures, strain sticks were installed on the middle phase, while the top phase was pushed out with a hotstick to increase available distances between worker A and the top phase. The top and bottom phases were grounded during tests to simulate worst-case worksite stresses.

Tests were conducted to confirm barehanding performance on the top phase when appropriate distances to the middle phase and the tower are observed. The test setup consisted of a simulated metal bucket positioned at the level of the top phase. The metal bucket was on an insulating support and was bonded to the energized top phase. The bottom phase was grounded and the middle phase was left electrically floating for these tests.

The middle phase was left electrically floating. The ring-to-ring distance on the energized top phase was 0.98 m (38.5 inches), the shortest distance between the head of worker “D,” who was bonded to the top phase, and the grounded top ring of the top phase was 0.94 m (37 inches) (Fig. 5), and the shortest distance between the energized metal bucket and the grounded middle arm was 1.25 m (49 inches). The resulting U₅₀ is 544 kV (corrected). This is somewhat lower (by about 4%) than the U₅₀ value of the top phase without the worker.

It is interesting to study the CIGRÉ equation for the case of a string of insulators with broken units and arc rings. Resulting data clearly shows the following trends:

• Actual U₅₀ values are governed by the arc rings for a small number of broken units (up to about three)

• The experimental data points fall well above the CIGRÉ values for four and five broken units.

This last point provides further confirmation of the advantageous role of the arc rings in grading the voltage distribution along the string, even in the presence of broken units. The equation provides very conservative estimates of the sparkover performance of the string.

The maximum expected worksite stress for 138-kV lines is related to the maximum nominal system voltage. Based on standard industry calculations, tests that result in a U₅₀ of less than 340 kV (corrected) should be interpreted as undesirable from a live working viewpoint.

Testing has confirmed that conditions can be met for safe live work in most situations found on the 138-kV PSE&G steel lattice tower. Placing the tools and the workers at locations recommended by PSE&G does not appear to degrade the sparkover performance of the worksite. At the same time, several situations have been identified for which insufficient margin is available to accommodate live-line working.

Analysis of the sparkover behavior of broken insulators confirms the general applicability of the CIGRÉ formula. However, care must be used in applying the CIGRÉ formula to certain live working situations.

Testing continues to be an important part of PSE&G's live-line program, enabling the company to assess current work methods, evaluate new tools and procedures, and maintain worker trust and confidence.

Raymond Ferraro is a senior engineer in the Transmission Engineering Group of the Electric Transmission Department at PSE&G in New Jersey.

Tom Verdecchio has been employed at PSE&G for 27 years and is now the safety live-line coordinator in charge of live-line work on all high-voltage transmission lines at PSE&G.

George Gela is a senior engineer for EPRIsolutions — Lenox, with responsibility in the areas of high-voltage testing, live working, compact and upgraded transmission lines, and application of fiber optics in high-voltage environments.