This is an article that I wrote for the January/February 1999 issue of Electric Perspectives magazine. Though the article is somewhat dated, physics surrounding solar weather and its impact on the earth and earth systems is just as valid today. I’ve changed having grown older since writing this piece, but the universe hasn’t changed measurably.
STORM WARNING – Come 2001, the electric power system, as well as telecommunications networks, will have to contend with the uncontrollable – invisible blasts of matter hurled at earth from a raging storm on the sun.
Storms of varying intensity occur on the sun's surface at any time, but they happen more frequently in the active season of an 11- ear cycle - the solar maximum - with the biggest storms erupting during or immediately after the peak of the cycle. More than ever, utilities are watching the sun closely as 2001 - the next solar maximum - approaches.
Solar storms and the weather conditions in space they cause are nothing new. Galileo first discovered variable "spots" on the sun when he first trained his telescope on it. For residents of the far north, the aurora borealis - the Northern Lights - is a common sight.
But with each solar maximum, the danger from solar storms increases. First, our technologies have advanced faster than our knowledge of how space weather might adversely affect them. Most modern communications systems, for example, depend heavily on satellites and their microchip circuit technology, which can be damaged by high-energy subatomic particles hurled from the sun. These systems today depend more on satellites, and so are at greater risk. Between the last solar maximum in 1989 and today, hundreds more satellites have been placed in space - and several have suffered costly equipment failure during solar storm activity.
Space weather effects don't stop at the outer fringes of the earth's atmosphere. It can have an impact on the North American electric power grid. The interconnection of the grid and the increase in long-distance, wholesale power delivery in the last 10 years or so have increased the grid's vulnerability. During the last solar maximum in 1989, geomagnetically induced currents (GICS) caused a power failure in the Hydro Quebec system - and basic transmission and transformer technology has changed little since then. According to a recent study by the Oak Ridge National Laboratory, a three-day power outage in New England alone would cost businesses six billion dollars.
Of course, many safeguards exist when it comes to preventing blackouts. But GICS have indirect effects, too, creating high harmonic flows and volatile VAR requirements that can inflict unanticipated and hard-to-measure damage to large transformers. A burned-out transformer can shut down your plant. A shut-down plant means lost wholesale revenue.
Still, utilities may not be looking at the sun as hard as they should be. Cost-cutting tends to pull funding from some GIC monitoring sites. Corporate memory (as well as staff) tends to wane over the course of Il years. And the problem, hard to define and assign probability to, tends to stay off radar scopes. After all, how can an event occurring 93 million miles away possibly shut down power on earth?
What's the Problem?
Welcome to the world of space physics. Most of the general space weather that affects earth starts at the sun, where hot fusing gases churn amid massive magnetic and electric fields. The extreme heat at the surface reaches more than 2 million degrees Fahrenheit, causing hot gases to explode continually into space. The explosions produce a steady solar wind composed of charged subatomic particles, electric and magnetic fields, dust, and—yes— cosmic rays.
This wind strikes earth's magnetic field (magnetosphere) at an average velocity of 2 million miles per hour—it is forceful enough to alter the magnetosphere's normal shape. Without the constant pressure of solar wind rushing against the earth, the magnetosphere would be round. The wind, however, presses in its sunward side and blows it back around the earth, creating a long tail out into space. Despite the deformity, the magnetosphere manages to provide a shield that prevents most of the wind's charged particles from entering the atmosphere. The small quantities of particles that invade enter at the magnetic poles and become embedded in the ionosphere. This outermost layer of the atmosphere is a blanket of charged particles, starting 25 miles up. Within the ionosphere are two moving bands of charged particles, called the auroral electrojet, which circulate around the polar regions.
The movement of the electrojet transfers magnetic energy to the earth's surface directly below -it's like a giant electric transformer. The magnetic field passing through the earth's surface induces an electric current in the ground, which travels along all the conductive matter it finds.
The level of GIC during normal space weather does not significantly affect the power grid. But conditions can change dramatically when stormy space weather develops.
Space weather storms originate when something snaps within the internal churning processes on the sun -one of the sun's huge tubular loops of magnetic field twists and snaps, and three types of space weather occur. First, solar rain - a flash of X-rays traveling at the speed of light - arrives at the earth within minutes. The rain is followed in one or two hours by spiraling high-energy particles—a solar tornado, which is hazardous mostly to satellites. Last is a massive burst of matter from the sun's plasma surface - a solar hurricane, or a coronal mass ejection. The hurricane poses the greatest threat to power systems: Hundreds of tons of plasma in a cloud millions of miles in diameter come streaming through space, though it takes two to three days to hit.
The blunt impact of the hurricane compresses the earth's magnetosphere on the sunward side even more than usual, extending the tail farther out on the opposite side. High-energy particles pour into the magnetosphere at both poles. For days after the cloud passes beyond earth, the magnetosphere wobbles back and forth until it regains its calm weather shape - imagine blowing on a large floating soap bubble and watching it shiver and shake until it regains its calm equilibrium in more stable air. The upset in the normal shape of the magnetosphere and the resulting oscillation is called a geomagnetic storm. The shifting magnetosphere distorts the size and pattern of the auroral electrojets, producing several hundred million amps of electrical current. A moderate solar storm in January 1997, for example, poured 1,400 gigawatts of electric power into the atmosphere over two days—almost double the electric power producing capability of the United States. The amount of electric current in the electrojet and the degree of pattern-shifting directly correspond to the level of current induced in the ground below.
Why We Care
"North America is most affected by solar storms," says John Kappenman, former head of transmission power engineering at Minnesota Power and now president of Metatech, a solar storm prediction service. "We are closest to the magnetic north pole, and our power grid extends into the high latitudes." According to Kappenman, power grids in Canada and much of the northern U.S. receive the brunt of a solar storm. First, they fall directly beneath the auroral electrojet. Second, they are on surfaces consisting of igneous rock, which resists the flow of electric current and makes these areas even more susceptible to space weather. The induced ground current cannot travel through the rock, so it naturally takes a path of least resistance - the shoreline of large bodies of water, railroad tracks, pipelines, communication cables, and power distribution systems.
Entering the power distribution system, GICS can cause massive power network failures and permanent damage to equipment in power generation plants and substations. The currents appear on the neutral connections of transformers located at opposite ends of long transmission lines. (Current magnitudes depend directly on the level of change in earth-surface potential - typically 2-10 volts per due to the electrojet-to-earth magnetic coupling, and indirectly on the direct current resistance of paralleled transformer windings and power lines. GICS have been measured exceeding 100 amperes on transformer neutrals.)
The frequency of these potentially high ground currents is on the order of a few millihertz - one full cycle taking roughly five minutes. The current magnitude and long rise-and-fall time drive transformers into saturation, causing them to overheat, reducing their efficiency, and perhaps leading them to catastrophic equipment failure. Moreover, a saturated transformer becomes a source of unwanted high-level harmonic frequencies and creates an increased draw of inductive voltampere reactive units (VARS).
The unwanted harmonic frequencies adversely affect the power distribution system in several ways: They
overload capacitor bands,
cause relays to trip,
produce over voltages on long-line energizations,
produce higher arc currents during single-pole switching,
produce higher circuit-breaker recovery voltages, and
overload harmonic filters on high-voltage direct current converter terminals, thereby sapping DC power in transmission.
Transformer-drawn inductive vars associated with major geomagnetic storms can cause excessive system voltage depression and unusually large swing in megawatt and VAR flows on transmission lines.
How great is the potential for equipment damage and service disruption? On March 13, 1989, during the last peak period of solar activity, an immense magnetic storm caused HydroQuebec's transformers to overheat, tripping protection relays across northeastern Canada. The Canadian power system (which Kappenman recognizes as one of the best and easiest to restore) couldn't hold up under the storm, and in only 90 seconds from storm onset, six million people were left in the dark for several days.
"That was a close call," says Kappenman. "Couple blackouts with severe terrestrial weather, and you get life-threatening situations." He believes the outage could have cascaded down through the northern United States as well, had the storm hit in midwinter or summer. The North American Electric Reliability Council logged more than 37 pages of power impact events during that storm. Compare this to two conventional natural disasters occurring the same year: 7 pages for the San Francisco earthquake and I page for Hurricane Hugo.
According to Kappenman, a powerful magnetic storm makes it difficult to handle outages on a localized basis, as one can with a conventional storm and even a natural disaster. "When a thunderstorm cell crosses a region, a few power lines may go down," says Kappenman, "but the problem is confined to a minor part of the whole system. When the earth's magnetic field goes into oscillation, large regions of the planet are affected at once." A failure, therefore, might not be due to one major system event, but be the result of many minor events happening across the grid at once.
But the minor things are not really minor, either. A transformer that is overheated, for example, may lose some of its operating life - unbeknownst to the utility. Replacement costs are high, especially if a breakdown occurs when you don't expect it.
In one fabled incident during the March 1989 storm, the windings in a step-up transformer at a nuclear plant burned out. No nuclear safety issues were involved, but the plant had to shut down, replace the 345-kilovolt transformer, and face the additional costs of starting up the plant again.
And in the years after the storm, Hydro-Quebec spent many hours in "storm mode." In 1991, the company spent more than 10 percent of the year that way, which also meant it had to back off its electricity transfers to give more room for VARS. In a robust wholesale market, this kind of action becomes a bottom-line issue.
"We have to work on improving our ability to recognize these events," says Kappenman. "You can help control the damage if you know they're coming."
In fact, there are several mitigation measures utilities take once they know a GIC event is underway. They can reduce key transformer loadings, allowing the transformers to operate at cooler temperatures. Or they can change equipment settings to free up more VARS. For the long term, they can put in place series capacitors on transmission lines and neutral blocking devices on transformers. But such total re-engineering of present systems is unrealistic. The second-best solution is to come up with methods to improve space weather forecasting and provide early warnings to the affected industries.
Finding Solutions
Within two years of the Hydro-Quebec outage, the Electric Power Research Institute (EPRI) funded the Sunburst Project, designed to gather hard data throughout the solar cycle on GICS and their effect on power company equipment. The program relied on power companies to voluntarily place monitoring boxes (at $40,000 apiece, which they had to pay for) on their primary power transformers. Space physicist Larry Zanetti, while he was at the Johns Hopkins University Applied Physics Laboratory, developed a mathematical approach that correlated ground current measurements with images of the auroral electrojet taken by a Swedish satellite. Kappenman and Zanetti believed that the data would provide insight into methods to predict GICS and provide early warnings to customers.
But right now, there are only 12 monitors in Canada and the United States combined. Several companies that originally were part of the program have dropped out. Part of the problem, according to Bill Feero, president of Electric Research and Management, which produces GIC monitors, was that there was a lot of activity right after the 1989 storm, but by the time monitors were in place (early 1991), there were no significant events to measure.
"The sun picks its own timeframe," says Feero. While the Il-year cycle is measurable (and has been measured since the 18th century), "you'll get four years of true activity, then moderate, then none - maybe five years of real activity altogether," he says. "Then it's five or six years of nothing." Sunburst has obtained significant data but has remained a small program with a low priority during the ensuing solar minimum years.
"The money is just not there anymore," responds George Anderson, a former senior power engineer with Potomac Electric Power Company, one of the first to participate in Sunburst. Anderson explains that though PEPCO is interested in the topic of magnetic storms (many of its key customers, like the Federal government and high-tech companies, require highly reliable power), even though it temporarily dropped out of the Sunburst program in the mid - 1990s. He explains that conditions are similar across the deregulating, cost-cutting industry.
Tom Molinsky, the supply-side enhancement engineer for Manitoba Hydro and the utility chairman of the Sunburst program, agrees. His funding, he says, is "pretty much guaranteed," because the utility is so far north and has a high probability of experiencing GIC activity - "lt is enough to keep management interested in it," he says. But for other utilities, he points out, there are issues like "deregulation and mergers and acquisitions. People ask, 'Is this solar stuff going to bother me?'"
And, in fact, the probability is that it won't. Molinsky points out that a large magnetic storm occurred last May, and the New England area experienced problems with voltage depression and a shortage of VARS. But not huge problems. "That's the general scenario in a GIC event," he says. "But you can blindly ride through it if you have enough surplus in your system, enough VARS. You might not even notice it. It's blind luck, sometimes, how close you can get to the edge of being very vulnerable."
The other issue is that in an 11-year cycle, some staffs change along with business direction, and the impetus for maintaining monitoring programs sags. "It's not an issue of the day for an executive," says Molinsky. "And a big GIC event is a rare thing." Molinsky is also a one-man GIC show, which is common in other utilities. It wouldn't happen at Manitoba Hydro, but "it's a very likely scenario that a program will collapse" if the staff person responsible for it went to a new job.
"It's a very difficult hurdle to overcome," Kappenman states. "The people in key positions keep moving on. New people have to be educated."
"It's hard to maintain something when there's nothing going on," says Feero. "And even in the high-probability period of solar activity, you can have low probability of occurrence.”
In addition, the release of monitoring information from a substation is theoretically competitive operational information, and some utilities are reluctant to parse it out.
How Do You Explain Probability?
Still, solar experts believe that at least one or two major solar storms equal to or greater than that of 1989 can be expected during each solar maximum. And with another peak solar activity period quickly approaching, interest in Sunburst has re-awakened.
The program recently launched the Sunburst 2000 system, a replacement for the data acquisition and communication equipment at the existing Sunburst remote sites (mostly substations). Rather than being research-oriented, the system is designed to provide real-time information, with faster computer processing and internet communication to disseminate information quickly. In the future, the system could use the information for the automatic control of power system devices to reduce susceptibility to GIC effects.
Also, according to Ben Damsky, EPR1's manager of power electronics systems and director of the Sunburst program, the internet capabilities could attract other utilities into the program, particularly foreign ones. Scandinavia, Russia, and countries around the southern pole also experience GIC effects.
In addition, the National Science Foundation formed the National Space Weather Program (NSWP), issuing its strategic plan in August 1995. NSWP is to provide space vigilance and real-time space weather reporting and forecasting. Now, as part of the program, space weather data are collected and reported by Space Weather Operations of the National Oceanic and Atmospheric Administration's (NOAA) Space Environment Center in Boulder, co, and the Air Force's 50th Weather Squadron at Falcon Air Force Base in Colorado Springs. Unfortunately, confidence in these reports is low in both the telecommunications and power industries.
According to Kappenman, forecasts of solar storms so far have only had a reliability rate of 30 percent. Even with advanced warning, the steps that the power industry must take to prepare for a possible magnetic storm are costly. The industry is reluctant to take the reports seriously after getting burned financially a few times by quickly reacting to unreliable forecasts. This lack of confidence is echoed by the telecommunications industry. But both industries stand to gain the most from accurate and timely space weather forecasting and have the most to lose should another big storm blindside them.
Protection of the nation's critical infrastructures has been an issue of primary importance to the Clinton administration. The President's Commission on Critical Infrastructure Protection in October 1997 ranked electric power distribution as the most important of eight other infrastructures, just ahead of telecommunications. The commission focus was on vulnerabilities to acts of terrorism and subversion, but it recognized the threat to the national power grid posed by space weather and recommended continued research and development to prevent solar-storm-related problems.
Last spring, the National Aeronautics and Space Administration took a step forward in space weather forecasting and reporting when it successfully positioned its Advanced Composition Explorer (ACE) satellite at a spot one million miles from earth and upstream in the solar wind. For the first time, space weather watchers, like JoAnn Joselyn, at NOAA's Space Environment Center, are getting real-time data. "The ACE satellite provides us with measurements of the solar wind that is about to strike earth," she says. "This knowledge is crucial to understanding the subsequent behavior of our magnetosphere."
Joselyn says the real-time data will dramatically improve "live" solar event reporting service to their customers but won't improve their long-term forecasting abilities. This is because the highly accurate ACE data can only describe the physical parameters of a solar plasma cloud "moments" before earth impact. These parameters constantly change from initial blast to arrival. According to Joselyn, ACE data may be used with existing modeling analysis to predict accurately the impact on satellites and monitoring instruments on the ground.
She adds that modeling does not yet exist to make accurate impact predictions from solar observations alone. When available in the future, such models would greatly enhance forecasting.
"The Space Environment Center is trying to keep up with new knowledge about space weather and apply that research to public needs as our vulnerability to space weather increases," Joselyn says. "We are learning from our brethren in the National Weather Service how to supply the basic products Metatech's Kappenman believes that sufficient data are now available to make power distribution system impact assessment modeling both a reality and an attractive "for hire" answer.
Whether a large solar storm in the next two years will have an effect on our infrastructures is anyone's guess. So is the extent of the effect. And some in the industry may be viewed as Chicken Littles, some as ostriches with their heads in the sand. For some, according to Feero, a perfectly good solution will be to "hunch their shoulders and duck." But the solar maximum is coming, and the sun may soon dramatically disclose the truth to all.