By David Kreiss, Kreiss Johnson Technologies
Editor’s note: This is the second article in a two-part series on non-operational substation data. The first article, “Non-operational Data: The Untapped Value of Substation Automation,” appeared in the September/October 2003 issue of Utility Automation.
In the battle to accomplish more with less, many utilities are turning to non-operational data such as fault records, voltage disturbance data, equipment temperature readings and lightning strike information. These and other data sets are routinely captured in the substation by digital fault recorders, protective relay devices, dissolved gas monitors and temperature sensors, but historically, they haven’t been put to good use.
In the 1980s, supervisory control and data acquisition (SCADA) systems became universally adopted by utilities to help them manage their T&D systems. Today, utilities are seeing the value of a non-operational data acquisition (NODA) system to capture and analyze non-operational data that cannot be collected, stored and analyzed through SCADA.
“New technology is making non-operational data much easier and faster to use than ever before,” said Bob Burnett, president of Fault Data Enterprises Inc. of Orchard Hill, Ga. “The key now is for utilities to learn to trust it.”
Burnett points out that while standard automated systems tell the utility what happened, the non-operational data explains why things happened. Utilities that harness this valuable information can restore outages faster, operate equipment more efficiently, and make more accurate decisions as to when infrastructure should be repaired, upgraded or replaced.
As a developer of automated utility technology, Kreiss Johnson Technologies (KJT) conducted a series of interviews with utilities this year to gauge the industry’s level of non-operational data use. The results show that many utilities recognize non-operational data’s potential value. These utilities have identified specific objectives, or “value events,” they wish to achieve through the use of non-operational data to produce information that will allow the utility to take a specific action that will save money.
The following is a summary of a select group of value events being pursued by North American utilities.
Georgia Power Company, Atlanta, Ga.
Value Event: Analyze fault characteristics to determine if lightning was an outage cause.
Georgia Power is one utility that has already incorporated non-operational data use into operational procedures. When a transmission line trips, the Georgia Power control center requests Mitch Cowan, or another transmission specialist, to dial up a digital fault recorder (DFR) and download the fault data. Using analysis software from Utility Systems Inc. (the DFR vendor), the transmission specialist graphs the fault and isolates its inception. He then accesses a lightning database supplied by Vaisala-GAI Inc. to correlate the fault with lightning strikes. If lightning was the cause, the line can be safely restored remotely and immediately. This saves on the cost and time required—up to six hours—for a crew to visually inspect the line.
In cases where the trip occurred in a transformer, Cowan examines the fault current on the low and high sides. If the current is coming out of the transformer in both directions, he knows the fault could be caused by a damaged lightning arrestor, in which case the transformer will not be restored until the storm has cleared the area. Again, this negates the need for a visit by a field crew while also allowing the utility to avoid destroying a multi-million-dollar transformer by exposing it to another lightning strike.
Outlook: Cowan reports that manual retrieval of DFR data by dial-up connection and the subsequent analysis generally takes from 30 minutes to two hours or more. Georgia Power and KJT are engaged in a pilot with the goal of automating this process so DFR data is continuously accessed—most likely by high-speed telecommunications connection—and cross-referenced with lightning strikes. Waveforms would be analyzed by computer to determine the fault type, with the resulting fault report delivered within minutes by e-mail or website to the control center.
“This could potentially reduce restoration times, in some cases, from several hours to just a few minutes,” Cowan said.
BC Hydro, Vancouver, British Columbia, Canada
Value Event: Recognize underground cable degradation due to insulation breakdown before circuit failure.
Failures can occur in underground distribution cables when exposure to ground water degrades the cable insulation. According to Ralph Barone in BC Hydro’s protection and control maintenance department, the ingress of water into voids in the cable insulation or cable splice can cause a fault which typically occurs at a voltage peak. As the fault current passes through, the water turns to steam and extinguishes the fault. The result is a fault that starts at a voltage peak and then falls to zero current within a quarter cycle, clearing itself without any protection operation. After several
of these, however, the fault doesn’t clear because the insulation has been damaged so much that the circuit fails.
Outlook: Barone says that Cooper Power Systems has integrated an algorithm into its equipment to recognize this unique fault signature. A challenge is that the device must sample quickly enough to identify the short-duration event. To make this information useful to the utility, an automated method must be developed to aggregate the fault data coming from numerous circuits so maintenance personnel can track the number of self-clearing faults each circuit experiences. Maintenance personnel can then take a proactive approach to replacing cables or splices before the circuit fails and customers lose power.
“All of this data and information can’t just come spilling into the control center,” Barone said. “Their duty is to operate the power system, and data shouldn’t be sent there just because it is convenient to do so. The information should be directed by e-mail, cell phone or pager to the person who should act upon it.”
Sierra Pacific Power Company, Reno, Nevada
Value Event: Accurately identify fault location for control center operators.
Sierra Pacific has installed protective relay devices throughout its service territory. However, manually interrogating these devices can be a time-consuming and frustrating process due to the dial-up phone links connecting substations with the control center.
Non-operational data provided by substation equipment monitors can be useful in making decisions about equipment repair or replacement.Click here to enlarge image
Although fault type and location information is available from the relay devices, operators do not always take advantage of it. A recurring and potentially costly problem has been operators basing assumptions on incomplete information. For example, when a storm is in the area, an operator might conclude a line trip was caused by lightning. There is a potential for an operator to remotely re-energize a line under that assumption, while a quick assessment of fault type and location might indicate that the fault is actually in an underground line section, therefore not weather-related and thus likely to cause more damage if re-energized.
Such location knowledge, if obtained quickly, can make the operator’s decision-making much more straightforward and less risky to the system and its equipment. In other situations, an operator might send a crew several hundred miles to inspect the wrong line. By simply knowing where a fault occurred, and gaining that information quickly, the operator can better diagnose the problem, thus avoiding costly damage or unnecessary crew time.
Outlook: Sierra Pacific participated in a pilot this year with KJT in which automated fault reporting software was linked to protective relay devices. Despite problems with noise on the phone lines, the pilot successfully retrieved the data and analyzed it to determine fault location, type and magnitude of current. According to Travis Johnson, supervisor of substation engineering at Sierra Pacific, this is a great start, but a high-speed communications network is needed to make it practical. Otherwise, Johnson said the next crucial step is to deliver analysis results to control center operators within 30 seconds so they can rapidly access the information and make the correct restoration decisions.
“Based on the cost of damage to a substation that can occur from a fault clearing mistake, this non-operational data application could pay for itself by avoiding one mistake,” Johnson said.
San Diego Gas and Electric, San Diego, Calif.
Value Event: Employ condition-based maintenance for optimal equipment replacement.
As utilities migrate from a time-based to a condition-based approach to substation equipment maintenance, non-operational data from circuit breakers and transformers—such as insulation values, operating temperatures, and wear and stress indicators—could also be useful for capital replacement and addition of equipment. Use of this non-operational data and other equipment performance data can guide the utility in optimizing the use of existing equipment and in choosing the best-performing equipment for future application. This will also help the utility to optimize the timing and location of equipment replacement and addition.
According to Patrick Lee, San Diego Gas & Electric’s (SDG&E’s) director of system protection, the utility is investigating the use of temperature monitors to determine the proper temperature relationship of the transformer and tap changer tanks. By creating a database cross-referenced by utilization history, load profile, ambient temperature, equipment make and model, etc., SDG&E expects to make better decisions on when to replace an existing transformer or add new transformers. Lee is also exploring the use of circuit breaker wear data in protective relays to correlate usage and operational stress with other performance data.
SDG&E is also planning to install gas pressure/density monitors on its critical circuit breakers that are insulated with sulfur hexafluoride. By correlating gas pressure with day/night temperature changes, the utility can identify gas leakage rate and determine corrective actions.
Outlook: Temperature and pressure monitors are commercially available. The challenge will be gathering the incoming data, integrating it with other related data, analyzing it and coming up with decision criteria. Most importantly, personnel responsible for maintenance, new equipment procurement and project planning must share the results.
“At $1 million for a power transformer replacement and up to $150,000 for a new circuit breaker, gathering equipment utilization and performance data will prove to be cost-effective for decision support—since transformers and circuit breakers are the most costly maintenance items in a substation,” Lee said. “By monitoring non-operational data, we can create a feedback loop between maintenance and capital addition personnel to improve the efficiency on equipment repair and purchase.”
Identifying Value Events
The interviews revealed that many utilities are already aware of the value of non-operational data. For utilities not tapping these resources, the excuses run the gamut of access, retrieval and analysis issues—all of which are slowly being overcome by new technology.
If a single conclusion can be drawn from the interviews, it is that utilities should not look at the data and ask, “What are we going to do with it?” Rather, they should identify value events by pinpointing significant, recurring problems in their operations and determining if non-operational data might be a solution.
The following procedures for leveraging this data emerged from the interviews:
- Identify value events that positively impact the bottom line, e.g., remotely restoring transformers and lines, replacing equipment before it fails, locating a fault in a grid, etc.
- Quantify each value event in time/dollars saved.
- Determine what information is required to take action that results in the value event.
- Determine what data is required to generate that information.
- Decide which departments and individuals must receive the information to act upon it.
- Determine when and how the information will be delivered.
Click here to enlarge image
As highlighted in the utility interviews, the technology is already available to accomplish most of these steps, and NODA systems are currently being deployed. In addition, the technology needed to access, correlate and report derived information is also available. As a next step beyond merely correlating non-operational data, KJT is working with several utilities in pilots to fine tune the final and most important piece of the non-operational data puzzle—applying artificial intelligence and fuzzy logic to automate the integration and analysis of data so that useable information is generated and delivered to decision makers.
David Kreiss is president of Kreiss Johnson Technologies, a developer of substation data acquisition and analysis software for the utility industry based in San Diego, Calif. The company specializes in web-based open architecture power analysis systems that employ artificial intelligence components for non-operational data analysis. David is a member of many industry standard setting committees including the IEEE1159, 1205, and the Standards Coordinating Committee 22. He is co-author of the Dranetz Power Quality Handbook and a regular author, speaker and panelist at industry conferences. Kreiss can be reached at firstname.lastname@example.org or 858 535-2088.