Coping with extreme solar storms

Solar flare emitted Dec. 19, 2014
Courtesy of NASA, Goddard Space Flight Center

The vulnerability of the world’s nascent communications infrastructure only began to be appreciated following Sept. 1, 1859. This is the date of the Carrington Event—a huge coronal mass ejection (CME) observed on that day by amateur astronomer Richard Carrington as a very bright solar flare.

Charged particles from the sun create geomagnetic disturbances (GMD) when they interact with the earth’s magnetic field. Rapid changes in the earth’s magnetic field strength induce an electric field of up to 15 or 20 V/km in the earth, causing destructive overvoltage on long-line electrical transmission and communications networks. In 1859, only telegraph lines existed, and many fires were reported in telegraph offices as well as the destruction of the newly laid intercontinental telegraph cable.¹

CMEs as large as the Carrington Event may happen once every 100 to 150 years, although significant CMEs also accompanied the 1921 Railroad Storm and the 1989 Hydro-Quebec Storm. Both of these events caused damage to the electrical grid, although even as late as 1921 the distribution of electrical power and our dependance upon it were not extensive. In contrast, “… the 1989 … storm blacked out the eastern half of Canada in 92 seconds, melted an EHV transformer at the Salem, NJ, nuclear power plant, and caused billions of dollars in economic losses.”¹

Geomagnetic storms are characterized by very long wavelengths. This is the reason that electrical and communications networks are damaged and that things with smaller dimensions, such as PCs and cars, are not.

Caution or paranoia

Because such a wide range of views is available on the Internet, sorting fact from theory can be difficult. On the one hand, there are reports of unfriendly contries optimizing nuclear bombs to have especially high gamma ray production. Detonating such a weapon a few hundred miles above the earth would produce high-altitude electromagnetic pulse (HEMP) and potentially could destroy electrical and communications systems for hundreds of miles in all directions. This is the scenario that supports “survivalist” websites—how to be one of the very few people who could exist with none of the facilities provided by modern civilization.

The electric grid is the most important of all our critical infrastructures. Without electricity, water pumps don’t work, factories don’t produce anything, refineries can’t run, people lose their jobs, banks close, and so on. Especially in cities, the situation would become very bleak very quickly with many websites predicting “the end of life as we know it.”

On the other hand, although the probability of a nuclear attack is not known, it is lower than that associated with the certainty of another Carrington-class CME at some time. Therefore, protection against the next large geomagnetic storm is the topic being addressed by a number of respected scientists. In fact, Congress established the Commission to Assess the Threat to the United States from Electromagnetic Pulse (EMP) Attack (the EMP Commission) to investigate and report on the EMP/GMD threat, which the commission did in 2008. Since then, House and Senate bills have been drafted to force utilitiy companies to protect the electric grid. None of the bills has progressed beyond the committee stage.

The plan so far

The severity of the GMD threat to the country’s electric grid is acknowledged by the U.S. Federal Energy Regulatory Commission (FERC). Among its duties, this independent agency regulates the interstate transmission of electricity. According to its website, FERC “protects the reliability of the high-voltage interstate transmission system through mandatory reliability standards.” It does this by requiring that the North American Electric Reliability Corporation (NERC)—an association of electric utility companies—produces a suitable standard.

As explained in a May 14, 2015, FERC press release, “… FERC directed NERC … to develop and submit new GMD standards through a two-stage process. FERC addressed the first stage in June 2014 by approving a standard on implementation of operating plans, procedures, and processes to mitigate effects of GMD.

“Today’s Notice of Proposed Rulemaking (NOPR) addresses the second stage and would adopt NERC’s proposed standard that sets requirements for transmission planners and owners to assess the vulnerability of their systems to a ‘benchmark GMD event,’ which NERC described as a ‘one-in-100-year’ event. If an entity does not meet certain performance requirements based on the assessments, it must develop a plan to achieve the requirements.”² A 60-day period followed this NOPR to allow comments from interested parties.

And there have been comments—many of them disputing the degree of severity NERC has associated with a “one-in-100-year” event. In Congressional testimony, Dr. George Baker, professor emeritus, James Madison University and principal staff member of the EMP Commission, cited key findings from the commission’s 2008 report. He said, “The operational procedure-based solutions that have been offered by NERC in their recently adopted EOP-010-01-1 standard are ineffective for a number of reasons. A nonexhaustive list of 10 pitfalls accompanying reliance on operational procedures to protect the electric power grid follows.”³

Dr. Baker listed pitfalls including reference to a 15- to 45-minute warning time that the Advanced Composition Explorer (ACE) satellite should provide when incoming CME is detected. ACE was launched in 1997 and is in an orbit about a million miles from earth. ACE is expected to remain operational through 2024. It is not at all clear what steps should be followed once a dangerous level of CME is spotted. Shed loads? Disconnect sections of the grid?

During his May 14, 2015, Congressional testimony on the EMP threat, Dr. Peter Pry also had many reservations. Dr. Pry is executive director of the EMP Task Force on National and Homeland Security and director of the United States Nuclear Strategy Forum, both Congressional advisory boards. In particular, he referred to a detailed review of NERC’s proposed GMD standard by John G. Kappenman and Dr. William A. Radasky.

In the review, the authors stated, “In this instance and others, a key feature of the NERC standard-setting process was to progressively water down requirements until the proposed standard obviously benefitted the ballot participants and therefore could pass. In the process, any remaining public benefit was diluted beyond perceptibility.”¹

This quote is particularly important because of the authors’ reputations: Dr. Radasky holds the Lord Kelvin Medal for setting standards and protecting European electronics from natural and nuclear EMP, and Kappenman, owner and principal consultant at Storm Analysis Consultants LLC and a proponent of the ACE satellite’s real-time capability, is recognized as an expert on space weather.

The organizational problem

Dr. Baker listed three reasons that little actual progress had been made in protecting the electric grid:

There are prevalent misconceptions about EMP and GMD threats and consequences.
Stakeholders are reluctant to act.
No single organization is the designated executive agent.

He addressed four of what he termed the most harmful misconceptions as well as NERC’s “… scientifically-flawed benchmark GMD threat description that enables most U.S. utilities to avert installing physical protection based on their own paper modeling studies. The benchmark GMD threat description is based on solar storm statistics over the last 25 years during which there were no ‘Carrington-class’ 100-year solar superstorms. The Carrington-class storm GMD levels are an order of magnitude higher than the largest storms in the NERC 25-year data window.”³

The technical problem

Large changes in the earth’s magnetic field caused by interaction with CME particles induce currents in the ground—ground-induced current (GIC). Field gradients of 300 nT/min can have large effects, and historical data shows that a rate of change as high as 5,000 nT/min is possible. Because of the very low frequencies involved, GIC can penetrate hundreds of miles deep. Modeling the earth’s conductivity in several different ways gave a range of induced fields from about 2 V/km to 20 V/km or higher when excited by a large simulated solar disturbance. This data agrees well with actual measurements. The GIC associated with this field enters power station transformers through the grounded neutral connection.⁴

Unfortunately, very high-voltage transmission systems are the most vulnerable. This is because they have lower resistance and run for longer distances. So, the accumulated voltage will be higher and the level of GIC current higher as well. Transformer design partly contributes to the severity of the problem. But, the simplest, most direct solution is to interrupt the neutral connection to ground by inserting a capacitor or small resistor. The capacitor or resistor blocks DC and very low-frequency GIC while allowing normally small AC neutral currents to pass to ground.⁵

GIC causes saturation and increased magnetizing current in a transformer, which in turn causes high levels of harmonics and internal heating. Severe storms with high levels of GIC have destroyed multimillion-dollar power station transformers. Reducing or eliminating GIC in the neutral lead is key to minimizing the additional MVARs the network sees.

In an interview with EE-Evaluation Engineering, Kappenman provided insight into the problem based on his 38 years of experience. While working at Metatech, a consulting engineering firm, Kappenman contributed to a series of extensively detailed reports for FERC and DOE on the effects of solar storms, both simulated and real, and proposed mitigation approaches.

During the last few years, the resilience of some parts of the electric grid has improved, but only coincidentally. Series capacitiors have been added to long transmission lines to improve their impedance—the capacitive reactance offsetting the line inductance. On a long high-voltage installation, this is not a simple undertaking.

In a paper describing the addition of series capacitors to 765-kV lines in South Africa, the authors listed characteristics of the capacitors as well as the operating conditions. An “Alpha” type capacitor was rated to carry 3,150 A, have a reactance of 15 Ω at 50 Hz, and withstand a load of 4,253 A for 30 minutes. Associated with each capacitor are various levels of protective devices. In particular, a fast-acting switch can bypass the capacitor in case of a high energy fault.⁶ To get an idea of the size and complexity of a series capacitor installation, two 400-kV capacitors are shown in Figure 1.

Figure 1. Two ABB 400-kV series capacitors installed at a Fingrid site in northern Finland
Courtesy of ABB

Series capacitors do block GIC, but only in the sections of the grid where the capacitors have been installed. Based on his Metatech simulations of series capacitors used by western state utilities, Kappenman estimated that GIC flow could be reduced by 15% to 22%. Given the level of GIC reduction actually required to cope with a storm similar to the 1989 Hydro-Quebec event, this amount of reduction probably is insufficient.

Kappenman simulated the effect of a capacitor or resistor in a transformer’s neutral lead and discussed the results in a separate Metatech report.⁵ As he clarified in the interview, it’s much cheaper to install a capacitor in the neutral line than it is to add series capacitors to the high-voltage transmission lines. Typically, the neutral blocking capacitor only has to be rated for 15 kV and a few hundred amps. Nevertheless, the capacitor still needs to be bypassed in the event of a fault, which could be as large as 20,000 A.

Because of his interest in a fast-acting bypass switch, Kappenman became associated with Advanced Fusion Systems (AFS) of Newtown, CT, a company developing extremely high-power electron tubes. Such a tube could be used as a bypass switch, and this is one of the applications discussed in a paper jointly written by Kappenman and AFS CTO and president Curtis Birnbach.⁷ Kappenman said that there is at least one other company interested in neutral blocking devices, and that company was using a spark gap as the bypass switch. The neutral blocking circuit suggested by Kappenman is shown in Figure 2.

Figure 2. Kappenman GIC neutral blocking circuit
Courtesy of Storm Analysis Consultants

The situation today

In view of the lack of action from Washington, some states, such as Maine and Texas, are actively taking protective steps.⁸ However, only a very few laws actually have been passed. Kappenman knew of only one utility that had installed a physical neutral blocking device and that test program began this year.

References