When lightning caused a pole-top fire in Flagstaff, Arizona, U.S., the community could have experienced an extended and costly outage. Instead, says Barbara Lockwood, director of energy innovation for Arizona Public Service (APS), robust self-isolating technology and a 900-MHz S&C Electric Co. radio communication system reduced the effect to a momentary flicker of lights, a mere blip. In fact, over several months in 2011, the utility's smart grid investment helped it to avoid some 600,000 customer outage minutes.
Similarly, when Hurricane Irene caused widespread outages on the East Coast in the U.S. last August, Pepco's smart grid telecommunications system proved resilient. The utility's advanced metering infrastructure (AMI) mesh network in Delaware allowed enough last-gasp messages to get through to help the utility not only predict the severity of the damage but also locate which pieces of equipment were most likely damaged, according to Karen Lefkowitz, Pepco's vice president of business transformation.
“Even a small percentage of last-gasp messages reaching our outage management system during large-scale outages greatly enhances our ability to send repair crews to the appropriate locations to begin restoration efforts,” Lefkowitz said. Although the utility received some 600 last-gasp messages from its smart meters, it was able to ping those meters and clear 582 of the outage events.
In contrast, other service areas pounded by Irene where the smart grid had not been fully embraced did not fare so well. News reports revealed customers often were shocked to learn their utility did not even know when they were out of power. Advantage: smart grid.
Reliability Up 50% at Pepco
Like utilities in other parts of the country, Pepco has had to adopt and adapt in dealing with the lack of spectrum availability for its smart grid communications system, a problem particularly acute in the Washington, D.C., metropolitan area.
According to Mike Kuberski, Pepco's manager of enterprise architecture, the utility went with a hybrid system for distribution automation (DA) that incorporates a Silver Spring Networks wireless field area network (FAN) operating in the unlicensed 902-MHz to 928-MHz spectrum. Backhaul is either through a public or private wireless system, depending on availability. In turn, that system connects to the utility's core network of multiple dedicated leased lines to bring data back to its control center.
According to Lefkowitz, the utility's DA work so far has led to a greater than 50% improvement in reliability. As such, the utility plans to upgrade 50 additional distribution substations by the end of 2013. It will replace analog relays with digital relays, add smart power meters, add distributed remote terminal units to expand visibility and control, and establish decentralized automation of field devices with an automatic sectionalizing and restoration control program at each substation.
Alabama Power had a similar experience. When storms struck on April 27, 2011, cutting power to more than 300 substations and destroying or significantly damaging six others, some 270,000 customers were left without power. Fortunately, the previous fall the utility had successfully merged its proprietary outage management system (OMS) with a Sensus FlexNet two-way wireless communications network to improve real-time situation awareness and grid stabilization, and to enhance outage estimation systems. As a result, the utility was able to restore most service two days earlier than it did during prior storms, including Hurricane Katrina.
Alabama Power had learned that, while its OMS could estimate where service was out, it could not tell system operators where service had been restored or, more importantly, what locations could actually take power. However, with the Sensus AMI system sending last-gasp signals over the FlexNet system back to the OMS, the utility not only could specify critical loads (hospitals, fire stations and traffic signals) for priority restoration, it also could tell when and where power was on without having to dispatch personnel, allowing it to prioritize restoration work.
The key was the robustness of the FlexNet telecom system, which continued to provide point-to-multipoint communication with nearly 1.5 million electric meters. Although it incorporates 150 antenna towers, each with battery backup and some with backup generators, the network remained largely intact throughout the storms.
What are the lessons learned? If both utilities and society at large are to realize these kinds of benefits from the smart grid — increased safety, reliability, efficiency and security along with a lower carbon footprint — utilities first need to invest in communications systems that also are efficient, safe, reliable and secure. But in North America, where geography and topography are as varied as population density, one size definitely does not fit all.
Most utilities already operate with a half dozen or so different legacy communications systems. So, it is no surprise that, in moving toward the smart grid, many utilities are developing and building advanced hybrid communications systems that employ any combination of telecommunications systems: broadband over power line, digital microwave, fiber optics, satellite, wireless technologies such as worldwide interoperability for microwave access (WiMAX), licensed and unlicensed radio, cellular and mesh networks based on unlicensed spectrum. Many of these utilities, like APS, Pepco and Alabama Power, are already seeing the benefits.
One other crucial lesson bears mentioning: Multiple application-specific, single-purpose networks are incompatible with the smart grid. In the past, utilities typically developed multiple communications systems in parallel with one another, said Michelle McLean, Silver Spring Networks' director of product marketing. “The metering guy didn't know the distribution automation guy,” for instance; and, as a result, utilities wound up with a whole set of parallel communications systems.
Silver Spring's idea, she said, was to develop a communications platform that could handle multiple communications needs — from operations to security — and accept multiple hardware and software applications.
No Silver Bullet
“You may decide to lead with DA, so buy a platform that does that well, but not only that,” said McLean. The multi-application platform should be built on the IPv6 Internet protocol, which offers considerable advantages over IPv4, the previous protocol. But, when choosing the type of network transport, cell, WiMAX or mesh, she said, utilities realize that sub-gigahertz spectrum is available, unlicensed, usually cheaper than cell and open for use with excellent propagation. As a result, McLean added, Silver Spring and numerous other vendors focus on mesh networks in the 915-MHz industrial, scientific and medical band.
Yet, mesh networks are no silver bullet. Inevitably, when utilities build out their networks, they run into pockets where mesh networks are not viable, making cellular communications the best choice. Still, McLean advised, “Mesh where you can; cell where you must.”
An example of Silver Spring's mesh architecture is in Pacific Gas and Electric's service area. Silver Spring built access points (red squares in the image) to get data from individual meters, or other end-point devices, back to the utility's data center.
“Essentially, it's the point where you either leave the utility's wireless FAN or NAN to join the utility's Ethernet link or connect with a cell carrier to get the data to the back office,” McLean said. “One access point can serve 5,000 to 10,000 meters.”
According to McLean, where PG&E runs into connectivity issues, for example, where topology or topography creates problems, Silver Spring provides relays (white circles in the image) to boost signal strength.
Barry Sullivan of the International Engineering Consortium (IEC) does not believe in silver bullet technologies, either. “Standards are being written to accommodate a variety of solutions,” he noted. “I don't see a narrowing down to a small number of technologies. The critical issue, of course, is making all technologies work together. They must support interoperability.”
Of course, none of this comes easy. Dean Siegrist, who leads Black & Veatch's utility telecom team, said simply, “Everyone wants an easy button, but there's no such thing in this industry. Strategic planning is key.” Still, he added, a hybrid, fully integrated system is most often the best solution. “Use a public network when available and a private network in other circumstances,” he advised. Still, he added, “The art is knowing how to integrate to get the best cost, performance and reliability.”
According to Tim Godfrey, EPRI senior project manager, when moving toward the smart grid, utilities basically fall into two broad categories. “One group has installed AMI and will build on that. They can do simple distribution automation with traditional capacitor banks, regulator adjustment, conservation adjustment, etc. They don't need the highest performance and, as a result, can use their existing networks.” Latencies, he said, generally range from one second to one minute. They require a large number of repeaters on the system, and designers must be aware of the system's limitation.
The second group, said Godfrey, already employs a FAN, which not only offers the higher performance required to include DA with AMI but also newer applications such as those necessary for the inclusion of distributed energy sources on their grids. And, he added, of course “some utilities are under mandates that require higher-performance networks.”
Going Hybrid Makes Sense
If a utility really wants to control its assets, especially if those assets are of strategic importance, said IEC's Sullivan, a hybrid network makes sense: “Build infrastructure to the substations and use a public network to reach each customer.” Still, he noted, because utilities are heavily regulated, the decision on which way to proceed can come down to capital expenditure versus operating expenditure funding. In that case, the discussion could depend on the utility's relationship with its regulator.
At Florida Power & Light (FPL), however, the driver to upgrade telecommunications was neither the smart grid nor direct mandates. It was simply reliability. After the utility experienced a number of hurricane-induced outages, it sought a more robust telecom system, and the key, according to Ron Critelli, director of transmission engineering and technical services at FPL, was redundancy.
Because it never wanted to lose communication with its substations, FPL began installing both landlines and private wireless communications systems to its substations in the late 1990s and early 2000s. “We saw a pretty significant gain in reliability,” Critelli noted. “Both systems are monitored, so we always know which is up and which is down.”
Operated by AT&T, the landline is preferred because it offers more bandwidth, Critelli said. But, because the utility seeks even more bandwidth for increased substation security reliability, including video, FPL is converting from AT&T's existing frame relay network technology to multiprotocol label switching technology. “AT&T forced our hand because it's sunsetting frame relays,” he said, “but we need more bandwidth in any case.”
The utility uses local cell carriers, such as AT&T, Verizon and T-Mobile, for wireless communications. “There have been a lot of improvements in cell since the hurricanes,” Critelli added. “Wireless coverage is getting better and better.”
Fiber in Vermont, Wireless at OG&E
In Vermont, where the service area is both small and rural by comparison, the state's transmission utility took a different approach. Vermont Electric Power Co. (VELCO) has announced an agreement with IBM to build an intelligent fiber-optic and carrier Ethernet communications and control network across the state to enhance reliability.
The network will span more than 1,000 miles (1,609 km) to connect transmission substations to Vermont's distribution utilities and provide “an innovative model for the rest of the country.” The communications system will relay information on usage, voltage, existing or potential outages, and equipment performance.
Like VELCO, Oklahoma Gas & Electric (OG&E) serves a largely rural area, but, unlike VELCO, it is considerably larger, roughly 30,000 sq miles (77,700 sq km), which made a wireless system for AMI and DA telecom almost a foregone conclusion.
AMI data for dynamic pricing is collected from the meters by a Silver Spring Networks radio frequency mesh network, which operates in the unlicensed 900-MHz spectrum. It connects to a private wireless Motorola WiMAX network operating at 3.6 GHz. In some cases where WiMAX coverage is limited, AMI backhaul service is leased from local cellular carriers, for example, AT&T or Sprint.
From the WiMAX layer, the data is transported by an Alcatel-Lucent microwave backbone, operating at 6 GHz, to OG&E's data center. This system also provides substation and DA connectivity for the utility. Alcatel-Lucent is handling the build-out of OG&E's wireless backhaul network, and the three-year project is scheduled for completion in fall 2012.
According to Scott Milanowski, director of grid intelligence at OG&E, after reviewing the total cost of ownership and risk assessment of a range of options, the utility decided ownership and operation of a private network was the best option to meet its needs going forward.
The project is complex, with an aggressive schedule and rigorous technical specifications. For example, OG&E set strict latency and bandwidth requirements for the automatic reclosers and capacitor controllers on the grid — the DA piece of the project — and those switches have to send information back and forth quickly enough to the distribution management system to deliver the reliability improvement and grid performance goals the utility envisioned.
Milanowski said the network not only will have enough bandwidth to support every necessary smart grid component OG&E plans to install, but it also will meet the utility's future growth needs. The network is a critical part of a complex system of integrated technologies that will allow OG&E to meet its smart grid program goals to engage customers, mitigate cost increases and reduce peak electricity demand, while increasing operational efficiency and reliability.
Strategic Planning is Key
No matter what combination of smart grid communications networks utilities have selected to adopt or adapt, the key to successful implementation is strategic planning. The planning should commence with four goals in mind:
All devices must be interoperable within an open standard environment.
The system must be extensible to protect the utility's investment.
Security must be built into the architecture.
There must be a plan to oversee and manage all those devices.
Think Inside the Boxes
Another concept utilities appear to be coalescing around is the choice of Internet protocol (IP) for end-to-end network layer technology, which not only offers the required levels of reliability, redundancy and availability but also supports legacy systems and applications.
Like Silver Spring's McLean, Mark Madden, Alcatel-Lucent's regional vice president for North American energy markets, supports an IP approach. “IP can carry legacy protocols through a variety of methods such as tunneling via MPLS, a proven technology that is being deployed broadly by the utility industry,” he said.
Still, for those utilities seeking to install wireless systems, frequency availability can be a challenge. Unlike Canada, which has allocated 1800 MHz to 1830 MHz to support its electric grid, the U.S. still has no overall policy to ensure adequate spectrum for smart grid applications. That said, with the military abandoning the 700-MHz spectrum, there may be hope. Officials in Michigan recently asked the Federal Communications Commission to grant a 700-MHz broadband waiver in the state to build a shared long-term evolution (LTE) network for first responders that also would support commercial utility users.
According to Alcatel-Lucent's Madden, LTE is an emerging technology with great promise, though it is not widely deployed in the U.S. because of spectrum difficulties. Still, he said, “LTE offers better spectral efficiency than WiMAX for both fixed and mobile service. It almost looks like MPLS quality.” Accordingly, he noted, “We expect LTE to overtake WiMAX.”
Technology could help to resolve the scramble for frequency, too, according to Doug Houseman, vice president of technical innovation at EnerNex Corp. Highly directional, phased array antennas that use small segments of the spectrum, like Trilliant's SkyPilot technology, could produce tailor-made solutions for the urban environment, he said.
Four Situations to Consider in Smart Grid Telecom Planning
But, he cautions, utilities still need a sound methodology to determine which type of telecom system to choose. First, said Houseman, they must determine which smart grid applications they want to install, the level of assurance of delivery they want to achieve and the latencies required to achieve it. Once that is done, he said, they should draw four boxes to evaluate their communications options and needs in certain situations.
For example, when a utility is in restoration mode after a major outage, “This is the hardest [situation] because we only have part system, and you need to know what's going on,” said Houseman. Too frequently, he said, “Nobody looks at the bandwidth requirements necessary to bring the system back.” It also is important for the utility to plan for firmware download (that is, when field devices, including security keys, need to be updated). “Instead of pulling data from field as usual, we need to send large messages out, but most utilities don't consider this need,” Houseman cautions.
If a utility goes through this process and considers a wireless network the best solution, Houseman noted, “It should get involved with the Utilities Telecom Council. If they don't,” he warned, “they may find their answer handed to them by a government agency” and have no choice in the matter.
Lee Harrison ([email protected]) has been writing about the power industry since 1978. He has been an editor for Business Week, a researcher with EPRI and a freelance writer, writing articles for The New York Times and the EPRI Journal. Harrison holds a bachelor's degree in engineering from Northeastern University and a master's degree in journalism from Columbia University. He is a former writing instructor at Massachusetts College of Liberal Arts.
How will the utility keep things going as they are right now?
When the system crashes, what will the real-time information needs be?
What bandwidth and equipment will be needed to bring the system back?
Instead of the utility pulling data from the field as usual, large messages need to be sent to its field devices. How will the utility do this?
Alabama Power | www.alabamapower.com
Alcatel-Lucent | www.alcatel-lucent.com
Black & Veatch | www.bv.com
EnerNex | www.enernex.com
EPRI | www.epri.com
Florida Power & Light | www.fpl.com
IBM | www.ibm.com
International Engineering Consortium | www.iec.org
Oklahoma Gas & Electric | www.oge.com
Pepco | www.pepco.com
Pacific Gas and Electric | www.pge.com
S&C Electric Co. | www.sandc.com
Sensus | www.sensus.com
Silver Spring | www.silverspringnet.com
Trilliant | www.trilliantinc.com
Utilities Telecom Council | www.utc.org
VELCO | www.velco.com