The “pressure” analogy is probably more useful more often, but this analogy seems sufficient for this model.  Models are not reality, voltage is not pressure — voltage is LIKE pressure, and voltage is LIKE pressure in more ways than voltage is like depth, but, in THIS explanation, “depth” is sufficient.

First off: The oil in transformers is used to insulate the high voltage coils (they’re just bare wire) and cool them by conducting heat to the metal case via convection.  They used to use Polychlorinated biphenyl PCB’s because they are relatively flame resistant and a good insulator.  Trouble is some PCBs are horribly toxic.  So now they use other somewhat more expensive stuff.

Second: Line Transformers all have a thermal circuit breaker and a fusible link to protect the transformer if it’s over loaded (drawing too much current).  If you short the secondary side of the transformer or there is a phat short circuit inside then the breaker trips, failing that hopefully the fusible link melts.  Lot of outages are due to breakers tripping.  In which case the lineman climbs the pole and resets the thing and suddenly you have access to p0rn again.

Third: What happens when a transformer blows up is a arc caused by a voltage surge, age, whatever, etc between a set of coils.  In that case the magnetic field induces massive current through the shorted coils.  They heat and they melt, the oil inside carbonizes and no longer insulates.  It’s possible for this to go on for a long while as power is dumped inside the transformer via arcing* instead of being transferred through it. As long as the amount of current being drawn isn’t too much the breaker won’t trip.

*If you’ve ever seen an arc welder in action, that’s what’s going on.

Forth: Circuit Breakers.  They all have a minimum current at which they will trip.  and a maximum current that they can interrupt.  Higher than that and they usually fail.

The melting point of pure copper is 1,095 degC (or 1,981 degF for those people still living in the Pleistocene Era).  

If the current is high enough to melt the transformer’s windings/wiring – the molten/vaporized copper will ignite the oil that touches it – assuming that the ignition point of oil is less than 1,095 degC, which is definitely the case. 

The now-ignited oil will burn at a higher-than-its-ignition-point temperature. This will cause the oil molecules right next to the copper-touching oil, to ignite.  Rinse, repeat. 

This ignited (as in ‘small-explosion’) oil will cause the rest of the oil be turned into a mist.  This increases the surface area of the oil exposed to oxygen, increasing the rate of reaction – Kaboom!

If you check the video, you can see the explosion happens in a few phases.

The colour of the light given off by the chemical reaction / explosion tells you what temperatures are achieved – the blue part of the flame in your propane torch is somewhere around 3,000 degK. That substation very likely achieved much higher.

Jokes aside, is this really just a skewed world view or are European installations less likely to fail so spectacularly? If so, what is different in the European and American transformer and fuse designs? This would probably make a very informative comparison.

I would suggest, however, that there are good transformer analogies if look at air powered gas booster pumps — http://www.hydratron.co.uk/media/html/gas_pumps.htm – or air amplifiers — http://sprague.cwfc.com/Products/spokes/Amplifiers.htm

In each case, these convert one pressure/volume to another pressure/volume in part by trading off one for the other. In the case of the gas booster pumps, it does it without actually connecting the “primary” gas supply with the “secondary” boosted gas. There is still a mechanical connection, but there is no connection of the “working fluid” on each side of the device, in much the same way that a properly operating transformer does not create regular electrical connections between the primary and secondary windings.

The transformer does not create a mechanical connection, true. But this is an analogy, and the gas booster pumps don’t create a magnetic connection because they don’t work in the electromagnetic domain. I will note that most transformers do not have an active mechanical connection, but they do have some form of mechanical structure to ensure they are correctly constructed and aligned.

The step-down side exists as well. The mixing version (an impure analogy, I agree) would be the Dyson ring fans, or a locomotive stack blower (had to look that one up), both a form of venturi blower. There are simple fan blowers than use high-pressure air jets to power the spinning blades. If you want the non-mixing aspects, then there air motor powered blowers — the blower discharge is separate from the air motor discharge.

There are a couple of things that bother me about this post. You said

Electricity has to flow along it from the power plant, to the customers, and back around to the power plant again.

This isn’t correct. This may seem to be correct if you were talking about a 1 phase grounded system, but it doesn’t apply to 3 phase ungrounded systems that power plants use. And considering that there will be transformers that will electrically isolate and convert from 3 phase ungrounded to 1 phase grounded, it can’t be true.

Second, your mechanical transformer analogy doesn’t work. It implies that there is an electrical connection between the coils. That is not true, the connection is through magnetic flux. And it doesn’t explain that the power input is almost equal to the power output. I’ve worked in electrical generation for many years. There is no decent mechanical analogy for a transformer that I have heard. Trying to make one work will only confuse people about the key characteristics of a transformer which include electrical isolation and near conservation of power.

Edit: I meant to say “or there are hot spots”. Most transformers explode due to old age under nominal conditions when hot spots (part of the coil that has heated up far beyond normal due to degradation) produce too much acetylene. Another option which doesn’t occur often is the build up of static charge due to the movement of the transformer oil. One interesting potential failure mode (which would be taken care of due to protective features) is an overcurrent condition caused by solar storms on transformers with a neutral ground. Power plants sometimes have to limit output during solar disturbances due to this additional current.

Why a transformer fails really involves what is causing it to be overloaded.  If it is a faulty disconnect circuit then the transformer simply has to deal with excess power input or draw leading to heat and breakdown of the mineral oil.  If it is from an external source of power input, like lightening, then it is very possible that the internal wiring becomes shorted and destroys everything by simply becoming the faulty part and drawing excess power.  (Of course the same thing can happen if a high enough voltage becomes present on the lines going to the transformer.)

The pretty blue arc isn’t the transformer itself failing, as much as it is the excess voltage arcing between the conductors.  (Voltage that the safety system should have disconnected.)

Maggie’s been consistent about “voltage as depth” and the consistency is probably more important from a science journalism point of view than any nitty-gritty details about why physics profs like to use the pressure analogy instead.

Current (or amperage) is what melts fuses. So, when the ultra-high current surge (equivalent to volume-flow of water) is high and fast enough, it melts the transformers’ wiring, or even vaporizes them.

Big substation transformers are usually cooled by oil – as in “the transformer is sitting in an oil bath”. Molten/vaporized metal meets oil…the oil explodes. Usually in a very spectacular fashion.

Because I know you’re asking: If the power-plant-end is connected to ground, it’s the power plant and transmission lines which are connected to ground. These can ‘usually’ survive the trauma, since amperage is lower (y’know, ‘cuz the voltage is much higher). In really bad freezing rain conditions, grounding the power transmission lines is a method used to melt ice off of the lines – and is a procedure that would be performed at a substation.

Why, for our YouTube entertainment of course! You mean those things actually do something else?

I think depth is a reasonable part of the rivers analogy.