Spanish grid saying solar is a likely cause (though the PM seems to be disputing this 'in his view', but may be a poor translation from BBC).
My personal hunch is they were struggling to balance the grid as peak solar approached with too much supply. The last snapshot of data showed less than 1GW of CCGT which is not a lot for a grid of this size (the UK runs 2.5GW CCGT 24/7 for inertia and balancing reasons, even if there is excess renewables available - they will be paid to be turned off). Spain does also have hydro which can respond to the grid in fairness, but I am unsure how well prepared they would be to do this if CCGT is the main balancing source.
I also saw data showing Spain was exporting to France heavily at the time. I suspect if that interconnector went down (as rumored), you'd have potentially 1GW+ more load with nowhere for it to go.
This may have then caused the disconnection events (the spanish grid operator said there were 2), as various plants trip as they saw the grid becoming unstable. They recovered from the first but couldn't recover from the 2nd, there isn't a huge amount of detail on these.
This is basically exactly what happened in the 2019 outage in the UK (and from comments on HN yesterday was also a major factor in the Brazilian outage recently) which affected SE England and London (but didn't cause a full grid collapse, I suspect averted because of more CCGT online to balance).
In that event:
1) 400kV cable goes down. Power is rerouted. Was a lightning strike which is pretty common.
2) This caused a very brief change in the voltage/frequency on the immediate substation that was taking power from Hornsea 1 offshore wind farm to the national grid
3) Hornsea 1's voltage detection equipment was too sensitive and tripped, dropping 700MW of supply. It shouldn't have tripped for such a short drop while they rerouted grid power to bypass the out of service 400kV line.
4) This then caused a cascade of other generating assets to trip (also too sensitive), causing a major drop in grid frequency and this then triggered various distribution points in England to disconnect automatically to try and drop load by 5%. Furthermore all 'variable' generation is requested to immediately raise output to balance the system while load shedding goes on.
5) The load shedding operation did not work as expected. They expected drop was 850MW but only 350MW was achieved. This is likely due to embedded solar which they hadn't taken into account in their calculations (or one of the reasons). This causes frequency to drop further and another level of load shedding to occur.
6) Frequency recovers.
Even worse, there was another potential near miss here - the DNO started reconnecting customers before they had the goahead to connect from the grid. It is lucky that another outage didn't occur because of this as the grid was not fully checked and they should have waited.
Anyone saying that renewables are not affecting grid stability are unfortunately just wrong. I am huge proponent of renewables but we have the ratio of solar/wind/batteries/other inertia (flywheels) really wrong. We should not be running so much variable capacity with so little storage.
My personal hunch is they were struggling to balance the grid as peak solar approached with too much supply. The last snapshot of data showed less than 1GW of CCGT which is not a lot for a grid of this size (the UK runs 2.5GW CCGT 24/7 for inertia and balancing reasons, even if there is excess renewables available - they will be paid to be turned off). Spain does also have hydro which can respond to the grid in fairness, but I am unsure how well prepared they would be to do this if CCGT is the main balancing source.
I also saw data showing Spain was exporting to France heavily at the time. I suspect if that interconnector went down (as rumored), you'd have potentially 1GW+ more load with nowhere for it to go.
This may have then caused the disconnection events (the spanish grid operator said there were 2), as various plants trip as they saw the grid becoming unstable. They recovered from the first but couldn't recover from the 2nd, there isn't a huge amount of detail on these.
This is basically exactly what happened in the 2019 outage in the UK (and from comments on HN yesterday was also a major factor in the Brazilian outage recently) which affected SE England and London (but didn't cause a full grid collapse, I suspect averted because of more CCGT online to balance).
In that event:
1) 400kV cable goes down. Power is rerouted. Was a lightning strike which is pretty common.
2) This caused a very brief change in the voltage/frequency on the immediate substation that was taking power from Hornsea 1 offshore wind farm to the national grid
3) Hornsea 1's voltage detection equipment was too sensitive and tripped, dropping 700MW of supply. It shouldn't have tripped for such a short drop while they rerouted grid power to bypass the out of service 400kV line.
4) This then caused a cascade of other generating assets to trip (also too sensitive), causing a major drop in grid frequency and this then triggered various distribution points in England to disconnect automatically to try and drop load by 5%. Furthermore all 'variable' generation is requested to immediately raise output to balance the system while load shedding goes on.
5) The load shedding operation did not work as expected. They expected drop was 850MW but only 350MW was achieved. This is likely due to embedded solar which they hadn't taken into account in their calculations (or one of the reasons). This causes frequency to drop further and another level of load shedding to occur.
6) Frequency recovers.
Even worse, there was another potential near miss here - the DNO started reconnecting customers before they had the goahead to connect from the grid. It is lucky that another outage didn't occur because of this as the grid was not fully checked and they should have waited.
Anyone saying that renewables are not affecting grid stability are unfortunately just wrong. I am huge proponent of renewables but we have the ratio of solar/wind/batteries/other inertia (flywheels) really wrong. We should not be running so much variable capacity with so little storage.