Switching in power systems is most often thought of as something carried out by heavy-duty specialized devices such as circuit breakers. One therefore tends to forget that a large number of switching operations on high-voltage networks are in fact performed by disconnectors – open-air devices that outnumber virtually every other piece of apparatus at a HV substation.
These are relatively simple mechanically, with two contact blades that slowly move away from each other whenever instructed. Their basic function is to make a visible separation between different parts of the electrical supply system, typically to isolate a particular component for maintenance.
Disconnectors are often involved in ‘live’ switching as well. This means that there is some current involved, even when the relevant network section is not loaded. The current in this case flows as a result of ‘parasitic capacitance’ and, though only a few tenths of an ampere, it can appear quite spectacular. At separation of the contacts, the current will simply not cease – mainly because the voltage across the blades at this point immediately breaks down the hot air gap. Only when a substantial contact distance is reached (namely one that the available voltage cannot bridge), does current stop flowing.
During the intervening seconds, an electric arc becomes visible – namely, ‘electricity in air’. The length of this free burning arc can attain several meters because the arc heats up the air and thereby creates a more conductive path for itself. Only wind has the potential to destroy the ‘party’, much like in breakers where far more powerful fault arcs are cooled by blasts of gas.
The nature of these free-burning arcs was recently studied at the University of Eindhoven as part of a PhD project whose goal was to find ways to increase disconnector current switching capabilities. Based on tests at various laboratories, it was discovered that disconnector arcs are highly repetitive in nature. There is also a natural tendency for any arc to self extinguish at current zero. However, the poor insulation of the heated air immediately upon extinction usually cannot resist the voltage imposed by the circuit and the arc re-ignites for another cycle.
The research helped clarify that the interaction between the arc and electrical transients of the circuit can be quite intense. This interaction then has an impact when it comes to testing. In the previous IEC standard on disconnector switching, the test-circuit was not defined. After recognizing its potential impact on disconnector arc behaviour (especially duration), a new IEC document (TR 62271-305) was developed which details the test-circuit and therefore makes testing less arbitrary.
Our research work also showed that the electricity flow through air appears in two distinct modes: a short-lived ‘machine gun’ like firing arc across the shortest point between contacts; and a highly erratic one that burns softly like a huge flame and expands rapidly upward in air, often covering meters in length.
The repetitive nature of the ‘machine gun’ type arc (though pleasing to the eye and ear of any well-rooted HV engineer) has its drawback. The rapid and frequently collapsing voltage imposes steep surges on nearby equipment, sometimes leading to reports of damage. The erratic arc version can also reach too close to neighbouring conductors, thus initiating faults.
Disconnector arc switching becomes especially complex when the arc operates in the seemingly safe environment of a closed GIS installation and needs to switch even lower current, i.e. few milliamps. While there is no longer any ‘sound and light show’, there are still neat short arcs that are even more repetitive than in air. And there lies the problem!
The successive breakdown events in an SF6 environment create very rapid trains of transients that roam with virtually unlimited freedom throughout the enclosure. At any opportunity to exit, e.g. at bushings, they partly escape the GIS and continue on outward. Now they are called very fast transient overvoltages (VFTO) and become a serious concern – especially as the equipment voltage rating increases. For UHV equipment of 1100 kV and 1200 kV, VFTO transients reportedly reach even higher levels than lightning impulses.
GIS disconnector testing has been under discussion for a long time and, with the advent of IEC 62271-102, the ‘dust seemed to have settled’. Nevertheless, testing of such ‘bus-charging’ switching duty remains a real challenge in every respect. Transients with mega-hertz frequencies have to be measured inside the GIS enclosure and laboratory HV sources then have to be protected against the evil of a ‘hit-and-run’. This calls for very special skills and equipment. KEMA, for example, has already conducted testing with equipment up to 550 kV but we are increasingly aware of the risks during installation.
And so it is that specialists when it comes to testing switchgear with hundreds of kA can find themselves challenged by disconnector switching with no more than a few milliamps.
Professor Rene Smeets