I had to determine why a single hydro electric generator kept on going off-line last week. Turned out to be a temperature sensor in the generator gone bad or high resistance connection in that circuit. Due to improper historical logging of alarms, older protection relays, etc, it took a visit to the power plant and a few hours of downloading and analyzing event files and correlating signals with the schematics.

So a massive outage like NYC has just experienced will take weeks to collect the event files and analyze them. Probably the time was wrong in some places as well so it will be hard to figure out the exact sequence of events.