Re: Grand Apagon - Electricity (not) in Spain

On 30/04/2025 1:07 am, Martin Brown wrote:

On 29/04/2025 15:42, Bill Sloman wrote:

On 29/04/2025 10:24 pm, Martin Brown wrote:

Spain suffered a very spectacular near total loss of its national grid yesterday taking parts of France and all of Portugal down with it. This is an unprecedented failure of a supergrid system by cascade failure.
>
It seems likely they had got the effect of widespread solar PV has on load shedding wrong (much like happened in the UK) and so it failed completely. Two events a second apart delivered the coup de grace.
>
They seem to have ruled out cyber attack and the electricity company is now trying to blame "the wrong sort of temperature variations"...
>
Their 400kV lines seemed to be taking the blame with the national power company blaming exceedingly rare atmospheric phenomena due to "large" temperature differences in central Spain. They claimed that the magical sounding "induced atmospheric vibration" was to blame.
>
https://www.theguardian.com/business/2025/apr/28/spain-and-portugal-power-outage-cause-cyber-attack-electricity
>
Another marginally plausible explanation given was that different impedances on cables at radically different temperatures on different paths messed up the phasing (but the numbers don't look right to me).
>
Anyone have any idea what actually happened?
>
The only one I am aware of that can take 400kV supergrid down is cables clashing together in older pylon configs where they are exactly one above the other and resonance effects allowing large amplitude standing waves to build up in the spans can occur in 70+mph winds.
>
Most UK ones now have a longer central pylon spur so that the lines are more widely separated and up-down motion cannot allow them to touch.
>
They do sing quite impressively in a gale though. The little weights at each end are apparently there to prevent such standing wave resonances damaging the pylon structure. Without them some pylons did fall down in the distant past during the most extreme of winter storms.

>
The Guardian's science and technology reporting has never been great.

 It is about the best of the broadsheet newspapers most others are worse.
They are *only* reporting what the Spanish authorities have said.
 This is Reuters latest update:
https://www.reuters.com/world/europe/large-parts-spain-portugal-hit-by-power-outage-2025-04-28/
 Some of the blame almost certainly is with the high reliance on solar PV and dodgy load shedding algorithms that don't take proper account of it.
 

The idea that renewable sources make the grid frequency harder to manage sounds like total nonsense.

 It is pretty much accurate. The local feed in for domestic PV track whatever frequency they see on the network. The big problem is that without the large spinning generators and the energy stored in that angular momentum the frequency is able to shift rather too quickly.

If you let it. There's a lot more energy stored in grid scale battery, and you can deliver it exactly when you need to. If the algorithms that control whend an where it gets delivered aren't good enough you can obviously still have a problem, but a high power controllable source must be able to do better than a spinning chunk of metal.
Finding the people who can design a system that lives up to this isn't trivial, but the South Australian system seems to be have been able to deliver. It did cope with a major power line getting broken when very high winds knocked over a couple of pylons.

There are frequency stabilising BESS facilities around that make their living doing just that, but the domestic kit tracks line frequency and if that goes out of allowed bounds then there is trouble.
 

When South Australia bought the first grid scale battery from Tesla in 2017
>
https://hornsdalepowerreserve.com.au/
>
half of it got devoted to grid frequency management. The inverters that convert the DC current from the battery into 50Hz AC can be tweaked millisecond by millisecond to make sure that the voltage on the grid is a nice smooth sine wave. The previous grid management hardware got retired - you don't need rotating hardware to do the job.
>
They made ten times as much money out of providing this service than they did out of using the the other half of the battery to buy up grid power when it was cheap and sell it back the grid when the solar panels and wind farms weren't producing enough power (when you can get a higher price for it).

 We will have to wait and see what the engineering explanation of how it all went pear-shaped in 5s but it looks to me like a re-run of the network stability problems that took down a big chunk of the UK one sunny august evening in 2019 after a thunderbolt hit a minor power station.
 https://ukerc.ac.uk/news/august-9-investigations/
 They try very hard not to mention the fact that their load shedding algorithm was shedding a lot of local solar PV generation along with the load so that it was playing catch up and never quite getting there.

So nobody ever simulated that particular kind of disaster when they ran the acceptance tests? That's incompetence.
--
Bill Sloman, Sydney