Arc flashes, for the uninitiated, are short-circuits caused by electricity flowing through through air (partially ionized) between two live conductors. The arc of electricity causes an arc blast, which is a huge fiery shockwave that can (and does) blast through metal and all surroundings
One scenario is equipment failure. On-site person was not familiar, called for back-up, back ups did not don proper PPE nor follow proper protocol. Another scenario is the equipment failed in a manner not envisioned by the designers of the protocols and were really unlucky.
In any case, this is really sad and bad and hope they recover.
It may be something simple as a shor tbetween the phases. In power engineering these are called fault conditions and the currents can run up to tens of thousands of amps before the circuit breaker acts. Not sure if you can imagine the magnitude but that is why the standard PPE in working with on-line low voltage (<1 kV) switchgear is a full body silver arc flash suit.
Now imagine if the datacenter guys need to wear that suit before going down the racks...
There are unknowns to us. But, no I don't imagine regular detacenter folks walking around in PPE --unless they are going in to these segregated areas to address a known issue. Maybe it was totally unpredictable --like someone just walking in proximity to said equipment just as it was about to catastrophically fail, but in most accidents there was a known bad condition, techs go in to check things out and then go back in to bring in the certified electrician or whatever and in those instances people are supposed to don proper equipment and follow certain procedure typically involving cutting power away from the faulty equipment, though sometimes people cut the wrong power or altogether forget that step in the SOP. Sometimes the diagrams are outdated --CBF is old now, I think?
You are forbidden by law (and mandated to report) such arcs in Europe. In Germany people who work in such environments have to have a few special certificates and PPE that looks like they are trying to defuse a bomb.
Arc flashes like that can burn with 10000 to 15000 degrees celsius and if you are close by not only will they send you flying, you are probably dead before you hit the ground.
And trying to extinguish such an arc can be quite a challenge.
> You are forbidden by law (and mandated to report) such arcs in Europe.
You are forbidden by law to have an accident? This misunderstanding is probably because of my lack of knowledge, but what do you mean?
> And trying to extinguish such an arc can be quite a challenge.
I only read about them in this thread, so I’m probably missing something again, but isn’t it an instantaneous thing and not something you need to extinguish?
You are forbidden by law to have an accident of this nature. Yes. If you do cause an accident of this nature you are criminally negligent no matter the circumstances leading up to the accident because it's your duty to ensure it doesn't.
Of course you need to keep in mind that courts in Europe tend to not hold to the letter of the law quite as strongly as US courts seem to do. Laws like these are mostly so the judicial system gets booted up when something awful happens. Might still turn out that no one was responsible for the accident. Might turn out the boss of the guy throwing the switch was responsible. Or the equipment manufacturer.
It sounds like an accident of this nature would be quite difficult to cover-up, and given that doing so would be a criminal offense I imagine it's an easy decision for those involved to just go ahead and file the appropriate reports.
> Making this sort of accident legal makes it definitive that people won't report it. Because there's no requirement to.
Huh? No. Make it a legal requirement to report it. Then there is all the incentive to report it and no incentive not to.
If it's illegal to have the accident, then attempting to cover it up is the most rational course of action, because if you can not report and get away with it then that's the only way you aren't punished.
If instead it's just a legal requirement to report it, then reporting it is the most rational course of action, because you can report and then aren't punished.
Right, but inserting a legal incentive to do everything to prevent the accident is considered more important than inserting a legal incentive to report when it happens. If there was no consequences to having the accident, sure, you'd know all about the ones that happen, but there wouldn't be much incentive for investing in prevention
We are talking about accidents that mostly happen on the ownership border between house owners and grid maintainers. If you can cover up an arc flash there in such a way they will not notice on their next visit without dying, you might deserve to get away with it.
>You are forbidden by law to have an accident? This misunderstanding is probably because of my lack of knowledge, but what do you mean?
There are systems (permits to work, lock out tag out, risk assessments, etc.) in place to maintain worker safety mandated by law. An incident of this kind typically means that there was a failure with one of more levels of the system... Someone did something they weren't supposed to, they turned off the wrong equipment, they didn't turn off the equipment before working on it, etc.
> And trying to extinguish such an arc can be quite a challenge.
Once established, an arc can continue as long as current is still flowing. It's possible for arcs to continue and still be below the trip current of the upstream circuit protection. In non-catastrophic scenarios, you have mechanisms to "quench" an arc... one method is to use compressed air.. like blowing out a very hot, angry candle on a birthday cake.
> isn’t it an instantaneous thing and not something you need to extinguish?
I'm not familiar with the details of this accident, but in general, not necessarily. Air is normally an excellent insulator. But when voltages are high enough, the molecules ionize, become conductive and transmit current. This is a path-dependent process, meaning that once it begins, the voltage can drop below what is required to begin the process while maintaining current.
So an arc caused by a system with low capacitance (or total count of free electrons on the negative side) may quickly transfer all the free electrons, lowering the voltage enough to close the circuit. But anything connected to mains power will have an essentially infinite source of more electrons, meaning there won't be a voltage drop big enough to close the circuit. I'm being a little handwavy with the physics concepts, but hopefully the point is clear.
This can be thought of as a sort of hysteresis: the dependence of the state of a system on its history.
Edit: If you want to go down a rabbit hole, check out the solutions used in power substations as circuit breakers. It's not enough just to mechanically separate the connection, so they use systems like these:
Arcs are rated by calories per square centimeter. There's a lot factors that go into arc ratings, more than just voltage. To give a frame of reference, a flame from a lighter on the tip of your finger for 1 second is considered 1 calorie. Arc flash PPE ratings start at 4 calories and can go above 100.
The high cal rated PPE does indeed look like bomb squad gear.
Energy is used rather than power as it's the cumulative effect in energy incidents that causes damage. In most of the world, the unit is J/cm2, though.
Ah, all right, I understand if it's used to rate accumulated damage for incidents. For ongoing arcs, or comparing them however, I think it's a weird.
A small arc, not even dangerous to touch, could have the same energy rating as a gigantic one if you measure the first over a year and the second over a millisecond.
> You are forbidden by law (and mandated to report) such arcs in Europe.
When you say in Europe, normally that refers to an EU Regulation or Directive (since other legislation varies wildly in member states). What are you referring to here?
> The pressure wave created by an arc flash explosion has the force of thousands of pounds per square inch. This is powerful enough to knock down or throw nearby workers and at the same time cause damage to the eardrums, lungs, brain and other vital organs.
(I didn't know that, and just found that explanation while looking for an answer to your question.)
A different risk is that electric current passing through the body could cause muscle convulsions, which could also cause someone to move or fall suddenly or erratically.
The heat of the arc rapidly expands air (like a lighting bolt), the heat of the arc also vaporizes the metal conductor, and the vaporized metal expands rapidly, and enormously.
It's basically a small explosion.
The hot metal then ignites in air adding even more heat to the explosion.
But ac will break the current at the alternating frequency and dc won't, so it is much harder to break a circuit with a high direct current. That's why monster movies have those huge knife switches.
Basically, the dielectric constant of air is not infinite. If the electrical potential is high enough between two points, air can conduct electricity.
If you put a small bit of conductor near two legs of a high voltage system, and that conductor has a path to ground, you can get now get an arc between a voltage source and the conductor (e.g., someone with a screwdriver loosening a screw is too close to a high voltage leg). This initial arc ionizes the air, which makes a pretty great conductor, so the amount of current flowing through the air increases really quickly as if it's flowing through an open circuit. The arc is superheated, can burn through steel (or flesh) quickly, and expands.
One of the hazards of working on high voltage is gasping in shock after an electrical arc - you inhale superheated plasma and it cooks your lungs instantly, so you suffocate.
Anyway, to prevent this, you don't cut corners as an electrician - you have proper insulation on all your tools and body parts to prevent this sort thing. You isolate legs of high voltage circuits, or de-energize the circuit you're working on with proper lock-out tag-out procedures.
I recall a similar incident, causing us to enact our DR plan a decade ago. A truck driver had a medical issue and drove into the transformers outside a Rackspace facility. Emergency crews required everything to go dark for the rescue. High-power electric delivery is the Achilles heel of the cloud.
Doubtful they'll be okay long term. Even if they survive the burns, electrocution does a number on most of your organs including your kidneys. I'm in no way to evaluate the veracity of the claims (which are mostly: electricity damages the body in a lot of different ways), but here's a graphic description of the damage electrocution does to the body:
What I'd read in a non-academic context is that the kidney damage suffered from electrocution takes more than a few years off your life expectancy. I couldn't (quickly) find an authoritative source discussing long-term survival but burns and acute rhabdomyolysis (like the paper above is talking about) on their own are no joke.
It'll be interesting to see if Google opens up about what went wrong.
It's unlikely to be both an arc flash and electrocution, and to survive for any length of time.
Electricity powerful enough to cause an arc flash would instantly kill if it went through a body. It was more likely a tool shorted two conductors, and the arc flash harmed them, but no electricity went through them.
I was testing a light bulb socket one time, and accidentally shorted it - the short set fire to the metal of the bulb socket (I assume it was aluminum), which proceeded to burn until it was all consumed, despite me pulling the plug.
It produced a jet of flame and sparks many feet tall. I had no idea this was even something that could happen.
I can't even imagine how bad a powerful arc flash would be.
Aluminium burns too, just harder to get going. Almost all the metals can be coaxed into burning in various ways, all metal fires are an absolute disaster when they happen.
Can you point to an aluminium fire that wasn't based on the aluminium being powdered, and where there were no fire enhancers (O2-enriched atmosphere, oxidizers like permanganate) present, and that sustained itself? (Because that's the conditions ars was talking about. Maybe if in ars' case the aluminium part was mostly liquid due to the short before it started burning that would explain it, but then they didn't describe it that way.)
It was very thin metal, it was the shell of a cheap light bulb socket, and it was enhanced by the electrical short - I don't remember the circuit breaker tripping.
It burned until there was no more silver color metal left, leaving just the copper wire leads. I assume it was aluminum, I think steel would be harder to burn.
I do remember unplugging it during the fire, but it didn't take very long to burn, so I might have done it after it was over.
Again, from all sources I've ever seen, you can't ignite aluminium blocks and have a sustained fire, whereas you can with magnesium. If the aluminium is liquid then perhaps by spraying it around because of the spark it would give the effect you describe, but it sounds dubious that it melted so quickly in a large enough amount, and it seems somewhat unlikely to me that once it started the sprayed liquid aluminium would stay close enough to melt and burn the remainder, too. When you see pictures of large arc flashes, a lot of metal is sprayed from the location of the arc and then burns, but it burns flying away so it cannot sustain the reaction; the cause of the metal burning and flying away is the arc. In your case you didn't have enough electric power to get such a strong arc (the 10-15 A you get from house hold power before tripping the fuse is not that much), so the energy to sustain burning up the remainder of the metal must have come from the metal fire itself. That is definitely possible with magnesium, but from all I've seen it is not with aluminium.
Magnesium can be set alight even when multiple cm in thickness [1], and lighting up thin strips is much more spectacular[2]. You can hold aluminium foil to a blow torch, what happens is it will melt and oxidize, but not produce any flame or light whatsoever[2]. If aluminium foil doesn't do it, I don't see how a thicker sheet would do it. And all I've found confirms this. (BTW steel wool however does burn continuously. But I agree that a steel lamp base burning this way doesn't look feasible, either.)
Since you say that it was thin, it must have been an alloy in either case. It seems Magnesium alloys need to be cast. I've got a laptop with a magnesium base which looks like sheet metal from the outside but when you open it up, on the inside it has all sorts of structures coming out of the "sheet", clearly confirming it was cast. I've never seen aluminium parts made that way, aluminium cast parts seem to always be relatively thick, not like sheet metal. Aluminium can be pressed, however (think beer cans), but then it doesn't have a fancy structure. I'm not a specialist in these matters in any way, but maybe this might give some further ideas.
Google "aluminum foundry fire", there is plenty of examples.
It does not burn easily but once aluminum has ignited it can continue to burn under normal atmospheric conditions, especially if it begins to melt and boil.
Yes, what I've found indicates that aluminium burns when it boils, which is at 3073 °C, which doesn't happen in normal fires (house, car, forest fires), which is why it's safe in normal fires. It is melting at 660°C, at which point it doesn't burn unless perhaps if finely distributed by way of pressured air bursts or similar.
Sparks are hotter than 3073°C, but usually they don't ignite aluminium either--there not being enough aluminium vapour to sustain the reaction.
I've done what you said and googled, but have not found a case that indicates non-dust aluminium was burning.
That's seriously amazing, do you know if there are solid state breakers for high voltage? 480v is of course a lot, but if these guys were working on a distribution box outside the actual DC, I wouldn't be surprised if it wasn't fully stepped down yet.
We had an arc flash incident at a Digital Realty DC years ago. The project was to build out a data hall to be used by a prominent social networking website. Someone attempted to incorrectly terminate a wire on the UPS battery bank.. huge number of lead acid batteries all wired together. The technician had a sizeable portion of skin burnt off one arm.
My sympathy with the injured workers and their families.
The technical aspects of the accident are fascinating, of course, but the main point here is how this people will learn to adapt to their new life. And they will. They still have a loving family and friends and I'm sure that Google will do their best to support them also. They probably will need skin grafts and surgery but, as they were found breathing and conscious, I'm confident in that they will survive this horrible experience. Life deserves to be lived even if personal circumstances are hard.
