By Stephen Gooding, Georgia Power Co. and Greg Hataway and Timothy Knox, Schweitzer Engineering Laboratories Inc.
Faults on electric power grids are never good, but sometimes are unavoidable due to inclement weather or other unforeseen occurrences. The causes of such faults are typically not too difficult for grid operators to identify. Clear sky, also known as blue sky, faults, however, are a different story. The cause of faults on transmission and distribution lines that occur for no apparent reason can be difficult to locate and diagnose; and can be frustrating for the electricity provider, as well as its impacted customers.
In the summer of 2014, Georgia Power Co. experienced several recurring clear sky faults and decided to investigate. The company installed two transmission protection relays with traveling wave fault location-one on each end of a two-terminal 230 kV transmission line that was approximately eight miles long. The relays were configured to protect the line using the distance protection function and they were connected to a GPS satellite-synchronized clock to ensure accurate time records. Because they were temporary, Georgia Power did not install a communications link between the relays, which is necessary to provide an automatic traveling wave fault location. Instead, the company set up the relays so that if a configured differential communications channel existed, following a disturbance, the relay would send the traveling wave arrival information to the remote terminal unit. When the relay received the remote travelling wave information, the traveling wave algorithm automatically calculated the fault location.
Event on the Line
On Aug. 9, 2014, an event occurred on the Georgia Power transmission line. The installed relay at each terminal generated an impedance-based fault location for this event, but without the communications link, the relays could not automatically generate a fault location using the in-relay traveling wave function. The relays, however, still captured traveling wave oscillography with the wave front arrival times. These data, along with line length and the wave propagation velocity, were sufficient to allow a simple hand calculation of the traveling wave fault location.
The utility assumed that because a storm had been present at the time, the fault was due to a lightning strike and therefore was unrelated to the recurring clear sky events. Even so, the staff still wanted to send a line crew to investigate potential damage. The utility contacted Schweitzer Engineering Laboratories Inc. (SEL) and provided the company with the event reports and relevant line diagrams to determine the fault location.
|FIGURE 1: Example fault on a transmission line. Traveling wave data were recorded at both ends of the line by the transmission relays.|
Calculating the Traveling Wave Fault Location
In addition to the oscillography captured by the relays, an accurate traveling wave fault location requires two pieces of information: the line length and the line’s wave propagation velocity. Any variance in these values will impact the accuracy of the calculated fault location.
Georgia Power believed the line distance was 7.84 miles. Using the wave propagation velocity setting in the relay and the first wave arrival times at both terminals, its engineers determined that the fault distance was 3.2677 miles from the northern substation. In contrast, the impedance-based fault location on the event report indicated a fault distance of 3.06 miles from the northern substation.
While analyzing the fault data, the engineers also noted that the fault happened at almost the 90-degree point on the voltage wave, as seen by the current’s minimal direct current (dc) offset. The resulting fault current magnitude was therefore relatively low compared to fault currents with dc offset.
Finding the Fault
Structure 21 (transmission tower) on the charts was identified as a likely location to investigate based on the fault data analysis. Assuming the towers were evenly spaced, Georgia Power calculated that the structure was 3.2158 miles from the northern substation, which was only 274 feet from the traveling wave fault location indicated.
Evidence of a flashover is typically difficult to find and this case was no different. Georgia power had to use a telescope to examine each individual bell on an insulator at the tower. This task was potentially even more difficult due to the lower fault magnitude, which could cause less damage and leave less evidence. In addition, the high-speed protection of the transmission relays, while important for reducing the impact of faults on the system, further reduced the potential damage by limiting the fault duration. Despite these challenges, the line crew successfully identify evidence of a flashover on one of the insulators (Figure 2).
|FIGURE 2: Even looking through a telescope, evidence of flashover on an insulator at Structure 21 was minimal.|
“The traveling wave algorithm put the line crew right on the correct structure,” said a utility worker. “Fast line clearing limited damage but made it very hard to see the evidence. Without the traveling wave fault location, finding the flashover point would have been very difficult.”
Fault Location Accuracy Comparison
Traditionally, protective relays employ impedance-based fault location methods, using voltage and current measurements along with line impedance to locate line faults. Although the impedance-based method works accurately in most applications, this method is affected by fault resistance, zero-sequence mutual coupling and series compensation. In addition, advanced high-speed transmission relays clear faults in less than a cycle, which provides limited fault data and contributes to additional error in an impedance-based fault location.
When traveling wave fault location is used, the relay measures the arrival time of the traveling wave fronts that originated from the fault. The data, which are not affected by the factors mentioned earlier, are used to calculate the location. The accuracy of the traveling wave fault location, however, depends on the accuracy of the line length and wave propagation velocity.
With the best-known line data available to the utility, the traveling wave fault location indicated a distance of only 274 feet from the actual fault location, or within one tower span, and was sufficient to guide crews directly to the correct structure. In contrast, the impedance-based fault location indicated a distance that was approximately 823 feet away from the actual location. This means the traveling wave fault location for this fault event was three times more accurate than the impedance-based location. It should be kept in mind that this was a short transmission line-only 7.84 miles. Longer lines would further compound the potential inaccuracy of impedance-based fault location because the traveling wave measurements are not adversely affected by distance. Table 1 includes a comparison of the fault location accuracy based on these methods for this fault event.
Figure 3 shows the traveling waves captured by the relay at both terminals using event visualization software. The software also can aid the user in selecting the appropriate traveling wave fronts to measure the wave velocity. Based on the selected traveling waves and a known fault location, the software can automatically calculate the wave velocity, which can improve accuracy.
|FIGURE 3: Event visualization software showing the traveling waves with a Bewley diagram.|
Locating faults and taking the necessary steps to fix the damage increases power system reliability. Temporary faults with high-speed line clearing and minimal flashover energy can, however, leave little evidence, making it difficult for line crews to identify the exact fault location, using up valuable time. In the past, utilities sent crews into the field to travel the line and inspect multiple tower structures with a telescope to discover the fault location. With accurate traveling wave fault location provided by the transmission relays, utilities can increase the efficiency of repair crews by sending them to the correct structure to evaluate the damage. Even without a communications link, as in this study, the transmission relays can provide the necessary wave arrival information to manually calculate the fault location. Traveling wave fault location is more accurate than traditional impedance-based methods and has the potential to save utilities’ substantial time and money.
|Courtesy: @GeorgiaPower Facebook|
Other faults have occurred on Georgia Power’s line since this first event, and each time, the relays calculated the fault locations accurately to within one tower span. The most recent event happened in May 2016 on the same transmission line. In that incident, Georgia Power experienced an A-phase to-B-phase fault. The fault carried 18.8 kA on A-phase and 16.89 kA on B-phase, and caused a 36 percent voltage depression at the McDonough substation. Using traveling wave event records in the SEL-411L-1 relay, the engineer calculated a fault location of 5.23 miles from the McDonough substation. With this information, a helicopter patrolled the line and found that a tree fell through the line at Structure 38, which is located just 5.22 miles from the McDonough substation.