This post is for those who enjoy getting into the weeds. It's not a casual read, but it is quite significant. It was contributed to the blog by Bill Harris of the Foundation for Resilient Societies Board of Directors, whom I met at the Space Weather Workshop I've been blogging about since April.
The gist of Mr. Harris' piece is that the power companies are manipulating the North American Electric Reliability Corporation (NERC) in order to produce regulations that protect the power companies from liability without creating robust protection for the grid and the American people. Mr. Harris specifically mentions Frank Koza, Chair of the Standard Drafting Team at NERC and Executive Director of Infrastructure Planning at PJM Interconnection. Mr. Koza's response can be found on my blog here.
Mr. Harris is an international lawyer specializing in arms control, nuclear non-proliferation, energy policy, and continuity of government. He worked on Hot Line upgrades, creation of linked Nuclear Risk Reduction Centers (1982-84), and was a co-drafter of arms limitation treaties in 1986-87, 1991, and 1993. Mr. Harris worked for the RAND Corporation and in a variety of assignments for the U.S. Government. Mr. Harris holds a B.A. from Harvard College and a J.D. from Harvard Law School. Mr. Harris serves as Secretary and attorney for Resilient Societies.
DOES THE PUBLIC INTEREST REQUIRE BETTER SCIENCE-BASED STANDARDS TO PROTECT THE ELECTRIC GRID FROM SOLAR STORMS?
There are several elements of concern that the North American Electric Reliability Corporation (NERC) is low-balling the threat of a severe solar geomagnetic disturbance to the bulk electric system:
It appears that the U.S. regulators of the bulk electric system, the Federal Energy Regulatory Commission (known as “FERC”), made a political deal to coax the electric utility industry into committing to develop a “reliability standard” for severe space weather. The industry was acutely aware that a solar storm could destroy critical equipment due to experience with the March 1989 storm that in 92 seconds damaged reactive power equipment and high voltage transformers. So then-FERC Chairman Cheryl LaFleur (via FERC Order No. 779 in May 2013) allowed the industry standard-proposing entity, NERC, to set its own “standard” for the 1 in 100 year solar storm.
Not surprisingly, NERC picked a less severe event, the March 1989 solar storm, as the 1 in 100 year geomagnetic disturbance event.
Why would NERC “low ball” the threat? If the industry prefers no standard (there was none between 1989 and year 2015), or a low standard (under development since May 2013), the result can be a shield against financial liability for an extended electric blackout.
What is the problem? In the last 100 years, there have been far higher magnitude solar storms that affected electrical systems: the May 13, 1921 New York Central Railroad Storm, which took out the train switching system south of 125th street in Manhattan, and caused a destructive fire at a 57th Street station; and another storm on March 25, 1940 (Easter Sunday), that caused electrical damage in North Dakota, Winnipeg, Canada, and Bangor, Maine.
Underestimating the magnitude and mis-locating the geomagnetic peak (more southerly latitudes are likely to be the epicenter of higher magnitude geomagnetic fields during the largest geomagnetic storms impacting earth) can result in failure to protect high voltage transformers and especially vulnerable generators at mid-latitudes and along the Atlantic, Pacific, Florida, and the Gulf of Mexico coastal regions.
NERC’s GMD Task Force and especially its Standards Drafting Team sought out parameters that lowered the baseline 100 year Event and its modeling components.
In a year 2015 NERC appeal to NERC’s Director of Standards and then to a subcommittee of the NERC Board of Trustees, Resilient Societies asked NERC to revise their analytically faulty GMD Event Standard and underlying Event Model. NERC refused to do that. Three of the key failures of the NERC solar storm model include:
1. Selecting data inputs and modeling components for the Benchmark Event from a more benign part of the Northern hemisphere: Scandinavia and the Baltic States in Northern Europe, rather than data and model variables derived from empirical records from solar storm experience in the North American electric grid. The Standard Drafting Team of NERC, chaired by Frank Koza of PJM Interconnection, opted to use data from Northern Europe (Scandinavia and the Baltic states) rather than data from North America in developing their solar threat Benchmark Event.
Why does this matter? First, the “coastal effect” of solar storms varies by the resistivity of sub-oceanic and land mass geology where electrojets carried in high salinity oceans cross into high resistivity land masses. Second, the more southerly waters between Finland-Sweden and the Baltic States tend to dissipate geoelectric currents during solar storms. Applying this artificial reduction in geoelectric fields from Northern Europe to the land masses of North America leads to false assurances than transformers at lower latitudes in the United States are immune to damage from severe solar storms. NERC’s selection of the Northern European data to be substituted for the North American data for a standard to be applied to North America, is unscientific and perhaps reckless. By cherry-picking a more benign environment for a standard to be applied to another continent which has lesser southerly attenuation of geoelectric fields and a more severe “coastal effect.”
2. Selecting a linear variable, the so-called alpha factor, to model a claimed “decrease” in the effective geoelectric fields that result from geomagnetic latitude; when a study by Los Alamos National Laboratory (Rivera and Backhaus, 2015) use North American data to develop a power function model. The NERC alpha factor within the GMD Benchmark Model has more rapidly declining geoelectric fields when measuring at more southerly latitudes: they (falsely) assume that the more southerly locations will experience a rapid fall off of geoelectric fields, hence that critical equipment like transformers and generators will survive without need of any protective equipment. In contrast, the Los Alamos model estimates that the March 1989 solar storm peaked over New York state, not Canada; and that a higher magnitude solar storm would be likely to cause peak geoelectric fields further south, then attenuate more slowly when measuring at more southern locations. NERC declined to accept the Los Alamos correction to changing geoelectric fields as a function of geomagnetic latitude. Overall, the NERC Benchmark model does not even require “assessments” for hardware protection at sites such as Wiscasset, Maine, or Salem, New Jersey, where high voltage transformers have suffered damage or total loss in solar storms well less than a true 1-in-100 year solar event.
3. NERC's refusal to accept that there is a significant “coastal effect” upon the vulnerability of critical electrical equipment to solar geomagnetic storms. NERC’s Benchmark Model cannot explain why there are high insurance claims (Lloyd’s database, 2013; Zurich Re database, 2014 and 2015) for electric equipment damage in North America within coastal counties. Resilient Societies has noted that in just moderate solar storms, transformers sited on salt marshes or otherwise close to the Atlantic Ocean have suffered damage during merely moderate solar events. NERC’s Geomagnetic Disturbance Task Force refused to collect and openly shared data for North America, or to model a “coastal effect.”
The Seabrook, New Hampshire case study.
As an example, Resilient Societies submitted to NERC a report on Seabrook Station in January 2012. NERC failed to respond on the merits or add a “coastal effect” to the NERC GMD Benchmark Model. In January 2012 Bill Harris of Resilient Societies compared databases of transformer outages and solar storms impacting earth, then analyzed a matching event: what happened at Seabrook (NH) Station, a nuclear power plant, in November 1998? There was a North to South solar storm impacting earth on Nov. 8, 1998; then a rapid reversal of the electrojets , and a South-to-North solar storm on November 9-10, 1998. This was a far weaker solar storm that the March 1989 Quebec storm, yet transformer damage resulted. Electrojets from the high-saline Atlantic Ocean, when encountering high-resistivity granite, will enter the steel framework of the power plant via its pile-driven steel embedded in the granite subsurface at Seabrook Station. On November 10th, 1998, a 4 inch stainless steel bolt vibrated loose inside the Seabrook 345kV transformer. (NERC, unlike Chinese scientists, declines to model vibrational impacts on critical equipment). The stainless steel bolt relocated into the low voltage windings of the main transformer, overheated and caused the low voltage windings of the Phase A Seabrook transformer to melt. FLIR imaging showed the damage, and the plant was shut down for 12.2 days on Nov. 10, 1998, with a spare for the Phase A transformer now utilized. Seabrook engineers in January 2012, reviewing the records from year 1998, said the Nov 1998 outage appeared not to be the result of a solar storm, because the geomagnetic currents would enter the high voltage end of the transformer, yet the damage was to windings at the low voltage end of the (345kV/24kV) transformer. Bill Harris checked with a national transformer expert, John Kappenman, who explained that the transformer that melted at the Salem-1 nuclear power plant on March 13, 1989, also had the GIC [geomagnetic induced currents] travel from the high voltage end to the low voltage windings, and photos showed the melting at the low voltage winding end. Later, Bill Harris explained to Seabrook engineers why this was a solar GMD event, with photos from Salem-1 melting the low voltage windings of the main transformer. NextEra Seabrook, after some delay, now claimed the damage in November 1998 was caused by a manufacturing defect in the 4 inch stainless steel bolt. Bill Harris responded that the proximate cause was the solar storm, not the defect in manufacture: that same stainless steel bolt had stayed in place for about 3000 days of transformer operation at full power; it only failed during a “sudden impulse” reversal of the electrojets, and resulting harmonic-induced vibrations, at a coastal salt marsh during a solar storm.
Experts on coastal amplification of solar storms estimate at minimum a doubling of geomagnetic currents near salt-water coasts (J. Gilbert, 1972), or a factor of between 4X and 7X (Boteler, Natural Resources Canada, 2013, 2015). Footnote: When NextEra Seabrook ordered a replacement transformer from Siemens in Linz, Austria in 2013, the replacement transformer shipped to coastal New Hampshire and installed in October 2015 utilized synthetic oils for extra cooling protection against overheating (in solar storms or otherwise). But NextEra declined to purchase a neutral ground blocker, at a cost of merely $400,000. So geomagnetic induced currents can enter the high voltage transmission network, and require extra reactive power, cause line sags, and cause transmission line overheating and harmonic-induced damage to grid equipment in Maine, New Hampshire and Massachusetts. Without a reliability standard requiring elimination of geomagnetic induced currents, Seabrook Station under pressure from the parent NextEra holding company, kept costs down and failed to protect against solar storm damage to the regional grid.
But with no acceptance of a “coastal effect” combined with a falsely modeled rapid-decline in geomagnetic induced currents (“GICs) at more southern latitudes, even the State of Maine appears to “require” no protective equipment, such as neutral ground blockers for large power transformers. More southerly utilities are generally free to postpone hardware protection from severe solar storms; yet these same electric utilities will receive liability shielding if FERC adopts the weak NERC standards now under FERC review.
Then why did the large power transformer at Maine Yankee, sited along the south Maine coast in Wiscasset, experience damage to its Generator Step-Up (GSU) transformer on March 13, 1989? Managers at Maine Yankee did not announce the damage publicly, but within 2 weeks of the March 13, 1989 solar storm, Maine Yankee managers asked permission from the Nuclear Regulatory Commission to install a 2nd parallel GSU transformer. One that could take over if their main transformer malfunctioned. In a small solar storm in April 1991, the operating GSU transformer at Maine Yankee, earlier damaged in the March 1989 solar storm, caught fire and with hydrogen gases emitted, exploded. So the backup transformer that NRC had approved, took over. But this coastal site was vulnerable to solar storm damage.
Quietly, Maine Yankee replaced both GSU transformers by year 1993. They did not announce the vulnerability of coastal-sited transformers; but when they challenged a Town of Wiscasset tax increase due to the new transformer installations, the official Maine Supreme Court Reporter showed that 4 GSU transformers had been deployed at this vulnerable site, when only one was needed to provide a generator step-up transformer for Maine Yankee.
At the Broomfield, Colorado Space Weather Conference on April 27, 2016, Frank Koza of PJM Interconnection, Chair of the Standard Drafting Team at NERC, implied that very few large power transformers would be lost or severely damaged in a severe solar storm. But his own PowerPoint illustrations showed high geoelectric currents near ocean-coastal sites and also on the eastern coast of Lake Michigan. When questioned by a Board Member from the Foundation for Resilient Societies, Mr. Koza now welcomed future modeling of “coastal effects” upon critical equipment. But this comes after NERC has refused to include a “coastal” element in the GMD standard and has asked FERC to approve a hardware assessment standard that does not include any “coastal effect” upon critical grid equipment.
The public interest requires actual protection of the electric grid, not liability protection combined with defective grid modeling.
Bill Harris' original post ended here. Upon further reflection, Mr. Harris offered the following. As his phrasing reads better than what I could do, I elected to leave these comments in Mr. Harris' own words:
What I did not address, and what you might wish to complement, as your own addition to the Bill Harris post, is that a solar storm benchmark model that discourages purchase of protective equipment to “operate through” solar storms has adverse effects upon deterrence against man-made EMP attack and for defense of the U.S. electric grid against EMP attack if an EMP weapon is utilized by a foreign adversary.
If the U.S. adopts an assessment standard to mitigate solar geomagnetic storms, and that standard prevents cost-recovery for equipment that would protect transformers at mid-latitude, even high latitude in much of the U.S., then the same utilities that decline to protect transformers against solar storms will also lobby against protection from man-made EMP attack. The NERC benchmark model requires assessment only if a transformer would experience 75 amps per phase during the GMD Event. This will limit assessable transformers to a few dozen at most. If FERC does not allow cost recovery for transformers with lower projected amps per phase of transformers, NERC and FERC will actually create barriers to deterrence of EMP attack.
If the “coast effect” for solar storms (first identified in Australia in the year 1926 and well-studied for nearly a century) is disregarded in the solar standard now before FERC, then coastal generating facilities are likely to remain without hardware protection. About 39% of the U.S. population lives in coastal counties, and these counties produce a higher share of GDP and employment. See NOAA Website on Coastal Economy. So loss of electric generation in coastal counties will have high risks to human life and overall national economic activity.
Without cost recovery to protect high voltage transformers and generators from severe solar weather, it will be difficult to deter EMP attack by foreign adversaries, and impossible to defend critical infrastructure if a high altitude EMP or HEMP attack takes place.