The service territory of Florida Power & Light (FPL) encompasses 27,000 square miles, generally along the coast, serving approx. 4.8 million customers. Its Distribution Business Unit (23 kV and below) is responsible for more than 1 million poles and some 60,000 miles of line.
Florida has hurricanes, with associated torrential rains and high winds. Even routine summers come with high temperatures, high humidity and almost daily lightning storms. While the majority of new residential construction is now underground (UG), the older two-thirds of the system is overhead and vulnerable to these climatic factors. These distribution lines compete with nearby trees as high points on the horizon and are attractive targets for lightning. This article, based on a contribution at the 2015 INMR WORLD CONGRESS by retired FPL engineer, Larry Vogt, reviews how FPL deals with lightning.
On average, lightning strikes the FPL service territory more than 300,000 times a year. Lines in areas with high ‘isokeraunic’ levels (# of strikes/sq. km) could experience 2-3 direct strikes per mile, annually. Typical lightning can deliver currents ranging from 10,000 – 200,000 amps and those currents can induce voltages exceeding one million volts. When lightning hits a distribution line, a high-energy traveling wave is created on either side of the strike point as the lightning charge travels to the nearest ground.
FPL’s overhead distribution system has insulation levels between 95 kV and 150 kV BIL. While a surge arrester can clamp voltages at 40 to 50 kV (far below those insulation levels), a flashover can still occur before the wave even reaches the arrester.
Ways to Deal with Lightning
A logical best way to avoid lightning-related problems might be to convert the entire overhead distribution system to underground (UG). It’s a great idea but converting those lines to underground is generally about 10 times the cost of a new overhead line. Consequently, the overhead system is likely to stay around for a while. Regarding the FPL overhead distribution system, there are basically two widely used means to mitigate lightning problems: 1. Install surge arresters at potential flashover points (insulators, switches, etc.), and; 2. Use framing methods (high BIL) that minimize the flashover scenarios.
While it is common utility knowledge that using an overhead ground wire with stand-off insulators is an effective means to mitigate flashovers to lines, it becomes far less effective when transformers are mounted on those same poles. In highly populated areas, transformers are needed quite often and the high BIL of the line quickly becomes compromised. Consequently, FPL has used surge arresters as its primary defense against lightning-related outages for decades.
Est. Population of Surge Arresters on the FPL Distribution System: 1,545,000
Expected life: 20-25 Years
Annual Usage: 40 – 80 k
Arrester removals: 7000 annually
Arresters have been added to protect the system, but it seems they have problems too. For example, in 1998 FPL began a thermo-vision program using infrared cameras to look for hot spots on the overhead distribution system. The crews discovered hot connectors and splices but they also found something that was not anticipated – hot surge arresters. Some of those units were quite a bit hotter (+40°F) versus adjacent arresters.
After investigation, it was determined that these warmer than ambient temperature units were likely showing early signs of failure and moisture intrusion seemed to be the primary cause. Many units were removed for analysis in a laboratory and attempts were made to correlate temperature rise with time before failure. It seemed an obvious correlation but the weather variation in the field made this extremely difficult to evaluate. While it was known that moisture ingress and rising temperature were not good signs, it was not possible to establish whether a 15°F rise meant an expected residual life of 3 days, 3 months or 3 years. Regardless, it was felt that removing any arrester found to be more than 10°F above an adjacent unit was a good decision.
Based on the thermo-vision program, several hundred suspect arresters were identified and removed annually. Using infrared cameras, approximately 500 primary feeder sections were inspected each year, i.e. about 1/6 of the primary system. Results indicated that there were many arresters failing before detection and it was also realized that different scenarios may have caused ‘over-heated’ arresters, e.g. actual lightning duty, moisture ingress, physical damage, ageing, etc. Basically, this was a situation where a component the utility was counting on to protect the system was in fact creating reliability concerns of its own.
Over the years, FPL crews categorized outage investigations by what they found while physically investigating the outage. In general, the crews start at a substation and travel the line until trouble is detected. They would then clear up the trouble and report to a dispatcher who logged the probable cause. If the outage occurred during a severe lightning storm or high wind event, it was identified as ‘weather-related’ (i.e. blown arrester, flashed insulator, debris, etc.). If the outage occurred during routine weather, the cause was typically identified as related to the damaged equipment itself, or other logical categories. For example, it could be that a tree branch had fallen onto the line (i.e. vegetation) or a squirrel had ‘burned’ an insulator (i.e. wildlife), or an arrester bracket had tracked (i.e. failed arrester) or that there was a damaged recloser (i.e. failed equipment). Often, the component would be too damaged from the fault current to determine the actual root cause of failure. Since most of these units were fairly new, this made the engineering department wonder, ‘How many of those equipment failures were actually initiated by lightning?’ It appeared that lightning was likely related to a multitude of additional outages well after the weather event.
Arrester Related Outages
Based on the thermo-vision program, it became possible to explain why arresters were sometimes failing on ‘blue sky’ days. It also came to be understood that arresters typically fail for one of two reasons:
While arresters can handle some strikes effectively, their energy dissipation capacity is sometimes exceeded.
• Moisture Ingress
In older porcelain arresters, the top seal often fails with age and allows moisture to enter. Even polymeric type arresters tend to have sealing issues, allowing moisture to creep in.
Seemingly, however, it appears that arrester failure should cause only a momentary outage. As any unit shorts out for whatever reason and causes the disconnector to operate, the system should restore immediately. But there were sometimes extended outages related to arresters so there was need to investigate further. A review of 50 arrester related outages resulted in the following conclusions: 66% of the outages had to do with how the arrester leads were formed. When the disconnector operated, the lead came into contact with the feeder primary or was formed in such a manner as to not allow separation; 22% of the outages were due to the bracket tracking over.
The tracking evident on the glass-filled polyester bracket in this photo could have been detected as the line inspector traveled down the feeder. More likely, however, the worker saw only the blown arrester and continued on, assuming the bracket insulation was adequate. Even though the quantity of these events was not large by conventional standards, the bracket tracking issue caused extended outages. These outages eventually became significant enough to require a track-resistant, high BIL bracket, on all purchase orders issued after 2001.
Arrester-framing countermeasures included awareness presentations during crew safety meetings and changes to the standards. Since the ground disconnector eventually operates, the ground lead must be:
• Short enough so that it will not come into contact with another phase;
• Long enough to allow a minimum separation of 6 inches between the bottom of the arrester and ground.
Distribution Line Framing
In an effort to analyze different types of distribution line framing, FPL supported a Triggered Lightning Project initiated by the University of Florida. The project was set up at Camp Blanding in the early 90’s. While Professors Uman and Rakov were studying lightning, FPL added to their analysis by supplying the necessary materials and constructing a small section (6-8 poles) of typical feeder poles and lines. Crews used FPL’s framing methodology, i.e. vertical (insulators mounted on one side of the pole) and modified vertical (one phase on top and two on one side). Modified vertical had been the FPL standard now for over 30 years.
As natural thunderstorms developed in the area, the team used weather sensors to determine a high probability of lightning strike. When the timing was right, they sent a small rocket into the air with a leader wire attached. The other end of the wire was attached to a strike point on the line under test. As might be expected, the arresters typically failed near the strike point, but it seems they also failed on poles located more than 800 ft. away. While these experiments were interesting, it was difficult to say just how relevant this test line section was to an actual line in service. While the line section had equipment attached along with grounds, etc., it was not energized at rated voltage. Obviously, fault current was not part of the arrester failure mode. Some simple conclusions, however, were drawn: primarily, it was learned that the FPL standard heavy-duty arrester was no match for a direct strike. The following points were also concluded:
• Modified vertical framing flashes over 90% of the time with direct strikes to the line. Note: Direct strikes range from 10% to 20% of all strikes depending on line location. Rural areas have the highest exposure, while buildings and trees provide more shielding in urban areas.
• An average of 5 return strokes per strike were found, with continuing current flowing between return strokes. The energy created from the continuing current is likely a major factor in arrester failure.
• Flashovers can occur even if the arresters do not fail.
• With direct strikes, arresters are called to dissipate above 84 kJ of energy, yet they are typically designed to handle only 20-30 kJ. A distribution class arrester will fail more than 50% of the time if its rated energy dissipation capacity is exceeded.
• 80-90% of lightning strikes are indirect or nearby strikes. The arresters see far less energy and consequently almost always survive. Hence, Distribution class arresters provide adequate protection for indirect strikes.
• Overhead ground-wire framing provides the best protection for direct strikes and, combined with arresters (for indirect strikes), provides the best overall protection for lines.
|Type of Line Framing||Expected Number of Flashes||Total # Expected Interruptions|
|Vertical w OHGW||1.963||0.692|
|Vertical w/OHGW & Arresters||1.963||0.258|
A vertical w/OHGW framing and arresters was tried every 8 spans on 4 different feeders, each for a two-mile section. While it is very difficult to actually determine how many nearby strikes would have normally affected each feeder, a polygon was made of one feeder for a period of two weeks. In this study, 16 strikes were recorded within 50 ft. of the feeder line. There were no recorded flashovers in SCADA.
Each pole has a driven ground of 25 ohms or less. The overhead ground wire is connected to the driven ground at each pole using 300 kV BIL stand-off insulators. Note: this framing is only effective when there are no transformers mounted on the pole. Obviously there is a substantial cost to reframing existing lines. Because of the damage caused by multiple hurricanes in 2004 through 2006, however, FPL funded an extensive hardening effort. Since this involved replacing many older wood poles, these framing changes were incorporated into that program for only minimal extra cost.
• Arresters will eventually fail from lightning or age/moisture ingress issues. Care must be taken during installation to ensure that, when the disconnector does operate, the ground lead will not come in contact with primary conductors.
• The best protection for direct strikes is achieved with OHGW.
• The best protection for nearby strikes is achieved with arresters.
• The best overall line protection is a combination of overhead ground wire and arresters.