More than ever, utilities are focusing on distribution system reliability, including outage management and the utility’s responsiveness during outage situations. As utilities plan for deployment of advanced metering infrastructure (AMI), outage management benefits often appear as key business drivers behind their investment.
Today, nearly all major AMI systems offer some level of outage management support. Still, not all AMI systems use the same technology, meaning system features differ. When evaluating an AMI system’s ability to support a utility’s outage management initiatives, there are a number of factors to consider.
Outage 101
First, AMI is an outage management tool, not to be confused with an outage management system (OMS). AMI supports the outage management system.
While the AMI system understands the AMI network layout, the OMS typically knows the distribution system model. This includes detailed information on distribution system devices, such as transformers and protective devices including fuses and reclosers. Both systems typically use the same source for device location specifics, the utility’s graphical information system (GIS).
When an outage occurs, the utility wants to quickly figure out what caused it and where to send crews to fix affected equipment so it can quickly restore power to customers. A fault on a line, a typical source of power outage, will cause protective devices ahead of the fault to operate, thus avoiding equipment damage from high fault currents. The OMS is mainly concerned with mapping outage notifications to the distribution network model to infer which protective devices have opened. With this knowledge, the utility can quickly determine the extent of the outage because it knows power downstream of these protective devices has been interrupted.
Figure 1 shows a typical radial distribution circuit and possible fault locations. Outlined below are the clearing mechanism and customer impact.
Fault at F1:
The fault is in the customer’s premise, and is cleared by an in-home circuit breaker or main fuse. Only one customer is affected.
Fault at F2:
The fault is on the distribution line between a fused transformer and several customers’ premises, and would be cleared by the transformer fuse.
Fault at F3:
The fault is on the distribution-system lateral, and is cleared by a fuse for that lateral line. A hundred or more customers may be affected.
Fault at F4:
The fault is on the distribution line, and would be cleared by a line recloser or station circuit breaker with reclosing relay. Several hundred customers may be affected.
Fault at F5:
The fault is on the transmission line, and would be cleared by a station circuit breaker. More than a thousand customers may be affected.
The value provided by the AMI network varies, depending on the location of the fault in the system. AMI networks offer particularly high value with isolated faults, such as those at F1 and F2. These smaller outages may occur when no one is home to report the outage. Many AMI systems offer endpoints with a “last gasp” transmission capability to tell the utility that the endpoints have lost power. This last-gasp transmission serves as a surrogate for the customer’s call, often allowing the problem to be fixed before the customer even becomes aware of the outage. AMI systems also work well in helping the OMS and dispatcher understand and efficiently respond to widespread outage conditions, such as those that would be seen with faults at F3, F4 or F5.
Prior to AMI, an OMS often couldn’t see the meter endpoint, the last device on the system before power hits the customer’s outlets. Most utilities could only see if power was flowing to devices installed at substations, which is where communication networks for most SCADA systems end.
Beyond that, dispatchers had to wait for customers to call in outage notifications in order to determine the extent of the outage and the location of the cause of the outage. Without AMI, as much as 90 percent of notifications in a large outage may come from call-ins. With AMI, the metering endpoints become a valuable, perhaps even prevalent, additional source of outage notifications feeding into the OMS.
Typically, metering endpoint last-gasp messages are transmitted over the same communications channels used for sending metering data to the utility. These channels include a local area network (LAN) — such as a mesh radio network — that carries information to a data collector, also known as a concentrator or gateway. The collector then links back to the AMI head-end system, which is integrated with the OMS via a wide area network (WAN) such as cellular. The last-gasp messages help the OMS to identify which section of the line is faulted and which protective device has operated.
Just as important to the utility as outage notifications, if not more so, is notification of power restoration, and the ability to verify power restoration to endpoints. The restoration notifications help dispatchers verify that customers in the area are back in service. That way, dispatchers can efficiently manage work crews without having to send them back to a restored area because an isolated outage was missed. Systems that support automatic acknowledgement of restoration messages before reporting a restoration are best, as they help avoid false reports when power returns but quickly goes out again.
AMI systems typically allow the utility to “ping” the meter, thereby verifying whether it’s energized. Not only does this allow dispatchers to poll meters and verify restoration strategically, it also eliminates false-outage truck rolls when the outage is really inside the customer premise. According to the late AMI consultant Ed Malemezian, some utilities say that as many as 40 percent of power-out customer calls are due to problems on the customer’s side of the meter. When utilities can ping the meter, they can quit wasting resources on unnecessary trouble calls.
RF-Based AMI Technologies
Recently, RF-based AMI technologies have gained significantly in popularity. The two dominant types of radio AMI networks are mesh networks and tower-based systems. Most field-proven systems use 900 MHz range frequencies for LAN communications. A variety of technologies are in use for communicating with head-end systems over the WAN. Radio signal penetration and propagation are important concerns for the LAN; communications flexibility and adaptability are key concerns for the WAN. In both mesh and tower systems, metering endpoints communicate with collectors and the collectors, in turn, communicate with the AMI head-end system.
Tower-based systems typically use data collectors mounted atop high towers or buildings establishing a direct point-to-point connection to endpoints. Sometimes these systems use supplementary repeaters when certain endpoints cannot “see” the tower. As some tower systems evolved from older one-way technology, they may, depending on the particular system, use endpoints that “bubble” up data, with special algorithms implemented for certain two-way functionality. Tower systems typically use higher power radios than mesh systems and so they often use licensed frequencies. This may result in collisions due to narrower communication bandwidths. As a result it may be difficult to get immediate notification from all reporting endpoints during large-scale outage and restoration conditions through to the head-end. Strategic pinging of meters to confirm power restoration can be done with a direct collector to endpoint communication.
Mesh systems use collectors mounted within the service territory, either in meter based forms or in standalone forms which may be mounted on poles or buildings. Each collector manages a network of endpoints that may have multiple levels of devices forming a mesh network below each collector. This network typically includes meters, repeaters and other devices. To get data through the network, some mesh AMI systems establish and maintain optimized communication paths to endpoints using periodic network polling algorithms. There are also systems that build out communication paths in real time as communications occur. Systems that build out communication paths in real time offer flexible communication paths, but the reliability and uncertain nature of the paths may not yield consistent results. Proactive systems that establish and maintain optimized communication paths are reliable, but the time to establish a new communication path may take longer than in an ad-hoc network. Some mesh networks provide the benefits of a proactive network for normal communications with the benefits of ad-hoc communications for outage reporting.
Outage and Restoration Notifications from Endpoints
A point of debate among utilities when considering an AMI system as an outage management tool is the number of endpoints that need to be heard. With AMI the amount of available devices to report outages is significantly increased from that previously available. Some utilities that adopted AMI early on found that having all devices report in during large scale outages could slow down the fault isolation process. Consequently, AMI system vendors began implementing “storm mode” where outage reporting was either turned off or limited when large outage events were expected to occur. Generally, the OMS wants to know when an outage has occurred for both small and large outages, but doesn’t require every home to respond during a large outage in order to isolate the fault.
Both tower and mesh networks can typically get a large percentage of outage notifications back to the head-end system. With tower-based systems, the density of endpoints on the tower needs to be managed so that endpoint notifications don’t create a bottleneck. With mesh systems, the ability to configure for multipath broadcast propagation may be important. In general, flexibility in how outage and report notifications are propagated to the head-end system is important as it allows the utility to limit outage notifications in large scale events without worry of false reports or losing the ability to accurately determine the extent of the outage. Due to the distributed architecture of mesh networks, they may inherently provide more configuration options for outage and restoration notification strategies than tower based systems.
When power to parts of the utility’s distribution network is restored, there may still be devices in the area that are out, so knowing what is restored is essential. With tower-based systems, since restoration is generally done incrementally across the distribution network, bandwidth overloading is not as large of a concern as during outages. With mesh networks, considering that restoration messages can usually be heard well beyond the normally used communication paths, all restoration messages are also typically able to be processed. This is particularly true if the mesh supports routing of the reports through different network branches and if collectors are battery backed up so they can process reports even when power at the collector itself is still out.
Other Considerations
Other considerations when looking at AMI for outage management support include the following.
Battery back-up:
Battery back-up options can be important in larger scale outages where collectors are more likely to lose power. Leading edge systems may even support supplemental power sources for battery charging, such as solar power, which can be particularly beneficial in extended outages.
Programming flexibility:
Utilities have different outage strategies. Some may want to be notified of every momentary outage. Others may not want to be notified of an outage where the fault clears itself, allowing the recloser to eliminate the outage automatically. Or, a utility might want to wait for several devices to report in to get a better idea of the extent of the outage before forwarding messages from data collectors to the head-end system. To meet these and other strategies, smart grid enabled AMI systems provide configuration options in endpoints and data collectors, allowing utility managers to determine exactly how the AMI system will manage outage and restoration information.
Reliability Indices support:
When choosing an AMI technology, it is important that the technology be able to differentiate between momentary and sustained outages and to filter out momentaries so the OMS does not receive unnecessary outage reports. However, the system must keep track of these outages to allow the utility to calculate important reliability indices such as SAIFI, SAIDI and MAIFI. Advanced AMI endpoints store the information required to calculate these indices in the meter, ensuring that the data is always available and not dependent on system recognition of each event. Reliability indices are useful in categorizing system data to meet the specific needs of the utility on a feeder or system level. If a particular feeder is having problems with frequent interruptions, SAIFI could be most important. SAIDI may be more important where continuity of power is a high priority. MAIFI, which indicates momentary interruptions, can help identify potential areas for proactive asset investment.
Smart Grid or Advanced Grid Infrastructure (AGI) support:
Leading edge AMI systems allow integration of smart grid AGI devices into the AMI network. These include devices such as distribution feeder monitors and line fault current indicators (FCIs). Integrating these devices into the AMI network allows valuable additional information to flow into the OMS. This can be especially important when isolating faults and responding to larger scale outage events. Adding FCI information to the outage management process can enable utilities to narrow down possible fault locations, thereby reducing overall fault investigation time. AGI device information can also lead to greater system reliability allowing the utility to avoid some outages altogether via proactive O&M planning. Some examples include early detection and mitigation of potential vegetation overgrowth, excessive transformer loading, high VAR flow and phase load balancing.
Conclusion
AMI is revolutionizing utility outage management and system reliability strategies. With a well developed AMI and OMS integration, utilities can significantly improve outage responsiveness and workforce utilization during outage conditions. Additionally, AMI systems allowing the use of smart grid AGI devices offer unique opportunities to utilities for improving system reliability and avoiding outages from equipment failure and poor asset utilization. The value realized by properly leveraging the capabilities of well designed AMI systems for outage management support and system reliability initiatives allows utilities to improve their customer service while realizing impressive returns on investment and moves utilities ever closer to realization of the smart grid of the future.
About the Authors
Raymond Kelley is the Vice President of Software Development responsible for the design and development of Elster Integrated Solutions’ fixed network and mobile AMI solutions. Raymond joined Elster (formerly ABB) in 1992 as a system architect and lead developer for ABB’s distribution and outage management systems. In 1996, he became the development manager for a group that designed and developed a large scale AMR Data Collection and Meter Data Management System. Raymond currently directs the development of Elster’s EnergyAxis® MAS System, Route Manager System, and Meridian System. Raymond has 20 years experience in system architecture, design, and development of large scale data collection and control systems including 6 years at AT&T Bell Labs. He holds a B.S. in Electrical Engineering from The Citadel, and an M.S. in Computer Engineering from Clemson University. Raymond is a member of IEEE, Tau Beta Pi and has published several articles and holds multiple US and foreign patents within the AMR/AMI domain.
Ron D. Pate is the Sr. Product Manager responsible for the electrical meter based components of Elster’s EnergyAxis AMI system. He is an electrical engineering honors graduate of North Carolina State University and has worked for Westinghouse, ABB, Ohio Transformer, Grand Eagle and Elster. He has worked with electric utilities extensively since joining Westinghouse in 1989 and has been involved in the engineering development and marketing of instrument transformers, power transformers, electricity meters, and metering systems.
