Ultimate Guide to Electric Power Engineering: Transmission System--Geomagnetic Disturbances / Impacts on Power System Operation (part 1)

Home | Articles | Forum | Glossary | Books




1. Introduction

Recent analysis carried out for the EMP Commission, Federal Emergency Management Agency (FEMA), Federal Energy Regulatory Commission (FERC), North American Electric Reliability Corporation (NERC), and the U.S. National Academy of Sciences has determined that severe geomagnetic storms (i.e., space weather caused by solar activity) has the potential to cause crippling and long-duration dam age to the North American electric power grid or any exposed power grid throughout the world (NRC 2008, Kappenman 2010, NERC/US DOE 2010). The primary impact to the power grid is the risk of widespread permanent damage to high-voltage transformers and other power delivery and production assets, which are key, scarce, and difficult to replace, of the high-voltage power network.

These storm events can have a continental and even planetary footprint causing widespread disruption, loss, and damage to the electric power supply for the United States or other similarly developed countries around the world. It is also estimated to be plausible on a 1 in 30 to 1 in 100 year time frame. In short, this is potentially the largest and most plausible natural disaster that the United States could face, as the loss of electricity for extended durations would mean the collapse of nearly all other critical infrastructures, causing wide-scale loss of potable water, loss of perishable foods and medications, and many other disruptions to vital services necessary to sustain a nation's population. The severity of the threat geomagnetic storm impacts to present-day electric power grid infrastructures around the world have grown as the size of grids themselves have expanded by nearly a factor of 10 over the past 50 years, while at the same time they have become much more sensitive as higher EHV voltages and designs of transformers have evolved that react proportionately more to GIC exposure. These aspects of current design practices of electric grids have unknowingly and greatly escalated the risks and potential impacts from these threat environments. There has been no power grid design code that has ever taken into consideration these threat concerns.

Reliance of society on electricity for meeting essential needs has steadily increased over the years. This unique energy service requires coordination of electrical supply, demand, and delivery-all occurring at the same instant. Geomagnetic disturbances that arise from phenomena driven by solar activity commonly called space weather can cause correlated and geographically widespread disruption to these complex power grids. The disturbances to the earth's magnetic field causes geomagnetically induced currents (GICs, a near DC current typically with f < 0.01 Hz) to flow through the power system, entering and exiting the many grounding points on a transmission network. Geomagnetically induced currents 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 quiescent magnetic field. These fluctuations induce electric fields across the earth's surface-which causes GICs to flow through transformers, power system lines, and grounding points. Only a few amperes (amps) are needed to disrupt transformer operation, but over 300 A have been measured in the grounding connections of transformers in affected areas. Unlike threats due to ordinary weather, space weather can readily create large-scale problems because the footprint of a storm can extend across a continent. As a result, simultaneous widespread stress occurs across a power grid to the point where correlated widespread failures and even regional blackouts may occur.

Large impulsive geomagnetic field disturbances pose the greatest concern for power grids in close proximity to these disturbance regions. Large GICs are most closely associated with geomagnetic field disturbances that have high rate-of-change; hence a high-cadence and region-specific analysis of dB/dt of the geomagnetic field provides a generally scalable means of quantifying the relative level of GIC threat.

These threats have traditionally been understood as associated with auroral electrojet intensifications at an altitude of ~100 km, which tend to locate at mid- and high-latitude locations during geomagnetic storms.

However, both research and observational evidence has determined that the geomagnetic storm and associated GIC risks are broader and more complex than this traditional view. Large GIC and associated power system impacts have been observed for differing geomagnetic disturbance source regions and propagation processes and in power girds at low geomagnetic latitudes. This includes the traditionally perceived impulsive disturbances originating from ionospheric electrojet intensifications. However, large GICs have also been associated with impulsive geomagnetic field disturbances such as those during an arrival shock of a large solar wind structure called coronal mass ejection (CME) that will cause brief impulsive disturbances even at very low latitudes. As a result, large GICs can be observed even at low- and mid-latitude locations for brief periods of time during these events. Recent observations also confirm that geomagnetic field disturbances usually associated with an equatorial current system intensifications can be a source of large-magnitude and long- duration GIC in power grids at low and equatorial regions. High solar wind speed can also be the source of sustained pulsation of the geomagnetic field, (Kelvin-Helmholtz shearing) that has caused large GICs. The wide geographic extent of these disturbances implies GIC risks to power grids that have never considered the risk of GIC previously, largely because they were not at high-latitude locations.

Geomagnetic disturbances will cause the simultaneous flow of GICs over large portions of the interconnected high-voltage transmission network, which now span most developed regions of the world. As the GIC enters and exits the thousands of ground points on the high-voltage network, the flow path takes this current through the windings of large high-voltage transformers. GICs, when present in transformers on the system, will produce half-cycle saturation of these transformers, the root cause of all related power system problems. Since this GIC flow is driven by large geographic scale magnetic field disturbances, the impacts to power system operation of these transformers will be occurring simultaneously throughout large portions of the interconnected network. Half-cycle saturation produces voltage regulation and harmonic distortion effects in each transformer in quantities that build cumulatively over the network.

The result can be sufficient to overwhelm the voltage regulation capability and the protection margins of equipment over large regions of the network. The widespread but correlated impacts can rapidly lead to systemic failures of the network. Power system designers and operators expect networks to be challenged by the terrestrial weather, and where those challenges were fully understood in the past, the sys tem design has worked extraordinarily well. Most of these terrestrial weather challenges have largely been confined to much smaller regions than those encountered due to space weather. The primary design approach undertaken by the industry for decades has been to weave together a tight network, which pools resources and provides redundancy to reduce failures. In essence, an unaffected neighbor helps out the temporarily weakened neighbor. Ironically, the reliability approaches that have worked to make the electric power industry strong for ordinary weather, introduce key vulnerabilities to the electromagnetic coupling phenomena of space weather. As will be explained, the large continental grids have become in effect a large antenna to these storms. Further, space weather has a planetary footprint, such that the concept of unaffected neighboring system and sharing the burden is not always realizable. To add to the degree of difficulty, the evolution of threatening space weather conditions are amazingly fast. Unlike ordinary weather patterns, the electromagnetic interactions of space weather are inherently instantaneous. Therefore large geomagnetic field disturbances can erupt on a planetary scale within the span of a few minutes.

2. Power Grid Damage and Restoration Concerns

The onset of important power system problems can be assessed in part by experience from contemporary geomagnetic storms. At geomagnetic field disturbance levels as low as 60-100 nT/min (a measure of the rate of change in the magnetic field flux density over the earth's surface), power system operators have noted system upset events such as relay misoperation, the offline tripping of key assets, and even high levels of transformer internal heating due to stray flux in the transformer from GIC-caused half-cycle saturation of the transformer magnetic core. Reports of equipment damage have also included large electric generators and capacitor banks.

Power networks are operated using what is termed an "N-1" operation criterion. That is, the system must always be operated to withstand the next credible disturbance contingency without causing a cascading collapse of the system as a whole. This criterion normally works very well for the well-understood terrestrial environment challenges, which usually propagate more slowly and are more geographically confined. When a routine weather-related single-point failure occurs, the system needs to be rapidly adjusted (requirements typically allow a 10-30 min response time after the first incident) and positioned to survive the next possible contingency. Geomagnetic field disturbances during a severe storm can have a sudden onset and cover large geographic regions. They therefore cause near-simultaneous, correlated, multipoint failures in power system infrastructures, allowing little or no time for meaningful human interventions that are intended within the framework of the N-1 criterion. This is the situation that triggered the collapse of the Hydro Quebec power grid on March 13, 1989, when their system went from normal conditions to a situation where they sustained seven contingencies (i.e., N-7) in an elapsed time of 57 s; the province-wide blackout rapidly followed with a total elapsed time of 92 s from normal conditions to a complete collapse of the grid. For perspective, this occurred at a disturbance intensity of approximately 480 nT/min over the region ( FIG. 1). A recent examination by Metatech of historically large disturbance intensities indicated that disturbance levels greater than 2000 nT/min have been observed even in contemporary storms on at least three occasions over the past 30 years at geomagnetic latitudes of concern for the North American power grid infrastructure and most other similar world locations: August 1972, July 1982, and March 1989. Anecdotal information from older storms suggests that disturbance levels may have reached nearly 5000 nT/min, a level ~10 times greater than the environment which triggered the Hydro Quebec collapse. Both observations and simulations indicate that as the intensity of the disturbance increases, the relative levels of GICs and related power system impacts will also proportionately increase. Under these scenarios, the scale and speed of problems that could occur on exposed power grids has the potential to cause wide spread and severe disruption of bulk power system operations. Therefore, as storm environments reach higher intensity levels, it becomes more likely that these events will precipitate widespread blackouts to exposed power grid infrastructures.


FIG. 1 Four minutes of geomagnetic field disturbance on the March 13, 1989, Superstorm that triggered the Quebec grid collapse, from 7:42UT to 7:745UT.

3. Weak Link in the Grid: Transformers

The primary concern with GIC is the effect that they have on the operation of a large power transformer.

Under normal conditions, the large power transformer is a very efficient device for converting one voltage level into another. Decades of design engineering and refinement have increased efficiencies and capabilities of these complex apparatus to the extent that only a few amperes of AC exciting current are necessary to provide the magnetic flux for the voltage transformation in even the largest modern power transformer.

As GIC levels increase, the level of saturation of the transformer core and its impact on the operation of the power grid as a whole also increases.

However, in the presence of GIC, the near-direct current essentially biases the magnetic circuit of the transformer with resulting disruptions in performance. The three major effects produced by GIC in transformers are (1) the increased reactive power consumption of the affected transformer, (2) the increased even and odd harmonics generated by the half-cycle saturation, and (3) the possibilities of equipment damaging stray flux heating. These distortions can cascade problems by disrupting the performance of other network apparatus, causing them to trip off-line just when they are most needed to protect network integrity. For large storms, the spatial coverage of the disturbance is large and hundreds of transformers can be simultaneously saturated, a situation that can rapidly escalate into a network wide voltage collapse. In addition, individual transformers may be damaged from overheating due to this unusual mode of operation, which can result in long-term outages to key transformers in the net work. Damage of these assets can slow the full restoration of power grid operations.

Transformers use steel in their cores to enhance their transformation capability and efficiency, but this core steel introduces nonlinearities into their performance. Common design practice minimizes the effect of the nonlinearity while also minimizing the amount of core steel. Therefore, the transformers are usually designed to operate over a predominantly linear range of the core steel characteristics (as shown in FIG. 2), with only slightly nonlinear conditions occurring at the voltage peaks.

This produces a relatively small exciting current (FIG. 3). With GIC present, the normal operating point on the core steel saturation curve is offset and the system voltage variation that is still impressed on the transformer causes operation in an extremely nonlinear portion of the core steel characteristic for half of the AC cycle ( FIG. 2), hence the term half-cycle saturation.

Because of the extreme saturation that occurs on half of the AC cycle, the transformer now draws an extremely large asymmetrical exciting current. The waveform in FIG. 3 depicts a typical example from field tests of the exciting current from a three-phase 600 MVA power transformer that has 75 A of GIC in the neutral (25 A per phase). Spectrum analysis reveals this distorted exciting current to be rich in even, as well as odd harmonics. As is well documented, the presence of even a small amount of GIC (3-4 A per phase or less) will cause half-cycle saturation in a large transformer.

Since the exciting current lags the system voltage by 90°, it creates reactive-power loss in the trans former and the impacted power system. Under normal conditions, transformer reactive power loss is very small. However, the several orders of magnitude increase in exciting current under half-cycle saturation also results in extreme reactive-power losses in the transformer. For example, the three-phase reactive power loss associated with the abnormal exciting current of FIG. 3 produces a reactive power loss of over 40 MVars for this transformer alone. The same transformer would draw less than 1 MVar under normal conditions. FIG. 4 provides a comparison of reactive power loss for two core types of transformers as a function of the amount of GIC flow.

Under a geomagnetic storm condition in which a large number of transformers are experiencing a simultaneous flow of GIC and undergoing half-cycle saturation, the cumulative increase in reactive power demand can be significant enough to impact voltage regulation across the network, and in extreme situations, lead to network voltage collapse.

The large and distorted exciting current drawn by the transformer under half-cycle saturation also poses a hazard to operation of the network because of the rich source of even and odd harmonic currents this injects into the network and the undesired interactions that these harmonics may cause with relay and protective systems or other power system apparatus. FIG. 5 summarizes the spectrum analysis of the asymmetrical exciting current from FIG. 3. Even and odd harmonics are present typically in the first 10 orders and the variation of harmonic current production varies somewhat with the level of GIC, the degree of half-cycle saturation, and the type of transformer core.


FIG. 2 Transformer saturation characteristics for normal operation and for half-cycle saturation due to the presence of GIC.


FIG. 3 Transformer excitation current characteristics for normal operation and for half-cycle saturation due to the presence of GIC.


FIG. 4 Transformer increased reactive power demands (MVARs) due to GIC for a typical 500 kV transformer for single phase and three-phase three-legged core type.


FIG. 5 Example of even and odd harmonic spectrums of half-cycle saturated excitation current for the excitation current waveform shown in FIG. 3.

With the magnetic circuit of the core steel saturated, the magnetic core will no longer contain the flow of flux within the transformer. This stray flux will impinge upon or flow through adjacent paths such as the transformer tank or core clamping structures. The flux in these alternate paths can concentrate to the densities found in the heating elements of a kitchen stove. This abnormal operating regime can persist for extended periods as GIC flows from storm events can last for hours. The hot spots that may then form can severely damage the paper-winding insulation, produce gassing and combustion of the transformer oil, or lead to other serious internal and or catastrophic failures of the transformer. Such saturation and the unusual flux patterns that result are not typically considered in the design process and, therefore, a risk of damage or loss of life is introduced.

One of the more thoroughly investigated incidents of transformer stray flux heating occurred in the Allegheny Power System on a 350 MVA 500/138 kV autotransformer at their Meadow Brook Substation near Winchester, Virginia. The transformer was first removed from service on March 14, 1989, because of high gas levels in the transformer oil which were a by-product of internal heating. The gas-in-oil analysis showed large increases in the amounts of hydrogen, methane, and acetylene, indicating core and tank heating. External inspection of the transformer indicated four areas of blistering or discolored paint due to tank surface heating. In the case of the Meadow Brook transformer, calculations estimate the flux densities were high enough in proximity to the tank to create hot spots approaching 400°C. Reviews made by Allegheny Power indicated that similar heating events (though less severe) occurred in several other large power transformers in their system due to the March 13 disturbance. FIG. 6 is a recording that Allegheny Power made on their Meadow Brook transformer during a storm in 1992. This measurement shows an immediate transformer tank hot-spot developing in response to a surge in GIC entering the neutral of the transformer, while virtually no change is evident in the top oil readings. Because the hot spot is confined to a relatively small area, standard bulk top oil or other over- temperature sensors would not be effective deterrents to use to alarm or limit exposures for the transformer to these conditions.

Designing a large transformer that would be immune to GIC would be technically difficult and prohibitively costly. The ampere-turns of excitation (the product of the normal exciting current and the number of winding turns) generally determine the core steel volume requirements of a transformer.

Therefore, designing for unsaturated operation with the high level of GIC present would require a core of excessive size. The ability to even assess existing transformer vulnerability is a difficult undertaking and can only be confidently achieved in extensive case-by-case investigations. Each transformer design (even from the same manufacturer) can contain numerous subtle design variations. These variations complicate the calculation of how and at what density the stray flux can impinge on internal structures in the transformer. However, the experience from contemporary space weather events is revealing and potentially paints an ominous outcome for historically large storms that are yet to occur on today's infrastructure. As a case in point, during a September 2004 Electric Power Research Industry work shop on transformer damage due to GIC, Eskom, the power utility that operates the power grid in South Africa (geomagnetic latitudes -27° to -34°), reported damage and loss of 15 large, high-voltage transformers (400 kV operating voltage) due to the geomagnetic storms of late October 2003. This damage occurred at peak disturbance levels of less than 100 nT/min in the region.


FIG. 6 Observed Meadow Brook transformer hot-spot temperature for a minor storm on May 10, 1992.

4 Overview of Power System Reliability and Related Space Weather Climatology

The maintenance of the functional integrity of the bulk electric systems (i.e., power system reliability) at all times is a very high priority for the planning and operation of power systems worldwide. Power systems are too large and critical in their operation to easily perform physical tests of their reliability performance for various contingencies. The ability of power systems to meet these requirements is commonly measured by deterministic study methods to test the system's ability to withstand probable disturbances through computer simulations. Traditionally, the design criterion consists of multiple out age and disturbance contingencies typical of what may be created from relatively localized terrestrial weather impacts. These stress tests are then applied against the network model under critical load or system transfer conditions to define important system design and operating constraints in the network.

System impact studies for geomagnetic storm scenarios can now be readily performed on large complex power systems. For cases in which utilities have performed such analysis, the impact results indicate that a severe geomagnetic storm event may pose an equal or greater stress on the network than most of the classic deterministic design criteria now in use. Further, by the very nature that these storms impact simultaneously over large regions of the network, they arguably pose a greater degree of threat for precipitating a system-wide collapse than more traditional threat scenarios.

The evaluation of power system vulnerability to geomagnetic storms is, of necessity, a two-stage process. The first stage assesses the exposure to the network posed by the climatology. In other words, how large and how frequent can the storm driver be in a particular region? The second stage assesses the stress that probable and extreme climatology events may pose to reliable operation of the impacted network. This is measured through estimates of levels of GIC flow across the network and the manifestation of impacts such as sudden and dramatic increases in reactive power demands and implications on voltage regulation in the network. The essential aspects of risk management become the weighing of probabilities of storm events against the potential consequential impacts produced by a storm. From this analysis effort, meaningful operational procedures can be further identified and refined to better manage the risks resulting from storms of various intensities.

Successive advances have been made in the ability to undertake detailed modeling of geomagnetic storm impacts upon terrestrial infrastructures. The scale of the problem is enormous, the physical processes entail vast volumes of the magnetosphere, ionosphere, and the interplanetary magnetic field conditions that trigger and sustain storm conditions. In addition, it is recognized that important aspects and uncertainties of the solid-earth geophysics need to be fully addressed in solving these modeling problems.

Further, the effects to ground-based systems are essentially contiguous to the dynamics of the space environment. Therefore, the electromagnetic coupling and resulting impacts of the environment on ground-based systems require models of the complex network topologies overlaid on a complex geo logical base that can exhibit variation of conductivities that can span 5 orders of magnitude.

These subtle variations in the ground conductivity play an important role in determining the efficiency of coupling between disturbances of the local geomagnetic field caused by space environment influences and the resulting impact on ground-based systems that can be vulnerable to GIC. Lacking full understanding of this important coupling parameter hinders the ability to better classify the climatology of space weather on ground-based infrastructures.

cont. to part 2 >>

Top of Page

PREV.   NEXT   Guide Index HOME