An average of once every one hundred years the sun takes aim at earth and launches a ginormous coronal mass ejection(CME). Less than a day later, it arrives as a cloud of charged particles and hits the earth’s magnetic field. It has a southern polarity and, therefore, “couples” with the earth’s magnetosphere, creating swirling “electrojets” of charged particles 100km above the earth. These produce geomagnetically-induced currents (GIC) in the earth itself. These currents flow into the grounding mechanisms of large Extra High Voltage (EHV) transmission towers. The current then flows through the transmission lines and into the EHV transformers in the system. This quasi-DC current (in an AC system) produces “half cycle saturation” that overheats and permanently damages those $5-10 million boxcar-sized transformers. The current also produces harmonics that can damage or trick other components in the system, resulting in a collapse of the grid. The most important EHV transformers are the Generator Step Up (GSU) units located at nuclear or equally large coal generating plants. When these transformers fail, there is no way to get power from the plant to the grid. The icing on this cake is that it takes about a year to order, manufacture, and install a replacement EHV transformer (when the grid is up everywhere).
In 2008 a leading geomagnetic storm researcher developed a model to simulate the effects of another 100-year storm on the modern electric grid. The study, “Severe Space Weather Events; Understanding Societal and Economic Impacts” done by Metatech, an electric industry consultant (working for the Congressional EMP Commission and FEMA, not the Sierra Club), predicted that in a geomagnetic storm equivalent to the 1921 “100-year storm” approximately 365 EHV transformers would fail. The grid could not compensate for that many failures, leading to collapse east of a line from Chicago to Memphis to Jacksonville, FL and in the Pacific Northwest. The estimate for full recovery is four to ten years at a cost of trillions of dollars. The above is not the worst case scenario. The modeled storm is centered over southern Canada. If it is farther south, the predicted damage is over 600 EHV transformers. The model only examines transformers down to 345kV. Many 230kV transformers will fail, too. Also, storms larger than the 100-year storm have struck us in the past.
This vulnerability was first documented with the 1989 Quebec Hydro storm. A moderate-sized storm took Quebec and part of the Northeast grid offline in 92 seconds, put six million people in the dark, and immediately damaged two EHV transformers in Quebec and one in the U.S. The two Quebec transformers were not damaged directly by the GIC but by “the uncontrolled operation of circuit breakers in rapid succession” causing overloads as the grid collapsed. [Kappenman, Meta-R-319 p.2-12, 2010.] Also noteworthy is that 11 nuclear power plant GSU transformers needed to be replaced over the next two years, indicating that even if GIC doesn’t immediately kill a transformer, it can greatly shorten its life. The cost of this storm has been estimated at $360-645 million [Tsurutani, et al, Journal of Geophysical Research, 3July2003 online]. In the “Halloween Storms” of Oct-Nov 2003, a series of storms destroyed 14 EHV transformers in South Africa [NERC, HILF, 2009]. These storms produced lower GICs, but they lasted for several days. The transformers failed over a period of 10 months following the storms, so there was no massive blackout during the storm. Instead, there were brownouts and rolling blackouts as transformers failed. This storm was also important because previously these latitudes from the magnetic poles (equivalent to southern California and Florida) were thought to be safe from damaging GICs.
Another important event occurred in 2003. A high voltage line touched a tree and precipitated a collapse of the grid from Ohio to New York City. This put 50 million people in the dark. It was important, since it showed that even 14 years after the 1989 storm we could not prevent cascading collapses of the grid. The minimum cost estimate for the blackout is $6 billion [CENTRA, 2011]. Later, we will see that blackouts cost much more than hardening the grid to prevent grid damage and collapses.
There have been other more powerful storms, but they pre-date a modern electrical grid. The 1921 “Railroad Storm” is named for the impacts to railroad signaling and switching devices, as well as the trans-Atlantic cable and telegraph systems. It has been estimated to have been ten times the intensity of the 1989 Quebec Hydro storm [Meta R-322 p. 7-5]. It is considered the “100-year” geomagnetic storm and is the basis of the Metatech modeling from 2008 and 2010.
The largest recorded geomagnetic storm was the “Carrington event”. The name is from the astronomer who was actually looking (indirectly) at the sun in 1859 when the CME erupted. Nitrates are produced in the atmosphere above the poles by geomagnetic storms and settle to the polar ice. Measurements from ice core samples from 1561 to 1994 show that the 1859 storm was the most intense in that 433 year time span [McCracken, 2001].
EHV transformers are large, custom designed, and very expensive, so there are few spares. A representative of one, large, electrical provider estimated its number of spare EHV transformers “would be a single digit percentage” [comment of Mr. Heyeck of American Electric Power at the FERC Technical Staff Conference, April 30, 2012]. By 2009, almost no EHV transformers were made in the U.S. However, because 70% of our “fleet” of 2148 EHV transformers is at least 25 years old and 50% is at or beyond its 40-year design lifetime, demand has been increasing since 2002 [Kappenman, Meta-R-319, 2010]. This means that around 1074 new vulnerable transformers, an investment of over $5 billion, will be installed shortly. So, four new EHV/HV transformer plants have been constructed in the U.S. since 2010.
In 2010 the U.S. Department of Energy’s (DOE) Office of Electricity Delivery and Energy Reliability issued “Large Transformers and the U.S. Electric Grid” that stated there were six plants producing large transformers in the U.S. Those plants satisfied only 15% of domestic demand. The other 85% was imported. From 2007-2011 an average of 500 Large Power Transformers (LPTs) were imported each year.
Even with four new transformer plants, we still have limited production capability in the U.S. If 365 EHV transformers go down, as the modeling suggests, many will stay down a long time. Although the normal lead time for an EHV transformer is about 12 months, it can be 20 months in some cases [DOE, 2012]. Will they even be able to produce replacement transformers with large parts of the grid down? How long will it take under those conditions?
Ramping up production is seriously impacted by the raw materials for EHV transformers. Even when produced here, many of the materials come from overseas. Copper and “electrical steel” will become very sought after, and not just in the U.S. There were only 13 manufacturers in the world of electrical steel and only a handful of them capable of producing the high-permeability core steel used in LPT cores. Only one is in the U.S. [DOE, 2014]. If the storm affects the whole northern hemisphere, or even the world, know that even though China is a huge transformer producer, it still has to import transformers. Now try to envision the competition for imported transformers. One study suggested that long waits would ensue and “prioritizing delivery to customers would become a politically charged issue” [CENTRA, 2011, p 30]. Also, “If you don’t invest in [hardening] it’s hard to argue you should be first in line for replacement transformers.” [FEMA, Feb. 2010 workshop – Managing Critical Disasters in the Transatlantic Domain – the Case of the Geomagnetic Storm]
So, why has the grid become so vulnerable? First, nobody is responsible for the grid. The grid is really just a bunch of contracts and agreements between competing companies to move electricity among them. The system has three components– the generating plants, EHV and HV transmission lines, and the lower voltage lines that step down voltages and distribute the juice from the transmission lines to local customers. To move electricity most efficiently, transmission voltages are high to minimize resistance. The low resistance increases vulnerability to GICs. A 765kV line permits GICs ten times higher than a 115kV line [NERC, “HILF Event Risk to the North American Bulk Power System“, 2009]. The cost considerations also result in a preponderance of single phase and autotransformers, instead of the more durable 3-phase transformers. Finally, the desire to buy inexpensive power a long way away means there are more and more miles of EHV transmission lines. These lines are like antennae; the longer they are, the more current they collect. The age of the transformer fleet also increases its vulnerability.
In 2011 CENTRA Technology, Inc., on behalf of the Office of Risk Management, U.S. Department of Homeland Security, looked at likely consequences of power outages in 20 industries during the storm, one week later, and one month later. During the storm, there are widespread impacts due to the loss of power. Gas stations are unable to refuel vehicles, including freight haulers. Lack of power prevents people from getting their money or spending it. Dark traffic signals impede highway transportation. As backup generators come online, the impacts are reduced for services, such as hospitals, public water and sewer utilities, and emergency services.
The CENTRA report notes that, “…most continuity plans suffice for a period of days, not weeks”. After one week (or less) backup generators begin to run out of fuel. Nuclear power plants have backup power for “…up to 7 days, depending on location and circumstances” [Singh Matharu, NRC, in comments at the FERC Staff Technical Conference on GMD, April 30, 2012]. After that, how do they pump cooling water to the spent rod storage pools? The CENTRA report summarizes, “The concerns as time progresses after the storm grow from economic losses to major health and safety issues” (page 32). When I asked the director of a metropolitan utility how much fuel he had onsite for pumps for the water system, the answer was “about two days”. The answer was the same for chemicals for water and sewage treatment. Our economy is based more and more on “just in time” delivery.