The sprawling North American power grid resembles a large antenna, attracting electrical currents induced by giant solar storms. Severe space weather occurring during solar cycles has the potential to cause a large-scale blackout in North America.
by John G. Kappenman, Minnesota Power, Duluth, Minn.; Lawrence J. Zanetti, Johns Hopkins University, Applied Physics Laboratory, Laurel, Md.; and William A. Radasky, Metatech, Goleta, Calif.
Since the turn of the century, society has relied more heavily on electricity for meeting essential needs. Three large, interconnected grids that span the continent provide rapid response to the diverse energy demands of users in the United States and Canada. This unique energy service requires coordination of electrical supply, demand, and deliveryall occurring at the speed of light.
Disturbances caused by solar activity can disrupt these complex power grids. When the Earth's magnetic field captures ionized particles carried by the solar wind, geomagnetically induced currents (GIC) can flow through the power system, entering and exiting the many grounding points on a transmission network. GICs are produced when shocks resulting from sudden and severe magnetic storms subject portions of the Earth's surface to fluctuations in the planet's normally stable magnetic field. These fluctuations induce electric fields in the Earth that create potential differences in voltage between grounding pointswhich causes GICs to flow through transformers, power system lines, and grounding points. Only a few amps are needed to disrupt transformer operation, but over 100 amps have been measured in the grounding connections of transformers in affected areas.
Recent storms associated with Solar Cycle 22 (the 11-year sunspot cycle that began in 1986) have had an unprecedented impact on electric power systems. The great geomagnetic storm of March 13, 1989, plunged the entire Hydro Quebec system, which serves more than 6 million customers, into a GIC-triggered blackout. Most of Hydro Quebec's neighboring systems in the United States came close to experiencing the same sort of outage.
Less severe geomagnetic storm events in September 1989, March 1991, and October 1991 also hampered utility operations. GIC interactions with new technological devices such as large electric power controllers affected voltage regulation and caused undesired relay operations in the system equipment.
In contrast to today's more severe solar storm cycle, the preceding, relatively quiet 30-year period led designers of electrical systems to overlook the possible influences of GICs. Conventional threatssuch as high winds, ice loading, or lightningdid not cause the Hydro Quebec collapse. Rather, it was the consequence of a threat that had never been considered on a system-wide scale across the continental network.
Many portions of the North American power grid are vulnerable to geomagnetic storms. Much of the grid is located in northern latitudes, near the north magnetic pole and the auroral electrojet current and in regions of igneous rock, a geological formation with high electrical resistivity (see figure.) Systems in the upper latitudes of North America are at increased risk because auroral activity and its effects center on the magnetic poles, and the Earth's magnetic north pole is tilted toward North America
Power systems in areas of igneous rock (gray) are the most vulnerable to the effects of intense geomagnetic activity because the high resistance of the igneous rock encourages geomagnetically induced currents (GICs) to flow in the power transmission lines situated above the rock. Shown in cross-hatching are the auroral zone and the extremes that the aurora can reach during severe disturbances such as March 13, 1989.
The network depends on remote generation sources linked by long transmission lines to delivery points. The effects of GICs build cumulatively over a large geographic scale, overwhelming the capability of the system to regulate voltage and the protection margins of equipment. The Hydro Quebec outage resulted from the linked malfunction of more than 15 discrete protective-system operations. From the initial event to complete blackout, only one-and-a-half minutes elapsedhardly enough time to assess what was occurring, let alone intervene.
The Quebec outage did not, in this instance, cascade beyond the province's borders. But if the disturbance had occurred under higher load conditions (nearer to summer or winter peak demand conditions, for instance), cascading outages might have spread across a region of the northeastern United States extending to the Washington, D.C., area.
Extensive blackouts are the nightmare of the power industry. Once power is interrupted in large metropolitan areas, diversity of electric use on the network is lost. When power is restored, all thermostatically controlled electric loads come back on simultaneously. This stress, added to the higher demands of many devices such as motors and transformers, can draw up to 600% of normal load during restoration procedures.
Such a blackout is also likely to cause transient voltage stresses and permanent damage to network equipment such as high-voltage breakers, transformers, and generation plants, which makes them unavailable for restoring power. Hours or days may pass before power can be restored. Oak Ridge National Laboratory assessed the potential impact of a widespread blackout in the northeastern United States from a geomagnetic storm event slightly more severe than the March 1989 blackout as a $36 billion loss in gross domestic product. This figure does not account for the potential disruption of critical services such as transportation, fire protection, and public security. Other assessments placed the 1989 and 1991 geomagnetic storm effects in a category equivalent to Hurricane Hugo and the San Francisco earthquake in their relative impact on the reliability of the electric power grid.
Faced with operating power systems in Earth's uncontrolled natural laboratory, the power industry and the scientific community are working to develop a better understanding of the causes and effects of this phenomenon. The scale of the problem is immense: the physical processes entail vast volumes and uncertainties of the magnetosphere, the ionosphere, the Sun and the 93 million miles of interplanetary space to the Earth. A paucity of data limits our abilities to develop accurate models and understand the relationships between the power grid configuration, the underlying geological features, and their coupling to the fluctuations of the geomagnetic field.
The U.S. electric system includes over 6,000 generating units, more than 800,000 kilometers of bulk transmission lines, approximately 12,000 major substations, and innumerable lower-voltage distribution transformers. All can serve as potential GIC entry points from their respective ground connections. This enormous network is controlled regionally by more than 100 separate control centers that coordinate responsibilities jointly for the impacts upon real-time network operations.
Conductivities within the complex geological base underlying the network vary by 5 orders of magnitude. Power systems in areas of igneous rock are most vulnerable to the effects of intense geomagnetic activity.
Devices to block GIC flow have been investigated, but they are complex and expensive to install across a large area. With current technology, protecting the network from GICs would cost several billion dollars. Instead of making these costly investments, most utility companies rely on contingency strategies for weathering severe magnetic disturbances.
Choosing the best contingency procedures depends on being able to predict storm severity and how it affects the local system. With the poor reliability of current geomagnetic storm forecasts, utilities must either implement response measures prior to confirming a storm or wait for local confirmation. Since some operational changes require an hour or more, response may not be rapid enough after confirmation to prevent serious damage from the storm, however.
Electric system operators also may not know when to deactivate protective procedures because of the intermittent effects of geomagnetic storms. Severe activity may follow lulls in activity.
Real-time solar wind data transmitted from NASA's Advanced Composition Explorer in late 1997 will improve forecastsinitially providing a global, blanket warning. However, a nowcasting capability, which would provide regional assessment of the progressing storm, would boost protective strategies.
A geomagnetic storm produces large, multimillion amp auroral oval current systems centered over both of the Earth's magnetic poles. Utilities that are part of the Electric Power Research Institute (EPRI) Sunburst Project have collaborated with scientists at the Johns Hopkins Applied Physics Laboratory to track the location and structure of the auroral current in real time with low-Earth orbit satellites.
The monitoring obtained has been used to map the typical boundaries of the auroral zone and the extremes that it can reach during magnetic disturbances. In addition, such a monitoring system can detect the formation of auroral current structures over the North Atlantic several hours before North America rotates beneath them. Predictions of regional geomagnetic activity using this Earth rotation effect could potentially provide the equivalent of a hurricane tracking system.
A set of numerical models can accurately compute the electric fields induced in the Earth from the disturbed geomagnetic field and a map of the deep Earth conductivity. The induced current entering various points in a power network can then be determined from computed electric fields. The measured electric fields near Forbes, a major substation in northern Minnesota, and the measured induced currents flowing in the grounding points of transformers connected to two long transmission lines during a 30-minute period shows very good correlation.
Once the predictive models are validated and used routinely, utilities may be able to predict the potential impact of storms more accurately. Improvements may allow region-specific assessments of storm activity on a short-term basisallowing operators to make well-informed judgments about appropriate operational response measures.
The National Space Weather Program, sponsored by the National Science Foundation, integrates research, models, and algorithms of geomagnetic and particle event disturbances. Eventually, we may be able to predict space weather much as we do terrestrial weather. In the interim, however, existing models, as well as monitoring of solar wind, GICs, and auroral zone currents, will provide information for the upcoming solar maximum beginning at the turn of the century. Data from many scientific disciplines must be integrated and information distributed before the next solar maximumexpected in the year 2000 to 2002if we are to predict storm impacts and preserve reliable power system operation.
Source: Eos, January 29, 1997, p. 37.
John G. Kappenman is currently the head of Transmission Power Engineering at Minnesota Power. He is a graduate in electrical engineering from South Dakota State University.
He is a member of the American Geophysical Union and a senior member of the Institute of Electrical and Electronic Engineers. Kappenman has worked on lightning damage mitigation, the impact of geomagnetic disturbances on power systems, and the application of power electronics technology.
He is a recipient of the Walter Fee Outstanding Young Engineer Award and the Institute of Electrical and Electronics Engineers Prize Paper Award. Kappenman holds a patent for the invention of a static phase shifting transformer.
Back to Science for Everyone