|
That is the claim. It is quite common to see single phase arcing faults expand into l-g-l and then 3 phase faults in a single enclosure. There is also plenty of results of arc propulsion where an arc travels AWAY from the power source. The former is caused by increasing temperatures of the air around nonarcing components that the breakdown voltage of the air is compromised and additional arcing faults occur. The challenge is predicting when this form of propagation may occur, estimating both the size of the region of gases hot enough to initiate additional arcs and convection (how fast it moves). It would be nice to have at least a gross estimate because right now it appears that there is acknowledgement that this does not occur in some switchgear, definitely occurs in panelboards, and might occur in some MCC's under some circumstances. I have not seen any published cases though.
Personally I have seen plenty of arcing faults that propagated throughout a single section of very open mining switchgear which is kf the metal enclosed construction design. However even with compartments that are entirely open to each other even at 23 kv with the usual 36 inch deep cabinets, I have not seen any evidence of propagation by convection with around 10ka of available fault current. I know of two cases of confirmed magnetic arc propagation, both initiated by removing buckets off live mcc bus, in which the arc traveled to the non power source end of the mcc section (vertical). Arc magnetic propagation on overhead lines (looks like jacobs ladder) is common. But I have not seen arcing faults initiated from one SECTION of a mcc to another. Especially with reduced height buckets and high available fault current I can easily see propagation within the same section, but I'm dubious of the claim of moving from one section to another by convection of hot gases considering that this would be the norm for the relatively open designs of mjning switchgear at 5, 8, and 25 kv.
As to plasma, this is a bit confusing. Plasma is a state of matter where the outer shell electrons have essentially lost any affinity for an atomic nucleus and sort of float around in an energized state in a cloud. Electrons rarely get back to ground state. High temperatures are not a necessity. Fluorescent tubes essentially convert gas to plasma.
At room temperature and pressure, the breakdown voltage of air is around 300 volts/mm. If this voltage is exceeded, we get an arc. The voltage drop through the air rapidly heats it up. The exact temperature is hard to measure but some measurements and numerical simulations seem to suggest around 6000 F in the arc column. Needless to say we also get rapid expansion of the gas (arc blast) at least initially until things stabilize within around 10-30 cycles (different experiments show different results). This also results in a local decrease in pressurr once things stabilize at the new temperature. s the current (not voltage) crosses zero in an ac arc, the arc extinguishes. If the air temperaturr stays high enough then on the next voltage (not current) rise, the arc restrikes and we go again for another half cycle. Decreasing air pressure increases the breakdown voltage of air. Increased temperature decreases it.If the voltage is relatively close to the arc voltage of around 80-150 volts, the arc does not last long enough to sustain a high enough air temperature to restrike and will self extinguish within a few cycles. Pressure may also affect the dynamics but I haven't seen any research on this beyond puffer/air blast type circuit breakers and vacuum interrupters.
Some research published by Willis at Mersen has already shed some light (and more questions) on this. The published IEEE 1584 data set uses a test which is essentially open bus gear. Willis looked at various configurations of fiber board bus spacers which are commonly used. This caused a decrease in incident energy in low fault current cases because the arc stayed single phase. It also focussed and increased the amount of plasma and heat at close working distances but further away, there was no appreciable increase.
Obviously how an arc actually behaves in real world cases has major implications not only becausd of the issue of possibly propagating to adjacent phases or even entire compartments but worse still, the predicted incident energy may be understated for typical gear with fiberboard bus spacers.
As to how to respond to this, one has to use the best available information. Arc flash and blast are electrical hazards that have been known for a century. Prior to the 1990's we stood to one side, tried to avoid risks, etc. In the 1990's we had measurable results and responded with pickle suits (arc flash suits were originally green welder's overalls). Over time the understanding has been refined, gear has been improved, and ppe has gotten better. As I understand it the research has now gotten closer to quantifying arc blast and has improved the IEEE 1584 equations significantly including accounting for enclosure size and bus configuration (draft, not released yet). You can either take all the current engineering arguments and research data, none of which can be considered any kind of consensus, and react without any strong backing for your choices, or follow sound, well understood and agreed upon consensus engineering standards which have a long history of legal standing.
In a court room a lawyer can use all manner of technical and emotional arguments. But at the end of the day MOST of the time, judges and juries will pay much more respect to standards and regulations than wacky half baked theories and ideas. You cannot possibly hope to protect yourself against the paranoid ideas that are out there, and you can't for instance use any of the data from Willis' research to craft a sound safety policy because the particulars of the experiment are not comprehensive enough to make engineering conclusions from it.
Hence the reason I ask for examples, published case studies, etc., before I would subscribe to simulating mcc as if the main breaker did not exist.
|
