Breaking point

After a second review of my previous analysis and the official reports from REE and the Spanish government, I believe the following information I am now sharing marks the point where this issue clearly goes far beyond personal curiosity. 1.2 GW of DER generation was lost within a 10-minute window, 90 minutes before the blackout, for no known reason.

1.2 GW

Excerpt from the Spanish Government report (page 101):

"The 11:00 deviation, based on the available information, cannot be explained by generation behavior, but rather correlates with an unforeseen and sudden deviation in demand, which increases by about 1,200 MW with no known cause and without resembling nearby days. This “anomalous” increase in demand, marked in the graph below, is compatible with a temporary reduction in distributed and small-scale generation (including self-consumption) on lower voltage networks."

In the following image, you can see two elements A and B:

A) The official graph from the Spanish Government report shows an anomalous 1.2 GW increase in demand, equivalent to the output of a nuclear power plant. This indicates that the generation which had been matching that demand was suddenly lost. Crucially, this lost generation did not originate from large plants connected to the transmission grid, but rather from self-consumption and small-scale industrial solar inverters. The blue circle highlights the moment demand suddenly increases. As the inverters momentarily ceased to operate due to an unknown cause, the loads began drawing power directly from the grid, so demand increased.

B) This is a segment of the graph, starting at 11:00, that I created to identify Momentary Cessation and Trip patterns in the telemetry from thousands of residential (rooftop self-consumption) and small-scale industrial inverters on the day of the blackout. This telemetry data was shared with me by an anonymous, but reliable, source.

As the reader can verify, these patterns correlate exactly, including the two-peak pattern, with the anomalous 1.2 GW increase in demand. At 11:00, there were no reports of relevant oscillations.

But this is not the only precise coincidence.

Excerpt from Spanish Government report (p. 85)

"In the inverters connected to low voltage (Graph 36), alarm spikes associated with inadequate voltage levels are detected around 12:10 and 12:27, which coincide with a reduction in the aggregate generation represented by these inverters as a whole."

Same scenario, same significant peaks at 12:10 and 12:27 in my graph. This means that, once again, hundreds of megawatts of distributed generation were lost due to inverters entering momentary cessation or tripping.

Although this dataset of inverters may represent only a small portion of the total, it constitutes a statistically significant sample. As I previously mentioned, I chose to share these findings because I was confident that, when scaled to the full deployment across Spain, the percentages and other figures I’ve presented would remain consistent. I believe the figures for distributed generation losses included in the Spanish government’s report confirm this.

It's important to note that I created this graph and shared it with the relevant Spanish authorities many hours before the official government report was published, as I believed it could be relevant to the investigation.

For additional context, I previously explained the approach and background of my analysis in the following posts:

https://www.linkedin.com/pulse/report-season-here-ruben-santamarta-big6f

https://www.linkedin.com/pulse/quest-initiating-event-ruben-santamarta-kthlf

Conclusions

I believe there is sufficient evidence to justify an assessment by the relevant authorities, as no power grid incident of this scale has been reported in any other country.

I strongly believe that the relevant authorities should urgently require immediate access (if this hasn’t already been done) to the logging data from all inverters deployed across Spain.

This is essential to investigate (under any conceivable scenario) how it was possible for 1.2 GW of residential (rooftop self-consumption) and small-scale industrial photovoltaic generation to 'disappear' within a 10-minute window, just 90 minutes before the blackout, with a similar loss occurring just minutes before the total collapse of the system.