Anatomy of a Blackout: Findings from the Spain-Portugal Grid Collapse Final Report

An analysis of the European Network of Transmission System Operators for Electricity (ENTSO-E) Expert Panel’s final report on the April 28, 2025, Iberian Peninsula power grid incident.

On a mild, sunny Monday in late April 2025, the power grids of Spain and Portugal collapsed in less than 90 seconds. At 12:33 p.m. Central European Summer Time (CEST), more than 50 million people lost electricity in what the investigating Expert Panel called “the most severe and unprecedented blackout that had occurred in Europe in the past 20 years.” Portugal went dark for up to 12 hours; Spain for as long as 16 in some regions. A small area of southern France was briefly affected, though the rest of Continental Europe escaped unscathed.

Nearly a year later, a 49-member Expert Panel—drawn from regulators and grid operators across the continent—has published its 472-page final report, building on an earlier factual report released in late 2025. The document is exhaustive, technically dense, and notable for what it concludes: this was not the result of a single failure, but of many interacting weaknesses that converged catastrophically on a single spring afternoon. Here are the essential takeaways.

90 Seconds from Normal to Blackout

The morning of April 28, 2025, was unremarkable. Solar and wind generation were within normal ranges for the season. Spain was exporting about 5 GW of electricity. Voltage levels on the 400-kV transmission network were mostly within bounds, though variability had been increasing since about 9:00 a.m.

Two episodes of power system oscillations occurred in the half hour before the collapse. The first, between 12:03 p.m. and 12:08 p.m., was a converter-driven forced oscillation at 0.63 Hz, primarily affecting the Iberian Peninsula. The second, from 12:19 p.m. to 12:22 p.m., was a classic inter-area oscillation at 0.2 Hz—the well-known East-Centre-West mode of the Continental European grid, in which large areas swing against each other.

TSO operators in Spain and France took measures to dampen these oscillations, including reducing cross-border exports and reconnecting internal transmission lines. These measures worked for the oscillations, but had the side effect of pushing voltages higher in the Iberian system.

By 12:32 p.m., voltages had settled below 420 kV and the oscillations had subsided. What happened next unfolded with terrifying speed.

From 12:32:00 p.m. onwards, voltages at numerous nodes began rising. About 500 MW of large renewable generation in Spain reduced output, and because these plants operated in fixed power factor mode, their reactive power absorption dropped proportionally, further boosting voltage. An additional 208 MW of distributed wind and solar units either ramped down quickly or disconnected for unknown reasons. Net load in distribution grids increased by roughly 317 MW, partly attributed to small rooftop photovoltaic (PV) systems tripping on overvoltage protection.

At 12:32:57 p.m., a 400/220-kV transformer near Granada tripped on overvoltage protection, severing 355 MW of generation from the grid. At 12:33:16 p.m., two substations near Badajoz lost 727 MW of PV and thermo-solar capacity. Between 12:33:17 p.m. and 12:33:18 p.m., another 928 MW of wind and solar generation disconnected across several regions. In total, more than 2.5 GW of generation vanished in roughly 78 seconds.

Each disconnection removed reactive power absorption from the system, pushing voltages still higher and triggering the next round of disconnections. By 12:33:19 p.m., Spain and Portugal began losing synchronism with the rest of Europe. Automatic load shedding and system defense plans activated as designed, but could not arrest the cascade. The alternating-current (AC) interconnections to France tripped on out-of-step protection at 12:33:21 p.m. The link to Morocco had already disconnected on underfrequency. The high-voltage direct-current (HVDC) line to France, still pushing power in constant-power mode, was the last connection to sever at 12:33:24 p.m. All system parameters collapsed.

A Confluence of Vulnerabilities

The Expert Panel’s root cause tree identifies no single point of failure. Instead, the blackout emerged from the interaction of numerous factors, none of which alone would have caused a system collapse.

Voltage control was the central problem. The panel’s simulations demonstrate that increased reactive power margins could have prevented the cascade. But the reactive power resources available in Spain were insufficient, poorly coordinated, or too slow to respond to the rapid voltage rise.

Several specific weaknesses stand out. They are:

• Renewable Generators Operated in Fixed Power Factor Mode. Under Spanish regulations at the time, renewable energy sources provided reactive power proportional to their active power output, rather than responding dynamically to voltage changes. When active power dropped, whether from schedule changes, market signals, or protection trips, reactive power absorption dropped in lockstep, amplifying voltage swings rather than counteracting them.
• Conventional Generators Underperformed on Reactive Power. The panel found that the reactive power output of several large synchronous generators reached the required reference value less than 75% of the time. The regulatory framework lacked explicit criteria for dynamic reactive power behavior and imposed no economic consequences for non-compliance. In several cases, settings appeared to follow manufacturer defaults rather than the full grid-code envelopes for voltage support.
• Shunt Reactors Were Operated Manually. A substantial amount of reactive power capacity from shunt reactors was available but not activated during the voltage rise, partly because some reactors had been disconnected minutes earlier during the low-voltage phases of the oscillation episodes. Reconnecting them required manual decision-making and processing time that the operators simply didn’t have.
• Spain’s Wider Operating Voltage Range Left Little Safety Margin. Spanish regulations allowed the 400-kV grid to operate at up to 435 kV in transient conditions—significantly higher than the 380 kV to 420 kV harmonized range used elsewhere in Europe. Meanwhile, generators were required to remain connected only up to 435 kV or 440 kV depending on their commissioning date. The margin between normal operation and forced disconnection was very small or non-existent.
• Protection Settings at Many Generation Facilities Diverged from Requirements. The panel found that overvoltage protection thresholds at some generators were set below the voltage limits established under applicable rules, meaning plants disconnected prematurely. Some tripped instantaneously without the time delays that would have allowed transient voltage spikes to pass.
• Small-Scale PV Systems Were Invisible and Vulnerable. Data from two PV inverter manufacturers showed a clear correlation between transmission-level voltage rises and the tripping of rooftop PV systems on overvoltage protection. These sub-1-MW systems—mostly classified as Type A units under European rules—were largely unmonitored and uncontrollable by either TSOs or distribution system operators (DSOs); yet, their aggregate disconnections contributed meaningfully to the growing imbalance.
• Oscillation Episodes Created the Preconditions. While the oscillations did not directly cause the blackout, they shaped the system state in which it occurred. The measures taken to dampen them—reducing exports and reconnecting lines—changed power flows and raised voltages. The 0.63-Hz forced oscillation was traced to converter-driven instability in a specific area of southern Spain. The absence of power system stabilizers on some large generating units and insufficient damping from existing ones weakened the system’s resilience to the inter-area 0.2-Hz oscillation.

What the Defense Plans Could Not Do

The automatic load-shedding scheme (Low-Frequency Demand Disconnection, LFDD) and system defense plans in both Spain and Portugal activated exactly as designed. The panel confirmed their performance was consistent with applicable requirements; yet, they were unable to prevent the collapse. This is perhaps the report’s most sobering finding.

The defense plans were designed primarily around frequency-based threats, that is, generation shortfalls leading to underfrequency. The April 28 incident was driven by overvoltage-induced cascading disconnections, a phenomenon the existing defense architecture was not equipped to interrupt. Even with higher system inertia, the panel’s simulations show the loss of synchronism would not have been avoided, because the rapid reduction in synchronizing torque caused by the cascade of generator trips pushed the system past the point of no return too quickly. In other words, simply adding more rotating mass would not have been enough to keep the Iberian system in step with the rest of Europe. In the end, the panel explicitly recommended modernizing defense plans to address fast voltage-driven phenomena, noting that the current framework was “not sufficiently equipped to prevent a fast, overvoltage-induced cascading collapse.”

Painstaking Restoration Work

The restoration process was a mix of bottom-up black-start procedures and top-down re-energization from France and Morocco. Three main restoration areas were established in Spain. Several bottom-up restoration attempts failed, including in Galicia, East Asturias, Cantabria, Levante, Toledo, and Andalucía, requiring those regions to be re-energized from neighboring zones.

Portugal received continental frequency again at 6:36 p.m. when an interconnector with Spain was energized. Its grid was fully restored by 12:22 a.m. on April 29. Spain’s transmission system was fully restored by approximately 4:00 a.m. on the same day. The panel identified several issues that slowed restoration: failures of voice communication systems at DSOs and generators (while TSO systems remained operational), difficulties starting black-start units, problems maintaining stable electrical islands due to uncontrollable renewable generation, and insufficient observability of distributed energy resources during the restoration process. Beyond the grid, the blackout disrupted rail services, road traffic signals, and telecommunications across much of the peninsula, underscoring how tightly everyday life now depends on secure electricity supply.

The Recommendations

To prevent a repeat of this complex failure, the Expert Panel issued a comprehensive set of recommendations for implementation by TSOs and DSOs, power generators, and national regulators. The recommendations were divided into two main categories: fixes linked to root causes and broader system improvements. Here is a breakdown of what the report says should be changed to minimize the likelihood of a similar incident.

Part 1: Direct Fixes for Root Causes

These actions specifically target the technical and operational failures that led directly to the April 2025 blackout. Interestingly, the Expert Panel noted that the root causes of the Iberian blackout were remarkably similar to a 2024 grid incident in southeast Europe. Rather than reinventing the wheel, the panel’s first action was to amend two existing recommendations detailed in a 2024 report from that incident to specifically address the rapid, overvoltage-induced failures seen in Spain and Portugal. Specifically, the Expert Panel recommended:

• Update 2024 Key Performance Indicators for Rapid Voltage Change.Amend existing European guidelines to establish specific indicators that detect weakened grid states and the risk of rapid voltage changes before an incident occurs.
• Update 2024 Voltage Support Plans.Ensure TSOs update their national grid development plans to include sufficient reactive power (MVAr) support specifically designed to prevent fast overvoltage-induced cascading disconnections.

In addition to those updates, the panel issued several brand-new recommendations for the 2025 incident. These were broken into three categories.

1. Voltage Control

• Shift to Voltage Control Mode.Generators should abandon fixed power factor modes, which exacerbate voltage ramps during power fluctuations, and use dynamic voltage control modes wherever possible.
• Reactive Power Visibility.TSOs should ensure sufficient static and dynamic reactive power margins are available and visible in real-time to manage rapid voltage slopes.
• Automate Shunt Reactors.TSOs should implement automatic control of shunt reactors rather than relying on manual activation, which proved too slow during the fast-moving April incident.
• Harmonize Operating Voltage Ranges.Spain should strictly apply the harmonized European voltage range (380 kV to 420 kV). Operating with transient limits up to 435 kV leaves virtually zero safety margin before equipment automatically disconnects.
• Smooth Power Variations.Ensure that rapid changes in active power, which trigger rapid reactive power and voltage changes, are smoothed out via tighter connection requirements or controller settings.

2. Oscillations

• Continental Europe Damping Framework.Establish a framework across the entire Continental Europe Synchronous Area to monitor and improve the damping of inter-area oscillations. This includes sharing dynamic grid models and tuning power system stabilizers.
• Enhanced Monitoring and Early Warning.Expand the use of high-resolution monitors, such as phasor measurement units (PMUs), to detect, localize, and trigger prompt corrective actions against system oscillations in real-time.

3. Generator Disconnections

• Validate Protection Settings.Grid operators should verify that generator protection settings (like overvoltage thresholds) are properly aligned with system requirements so generation doesn’t trip unnecessarily during brief transient voltage spikes.
• High-Voltage Ride-Through for Solar PV.Update technical requirements so that small-scale (Type A) solar panels can safely ride through high-voltage episodes without disconnecting.
• Assess Everyday Disconnections.TSOs and DSOs should implement routine checks following any unexpected generator trip during normal operations. This allows operators to identify and correct structural weaknesses and faulty protection settings before they contribute to a major cascading blackout.
• Investigate Hidden Embedded Generators.Operators and manufacturers should urgently study the behavior of non-observable, small-scale distributed generation, such as rooftop solar, which played a massive, unpredictable role in the generation loss.

Part 2: Broader System Improvements

While not the direct cause of the April 28 blackout, the investigation uncovered several systemic weaknesses in how Europe prepares for, handles, and recovers from major grid events.

1. Voltage Control

• HVDC Reactive Power Support.Investigate ways to ensure HVDC systems continue providing critical reactive power support even if their active power connections are lost.

2. Oscillations

• Snapshot Grid Models.Create a standard procedure to generate a common grid model snapshot immediately after a major event to allow for accurate, synchronized post-event simulations across Europe.

3. Disconnections

• Streamlined Data Framework. Establish a permanent, standardized framework for data sharing between TSOs, DSOs, and power generators. In future incidents, high-quality data must be available quickly without ad-hoc legal negotiations delaying the investigation.

4. System Defense Plan

• Distributed Energy Resource (DER)–Aware Load Shedding.Modernize emergency load-shedding schemes so they recognize DERs. Current protocols often accidentally disconnect much-needed solar generation along with consumer load; defense plans could use “reverse-power blocking” to prevent this.
• Real-Time Defense Visibility and Joint Drills. Create a unified framework that provides grid operators with real-time visibility of power flows and DER status down to the feeder level. Furthermore, TSOs, DSOs, and power producers should conduct regular joint defense drills to validate their operational readiness, test response times, and ensure they can deliver standardized forensic data within 24 hours of a major event.
• Defending Against Fast Voltage Deviations.Update defense plans to explicitly identify and react to rapid voltage anomalies and oscillatory behavior before standard safety measures fail.

5. Restoration

• Mandatory Black-Start Testing.Make realistic black-start and island-mode tests mandatory every three years. Tests should include varying generator modes and specific demands (such as hydro pumps or batteries) to ensure readiness.
• Controlled DER Reconnection.Define clear rules for reconnecting distributed solar during a blackout restoration. Early, uncoordinated energization of power lines with high solar penetration can cause fragile restoration islands to collapse.
• Joint Restoration Training. Conduct common, cross-entity training exercises at least every three years involving TSOs, DSOs, and key grid users. These trainings should practice complex restoration scenarios, such as managing the frequency master role, controlling load pick-up, and safely resynchronizing isolated grid islands, to eliminate operational conflicts and misunderstandings that can severely delay bringing the grid back online.
• Blackout-Proof Communications.Ensure grid operators and key generators have at least 24 hours of reliable, blackout-proof voice communication and telemetry tools that are completely independent of public telecommunications networks.

The Broader Warning

The report’s concluding observations carry a warning that extends well beyond the Iberian Peninsula. Europe’s electricity system is undergoing a profound transformation: growing shares of inverter-based renewable generation, declining synchronous machine capacity, deeper market integration, and broader electrification. These trends place the system under increasingly challenging operational conditions.

The April 28, 2025, blackout exposed how the interaction between market design (schedule changes driving fast power ramps), grid code implementation (fixed power factor mode for renewables), protection coordination (settings diverging from requirements), and system architecture (insufficient reactive power reserves and manual switching of critical assets) can create cascading failures that unfold faster than human operators can respond. Beyond technical fixes, the Expert Panel included a stern note for national regulators. Having rules on the books is not enough. Regulators must proactively issue implementation guidance, actively monitor the effectiveness of grid regulations, and use economic incentives and penalties to ensure all grid actors, from massive transmission operators down to individual power plants, comply with their obligations to keep the grid secure.

The panel was careful to note that its recommendations are “voluntary in nature” and that monitoring their implementation falls outside the group’s mandate, but the message is clear: the vulnerabilities revealed through this incident are not unique to Spain. Every TSO should be asking whether their system could withstand a similar sequence of events—and acting on the answer before they have to find out the hard way.

—Aaron Larson is executive editor of POWER.