OK, I'll try not to make this too lengthy and boring, I spent much of last night looking for a photo which might help explain part of it because I definitely emailed it to the electricity people for them to forward to Endesa and convince them that a repair was necessary, but I couldn't recover it from the emails...
Re. The 3-phase, yes its the head. The pump has to raise the water 100m and a single phase motor isn't up to it. I think there is a general theory that a motor over a certain hp is more practical or efficient as 3-phase anyway. The depth seems hard to believe but when we have replaced it the truth is laid bare by how far up the track the pipe had to be pulled to get it all out!
I have to admit that I had not heard of a floating neutral before this happened the first time but I researched the symptoms after we had a small fire and took it from there but the utility took some convincing.
The manifestation is odd, the sockets in the house will still work although problems with 3-phase equipment may arise and it can also cause damage. Basically it is caused when the distribution transformer does not have a bushing tieing the Neutral and Earth together, relying on local Earths instead. The Neutral always 'floats' a bit, it seems usual for it to be about 4v above Earth potential and varies if the load is not balanced across the phases and there can be a distribution side fault without causing a power cut but if the fault is one of the phases shorted to ground, it causes the neutral at the transformer to float a lot and the supply voltage between each phase and Neutral to be maintained.
This means that L-N is still 230v but there are some nasty surprises which is why I think it is a dangerous situation - if you take a multi meter and measure N-E under these circumstances it is about 180v and L-E is about 410v - and you are not expecting the Neutral to have that kind of voltage present, bearing in mind anything over 50v can be enough to kill. The Live at over 400v to Earth is what bothers me because although the RCD would trip if you touched something live (I hope!), a domestic RCD is designed to trip at 30mA which is considered a safe current level but is that safe at 400v? I wouldn't want to find out.
As a 3-phase customer I can go to the meter or the distribution board at the well pump control and compare the phases which will show them all elevated with reference to Earth except one which will be very low - that is the one with the fault. The phase-phase voltage will be maintained as is the domestic voltage at the house sockets.
I had to demonstrate this to the engineers on the first occasion to get them to take it seriously although at least someone was sent out. Because the power is overhead on metal poles, the wind can damage the suspension points and allow a bundle to sag, if it then rubs on the top of the tower it will wear through the insulation until it contacts and if it is a Live conductor, hey presto. The photo I was looking for showed one arcing nicely.
Re. lightning our elevated and exposed position plus the overhead power makes us prone. The towers are like bloody lightning rods! Again I've got photos of overhead cables with the insulation blown off where there must have been a direct hit. We've lost to lightning countless light bulbs, two well pump motors, two plug in timers, three internet power supplies, two routers, two UPSs, a few circuit breakers, one dishwasher, one fridge freezer, one oven, a couple of lamp ballasts and a couple of extension leads.
The extension leads in fact were interesting to observe; because it was the first occurrence of this at the time the implications were unknown in advance but the lead was outside and plugged in, despite a power cut due to the storm, being plugged in made a difference. The insulation was burst off the lead at approx 1 metre intervals and one of the conductors was actually severed and melted where this had happened at a couple of points. The lead had passed under a gate and an arc had jumped from this to the gate and earthed through the bolt, exploding the concrete step. I didn't see any of this happen since we were taking cover somewhere but the evidence was there. Now if there is ever a storm nearby everything is unplugged or disconnected, such a strike will jump circuit breaker contacts and even fuses, even if the first strike caused them to blow, so now no chances are taken.