On Tuesday I wrote about Monday's emergency call to France: up to 1,200 MW of Emergency Assistance, the largest on a British interconnector since at least 2023, made because power could not get into London and the South East. In that piece I mentioned in passing that it was not the first call of the week. On Friday evening, 9 October, NESO had taken 270 MW from France for seven minutes, from 18:50 to 18:57, and published the notice at 20:04, after it was over. It came over IFA, the largest cable to France, which lands in Kent. That is the same cable NESO used for Monday's call, and for the 500 MW it took on 23 June: all three of NESO's requests for Emergency Assistance this year have come through it.
I held Friday back on purpose. Monday's story could be told from data published on the day: boundary limits, outage notices, interconnector flows. Friday's could not. Friday had something Monday did not, a frequency excursion, and that told me the whole system had been short of power for a few minutes. It could not tell me who was short. The answer is in Elexon's settlement data, which records what every metered generator actually produced in each half-hour, and which is only published a few working days after the event. The first run for Friday, the Initial Interim run of the P114 files, arrived at 06:45 on Wednesday 14 October. This piece is built on it. It points somewhere I did not expect to be looking.