It’s been three weeks since the blackout in Spain and Portugal on Monday, April 28, 2025.
Media stories at times follow the rules of drama.
In the Spanish blackout story, we’re waiting for a ‘reveal’.
That’s based on the assumption that one of several forensic investigations will tell us what happened.
Now, I’m as curious as anyone to know that.
But — speaking of the rules of drama — here’s a spoiler alert.
Don’t expect much out of the Spanish investigations.
Counter-intuitively, accident investigations often reduce our understanding of what happened when a complex system fails.
That’s because they draw our attention to the triggering incident.
Which people, being people, will call the ‘cause’.
Forgetting it was the system that failed.
We know a quite bit about the epistemology of accident investigations.
Accident investigations have — of themselves — been studied for decades by sociologists, psychologists, and even anthropologists.
Accident investigations produce a story of events linked together in a chain of cause-and-effect.
There’s nothing wrong with wanting to know that story.
It may be a good one.
We can confirm this analysis by looking at what Red Eléctrica has been doing since power was restored.
But knowing how something happened is not the same thing as knowing why something happened.
Or a very satisfying answer to the common question, “How could this have happened?”
Humans brains are biologically hardwired not to understand complex causality.
Which explains why certain fallacies dependably pop up in accident investigations.
Outcome knowledge, for example, permeates accident investigations like an inescapable fog.
In the 1980s, anthropologist Laurence Goldman lived with the Huli people of Papua New Guinea, and became fascinated by how the Huli resolved disputes.
Aside: Most of those involved pigs.
Among the Huli, every event is assigned an agent.
If not a human one, then a supernatural one.
There are no ‘accidents’. The Huli don’t have a word for that in their language.
A vestige of the Huli way of thinking lives on in Anglo-Saxon jurisprudence in the term 'Act of God'.
Legally, that’s an event, such as a weather event, not under human control.
Now, God seems to be doing okay these days.
But ‘Act of Gods’ are hanging on by a thread.
In the religion of climate change, weather events are now ‘attributable’.
In other words, that tornado was no accident.
The Huli would understand.
A scapegoat is a person who is blamed for the mistakes of others.
This all-too-modern practice was first described 3,500 years ago in the book of Leviticus.
When it involved real goats.
The lucky one got cast out into the wilderness. The other one got his throat slit at the alter.
In accident investigations, the scapegoat is often the victim of hindsight bias.
That’s the belief, often fallacious, that they could have or should have known what was coming.
Trial lawyers have known for years that to win over juries in accident cases the sequence of events must be laid out simply, like a chain of falling dominos.
The domino model dates back to 1931, when occupational safety researcher Herbert W Heinrich championed it was a way of understanding and preventing workplace accidents.
Lay juries ‘get’ the Newtonian physics of the falling dominos.
That’s good, but the model can introduce more subtle forms of fallacious thinking.
With the falling dominos, the cause-and-effect chain is so strong that jurors — or so the trial lawyers hope — believe the harm should have been foreseen by the guilty party.
And an accident ‘understood’ as the strong chain of events is reversible. Pulling one domino out of the chain would have prevented the accident.
If there’s an error of omission in the chain, juries often lose the plot.
A doctor neglects to order some medical test early in the course of the victim’s disease.
Jurors say to themselves: the doctor didn’t kill the guy. Cancer killed the guy. Which he got from smoking, or whatever the early dominoes say.
Starting in the 1990s, the diet of dominos was supplemented by Swiss cheese.
The ‘lines of defense’ against an accident were like slices of Swiss cheese.
You got an accident only in the rare event a set of holes in the slices lined up:
There are some telltale signs that the Spanish government under Pedro Sánchez, the leader of the left-liberal PSOE (Partido Socialista Obrero Español, Spanish Socialist Workers' Party), would be delighted to chalk up the blackout to an Act of God.
Aside: On a sailboat, a telltale is a piece of string on the mast that tells you which way the wind is blowing. In this instance, the political winds.
No single party in the Cortes Generales has a parliamentary majority.
Since 2018, Sánchez has managed to hang on as Prime Minister through a series of shaky coalition deals.
These have been the small regional parties (Basque or Catalonian) or Unidas Podemos (‘United We Can’), itself an electoral coalition of far-left parties.
The Partido Popular, the conservatives, won the largest share of popular vote in Spain’s last general election in July 2023, but not enough seats to form a government.
The next general election in Spain must take place no later than August 2027.
But an election could happen sooner should Sánchez lose a no-confidence vote in the Cortes.
A damning report on the blackout is the sort of thing that could bring that on.
The day of the blackout, Spain’s Ministry of Labor was quick to say the blackout triggered force majeure.
Force majeure suspends the obligations of both parties in a contract if something — like an Act of God — prevents them from fulfilling them.
The immediate impact on Spanish workers was they went on paid leave, and would have gotten it for up to four days if the blackout had gone on.
Prime Ministered Sánchez was quick to order a government investigation of the blackout.
But his choice of government agency to conduct that investigation may be also be telling.
It’s the one roughly equivalent to the U.S. EPA: the Ministry for the Ecological Transition and the Demographic Challenge (MITECO).
MITECO’s mission, according to its website, is to lead Spain’s fight against climate change.
MITECO also happens to house Spain’s meteorological bureau.
So a weather-related explanation may well be coming from it.
Whatever weather event it was, it was doubt made worse by climate change.
By telling a story with a clear chain of cause-and-effect, investigations into complex systems failures inescapably get things wrong.
Not in the sense that the story is factually incorrect.
But there are many paths up the mountain.
Lesson #1 in complex system failures is that the triggering incident doesn’t matter.
Meaning if it hadn’t been that particular thing, odds are it could have been something else.
The investigative story details one of many possible failure paths.
That won’t stop politicians from calling whatever made the first domino fall as the cause.
And assuring the public that fixing that one thing — if it can be fixed — will solve the problem.
Engineers, more partial to Swiss cheese, will suggest new locations for lines of defense.
Everybody will lose sight of the bigger picture.
Bloomberg reporter Javier Blas got it right when he called the Iberian blackout “The first big blackout of the green electricity era.”
What happened in Spain was a class of blackout we haven’t seen before, at least on a large scale.
Past blackouts resulted from a gross deficiency in electricity supply relative to demand.
They typically take place in summer — too many air conditioners running — or, more dangerously, in winter, when something like a freak freeze interrupts supply.
The Spanish blackout happened on a spring day, around noon, when the supply of electricity was abundant (perhaps over-abundant) and demand was low.
The analysis of the Spanish blackout as a complex system failure may sound complicated, but is really rather simple.
Red Eléctrica de España put Spain’s grid into a fragile state by putting on too many non-inertia sources of generation.
The tight coupling required for a complex system failure came from the self-protection settings of the devices connected to the grid.
Whatever exactly the precipitating incident was, after a critical mass of devices had taken themselves off the grid, there was a chain reaction, and every device took itself off the grid.
That’s it.
Time, from start to finish: 27 seconds.
Like a light bulb going out.
There’s an obscure metric used in the utility industry: System Non-Synchronous Penetration (SNSP).
SNSP is similar to — but not congruent with — renewables penetration.
Hydro is classed as ‘renewable’, but provides synchronous generation using those heavy turbines.
We’re used to discussing economic limits on renewables penetration. I did that with respect to California solar in a piece here.
As renewables penetration goes over 50% and heads toward 70%, so does ‘curtailment’ — throwing away electricity.
Batteries can — and already do — take some of the sting out of curtailment.
SNSP puts a different limit on renewables.
At present, the generally accepted number is 60%.
But, in the past few years, grid operators in the EU have been under pressure to increase this to 75%.
A grid with a lot curtailment just wastes money.
A grid with a high SNSP operates on the edge of chaos.
And batteries, of themselves, won’t help.
We can confirm this analysis by looking at what Red Eléctrica has been doing since power was restored.
Looking at Spain’s generation mix for today (May 14, 2025), solar photovoltaic is being limited to under 50%.
Here’s a link to Red Eléctrica’s web page. Note that the page gets updated often, so you need to know what time it is in Spain. To get a feel for a whole day, run the slider back and forth.
Spain’s nuclear plants are back in business. On average, they’ve been generating around 20% of Spanish electricity. They seem to get throttled back slightly during the solar peak hours.
Here’s a screen capture from my 14 May visit:
Policy-wise, the immediately sensible thing is for the ‘100% renewables’ crowd to reduce their ambitions.
Which is what Red Eléctrica has done, for now.
There are other options, which will undoubtedly get brought up in the various reports.
The tight coupling, for example, can be mitigated.
Electricity grids, as we know them today, are scary interconnected.
On June 4, 2022, a surge arrestor failed at a natural gas power plant in Odessa, Texas.
That dropped 333 MW off the Texas grid.
ERCOT, the Texas grid operator, should easily have been able to handle that.
Most grid operators prepare for ‘N-1’ events, in which some large source of generation suddenly goes offline.
The tolerable loss in gigawatts is usually set to be that of the largest generator on the system and might be 1.3 GW, the approximate output of a big nuclear plant.
But that planning is about resource adequacy, not grid frequency or inertia.
And it assumes there will be time for humans to react.
By a sort of frequency and voltage contagion — communicated through the grid — the Odessa failure knocked out a plant on the U.S. side of the border near Matamoros, Mexico.
That’s 710 miles from Odessa as the crow flies. It also knocked out the solar photovoltaic farms for 250 miles around.
Not to be stupidly pro-American, but the investigation into the Odessa incident by NERC and ERCOT was the sort we’d like to see come out of Spain.
Aside: NERC is the North American Electric Reliability Corporation which, like ERCOT a non-profit.
But for a variety of political reasons probably won’t.
Not that Texas doesn’t have politics.
Liability concerns shadow accident investigations like a dark cloud.
To such an extent that “legal and administrative concerns … contaminate the ‘facts’ which are subsequently used by scientific accident researchers and risk analysts,” in the words of Leo Tasca, a sociologist who studied maritime accident investigations.1
ERCOT faced hundreds of lawsuits seeking millions of dollars after the Winter Storm Uri blackout in February 2021.
ERCOT wiggled off that hook only by a 5-4 decision by Texas Supreme Court in 2023 that ruled that ERCOT, although a private corporation, is enough of a state agency to enjoy sovereign immunity.
In any event, NERC’s 2022 Odessa recommendations will no doubt show up in Spain.
Instead of just disconnecting to protect themselves, all those inverters at the solar farms can be re-programmed to try to ‘ride through’ frequency and voltage disturbances.
But no manufacturer of such devices wants to be singled out as the one potentially causing a problem.
The NERC and ERCOT investigation gave them witness-protection style anonymity that was almost cute. They were referred to as ‘OEM 1’, ‘OEM 2’, and so on.
Another common refrain, already being chanted by renewables zealots, is that Spain ‘just needs to add batteries.’
That’s a bit of sloppy thinking that needs to be clarified.
Batteries have no inertia. They just store power.
What is true about batteries is that they can be paired, at relatively little extra cost, with various electronics that can help out the grid.
Which requires maintaining some reserve of power that can be ‘injected’ into the grid if needed.
So a type of curtailment, just under a different guise.
Today’s inverters at solar farms sense the grid’s waveform to make their injection of power at the best point.
So they’re called grid-following.
Devices on the analog grid run in lockstep at the same frequency thanks to the fundamental laws of electromagnetism.
So the grid frequency is relatively easy to follow.
There’s nothing God-given about 50 or 60 Hz, by the way. Rather, frequency can be considered a signal.
The heavy rotating devices with high inertia work automatically — for a while — to keep the grid frequency up when the grid loses power.
Synthetic inertia is the New Hope of those who want to do away with coal, natural gas, and nuclear plants.
But first, a rant about flywheels, an ancient technology for storing rotational energy and a reductio ad absurdum of the Net Zero Project.
There are several of these in the UK, whose Net Zero folly knows no monetary bounds.
In the contemporary version, electricity is consumed to spin a heavy flywheel.
Flywheels do temporarily store and smooth out irregularly produced power, such as that from a wind farm. And they help provide grid inertia.
Which the UK needs because it has shut down its coal plants.
Whose flywheels were producing, not consuming, electricity.
But digital is better.
It has to be, because, well, it’s digital.
Just like with vinyl LPs.
In fast frequency response, an electronic device also follows the grid, but is programmed to inject power — possibly stored in a battery — in an effort to stabilize it.
Many more of those are no doubt in Spain’s future.
Helping the grid maintain its frequency shades into a discussion of electronic grid-forming, which is something else entirely.
That technology exists, but rightly faces all sorts of obstacles to widespread deployment.
Allowing those grid-formers to operate independently would certainly be a recipe for disaster.
But if they must be centrally controlled, what AI is going to do that? How will it communicate with them? What about cybersecurity?
The U.S. National Renewable Energy Laboratory (NREL) has been studying synthetic grid-forming technology for well over decade.
In a 2020 publication, “Research Roadmap on Grid-Forming Inverters”, the NREL estimated the technology was 10 to 30 years away for large power grids.
The roadmap the NREL authors propose sensibly starts small, and works larger.
First, the bugs need to be worked on ‘island’ grids.
A few of those are real islands that have lots of renewables: Ta’u in American Samoa; King Island off Australia; Maui in Hawaii; and Ireland.
Others are ‘grid’ islands, a description that fits ERCOT in Texas and — as we now know — Spain.
The important metric on the chart below is not annual renewables generation, but peak (open red boxes):
If the 100% renewables game is worth the candle — which voters, tax- and rate-payers ought to have a say about — here’s the pyramid that needs to be scaled:
I keep mentioning politics.
Spain’s are singular.
And will very likely to affect what we will see emerge from the blackout investigations.
The country, of course, has a rich — and at times tragic — history.
Spain has been a modern European-style democracy only since 1982. Franco died in 1975.
British writer Paul Preston has written an excellent history of modern Spain2.
Preston underscores three themes: (a) corruption, (b) political incompetence, and ( c ) social division.
Those didn’t go away after 1982.
An old-school corruption scandal, complete with embezzled funds and Swiss bank accounts, brought down what is now Spain’s conservative opposition party, the Partido Popular (PP), in 2018.
Corruption may be too strong a word, but the extent of the Spanish government’s ties to its power industry are likewise singular.
Three large conglomerates — Endesa, Iberdrola, and Naturgy — control 90% of the Spanish electricity market.
The three are vertically integrated and do everything from own solar farms to bill retail customers.
Iberdrola, headquartered in Bilbao, is often put forward as Spain’s ‘national champion’ in renewables.
Iberdrola is often portrayed as being in a ‘race’ with Enel — the Italian firm that owns Endesa — to become ‘the Exxon of Green Power’.3
The financial maze through which Spanish renewables are subsided is impenetrable.
The largest solar farm in Spain, Nunez De Balboa in Extremadura, has one of its major power purchase agreements with a Basque bank.
The Spanish grid is managed by a for-profit corporation, Red Eléctrica de España (REE), under a concession from the government.
REE’s parent company, Redeia Corporación, S.A., is 20% owned by Spain’s sovereign wealth fund, SEPI (Sociedad Estatal de Participaciones Industriales).
It’s also 5% owned by Spanish billionaire Amancio Ortega and 3.14% by Wall Street investment manager BlackRock. Ortega’s asset management fund Pontegadea Inversiones also owns 12% of the operator of the Portuguese electricity grid, Redes Energéticas Nacionais.
The presidency of Red Eléctrica is a clearly a political appointment.
The current president, Beatriz Corredor, was a PSOE member of Parliament until 2020 and before that (2010) served as the Housing Minister in a PSOE government.
Which helps explains why, just after the blackout, one of the first things the Partido Popular opposition leader, Alberto Núñez, did was demand that Corredor step down.
Spain has a relatively small anti-EU ‘far right’ party, Vox — 15% of the vote in 2023.
The PSOE, for its part, has been a dedicated follower of EU policy fashion. It particular, it’s ardently pro-renewables and anti-nuclear.
In 2023, the PSOE announced its intention to phase out Spain’s remaining 5 nuclear plants, starting in 2027.
After the blackout, PP opposition leader Núñez happily pounced on the nuclear phase-out issue, in an attempt to put a wedge between his party and the PSOE.
On the floor of the Cortes that brought a heated retort from Sánchez: “Those who link this incident with the lack of nuclear plants are either lying or demonstrating their ignorance.”
Sánchez was also very quick to say that the blackout was not caused by an ‘excess’ of renewable energy, nor a lack of supply to meet demand.
It should come as little surprise that Portugal’s Prime Minister, Luís Montenegro, demanded an ‘independent’ — read, non-Spanish — investigation into the blackout.
Montenegro suggested it be done by the European Union Agency for the Cooperation of Energy Regulators, ACER, based in Ljubljana, Slovenia.
The most promising investigation of the blackout is presumably the one underway by ENTSO-E, the European Network of Transmission System Operators (TSOs), headquartered in Brussels.
ENTSO-E investigations can take months, and are typically expected to be completed in six.
Red Eléctrica, being Spain’s TSO, is a member of ENTSO-E.
Which means the ENTSO-E investigation may prove to be the only path by which what Red Eléctrica knows about the blackout gets out.
But possibly not.
ENTSO-E’s official bailiwick is trans-border transmission.
If that was affected only as a side consequence of what went on inside Spain, ENTSO-E could document those effects, then draw a line at the border.
Forensic investigations do take time.
Logs need to be collected from many dispersed sites. Those will have different operators and owners.
Each piece of equipment may have different logs and different manufacturers, whose cooperation may be needed to get at it.
Then all that dispersed data needs put — very accurately — on a composite time scale, quite likely to millisecond resolution.
Only then can the event be played forward.
We’ll have a story. Now with details.
I’ll read it.
It’s got some promising dramatic elements.
There were the warnings, of course. All catastrophes needs a Cassandra.
And what was up with those spooky oscillations beforehand?
Stay tuned.
Leo Tasca’s Ph.D. dissertation at the State University of New York, Stoney Brook was called “The Social Construction of Error,” 1990.
Paul Preston, A People Betrayed: A History of Corruption, Political Incompetence and Social Division in Modern Spain, W.W. Norton, 2020.