> how likely is simultaneous failure of 9 out of 30 of them? Totally unlikely if you ask me.
It's quite possible because their failure probabilities aren't independent. A geomagnetic storm could take most of them out at once. Like respiratory pandemics, this is an event that happens every ~century.
But it will be known well in advance and can be mitigated by simply shutting the grid down for a few hours. Only places which will experience difficulties are those having a large fraction of nuclear power since a reactor can't be quickly restarted once shut down, so power will restore only in some days.
Do you have a cite? My impression was that we do not have sufficient preparations in place to do this.
(E.g., we may have satellites that can be taken as weak indicators, but they are not designed for this, and there is no clear decision authority starting from possibly ambiguous satellite data to prophylactically shutting down power for 300M people against their will. Hopefully, recent events have made clear that "leaders will get their act together and act decisively during an emergency" cannot be the default assumption.)
It's quite possible because their failure probabilities aren't independent. A geomagnetic storm could take most of them out at once. Like respiratory pandemics, this is an event that happens every ~century.
https://en.wikipedia.org/wiki/Geomagnetic_storm