It was nice to see the stars though.
* A power company failed to trim trees from its power lines properly. Four of their major transmission lines failed in one afternoon due to shorting out via tree.
* The software bug you mentioned caused a failure to alarm the power company of two of their line failures. The first and fourth failures were alarmed in real time.
* A separate issue rendered the regional grid operator's software modeling real-time grid instability inoperable for most of the afternoon. Crucially, if this had been running, the operator would have realized that the failure of one more line would have created grid instability requiring immediate action.
* The final line trip set off a cascade of overloads trips until Sammis-Star overloads and trips. This takes about 15 minutes, and analysis suggests that the Sammis-Star trip is when the blackout became inevitable.
* Sammis-Star triggers lines to fail one-by-one from east to west, until it severed every line in Ohio and Michigan. This causes a large power surge to go from Michigan through southern Ohio, Pennsylvania, New York, Ontario, back into Michigan and then into the final demand of Cleveland.
* Essentially every link along this spiral fails in the course of a few seconds, creating several grid islands. Whether these grid islands black out is dependent on the local mismatch between generation and consumption.
(Full story from https://web.archive.org/web/20120318081212/http://www.nerc.c...).