On 2 July and again on 10 August 1996 power outages affected millions of utility customers in several western states and adjacent areas of Canada and Mexico. With profound changes now sweeping the electric utility industry, the likelihood of such disturbances is expected to increase. Already, about half of the electricity generated in the United States is sold by the producer in the bulk power market before it reaches consumers—and this figure may rise to two-thirds with power industry deregulation. Indeed, some large utilities are now engaging in as many transactions in an hour as they used to conduct in a day. Such increased demand will further strain power delivery systems.

Fortunately, a variety of new technologies are becoming available that will help utilities maintain power system reliability while handling the larger volume of transactions. For example, power electronic systems can provide unprecedented control over electricity flow on transmission networks, preventing or containing the type of cascading disturbances seen recently. In addition, new sensor technology, faster communications between control centers, and advanced software tools can enable utilities to monitor system conditions in real time, letting them respond more quickly to disturbances and minimize their impact. Over the next decade, these technologies will facilitate increased power transfers through power delivery systems that are presently constrained, providing power at lower cost to a greater number of customers.

Until recently, the emphasis in power management was usually on adding individual devices to existing networks, to increase the capacity of specific lines and improve security at sensitive interfaces. Now, however, it is a question of how to optimize power flow through an entire network by integrating multiple technologies and coordinating control over wide areas. Such integration and coordination will be necessary for deregulation to proceed smoothly toward its twin goals: reducing electricity costs to U.S. consumers, while maintaining the high levels of reliability they have come to expect.

Flexible ac transmission system

A fundamental characteristic of ac transmission systems is difficulty in controlling power flow along specific "contract paths" through a network. Since the flow of electricity along each path is determined by the electrical characteristics of all the lines involved, "loop flows" may take power far away from the most direct route from a generator to a load center. For example, power from hydro plants in Ontario, Canada, destined for New York City may flow through transmission lines as far away as Ohio, threatening to interfere with the system operations of utilities that are not party to the intended transaction.

For a given transmission line, three key parameters determine power flow: terminal bus voltages, line impedance, and the relative phase angle between the sending and receiving end. To modify these parameters, a variety of mechanically switched devices are generally used, but none responds quickly enough to changing conditions to provide real-time flow control. Shunt-connected capacitors, for instance, help maintain voltage levels under heavy load conditions, while series-connected capacitors can lower impedance to increase power transfer on underutilized lines. Because of their slow response times, however, such compensators may actually degrade system stability after a disturbance. A phase-shifting transformer poses further problems because of the need for tap changing, which may take minutes to complete and—if operated frequently—involves high maintenance costs. These transformers are generally used to redirect power flow away from heavily loaded lines, with changes in phase angle made as seldom as possible to prevent too much wear on the tap changers.

Each of these conventional power flow controllers has a conceptual equivalent based on power electronics. In addition, with advanced thyristor technology, novel controllers that have no single conventional analog have been developed. These new systems furnish the foundation for the flexible ac transmission system (Facts), making it possible to redirect power in real time and provide virtually instantaneous response to transmission system disturbances. Most of the U.S. research required to develop Facts has been sponsored over the last two decades by the Electric Power Research Institute (EPRI), Palo Alto, Calif.

What is now considered the first Facts device—although that term had not yet been coined—was the static var compensator, which EPRI helped bring to market nearly 20 years ago. This compensator consisted of a fast thyristor switch controlling a shunt capacitor bank and/or a reactor, to provide auxiliary voltage support. It also contributed to system stability, though it was unable to control power flow directly. Conventional thyristors—silicon controlled rectifiers—formed the technological foundation for this device.

A later member of this first generation of Facts devices, the thyristor-controlled series capacitor, uses silicon controlled rectifiers to manage a capacitor bank connected in series with a line, enabling a utility to transfer more power farther on particular lines. Testing of the first, single-phase thyristor-controlled series capacitor was begun in 1991 by American Electric Power Co., based in Columbus, Ohio. In 1992, the Western Area Power Administration, based in Golden, Colo., installed a three-phase device so as to raise the capacity of a transmission line to 400 MW from 300 MW. The largest of these devices in the world, with a full range of features, such as power flow control and enhanced transient damping capability, has been operating since 1993 at Bonneville Power Administration, Portland, Ore.

Another first-generation Facts device, the thyristor-controlled phase angle regulator, has gone as far as detailed design, but never been constructed. The constraining issue is cost: replacing the mechanically switched steps of a phase-shifting transformer with silicon controlled rectifiers would add considerable capital expense to an already costly piece of equipment.

Next-generation Facts

The first demonstration of a second generation of Facts controllers began in November 1995 at the Tennessee Valley Authority (TVA). The systems installed by TVA can perform the voltage support and power control functions of first-generation controllers without the need for such large external circuit elements as a capacitor bank, shunt reactor, or phase-shifting transformer. By using an advanced configuration of gate turn-off thyristors, they can mimic reactors and capacitors electronically and so reduce the cost of Facts applications while substantially improving their performance.

The first full-scale static var (reactive volt ampere) compensator [Fig. 1] has been operating for more than a year at the Sullivan substation of the Tennessee Valley Authority, near Johnson City, Tenn. This second-generation Facts controller provides voltage support to a transmission line by generating or absorbing reactive power through an all-electronic shunt connection. It also can respond quickly to damp any big disturbance on the power system. By demonstrating this ±100-MVAr static var compensator, which cost US $10 million, the TVA has not had to build a 161-kV transmission line into the Johnson City area or a $20 million transformer bank.

A complementary second-generation Facts controller, the static synchronous series compensator, is in the design stage and expected to be selected for utility demonstration quite soon. This series-connected device could perform the functions of a thyristor-controlled series capacitor to increase or decrease the power flow along a specific line. It will probably be used in new installations, though with an existing thyristor-controlled series capacitor, the same task can be accomplished in less costly fashion if silicon controlled rectifiers are added to an existing capacitor bank.

Combining the static compensator and the synchronous series capacitor into a single device with a common control system represents the third-generation of Facts. The device is called the unified power flow controller. It will have the unique ability to simultaneously control all three parameters of power flow (voltage, line impedance, and phase angle). In this configuration, the series-capacitor component, connected in series with a line, injects an ac voltage with controllable magnitude and phase angle. The static-compensator component, connected as a shunt, supplies or absorbs the real power demanded by the series capacitor through a common dc link, and provides var control.

Both components can independently exchange internally generated reactive power with the line. The common control system uses continuous feedback to maintain a prescribed level of real and reactive power on a line, in response to instructions sent through a simple graphical interface.

The first utility demonstration of a unified power flow controller is being constructed at the Inez substation of American Electric Power. The Inez controller is being installed in two stages. The first section, a ±160-MVA static compensator, will be connected to existing lines at midyear. The second section is a ±160-MVA static synchronous series capacitor that will add power-flow control capability; it will be connected to a new 138-kV transmission line when both stages are completed by year-end. A spare shunt transformer will also be provided, which will allow both of the controller's components to be used for shunt compensation, if needed, up to a total of ±320 MVA.

The Inez area is the most heavily loaded portion of American Electric Power's transmission system, with power flows well above traditional surge impedance loading. The new controller will help the transfer of power into a coal-mining area in which loads have grown steadily. It will also provide voltage support to improve system reliability in the heavily industrialized tri-state area farther north, where the borders of Ohio, Kentucky, and West Virginia meet.

Hierarchical control of Facts

Applying Facts technology broadly will transform transmission systems. The transformation goes beyond simply changing the power flow on particular lines or easing the security constraints that create bottlenecks at critical network interfaces. The bulk power system itself will need to be reconceptualized as a more dynamic entity, with power flows fine-tuned for economic benefit on a network-wide basis and security issues addressed through operating practices that were inconceivable with mechanical controls. Achieving this transformation will be arduous. It will require fundamental changes in control strategies, development of real-time security assessment procedures, the addition of wide-area measurement capabilities, and greatly enhanced communications systems.

So far, installation of individual Facts controllers has had little impact on control center operations. But as multiple Facts systems are added to a transmission system, control activities will have to become more highly integrated and centrally coordinated to prevent unforeseen interactions. Coordination between neighboring control areas also must be tightened, so that operations in one area do not adversely affect those in another. Also, new strategies will be needed to accommodate the concurrent restructuring of the electric power industry itself, in which some previous control options—such as dispatch of a utility's own "must run" generation units—are less readily available.

Within every bulk power network, a basic hierarchical control framework already exists, with its various control functions distributed throughout the framework. Low-level, automatic control, for example, usually involves a preprogrammed equipment response based on local information (such as tripping a breaker in response to a line fault). This kind of action generally requires a swift reaction, without time for extensive data analysis. In contrast, control actions at the highest level must usually accommodate conditions over a wide portion of a network and regulate numerous lower-level decisions. Such activity involves more data, more analysis, more uncertainty—and more response time.

The addition of several Facts controllers to a system is likely to shift the ultimate burden of control toward the upper levels of this hierarchy. During steady-state conditions, for instance, making more frequent changes in the operating parameters of individual Facts devices may be desirable to take advantage of their multiple capabilities and to coordinate their combined effect on a transmission network. This enhanced parametric control can help optimize overall power flow, reduce the operation of mechanical equipment, and improve voltage management in response to changing load conditions. Centralized coordination is the only way to provide many of these benefits.

Higher-level control of multiple Facts devices will also be crucial during transient conditions. Unlike their mechanical counterparts, these devices can perform many operations per disturbance—and can provide continuous control throughout the disturbance—in response to local control signals or commands from a network control center. The result will be faster, better coordinated dynamic control over large power systems, enhancing stability and helping prevent cascading outages like those that recently swept the western United States.

Another impact that the spread of Facts will have on control strategies will be an improved definition of automatic generation and frequency control. Traditionally this has consisted of two major functions: economic generation dispatch (bringing generators on-line in an order that makes for least incremental cost), and network regulation (ensuring generator speed is adjusted to maintain adequate frequency and voltage support in response to changing load). Today's generators, though, may not be available for dispatch to control flows on the system. So, to optimize power flows in the future, the focus of automatic generation and frequency control may be shifted toward providing real-time network control by using Facts devices.

Minimizing generation and transmission costs, on the other hand, will more and more be left to market forces as deregulation progresses; that is, the dispatch of generators and the supply of ancillary services will be handled by contractual arrangements, rather than exercised through ownership control.

Toward on-line analysis

As power systems are operated closer to capacity, aided by the new generation of electronic controllers with subcycle response times, operations support will need analytical tools that use real-time information. These tools can enable dispatchers to schedule wholesale transfers on an hourly basis and help make best possible use of power system resources overall. Estimates show that they may additonally enable utilities to cut operating costs by up to 3 percent, saving billions of dollars each year for the industry as a whole. EPRI is currently developing an integrated set of such on-line security assessment tools.

One tool, the security enhancement system, deals with thermal limitations and is already commercially available. This computer program continuously monitors the condition of a power system and calculates how security would change in the wake of specific contingencies, such as the loss of a major transmission line. It can also recommend corrective actions for individual contingencies on a least-cost basis.

Another tool, the on-line voltage stability assessment program, will let utilities with voltage stability constraints operate their power systems at higher loads without risking voltage collapse. To be released commercially this year, it will be able to use real-time system conditions to calculate voltage stability limits dynamically. For large (regional) systems, the program will complete an assessment of voltage stability within 20 minutes, producing security indices for the operator as well as a list of contingencies that could lead to instability. It will also identify control measures to mitigate voltage problems.

A system allowing increased loading of constrained interfaces—those now being operated at reduced power because of the risk of instability—is called the dynamic security assessment program. The system also can be used to improve overall system reliability. Until recently, such real-time stability analysis was considered impossible, but advances in both computer hardware and computational techniques have now made it feasible.

The dynamic security assessment program will use techniques based on artificial intelligence to select contingencies relevant to a power grid in its present state and will then identify those contingencies most likely to destabilize the grid. To calculate stability limits, the program will be able to evaluate several hundred contingencies in less than 20 minutes. The program is scheduled to begin trials on-line this year. It is expected to allow timely assessment of the impact on system security of the increase in wheeling transactions that occurs as transmission grids are opened to competition.

Another important analytical tool can be used by utilities to comply with new regulatory requirements for assessing the available power transfer capability of their transmission systems. Called Trace (for transfer capability evaluation), the software package calculates what is the most power that can be transferred concurrently among two or more transmission control areas, subject to thermal, voltage, and interface limits. When an on-line version becomes commercially available this year, Trace will enable utilities to maximize energy transfers using near­real-time conditions, in support of a market for immediate transmission services. In addition, EPRI is developing an Independent System Operator Dispatch Model, which will help operators who are unaffiliated with a utility to dispatch, curtail, and determine the cost of transmission services efficiently, subject to reliability constraints.

Control center communications

The greater the complexity of network control, the more essential the communication of security-related data among control centers. An illustration of this need was evident in the Western system power outage of 10 August, 1996. Long before the point of no return was reached, two major power lines in the Pacific Northwest had been lost. Yet system operators down in California were unaware of the growing threat and kept on importing power from northern hydro plants heavily without increasing backup generation in their own territories.

Even when a third major line tripped, instantaneous notice of the event might have given operators a chance to bring small gas turbine generators (with 4-minute start-up times) on-line or to shift their power transfer patterns before the system finally began to break up, around 6 minutes later.

Ways to provide real-time data sharing among control centers automatically are now being implemented. EPRI has developed an internationally recognized standard—the inter­control center communication protocol (ICCP), or TASE.2, as it is known outside the United States—for data exchange between energy management systems at control centers. Links to power plants or substations can also be included, if desired, to integrate generation and network operations more closely.

ICCP is also designed to provide interoperability among the communications products of different energy management system vendors, a development that has already stimulated competition and lowered costs. A diverse mix of both hardware and software products is being developed to "plug and play" within the ICCP communications superstructure. Three major utility demonstrations of ICCP-based energy management systems are now under way: at Consolidated Edison Co. of New York, GPU Genco in Johnson, Pa., and Wisconsin Power and Light in Madison.

In addition, ICCP has been chosen as the technical basis for a new interregional security network being established by the North American Electric Reliability Council (NERC), Princeton, N.J. Each NERC region will have a communications node on the network, using an ICCP gateway protocol. Through this network, every region can have real-time access to security-related data from any other region, including information about disturbances as they are occurring.

One of the strengths of ICCP for this type of application is that it specifies the type of data to be exchanged automatically. The interregional network is scheduled for completion in 1998.

Wide-area measurement system

For real-time control and operation of power systems, better data as well as better communications among control centers will be needed. Data will be collected in significantly increased amounts and quality by monitors strategically located throughout a transmission network. Only recently has time-stamped phasor measurement technology with microsecond accuracy become available, capable of detecting angle swings and other critical system changes over a wide geographical area. The technology is currently being incorporated into a major collaborative project to set up a synchronized monitoring system for the western North American power system. Playing leading roles in this effort, which has been funded primarily by the U.S. Department of Energy and secondarily by EPRI, are the Bonneville and Western Area power administrations.

The phasor measurement unit at the heart of the new monitoring system is an advanced digital transducer that uses signals from a global positioning system (GPS) satellite to time-stamp simultaneous measurements of voltage and current magnitudes and phase angles at selected monitoring sites in a power system. Atomic clocks could also handle these tasks, but more expensively. Although each unit has local recording features, as a rule it will be operated in conjunction with a centralized monitoring system. When used this way, the unit is often paired with a portable power system monitor, which is an interactive measurements workstation with extensive networking capabilities, including real-time data sharing. The system monitor provides local intelligence for coordinating overall data recording, archiving, forwarding, display, and analysis.

Together, the phasor measurement unit and system-monitoring technologies will provide the backbone for the wide-area measurement system (WAMS) being implemented on the multi-state power system of the Western System Coordinating Council (WSCC). Participating in the effort, besides the Department of Energy, Bonneville Power Administration, Western Area Power Administration, and EPRI, are the U.S. Bureau of Reclamation and several national laboratories, notably the Pacific Northwest National Laboratory in Richland, Wash. WAMS is designed to supply the technology and infrastructure needed for real-time access to changing information emanating from a very large power grid, the goal being enhanced control, system efficiency, and capacity.

Already, more than a dozen phasor measurement units and portable power system monitors have been installed, and demonstration of system-wide data-processing capability is scheduled to begin later this year. During the two recent outages on the WSCC system, the installed units were able to record important data about frequency changes, line flows, and substation voltages. Although this information was not yet being shared in real time when the outages occurred, and thus was not available to help prevent the outages, it has permitted detailed and rapid analysis of what occurred [see "The role of real-time system monitoring in power grid control"].

Economies of precision

As the electric power industry undergoes its most profound transition in a century, many of the forces of change are focused on power delivery systems. Most of the regulatory and competitive forces involved are centrifugal in nature: breaking up the structure of traditional utilities and introducing new players into the business arena. In contrast, most of the technological forces now affecting transmission systems tend to be centripetal—opening up new areas for integration and hierarchical control. Economies of scale are also being replaced by economies of precision, as small generators become viable parts of distribution systems, and as power electronic devices create extraordinary opportunities for replacing numerous control centers with just a few centers that coordinate intelligent local control.

Many of the foundation technologies for integrated power delivery systems in the 21st century are already entering utility service. Others are progressing rapidly through various stages of research and development. These technologies can promise strategic competitive advantage to companies far-sighted enough to use them—lower costs, greater customer retention, enhanced use of assets, and improved planning and market management.

Spectrum editor: William Sweet