Rinnovabili publishes the exclusive first release of the analysis by Berizzi and Delfanti from the Department of Energy at Politecnico di Milano on the Spain blackout. This reconstruction is based on real-time monitoring data and recordings from experimental equipment linked to the MedFasee research project, coordinated by INESC P&D Brasil, with the participation of the Department of Energy at Politecnico di Milano.
By A. Berizzi, M. Delfanti – Department of Energy, Politecnico di Milano
The first major spain blackout in a renewable-powered grid
The blackout that affected the Iberian Peninsula on April 28, 2025, marks the first major incident in an electric power system primarily powered by renewable energy sources (some commentators, somewhat simplistically, referred to it as “the first blackout of the green era”).
Comprehensive analyses are currently underway to enable a detailed reconstruction of the event, led primarily by the association of European transmission system operators (ENTSO-E), which has launched an investigative process expected to last several months.
In the meantime, a few days after the incident, it is already possible to outline a plausible reconstruction. The one proposed in this article is based on real-time monitoring data (such as those from Red Eléctrica de España – REE – and the ENTSO-E transparency platform), as well as on recordings from experimental equipment used in the MedFasee research project coordinated by INESC P&D Brasil, with participation from the Department of Energy at Politecnico di Milano.
The article begins with a detailed technical overview of the Spanish grid, including its structure, generation mix, and cross-border interconnections, then focuses on the factors influencing grid stability. It presents a possible sequence of events that occurred on April 28, 2025. It also draws a parallel between the April 28, 2025 event in Spain and the similar blackout that occurred in Italy on September 28, 2003, followed by an outline of some specific measures adopted in Italy.
The article concludes with some preliminary insights and future perspectives, also in light of the profound transformation currently affecting electric and energy systems around the world.
Generation mix of the Spanish power system
The total installed capacity of the Spanish power system is approximately 125 GW, with over 60% coming from renewable sources such as wind, solar, and hydro, while the remaining share comes from nuclear and natural gas plants.
Specifically, among renewable energy sources (RES), wind power is the leading generation source, with around 30.8 GW installed, accounting for about 25% of total capacity. Solar energy has seen rapid growth, reaching about 30.1 GW of photovoltaic capacity, representing 21% of the total; in addition to photovoltaics, several gigawatts of solar thermal capacity are also installed. Lastly, hydropower contributes approximately 17 GW, supplying between 10% and 15% of the annual electricity production. Notably, there are about 3.3 GW of pumped-storage plants that also help store energy during periods of surplus generation.
Nuclear power plays a crucial role with 7 reactors (approximately 7 GW of installed capacity) providing about 19% of the generation, while natural gas plants—particularly combined cycle gas turbines (CCGT), essential for system flexibility—account for around 30 GW of installed capacity.
In terms of energy produced in 2024, renewables supplied about 59% of Spain’s electricity (Figure 1).
Interconnections with neighboring systems
Spain’s electricity grid is interconnected with those of neighboring countries at key points, as shown in Figure 2.
The main interconnections are:
Spain–Portugal. Spain and Portugal operate as a single synchronous zone, often referred to as the Iberian system (commercially reflected by the MIBEL, a unified electricity market). Several high-voltage (400 kV) lines link the Spanish and Portuguese grids, with a transfer capacity of approximately 4.5 GW.
Spain–France. This is the main link between the Iberian grid and the rest of continental Europe. It consists of several AC lines crossing the Pyrenees and one 1,400 MW HVDC connection (Santa Llogaia–Baixas). The total capacity of this interconnection is around 2.8 GW in both directions under ideal conditions. This capacity is relatively limited, representing less than 10% of peak load.
Spain–Morocco. Spain is also connected to Morocco through two 400 kV AC submarine cables across the Strait of Gibraltar, with a combined transfer capacity of about 1,300 MW. These links synchronize Morocco’s grid with the Iberian (and European) system, making Morocco part of the broader synchronous area of continental Europe.
Factors influencing grid stability
The various technical factors affecting the operation of the Spanish grid—similar to any grid under the ENTSO-E framework—are listed below.
Frequency control and reserves: To manage normal operating fluctuations, Spain, like other ENTSO-E members, uses automatic frequency control systems (primary and secondary reserves).
Voltage and reactive power control: The Spanish grid uses capacitor banks, reactor coils, and generator excitation controls to maintain voltage within limits. Voltage instability can occur if a large power transfer is suddenly lost or if there is a shortage of reactive power support, leading to voltage collapse in some areas.
Role of interconnections: The relatively weak interconnection with the rest of Europe is a known limitation of the Iberian system. In fact, in 2021, a significant disturbance in Spain caused the entire Iberian Peninsula to lose synchronism with the rest of Europe and led to a separation of the grid.
Grid topology and N-1 vs. N-2 events: The grid is planned and operated to handle an N-1 event (a single component out of service, such as a plant or a line) without load shedding. If two independent events result in the loss of two grid elements, the system may need to shed load—according to security criteria—to prevent a large-scale blackout. Load shedding should still avoid a domino effect, thanks in part to defense plans that use under-frequency load shedding and frequency derivative-based relays.
System inertia: Inertia is the power system’s resistance to frequency changes, typically provided by the rotating masses of synchronous generators. In Spain, inertia traditionally came from large thermal (nuclear, gas, coal) and hydro units. With the increase in renewables and the resulting limitations on dispatching large synchronous generators, grid inertia is decreasing. Wind turbines and solar PV are mainly inverter-based, which limits their contribution to system inertia—unless supported by specific inverter controls.
Short-circuit power: Inverter-connected sources also contribute less to short-circuit power than large rotating generators, making the grid more vulnerable to voltage collapse and instability triggered by transient voltages.
New stability tools: The Spanish TSO, REE, is integrating new tools to enhance stability: synchronous condensers, BESS capable of rapid frequency response, and smarter renewable controls (some of which are in pilot phases). For now, these measures appear to still be under development and not yet fully implemented in practice.
System status on the morning of April 28
Just moments before the blackout, with a load of around 30 GW, generation consisted of approximately 3 GW from wind, about 19.6 GW from photovoltaic solar (for a total of about 70% of generation connected through inverters), 3.5 GW from hydropower, 3.4 GW from nuclear power, and 2.2 GW from combined-cycle gas turbines. About 3 GW were absorbed by pumped-storage hydropower plants. The load trend and the generation employed are shown in Figure 3 (data after the event are considered unreliable).
The physical power flows recorded at the interface in question—between Spain and France—are available through the ENTSO-E transparency platform, as shown in Figure 4. In particular, it can be seen that Spain, and thus the Iberian Peninsula, was exporting more than 1 GW, though decreasing, to France. In the hours following the blackout, imports from France enabled a quick grid re-energization process.
Inter-area oscillations
Late in the morning on April 28, significant electromechanical oscillations were recorded by PMUs from the MedFasee project already mentioned. These are frequency oscillations around the nominal value, characteristic of geographically extensive systems in which portions of the grid can oscillate relative to one another. Specifically, two different events were observed: one at 12:04 (lasting 4 minutes, with a frequency of about 0.7 Hz, clearly associated with a generation loss), and another at 12:19 (lasting 5 minutes, with a frequency of about 0.2 Hz, the latter clearly attributable to an inter-area oscillation in the European electric system).
At present, the connection between these oscillations—recorded minutes before the incident—and the blackout itself remains unclear. A possible interpretation is that they indicate a general weakness or vulnerability of the system even before the main disturbance occurred.
A possible reconstruction of the incident
The following is based on data and information with a minimum degree of reliability available to date. Naturally, to definitively determine what happened, the exact sequence of events is needed, which is not yet known (data are recorded at different locations by REE’s SCADA systems). For now, only a few hypotheses considered plausible by the authors are presented and correlated with the available measurements.
In the minutes leading up to the blackout, as occurred, the Iberian system experienced various types of electromechanical oscillations. Of particular interest is the inter-area oscillation at 12:19, weakly damped, around 0.2 Hz. To counter this phenomenon, the Spanish TSO (REE) likely began reducing exports, as suggested by the Spain–France transfer data (Figure 4). This hypothesis is confirmed by the frequency trend (Figure 6): although the exact moment of the separation between France and Spain is unknown, there is no frequency increase observed in the Iberian Peninsula nor any significant frequency drop in the rest of the European grid (measured in Italy by MedFasee equipment), which would be expected if the disconnection had occurred during significant export from Iberia. Therefore, based on the frequency trend (a decreasing profile), it seems unlikely that the initial event was the complete loss of interconnection with France.
A more likely scenario is a sudden production deficit, corresponding to the observed frequency drop.
More specifically, while waiting for official reconstructions, Figure 6 shows that at 12:32:57, the frequency in the Iberian Peninsula begins to diverge significantly from the center of the European system. The rapid frequency collapse begins at 12:33:16 and ends with the blackout 6 seconds later.
As is well known, a deviation from the nominal frequency of 50 Hz indicates an imbalance between load and generation. This imbalance is initially moderate in the first 4 seconds but becomes increasingly severe in the final 2 seconds.
The distinction in frequency curves between the PMUs installed in Portugal and Italy indicates that at that point, the Iberian Peninsula had become electrically separated from the rest of the system, with a growing power deficit. This frequency decline, possibly partially counteracted at 12:33:21 by the activation of the grid defense plan, quickly leads to the blackout.
Disconnecting rotating generators at very low frequencies (e.g., below 47.5 Hz) is a necessary measure to prevent damage and keep them ready for later grid recovery.
Analysis of key events
The available data show that, likely a few minutes before 12:33, REE was carrying out operations on some generation plants, possibly in response to observed grid oscillations. It is reasonable to assume that these maneuvers involved a sizable and dispatchable plant, capable of affecting the export flows being reduced. It is possible that one of these actions (or an error during one of them) unintentionally caused an excessive power reduction (initially cited by REE among the causes of the incident), which led to a frequency drop. Additionally, between 12:25 and 12:30—before the blackout—a unit at a nuclear power plant in France, near Toulouse, became unavailable (source: EDF), suggesting the presence of a technical issue.
At this point, it is also plausible to assume the activation of protections on the interconnection lines with France. During major incidents, the European grid aims to support the affected area, but protections must block disturbances from spreading. This likely caused the spain blackout grid separation, triggered by relays on the interconnections—as confirmed by RTE and shown by diverging frequency curves in Figure 6. A similar separation had already occurred on July 24, 2021, though it did not result in a blackout.
The Spanish grid remained isolated and therefore dependent on its own stability, with an imbalance that was not initially severe. However, the frequency collapse indicates that the power deficit—unlike in the 2021 event—worsened, likely due to further generation losses.
Voltage trends also offer further insight (Figure 7). As observed, voltages tended to increase in the minute before the blackout. This could be attributed to a gradual load reduction on the grid (unlikely except perhaps in the final moments) or, more likely, to a progressive loss of voltage control, related to the disconnection of generation units that had previously been regulating voltage.
Several observations must be made here: the actual frequency collapse occurred very rapidly (with a rate of change exceeding 1 Hz/s at its worst) and, as noted, the entire phenomenon lasted 6 seconds. This suggests several critical aspects, listed below:
- The speed of the phenomenon suggests that:
- the initial imbalance was very significant and worsened during the event, and/or
- system inertia was low, as there were not enough devices capable of providing inertia (rotating machinery or synchronous condensers, possibly with flywheels),
- fast frequency regulation systems were absent or insufficient (e.g., FFR – Fast Frequency Regulation – from BESS units).
- The primary frequency regulation response (assuming reserves were sufficient and not already exhausted before the sudden collapse) did not have time to act effectively.
- The progressive worsening of the deficit and the resulting rapid frequency decline suggest that the defense plan response was likely insufficient. Possible reasons include:
- the emergency load-shedding plan did not activate in time, or
- the plan failed to produce the desired effect, possibly disconnecting distribution feeders that were injecting photovoltaic power into the transmission grid, thereby worsening the initial imbalance.
Parallels with the Italy blackout of September 28, 2003
Assuming the reconstruction above is correct (others exist but are less credible based on currently available data), there are significant similarities between the event recorded in Spain at the end of April 2025 and the blackout that occurred in Italy in late September 2003.
It is appropriate to draw a brief comparison between the two events, as the 2003 incident has been thoroughly analyzed from all technical perspectives by multiple independent commissions (notably, Politecnico di Milano—then the Department of Electrical Engineering—provided technical support to the independent regulatory authority, then AEEG). Furthermore, it is useful to examine what technical and regulatory countermeasures were adopted in the Italian case. Lastly, there are significant structural similarities between the Iberian and Italian systems:
- both are geographically significant “appendices” of the central European synchronous area (in relation to the natural center of gravity of the continental power system);
- both systems have either already integrated (Spain) or are planning to integrate (Italy) a large share of renewable generation (static generation).
The key shared features of the two incidents are summarized below.
Initial event. Both blackouts began under stressed system conditions. In Italy, the critical situation was triggered by a tree flashover (on the Lucomagno line [1]), while in Spain it was due to a planned reduction of internal generation (a fault on a line at the French border was ruled out by RTE).
Grid separation. In both cases, the initial stress led to the national grid disconnecting from the rest of the European system. In the Italian case, the disconnection occurred through line protections across the Alpine arc, involving multiple neighboring countries (France, Switzerland, Austria, Slovenia); in Spain, line protections likely acted correctly both on the northwestern side of the French border and on the southern interface, which already had limited capacity.
Generation-load imbalance. Large imbalances between load and generation caused frequency collapse in the isolated portion of the grid (lasting about 90 seconds in Italy, six seconds in Spain), followed by system shutdown. In both cases, low grid inertia was cited as a critical factor, as were the systems’ frequency regulation capabilities (in 2003, fast regulation functions were not yet part of the discussion).
Cascading disconnections. The sharp frequency drop triggered a cascade of generation losses. In Spain’s case, the broader adoption of inverter-based systems (virtually nonexistent in 2003) highlights the need to update or maintain protections (originally designed for local faults in MV networks) to meet today’s critical transmission system stability requirements, as well as defense plans. In both cases, untimely generator shutdowns and disconnections were a key concern.
Following the technical investigations into Italy’s blackout, several recommendations were made regarding system operations, both at the international level (then under UCTE) and nationally. At the UCTE level, key points included the need for mandatory emergency procedures, harmonization of the N-1 criterion (especially regarding maximum recovery time to secure N and N-1 conditions), better assessment of voltage stability, and improved real-time data exchange. The harmonization of minimum requirements for generating equipment, defense plans, and restoration plans was emphasized, as was the evolution of frequency/power control strategies. Strong encouragement was given to the adoption of WAMS systems for dynamic analysis and monitoring (thanks to WAMS, today we have a wealth of synchronized data on the recent disturbance).
At the national level, attention turned to stricter grid code requirements for generators, as well as improved coordination and structure of defense plans.
Generators and grid stability
From the comparison outlined above, it becomes clear that the role of generators is critical to avoiding disruptions of this kind, particularly:
- in the initial phase of the disturbance, especially if capable of providing system inertia, thereby making the frequency transient less abrupt;
- in the propagation phase of the incident, when their task is to remain connected even under degraded conditions, to allow defense plans to be deployed and ultimately to keep parts of the system in service with a view to quickly restoring normal operating conditions.
Focusing on these two potential weaknesses (providing inertia; avoiding unintended disconnections), it is worth noting that several measures have been implemented in Italy to enhance system resilience, even with a high share of static generation.
To provide inertia to the system, the installation of synchronous condensers has been planned and partially executed, playing a fundamental role in improving the stability and resilience of the Italian grid. Storage systems can also offer fast frequency regulation services (FFR), stabilizing the grid during rapid transients. Some of these systems have already been installed; more generally, over the next few years, according to the Scenario Description Document (DDS) by Terna-Snam (aligned with the PNIEC), Italy will deploy storage systems totaling 71.5 GWh by 2030 [1]. This value excludes existing pumped hydro and represents the storage capacity required to effectively integrate renewable energy into the electricity system.
On the front of preventing unintended disconnections, it is important to recall that historically, all generators connected to medium and low voltage networks were equipped with interface protections designed to disconnect them very quickly (in fractions of a second) in the event of even slight frequency or voltage anomalies (+/- 300 mHz). Similar protections were in place across other European countries.
These rules, once suitable for grids with limited renewables, became unsustainable as their share grew. To prevent disconnections like in the 2006 European split, Europe adopted new technical standards for static generators and launched a wide retrofit program.
In Italy, these new requirements have been enforced through Annex A70 of Terna’s Grid Code, which defines specific criteria for renewable generators, including frequency and voltage ride-through capabilities. Moreover, a wide-ranging and rigorous retrofit effort has been undertaken to ensure that both new and existing static generators are able to remain operational during voltage and frequency fluctuations.
Specifically for Italy, it is worth highlighting the strong collaboration between the regulatory authority (ARERA), the technical standardization body (CEI), and all relevant stakeholders (TSO, DSO, producers), which has equipped the Italian system with technical connection rules (Grid Code; CEI 0-16 and CEI 0-21) that can rapidly incorporate the need for adaptation to new technologies, serving as a continental benchmark.
Conclusions
The considerations developed in this article lead to some preliminary and certainly partial conclusions: only the analysis process recently launched by ENTSO-E will be able to provide definitive insights into the April 28 disturbance.
The incident stems from a system-level issue—not attributable solely to generation (including renewable generation), nor solely to the grid, but to the electric system as a whole, which is undergoing the most radical transformation in the past 80 years. Key elements of this transformation include: the increasingly vast geographical span of the European synchronous system (Morocco to Ukraine; Finland to Turkey); loads that are increasingly composed of static devices, with less inertia; and decarbonized generation that is predominantly inverter-based. These shifts make system operation more critical, as they have not yet been fully internalized into system management procedures.
Focusing specifically on the event that impacted the Iberian grid, the initial trigger appears to have been a significant loss of generation. Potential causes such as cyberattacks, fires along transmission lines, or intermittent weather—originally speculated in the media—have now been definitively ruled out by official sources. Based on current information, it is still unclear whether the incident resulted from an operator-commanded action (to reduce export flows), an unintentional occurrence (fault or anomaly), or a combination of both. In either case, the issue could have originated in conventional or renewable plants alike.
The turning point from a critical phase to a frequency collapse was the intentional disconnection of the Iberian system from the rest of the European grid. The subsequent cascade of disconnections involved a large share of inverter-based wind and solar generation, whose protection and control systems are highly sensitive to grid disturbances. This is unsurprising, given that 70% of the generation at the time came from those sources (you cannot “lose” generation that isn’t even connected).
However, it’s important to recognize that solutions for making inverter-based generation immune to grid disturbances have long been available and are already in use in many countries. Significant investments have been made to integrate these “new” technologies into power systems without compromising grid security.
These include updated technical connection rules that have fully incorporated the lessons of the 2006 incident that led to the separation of the European grid and revised requirements for low voltage ride-through capabilities—requirements that, according to ENTSO-E’s report on the 2021 Spain incident, had not been fully adopted by REE.
They also include careful operational planning that takes into account all aspects of grid security, supported by resources capable of rapidly assisting the system: fast frequency regulation, to be delivered by an adequate fleet of BESS systems; sufficient inertia, provided by connected rotating machinery and reinforced by enough synchronous condensers (possibly with increased inertia); and the implementation of synthetic inertia by capable power converters.
Finally, the real-time operation phase must be updated to include advanced grid monitoring technologies and appropriate algorithms that enable operators to always maintain a safety margin against dynamic phenomena—sometimes rapid—that can lead to more or less extensive blackouts.
In conclusion, expanding the discussion beyond this specific case (“the first blackout of the renewable era”), rather than blaming new technologies for such problems (which, as we have seen, are systemic and certainly include but are not limited to new technologies), it would be more constructive to fully integrate them into the system. These new solutions offer capacities and potential—such as synthetic inertia, fast frequency regulation, and grid-forming inverters—that are still underappreciated today.
