Empresas Publicas de Medellin (EPM) has been using arresters for the protection of transmission lines to improve its quality of service. Despite this practice, one of EPM's main customers, Empresa de Cementos Rioclaro, had repeated outages because of the high occurrence of automatic disconnect of power supply on a 115-kV transmission feed. When a study revealed that lightning was causing the problem, EPM installed Protecta*Lite arresters, supplied by Ohio Brass Co., Wadsworth, Ohio, U.S. After observing the positive results of this installation, the arresters were installed on other lines that had been experiencing similar problems.

The Problem The 115-kV Guatape-Rioclaro transmission line, at an elevation between 300 m-2200 m (984 ft-7218 ft), is 60 km (37 miles) long and runs through a rural area that has more than 200 thunderstorm days per year (Fig. 1). The transmission line is the only feed to the Cementos Rioclaro plant, which has control equipment that is highly sensitive to voltage fluctuations and power interruptions. The line was built in 1985. Operational records showed more than 120 outages per 100 km (62 miles) per year affected the production of the plant. One temporary outage of a few minutes, or even an automatic recloser operation or a voltage fluctuation caused by system switching, can cause an interruption of nearly 8 hours, resulting in the accelerated deterioration of the refractory furnace. The high maintenance costs and loss of materials were causes for concern.

The Diagnosis Once the problem was acknowledged, the Energy Design Department (EDD) and the Transmission Line Maintenance Department (TLM) developed a plan to revise the design and the physical inspection of the line. The first part of the plan confirmed the validity of the design. The line-inspection objective was to determine if the problem was caused by vegetation in the right-of-way (R/W), deterioration or contamination of the glass insulator strings, high ground resistance, proximity of objects near the line or vandalism.

To address these possibilities, the vegetation-free corridor below the line was extended an additional 8 m (26 ft) on each side, and glass insulators were replaced with porcelain on those towers with high ground resistance. Porcelain shows the presence of flashovers better than glass, since flashover burns on glass insulators can only be seen on their metallic fittings. In addition, counterpoises were installed on 95% of the towers to improve grounding.

Since laboratory tests revealed that the glass insulators were in good condition, fault-locating equipment was installed in early 1989 to identify the most problematic locations in the line. Because the problem persisted, the EDD recommended in 1991 that metal oxide arresters be installed in parallel with the insulator strings. Based on a statistical analysis of the automatic protective outages and the data from the fault-locating equipment, it was hypothesized that the outages were probably caused by lightning.

The Solution EPM acquired 81, 120-kV Protecta*Lite metal oxide arresters, having specifications that required 84-kV maximum continuous operating voltage (MCOV), a rating of 100 kV and an energy capacity of 184 kJ. To determine where to install the arresters, EPM contracted with the consulting firm Mejia Villegas S.A. (MVSA). Using information supplied by the TLM Department relating to ground resistance values, data from the fault-locating equipment and maintenance records, MVSA used software based on Monte Carlo techniques to simulate lightning characteristics, which were not known at the time. The study recommended the installation of the arresters on all phases at towers in five areas (Fig. 1). In February 1992, TLM installed the arresters with the line de-energized and completed the job in four 8-hour days (Figs. 2a & 2b).

The Results In a monthly follow-up, TLM monitored the discharge counters that had been mounted with the arresters and the fault-locating equipment. The data showed a high level of operation of the arresters and a 25% improvement in line reliability during the first two years. An analysis of the frequency of automatic outages (Fig. 3) between May 1985 and December 1995 encouraged EPM to acquire an additional 300 arresters to enhance coverage of the line. In addition, arresters were installed on other lines that had similar problems. The determination of where the new arresters were to be installed was based on using similar procedures followed by MVSA and based on the operational records of the first 81 units. For this case, after a laboratory check of each unit, the installation was performed using live-line techniques. A record of outages was obtained for different spans of the line before and after the arresters had been installed (Fig. 4) between 1992 to 1995. Table 1 shows a tabulation of results on different lines.

In December 1994, EPM began operating a lightning-location system. Based on the information collected from this system, continuous monitoring has been performed of lightning striking the line to disclose the time, place and parameters of the strokes. The data made it possible to link the automatic outages with the strokes causing the outage. For example, most of the strokes are of high multiplicity and high intensity. Based on this information, further simulations have been performed to diagnose line problems. In addition, the system has provided detailed information regarding lightning density on an annual basis that is useful for evaluating and understanding the operational behavior of the line (Fig. 5).

As an example of how the information obtained by the locator was used, an analysis was performed using the simulation program of the event when the three arresters on tower No. 26 got burned. The stroke that caused the problem had 10 subsequent discharges. The stroke's characteristics are shown in Table 2 involving intensity, times to crest, duration of discharge and time between discharges. The simulation showed that the energy discharge exceeded the arrester spec of 200-kJ capability. The ground resistance on the affected tower was the determining factor in the resultant value of dissipated energy. With values under 24 ohms, the arrester capacity would not have been exceeded. The ground resistance of the adjacent tower was found to be irrelevant within this range.

Although an evaluation of the economic benefits of these installations has not been made, the implementation has increased production and sales for Cementos Rioclaro and has improved EPM's image and its energy sales. TDW

Juan Carlos Nicholls Posada received his degree in electrical engineering in 1985 from the Universidad Pontificia Bolivariana of Medellin, Colombia. In 1994, he finished the specialization in transmission and distribution of electric energy from the same university. He joined EPM in 1986 and has been involved in maintenance of high-voltage lines and is presently manager of that department in EPM.

Jorge Henrique Cerezo Restrepo graduated in 1988 with a degree in electrical engineering from the Universidad Pontificia Bolivariana of Medellin, Colombia. He worked in MVSA from 1987 to 1989 when he joined EPM as a maintenance engineer in the high-voltage transmission line department of EPM.