Electric Power Networks Are Increasingly vulnerable to geomagnetic storms (or "space weather") created by solar activity.
The importance of electricity in our daily lives, powering an entire nation from individual households to industry, is unquestionable. Large electric power networks are designed, built, and operated with a very high degree of reliability, and with consideration for variations in terrestrial weather. However, this may not be true for “space weather” effects in the form of geomagnetically induced currents, even though geomagnetic storms created by solar activity have caused problems in the operation of high voltage power grids throughout the world. Parsons Brinckerhoff (now part of WSP | Parsons Brinckerhoff) is part of a working group for the International Council on Large Electric Systems (CIGRE), tasked with a study of the geomagnetic storm environment.
Geomagnetic Disturbances
The sun follows roughly an 11-year cycle of sunspot activity with several spots during the peak period. We are currently in 24th solar cycle since records began in 1755 and, based on the sunspot number predictions, it is postulated to be one with the lowest number of sunspots. Figure 1 shows the current prediction from NASA together with solar cycle 23 (1996-2008). Solar flares are another form of sun storm manifestation, with a massive burst of solar wind and magnetic fields rising above the solar corona or being released into space. This is commonly referred to as a coronal mass ejection (CME). When CMEs are directed to Earth and reach it, Earth’s magnetosphere may be disrupted, creating aurorae around the magnetic poles. These are commonly known as the northern and southern lights (aurora borealis and aurora australis) in the respective hemispheres. The scale used to categorize the intensity of geomagnetic storms is the K-index. The index ranges from zero to nine, zero corresponding to minimum geomagnetic disturbance.
Geomagnetically Induced Currents
When a geomagnetic storm interacts with Earth’s magnetic field it induces geoelectric fields at Earth’s surface (as shown in Figure 2), giving rise to voltages in pipes, telecommunication circuits, transmission lines, etc. The induced voltages are classified as quasi-DC due to their frequency being much lower than the grid system frequency. The currents resulting from these voltages are termed geomagnetically induced currents (GICs), and they enter and leave the network via transformer neutrals, which are connected to Earth, mainly at the extremities of the network. These flow through the transformer windings, producing extra magnetisation which in turn can saturate the core of the transformer. This produces extra eddy currents in the transformer core but, due to the large thermal capacity of extra high voltage transformers, it manifests itself only as localised heating (hot spots with possible damage to transformer windings). There will also be extra noise and vibration.
Another side effect of GIC is the generation of extra harmonics that can cause maloperation of protection relays, hence possible tripping of individual transmission circuits leading to system collapse. The most pronounced effect though is on the increased requirements for reactive power. This is due to the transformer’s operating point being shifted into the saturation region. A typical example of this would be a 600 MVA1 transformer with the excitation current jumping from 6 amps to 300 amps under the application of 75 amps GIC current, producing an extra 50 Mvar reactive power loss and subsequent voltage drop2. A system-wide increase in reactive power requirement puts excessive demand on the available reactive power compensation trying to support voltage levels.
Added to this are the extra harmonic currents being generated and possibly flowing into the capacitor banks. As a result, reactive power compensation devices may be overloaded and likely to trip under such strain, leaving the system with even more requirement for voltage support. With no such support, the eventual result is the tripping of transmission lines with a complete collapse of the system.
Hydro-Québec system collapse of 1989
The best example of such an event is the Hydro-Québec system blackout that took place in March 1989 in the Canadian province of Québec. This blackout was due to a magnetic super-storm with a CME estimated to be the size of 36 earths. At the time, two 735 kV transmission corridors were transmitting power from the north to the load centres in the south more than 1000 kilometres away. One corridor was running from the La Grande Complex in the James Bay area and the other from the Churchill Falls in Labrador, each with five transmission lines. Seven static compensation schemes were active in the La Grande transmission corridor for system stability and voltage control purposes.
An intense magnetic storm took place on 13 March 1989 around 2:45am which saturated the cores of transformers generating excessive, even-harmonic, currents that in turn overloaded and then tripped all the seven static var compensators4 in the La Grande transmission corridor. No preventive action could be taken during this tripping, as all took place within less than minute. The tripping of the last static compensator was followed after nine seconds by the tripping of one of the 735 kV transmission lines, resulting in automatic rejection of two generating units followed by the tripping of three further 735 kV transmission lines from the La Grande corridor as well as other faults. The last remaining transmission line was to trip next in the same transmission corridor which separated La Grande from the remaining Hydro-Québec system. The result of this was a rapid drop in system frequency (as seen in Figure 3) triggering the automatic load shedding system that started tripping all loads.
Even then, the loss of approximately 9500 MW generation from the La Grande complex was not offset, hence the result was a total collapse of the remaining system. As a result of the whole event some strategic transmission equipment was damaged on the Hydro-Québec system. Hence the cost of such an event is twofold: loss of revenue from the transmission or sale of electrical energy and the replacement of damaged equipment. Hydro Québec estimated the cost of the event to be close to $15 million.5
Mitigating the Effects of Geomagnetically Induced Currents
The U.K. didn’t entirely escape the effects of the storm of March 1989. During the night of the 13th/14th March, reports were received of spurious alarms from power stations, telecommunications networks, and supergrid transformers on the U.K. network. There was no actual loss of supply or tripping, but two 5-limb transformers, one on the eastern extremity of the network, and one on the western extremity, were taken out of service; the damage to these was attributed to GICs.7
As the electric power networks are getting more and more heavily loaded and complex, their vulnerability to geomagnetic disturbances is increasing. Following the collapse of the Hydro-Québec system, the National Grid of England and Wales (currently National Grid Electricity Transmission) undertook a risk assessment on the potential impact of solar cycle 23 and the effectiveness of possible mitigation strategies.8 At the time the system operator was heavily dependent on reactive power compensation for the transmission of active power. The results of the risk analysis indicated that severe geomagnetic disturbance could lead to an additional reactive power loss, hence a reserve of up to 3500 Mvar may be required. Furthermore, it was established that transformers in the coastal regions would be the key collection points for the GIC currents, with auto-transformers experiencing the highest flows and saturation.
It is of importance to note that 5-limb and single-phase transformers are much more vulnerable than 3-limb transformers. Most supergrid transformers in the U.K. are 3-limb type. The option of the installation of series blocking capacitors (which do not allow the flow of DC or quasi-DC currents) was overruled due to not being a viable mitigation strategy for the particular situation. Also an operational management strategy that would trigger action by local monitoring of GIC was discounted on the basis that the process would not be fast enough for preventive action to be taken. Instead, a predictive forecasting strategy was implemented relying on specific models of the transmission system to derive a forecast impact. This was further supported by a review of the various transmission system equipment capabilities. As a result, six strategically chosen substations were furnished with GIC monitoring equipment. The introduction of this predictive warning strategy proved to be successful in that many minor GIC events were predicted correctly and generally produced reliable forecasts two days ahead of the event.9
1MVA is the effective power of an AC power system, comprising real power measured in MW, and reactive power which is only present in AC systems, and is measured in Mvar.
2J. G. Kappenman and V. D. Albertson, “Bracing for the geomagnetic storms,” IEEE Spectrum, vol. 27, pp. 27-33, March 1990.
3J. G. Kappenman, “A perfect storm of planetary proportions,” IEEE Spectrum, vol. 49, pp. 26-31, February 2012.
4A device used to control reactive power, hence voltage, on the system.
5http://www.spaceweather.gc.ca/tech/se-pow-eng.php
6P Czech, S Chano, H Huynh and A Dutil, “ The Hydro-Quebec system blackout of 13 March 1989: System response to geomagnetic disturbance”, Proceedings of Geomagnetically Induced Currents Conference, 8-10 November 1989, Millbrae, California, EPRI Report TR-100450, 1992.
7P Smith, “Effects of geomagnetic disturbances on the National Grid System”, 25th Universities Power Engineering Conference, September 1990.
8, 9I A Erinmez, J G Kappenman and W A Radasky, “Management of the Geomagnetically Induced Current Risks on the National Grid Company’s Electric Power Transmission System”, Journal of Atmospheric and Solar Terrestrial Physics Special Addition for NATO Space Weather Hazards Conference, June 2000.
