When Following Faults, System Operators Face Special Challenges in Order to Restore System Operation Safely and Expeditiously. To make the best-possible decision, operators must know both the nature and location of a fault. A research and development project now under way at Consolidated Edison Company of New York Inc. (Con Edison; New York, New York, U.S.) demonstrates how using automated fault-location systems as a decision support tool improves overall system reliability, minimizes potential power interruptions and cuts operational costs.

RELIABILITY BENEFITS AND MORE

In general, the use of an automated fault-location system improves an operator's situational awareness of the transmission network. Because it provides the operator with accurate, real-time information, the major benefit is improved system reliability and compliance with operational standards. The value of having accurate fault-related data cannot be overstated. Its availability allows the operator to make informed decisions that ensure the safety and stability of the power system. This is especially important when a multiple-trip fault occurs and the operator must quickly restore system integrity.

Like many utilities in the United States, Con Edison has transmission lines tapped to distribution substations. A second benefit resulting from the use of an automated fault-location system is that it allows operators to correctly detect and locate faulted segments on transmission lines and, therefore, minimize power interruptions to distribution substations.

The third major benefit is that automated fault location provides a significant reduction in the time required to locate a fault. By automatically providing accurate fault location information, the repair crew's job is simplified to where they just need to fix the problem. The need to locate the problem is accomplished by the system. The same is true for underground cables: Automated and accurate fault location expedites cable restoration. The fault-location system also helps engineers when conducting the post-fault analysis, as it provides accurate information about what happened and where.

While the major benefits of an automated fault-location system are improved reliability, the system also allows for more-cost-effective delivery of least-cost power. Specifically, accurate detection of faults and fault locations helps expedite the restoration of transmission lines, avoiding or reducing the redispatch of expensive generation.

REQUIREMENTS AND ARCHITECTURE

Many of the fault-location methods available today, including intelligent electronic devices (IEDs), provide some kind of fault-location capability and help relay engineers in post-fault analysis. Generally, however, these systems do not provide the data that system operators need to make real-time decisions. Therefore, Con Edison focused on developing a system that would enable system operators to make informed decisions and would lead to the restoration of power in a timely and safe manner.

Con Edison's transmission network has many line configurations and structures that are difficult to handle with the traditional fault-location methods, including tapped lines, underground cables with heavy distributed-charging capacitance, series-reactor-compensated underground cables and mutually coupled lines. In addition, the company's underground cables are compensated by big series reactors, which limit fault currents to within the circuit breakers' fault-interruption capacity. These series reactors can be switched in and out of the network. The switching capacity, while useful for operational purposes, further complicates the calculation process since the algorithms must handle the series reactors being in or out of service.

The series-reactor impedance is about 10 times higher than the cable impedance. Therefore, to develop the fault-location algorithms, it was necessary to ensure that the much-higher-lumped impedance at one end of the cable did not introduce extra errors or lessen the algorithms' sensitivity.

After taking into consideration the difficult features associated with the transmission network and the need to integrate the automated fault-location system into other enterprise-level applications, the following list of specified requirements was developed:

• Accuracy (±0.5% in error)

• Automated and close-to-real time (5 minutes)

• Covering underground cables

• Covering tapped lines (feeders)

• Covering series-compensated lines (feeders)

• The ability to distinguish between a lack of data and no-fault conditions

• Open architecture that could be easily integrated into other enterprise applications (for example, EMS/SCADA, other visualization applications)

• Web based

• Powerful database platform — Microsoft SQL server.

The overall architecture of the automated fault-location system is shown in Fig. 1. Its core components are Grid Sentinel's web-based automated fault-location system, WebFL, and Con Edison's Transmission Visualization System, which is a web-based real-time tool that monitors transmission feeders and associated equipment status. When a feeder trips out, the Transmission Visualization System requests WebFL to perform a fault-location calculation. The results are posted on the Transmission Visualization System within minutes.

Design considerations for this demonstration of automated fault location included:

• Occurrence of fault or disturbance on the transmission network produces data files

• Data files are recorded in the substation digital fault recorder (DFR) and delivered to the DFR master on the corporate Intranet

• WebFL server detects the new data files, initiates an automated analysis and stores the fault analysis results in its database for viewing by engineering

• Transmission Visualization System detects fault occurrence from the real-time SCADA database

• Transmission Visualization System retrieves fault-location results from the WebFL database for display to system operators.

Figure 2 shows the interactions of the different applications and databases in the system architecture. The figure also shows the data converted to information and how that information is used to make a decision. While Con Edison's demonstration system only used DFRs, data files from other IEDs, including digital relays, also can be incorporated into the system. The WebFL system allows data files from multiple sources — regardless of the sampling rates and file length — to be correlated and analyzed.

VALIDATION RESULTS

During the evaluation period of Con Edison's project on automated fault-location systems, fault cases and no-fault cases were used to validate the algorithms. Testing fault cases validates algorithm accuracy. Testing no-fault cases validates algorithm capability to correctly discriminate faulted lines from healthy ones. While accurate fault location is important, it is just as important not to have false positives, reporting a line as faulted when there is no fault.

During the evaluation, four faults occurred on the monitored network. Of the four cases, three were on 345-kV series-compensated underground cables and the fourth was on a 345-kV overhead line.

As seen in Table 1, the fault location, or distance to fault, is shown as a percentage of the total feeder length from one of the feeder terminals. “Calculated location” indicates the fault location predicted by the fault-location system, and “actual location” indicates the fault location found by repair crews. In the first case, the series reactors were not in service. In the second and third cases, the series reactors were in service. The distance to fault in the second case was measured from the substation where series reactors were installed. The distance to fault in the third case was measured from the substation where no series reactors were installed. For comparison purposes, traditional single-ended algorithms — as used in typical DFRs or digital relays — were also applied to analyze these faults. The results are shown in Table 2.

The two numbers in the calculated location column represent the calculated fault location based on either of the two data files from the two feeder terminals. Similarly, the two numbers in the error column represent the corresponding errors for the two possible calculated location results. In the second and third cases, the first errors were unusually large, making the calculated fault locations irrelevant. These large errors resulted from calculations that used the data files at the substations where the series reactors were installed.

A comparison of Tables 1 and 2 shows that the new algorithms offer significantly improved fault-location accuracy:

Case 1. Traditional method error 9.18% versus new algorithms error 0.01% → A difference of about 900 times.

Case 2. Traditional method error 23% versus new algorithms error 0.12% → A difference of about 200 times.

Case 3. Traditional method error 6.35% versus new algorithms error 0.28% → A difference of about 23 times.

Case 4. Traditional method error 2.0% versus new algorithms error 0.3% → A difference of about 7 times.

LESSONS LEARNED

Con Edison's experience in developing an automated fault-location system showed that significant benefits are achieved when such a system is available to operators as a decision-making tool. While this experience should encourage other utilities interested in such systems to move forward, there are three lessons learned that should be noted.

1. Data format

   Common Format for Transient Data Exchange (COMTRADE) is a standard developed by the Power System Relaying Committee of the IEEE that helps make effortless data exchange possible. Presently, however, not all DFRs and digital relays are COMTRADE-ready. COMTRADE-ready means that the data files recorded in DFRs or digital relays are in COMTRADE format so users do not have to manually convert files from a proprietary format to the standard format. COMTRADE-ready should be an important selection criteria when purchasing or upgrading DFRs or digital relays. Con Edison's demonstration showed that even when the fault data is in COMTRADE format, some files were difficult to interpret.

2. Naming conventions

   In a utility, various departments rely on many different software programs and databases. These programs may need the same kind of power-system data, such as substation data, line parameters and generator data. To improve integration, all software programs and important databases should use the same naming conventions. Otherwise, the initial setup and integration of new programs will cause unnecessary work.

3. Open architecture

   For a seamless integration, advanced applications such as automated fault-location systems must be scalable and built on open architecture. The web-based nature, in combination with standard database platforms such as the Microsoft SQL server, makes it easy to integrate the fault-location system into other enterprise-level applications. In addition, the adoption of XML web services also offers many advantages for sharing resources between different applications.

Con Edison's automated fault-location system research and development project showed that such systems are an effective decision-making tool for system operators. By improving the operator's knowledge of what is taking place on the transmission network, these systems can significantly enhance a utility's reliability.

---

John Vasco is a section manager in the Relay Protection Engineering Section at Consolidated Edison Company of New York. He has a BSEE degree from the City College of New York and is a graduate of the General Electric Power Systems Engineering Curriculum. He has been in the relay protection area at Con Edison for more than 30 years. vascoj@coned.com

Ravindranauth Ramlachan is an engineer in the Relay Protection Engineering Section at Consolidated Edison Company of New York. He has worked in the relay protection area for that past four years and holds a MSEE degree from Stevens Institute of Technology in Hoboken, New Jersey, U.S. ramlachanr@coned.com

Jade Wong is a project manager for the Research and Development department of Consolidated Edison of New York. For the last five years, she has focused on the underground transmission network. Her previous experience includes SCADA and control room, man-machine interface. She has a BSEE degree from the City College of New York and a MSEE degree from Columbia University. wongj@coned.com

Table 1. Fault-location results with advanced multiterminal algorithms.

Case number	Voltage	Feeder type	Fault type	Actual location	Calculated location	Error
1	345 kV	Cable	AG	32.28%	32.27%	0.01%
2	345 kV	Cable	BG	16.80%	16.92%	0.12%
3	345 kV	Cable	BG	19.35%	19.07%	0.28%
4	345 kV	Overhead	CA	0.80%	0.50%	0.30%

Table 2. Fault-location results with traditional single-ended algorithms.

Case number	Voltage	Feeder type	Fault type	Actual location	Algorithm A		Algorithm B
					Calculated location	Error	Calculated location	Error
1	345 kV	Cable	AG	32.28%	49.6%	23.1%	17.3%	9.2%
2	345 kV	Cable	BG	16.8%	-409.0%	39.8%	426.0%	23.0%
3	345 kV	Cable	BG	19.35%	355.0%	13.0%	336.0%	6.4%
4	345 kV	Overhead	CA	0.8%	-10.1%	28.0%	10.9%	2.0%