Bringing together sensing, communications and control for high-value utility benefits.
The smart grid will anticipate problems, and it will self-heal when those problems arise. It will lower the utility industry's carbon footprint by enabling the broad inclusion of solar and wind energy, reduce line loss, enable conservation voltage reduction (CVR) and volt/VAR optimization (VVO), and, generally, make the entire North American grid a model of efficiency, safety and reliability.
That is the promise, but the realization of this 21st century wonder grid depends on three vital elements:
Millions of sensors collecting critical information on the state of the grid and its individual components
Sensing the Future
State-of-the-art communications networks
System controllers and vital hardware, software and systems making the necessary adjustments to the grid (power generation and storage, power flows, switching).
Dr. Andrew Phillips is leading the charge at the Electric Power Research Institute (EPRI) to develop sensors and sensing technology that will provide the critical data to enable the smart grid.
Many new sensors are available for niche applications, Phillips says. “We've had to do multiple field trials, in multiple locations, because one swallow does not make a summer; that is, we need to test sensors in at least four places, in different environments and applications, to see what works where.”
Realizing all the benefits of a smart grid will require data from more sources than just transmission lines. High on that list are high-voltage power transformers, which are some of a utility's most expensive assets.
Moving the Data Smartly
According to Phillips, EPRI has been working with Sandia National Laboratories and the University of Colorado for more than nine years to develop a “killer app” on-line transformer tester. “This is a real game-changer,” Phillips says. “If a high-voltage transformer fails, it takes a long time to get a new one.” And, current multi-gas sensors can cost upward of US$30,000.
At the heart of EPRI's device is a sensor on a chip that “will cost a few dollars,” says Phillips. Accordingly, “The cost is no longer the sensor. It's the peripheral hardware, communications, installing it and processing the data from it.” The first field trial of this killer app is planned for the first half of 2012.
Once data from all types of sources — including phasor measurement units (which monitor system voltages), currents and frequencies, weather sensors, line and equipment monitoring sensors, and other types of intelligent devices — is collected, it needs to get to the right end users or end-use applications with the required levels of accuracy, security and speed. Right now, that is a challenge.
“We've got a lot of information coming in now, and we're asking for more,” says Art Chavez of Tri-State Energy. “We need a communications system to deliver that information, so we're going to fiber optics. But it takes time and money to build the infrastructure with the capacity to transfer that amount of information, and we have to put in a system to manage the traffic. We're making strides, but we have to be patient and maintain the old system while installing the new.”
Another solution to smart grid communications gaining traction is the intelligent, wireless field area network (FAN), which is really a combination and extension of the following:
The familiar local area network (LAN), which connects computers within the same building, for example, Ethernet (generally 100 Mbps to 1 Gbps)
A metropolitan area network (MAN), which spans several buildings in the same city or town
A wide area network (WAN), which generally connects several LANs and is not restricted to a geographical location, although it might be confined within the bounds of a state or country. The technology, which offers high-speed communication, also is expensive. The Internet is an example of a worldwide public WAN.
According to EPRI, FANs will provide critical infrastructure for the second wave of information and communication technology (ICT)-enabled utility systems. The first wave, which ran roughly from 2007 to 2010 in North America and Europe, saw the introduction of smart meters, advanced metering infrastructure (AMI), which uses proprietary narrowband wireless or existing power lines to communicate meter readings as well as voltage and outage information, and some distributed energy resources (DER) — but no real integration with distribution supervisory control and data acquisition (D-SCADA) and other utility operations.
The second wave, which really just began worldwide this year, has three elements:
Infrastructure that is extensible or designed with future growth in mind
The ability to support a high level of DER penetration
The high reliability necessary for machine-to-machine (for example, sensor-to-controller) applications and more hierarchical/distributed control, including fast fault location, recovery, and automated ectionalization and restoration (ASR).
Tim Godfrey, EPRI's senior project manager, says some FANs will work alongside existing utility communications systems, while others will replace existing systems. “Utilities first have to decide whether it's easier to fund these new systems as CAPEX [capital expenditures] or OPEX [operational expenditures].”
Public utility commissions often look more kindly on recovering capital costs related to the smart grid than, say, monthly fees for cellular service. Of course, adds Godfrey, “Some utilities don't want to deploy a FAN; they're happy with their local carrier's 3G or 4G network. But others say, ‘We need five-nines of reliability (99.999%)’ and may push to deploy their own FAN.”
Such a separate second-wave wireless broadband network could have multiple applications:
Transmission of AMI backhaul data and directly connected smart meters
Automated fault location, isolation, service recovery
Dynamic/granular load modeling and control
Dynamic/granular situational awareness
Integration of distribution and transmission operations
New smart grid applications, such as integration of DER and electric-vehicle charging
Integrated field operations and support, mobile data and voice over Internet.
However, as Godfrey notes, there are some hurdles to overcome before U.S. utilities will be able to realize the full potential of smart grid communications. One is the lack of a dedicated communications spectrum.
Indeed, two years ago, American Electric Power, one of the nation's largest utilities, told the Federal Communications Commission (FCC) a “dedicated licensed spectrum is sorely needed by utilities.” But because “the FCC hasn't set aside a spectrum for utility-use only,” says Godfrey, utilities are looking at a number of options for high-performance communications networks.
What to Do with All the Data
One concept is to employ frequency agility to cover a service area, switching between channels and using multiple unlicensed bands to mitigate interference. Several vendors offer WiFi technology to do this, allowing utilities to build a meshed network, a network in which each node not only captures and distributes its own data but also relays data from other nodes, thus propagating the network.
According to Godfrey, meshed networks will always have higher latency (information travel time) than non-meshed, point-to-point or point-to-multipoint networks because packets have to be forwarded multiple times. He says 4G WiMAX and long-term evolution networks typically can provide latencies on the order of tens of milliseconds, which is good, “but there are exceptions under some circumstances.”
Still, he notes, “It is possible to deploy an unlicensed mesh network with performance adequate for most of the second-wave applications, but the exact latency numbers will vary depending on many factors.”
In any case, Godfrey says, “It is safe to say that the utility industry is moving towards the use of Internet protocol (IP) but not necessarily the Internet.” The public Internet may have a place in utility networks for some applications, but the security and reliability aspects must be considered upfront. Because of that, “utility FAN architectures may be completely independent of the public Internet or provide access only through a firewalled gateway,” he says.
Trials at ESB Networks Ltd. (ESBN), the sole distribution system operator for the Republic of Ireland, revealed that neither data loop carriers nor RF mesh fully meet the utility's demand for reliability, security and cost efficiency. Neither does general packet radio service (GPRS), a network overlay to the existing cellular network that allows 2G, 3G and wideband code division multiple access (WCDMA) mobile networks to transmit IP packets to the Internet.
“We have a number of issues to address,” says Teresa Fallon, manager of smart networks at ESBN. “About 800,000 people, or over a third of our customers, are in dispersed areas and will need to be supplied via a wireless solution. For the remainder, it is hoped that data loop carriers will work, but the technology that was available for this when these trials started in 2008 was not what it is today and, in any case, gave poorer results than what was expected.”
The best solution, she says, “was GPRS, however, cost and longevity of this solution would make it less than preferable.” So, she adds, “We don't have a solution as yet, but we are using the experience of the trials to direct us towards the ultimate solution. We are hoping progress in innovation in PLC [power line communication] technology will work for us, and, in the wireless area, we are looking at a number of options that will be evaluated as part of the final design.”
With PLC, the power line becomes a communications network — and a link to the Internet — through the superposition of a low-energy, low-frequency (1-MHz to 30-MHz) information signal on the power wave (60 Hz in North America and 50 Hz in Europe).
The Irish Commission for Energy Regulation recently concluded, “Ideally, the Internet protocol should be used at the networking layer for all smart metering solutions and the well-understood and standardized security mechanisms adopted by other security-sensitive industries, such as banking, should be deployed.”
Critically, very few smart grid applications place the same bandwidth demands on utility FANs as, for example, video downloads put on mobile carriers' wireless networks.
“In general, messages from sensors range from a few hundred to a thousand bytes, but the problem comes when there are thousands of sensors taking readings every second,” says Godfrey. “To adequately support existing applications and provide a margin for growth, a FAN should ideally provide a megabit-per-second bandwidth at its edge.”
While a FAN may offer utilities adequate reliability and security in most circumstances, can it be relied on to restore power quickly in a crunch — a widespread power outage, for instance? Some people are not so sure. One approach to ensure higher reliability for IEEE 802.16 (WiMAX) networks is the Greater Reliability In Disrupted Metropolitan Area Network (GRIDMAN), which holds the promise of improving MAN and wireless FAN reliability and robustness by orders of magnitude while also offering “four nines” (99.99%) of reliability.
GRIDMAN, as conceived by industry communications experts, would offer utilities immunity to a single point of failure by allowing the following:
Base stations to act as relays if backhaul is down
Mobile stations to act as relays to help other mobiles communicate with a base station
Mobile stations to form ad hoc networks if all base stations are down
Mobile stations to function as base stations, with limited capabilities, in case of primary base station failure.
In addition to the smart grid, GRIDMAN would have applications in public safety and disaster relief as well as other government and infrastructure applications. Two IEEE task groups are working to modify the 802.16 standards to serve smart grid applications: 802.16n (GRIDMAN) and 802.16p (machine-to-machine enhancements). Both task groups are expected to issue draft standards late this year, with final approval slated for 2012.
Due to security concerns, work being done by utilities in this area is highly confidential. EPRI, for one, is developing a FAN demonstration project that will open for utility participation in 2012.
With hundreds, thousands or even millions of sensors and other devices collecting data on the condition and status of tomorrow's smart grid, and sending it over whatever communications network a given utility is using, the obvious question is, How will all of that data be handled?
“More information is not always better unless this information can be dissected and analyzed quickly, so we as system operators can respond to the dynamic behavior of the system,” notes Tri-State's Chavez. “This, to me, is the tricky part.”
“EPRI's John Simmins says, “In a lot of ways, our ability to collect data is way ahead of where we are with regard to processing it. There are a lot of stories about utilities data mining to get information on specific events. They are then able to go back to the operations guys and say, ‘We know on that circuit when we switch this or that, this problem results.’”
Using that data in real time is altogether different. Simmins notes that part of EPRI's smart grid demonstration project is looking at common platforms that can span applications, interactive graphics and touchscreens so the organization can eventually develop a common information model to resolve human-factors issues, not only for system operators but also for field personnel who perform switching, repair outages and other tasks.
For all this to happen, cautions Simmins, “We need to develop standards that vendors can use.” Without standards, vendors cannot cooperate, he notes.
Still, smart grid demonstrations are moving ahead, focusing on using information to automatically and instantaneously reconfigure systems, including volt, VAR and watt optimization, says Simmins. “This is not pie in the sky. We have actual feeders being controlled at American Electric Power, Sacramento Municipal Utility District, ESBN and other places.”
Bringing It All Together
Of course, without an advanced grid operating (or energy management) system, it will be impossible to realize the full promise of the smart grid. While today's systems allow real-time monitoring and control with reclosers, protective relays, substation controllers, phasor measurement units and the like, tomorrow's smart grid will demand a much more sophisticated — and complicated — infrastructure, which in turn increases the potential for failure and cyber attack. While a number of utilities are conducting trials of new grid operating systems, EPRI sees an advantage in collaboration and is proposing a new Grid Operating System 3.0 project to make it happen.
So, yes, there is indeed a lot of work yet to be done to make the North American power grid truly smart.
Lee Harrison (email@example.com) 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 BS degree in engineering from Northeastern University in Boston and an MS degree in journalism from Columbia University in New York City. He is a former writing instructor at Massachusetts College of Liberal Arts.
A Smarter Grid with Energy Storage
In addition to grid sensors, automated sectionalization and restoration, and a bulletproof communications network, a key element of the smart grid — especially where renewable energy is concerned — is energy storage. Along with the development of robust, cost-efficient batteries, the challenge has been to blend energy storage seamlessly into grid architecture.
“The grid has to better facilitate the connection and operation of new generating technologies such as wind and solar,” says Art Chavez of Tri-State Energy. “System operators are having to learn how to manage all of these dynamics cohesively, even though the operating characteristics of each differ so drastically.”
“To make energy storage work, it has to be smart,” says Dan Vogler, Ionex's CTO. “And it's in the middleware — between the DC battery and the AC grid — where we add smarts to create smart storage and solve some of utilities' biggest problems, like maintaining power quality.
“Only storage can transcend time and inject load or generation to maintain power quality on the smart grid,” continues Vogler. “Previously, all utilities could do was upgrade their substations, but smart energy storage can provide a more cost-effective solution.”
The simplest type of application, says Vogler, is where the storage system's charge and discharge rate is predetermined. “But our approach is different,” he says. “With additional sensors and algorithms, and current transformers on the inside and outside, we can segregate input and output. And, by knowing the current as well as the voltage, we know — and can meter — kilowatt-hours in and out. In fact, our system acts like a simultaneous consumption and production meter.” There are two elements to every transaction, he adds, monetary as well as physical, and Ionex software creates a record.
“Because the current transformers give us the ability to know exactly what the grid is doing, we can react to system events and respond dynamically to meet problems in real time,” says Vogler. “If grid voltage or frequency is up and down, we can respond dynamically, creating generation or adding load as required.”
The SMUD Grid is Getting Smarter
The Sacramento Municipal Utility District (SMUD) is employing three different smart grid strategies — CVR, VVO and ASR — on about 90 circuits. The utility expects to see about a 6-MW reduction in peak demand, along with increased system efficiency, improved power factor, a reduction in system losses and energy consumption, and decreased outage frequency and duration, according to Lora Anguay, senior project manager at SMUD.
The key to all this — the installation of smart meters — is expected to be completed by the end of the first quarter of 2012. This will tie together all aspects of electricity delivery and consumption with two-way wireless communications.
While most utilities deliver power at the high end of the allowable voltage range (114 V to 126 V) to ensure they are meeting voltage requirements, says Anguay, “Smart meters and other smart devices along the circuit will give us the instantaneous voltage, allowing us to see if we can reduce voltage at the substation.” For every percent of voltage reduction, she adds, “we estimate a 0.7% reduction in energy consumption for the average customer.”
According to papers filed with the California Energy Commission, the utility will implement, in the near future (one to two years), an advanced operating system in 15% of its service territory to put into operation its CVR, VVO and ASR strategies.
Today, Anguay says, “If there is a fault on a circuit, troubleshooters go out to investigate, then call their findings in to the operator, who writes a switch order that goes back to the troubleshooter.”
To automate this process and begin to realize the benefits of a smart grid, the utility will install the following equipment:
SCADA at 35 substations (over 50% of SMUD's distribution substation transformers already have SCADA)
Automated line sensors, reclosers and capacitors with two-way communication on 90 distribution circuits (15% of its total)
Automated motor operators and fault indicators with two-way communication on 12 subtransmission circuits (22% of its total).
A wireless communications system that will connect devices to substations (communications from substations to the control center is hard wired, according to Anguay).
American Electric Power www.aep.com
Electric Power Research Institute www.epri.com
ESB Networks Ltd. www.esb.ie
Sacramento Municipal Utility District www.smud.org
Tri-State Energy www.tristategt.org