I didn't witness this, but it was a legend around the EE lab at school. One of the then-current professors had worked with 3-phase generators at a new power plant. These things were built like, bigger than, and probably stronger than, the proverbial tank; if you've seen one, you will get the picture. One of the jobs during new construction involved spinning up a generator and connecting it to the main buss to bring it online. In the olden days, synchronization was done by a set of 3 bulbs wired between the generator and the power buss, such that when the lights all went out there was no voltage differential and the generator could be safely switched in. It's not clear exactly how it happened, but the indicator was mis-wired such that it's display was 180° out of phase with the actual voltage. The operator carefully synced the lights, threw the switch, and...
The best description which came down of that event was that there was a hellacious screeching noise, and the generator ripped itself free of the floor mountings and rolled over. Most of the participants were fleeing in terror at that point, so further descriptions were vague.
One that I personally was involved with? OK. The school's power lab had a lot of very burly motor and generator equipment, which was used to train power engineers in 3-phase systems. A friend and I were hanging out before a class, and trying to make sense of some of the things we were learning. We had a heated discussion about the differences between 3-phase delta and wye systems, where the delta had a system neutral connection. He insisted that in the same distribution block, the neutrals of both systems came from a common source, should both be near ground potential and could be safely connected together (the power leads, of course, should be left isolated). I felt that was naïve, but couldn't refute his argument without further knowledge of how the secondaries were wired.
Sitting right next to us was the distribution panel for the lab's 3-phase power, with both a delta and a wye setup. I suggested, somewhat humorously, that if he was so sure he could just plug the neutrals together right there and prove it! He noted that there were circuit breakers for both systems - just run a big cable between the neutral sockets and flip the breaker, right? If there was current flow the breaker would pop and that would be it. So we switched off the breakers, ran the cable, and I insisted my friend be the one to flip the breakers on since he was SO certain.
I don't know exactly where the big BOOM came from (no, not from the lab where we were), but the lab and most of that wing of the building went dark. My friend, eyes as big as saucers, yanked the cable out and flung it into the box where the cables were stored. People came boiling out of the classrooms nearby, in a panic because the rooms had all lost power - and most of them had no windows. I thought we were going to be in big trouble, but thankfully no one seemed to connect us with the event. The professors were running around checking all the breaker panels, and finally found all the right ones to bring power back. It astonished both of us that a test panel, clearly used for high power experiments, affected so many essential circuits - I would have expected it to be split off near the incoming circuits out at the power drop. Perhaps the ground currents were so unbalanced that all the other breakers saw significant excess current?
There was no apparent damage, either, despite a significant number of circuits popping their breakers and lots of sensitive test equipment being run in the area. But no one who was in the EE department that day will forget the incident!