Michael Goggin, Grid Strategies
June 24, 2025
Last week, the Spanish government and Spanish grid operator Red Electrica released lengthy reports on the blackout that affected the Iberian Peninsula on April 28, 2025, along with a shorter press release summary of the government report. This post summarizes the reports’ findings on the causes of the event and recommended solutions, with interspersed observations on the implications of the blackout for electric reliability in the United States. Contrary to early speculation, the reports document that renewable generators were not the cause of the blackout, and in fact recommend that renewable and battery resources be a primary solution to prevent such events in the future.
The two reports agree that the primary cause of the blackout was a failure of conventional synchronous generators to adequately control voltage, with secondary factors including Spain’s weak electrical ties to the rest of Europe and the loss of some generation as the transformers that interconnect that generation tripped as system voltage increased.
Spanish grid operating procedures currently allow only conventional synchronous generators to provide voltage control. A primary recommendation of both reports is to allow renewable and battery resources to control voltage going forward. (pages 149-150 in government report and 18 in Red Electrica report) The government report notes this change will improve reliability, in addition to reducing cost and pollution. Wind, solar, and batteries are inverter-based resources (IBRs), in that they are interconnected to the grid via sophisticated power conversion equipment that allows them to quickly and accurately regulate voltage. Since 2016, the U.S. Federal Energy Regulatory Commission (FERC) has required IBRs to regulate voltage at levels comparable to that provided by conventional generators. In its report, the Spanish government also recommends “boosting electrical storage” for “increased strength and flexibility of the electrical system.” (page 151; this and all subsequent quotes translated via Google Translate, in some cases with non-substantive corrections)
The government report’s third recommendation is to “continue promoting the increase in interconnections with the European system.” (151) The report explains that “if the Spanish system had been more interconnected, the probability of the occurrence of the oscillatory phenomena on April 28 would have been smaller and they would have done so with less virulence.” (99) The government report concludes that more transmission capacity between Spain and its neighbors would reduce the likelihood of such an event happening in the future.
What happened?
Both reports explain that the primary cause of the event was the failure of conventional synchronous generators to adequately control voltage. Analogous to pressure in network of water pipes, voltage ensures the efficient flow of electricity on the power system. Voltage is primarily controlled by generators and other devices injecting reactive power to increase voltage or absorbing reactive power to decrease voltage. Reactive power control is needed in adequate quantities throughout the grid, and ensures the efficient delivery of “real power” that runs electrical devices. If left unaddressed, inadequate voltage control can result in power system collapse, as occurred in Spain. The government report explains that:
The ultimate cause of the peninsular electricity blackout on April 28 was a phenomenon of overvoltages in the form of a “chain reaction” in which high voltages cause generation disconnections, which in turn cause further increases in voltage and with it new disconnections, and so on. This phenomenon was preceded by high-amplitude voltage variations in short periods of time throughout the morning. Several factors contributed to this phenomenon:
- The system showed insufficient dynamic control capabilities to maintain stable voltage… (131)
The government report documents that a day in advance, the following generators were scheduled to provide dynamic voltage control on the day of the blackout: “10 thermal groups (3 nuclear groups and 7 combined gas cycles).” (21) One of the thermal plants scheduled to regulate voltage became unavailable, and there is some dispute in the reports regarding whether and how that plant’s voltage control should have been replaced. That plant was in the southern part of Spain where absorption of reactive power proved to be particularly important during the events leading up to the blackout. (131)
As a result, at the time of the blackout, “there were 11 thermal power plants coupled with an obligation to regulate voltage by set point: 4 nuclear power plants [redacted], 1 coal-fired power plant [redacted], and 6 gas plants[redacted]… in addition to hydraulic generation [redacted]…” (38-39)
As the government report notes, “throughout the morning” these plants struggled to control voltage. The following chart (Red Electrica at 2) shows repeated large fluctuations in voltage in the three hours preceding the blackout, with voltage varying by about 10%.
Red Electrica’s primary conclusion is that conventional “generation subject to Operating Procedure PO 7.4 did not comply with its dynamic voltage control obligations, which led to higher than expected voltage levels in the system. Furthermore, voltage excursions—both upward and downward—tend to be more pronounced due to this noncompliance.” (15)
The government report notes that despite these synchronous generators being paid to regulate voltage, “the operator has reported an insufficient absorption of reactive energy by practically all the generators in one or more periods in the morning of April 28.” (110-111) The report continues to note that:
However, it has been confirmed, with the information provided by the generators, that in the minutes and hours before zero, with already high voltage levels (above 410-420kV) in the 400kV network, several of the coupled thermal groups do not responded as expected in that context: either they absorbed less reactive power than expected by the system operator, not providing sufficient voltage control or, in some cases, generating reactive energy instead of absorbing it (thus contributing to worsening the overvoltage). (111)
Both reports note that a thermal plant in the southern part of Spain, where voltage control was most crucial on the day of the blackout, had some of the worst voltage control performance, exacerbating high voltage there. (government report at 131 and Red Electrica at 12)
During the late morning several grid oscillations emerge, some of which appear to be tied to a faulty power plant controller (which the reports identify as a problem that needs to be addressed), and others of which appear to be a common frequency oscillation pattern between Spain and the rest of Europe, primarily due to the weakness of the Iberian Peninsula’s transmission ties to France.
The grid operator contained these oscillations by reducing export flows to France and Portugal, placing transmission lines back in service, and switching static reactive power devices. (government report at 83) These steps are primarily intended to dampen the oscillations but also have the effect of increasing system voltage, which was low during the oscillations. As the Red Electrica report explains “All these measures adopted are intended to increase the system voltage, but they are necessary to correct the oscillation as they are measures that improve damping, which when it appears is the priority,” (4) and “All the actions taken allow the situation to be controlled, but they tend to cause system voltages to rise.” (5)
To counteract the increasing voltage, in the 5-10 minutes prior to the blackout the grid operatorreversed the switch for static reactive devices it had previously deployed when voltage was low during the preceding oscillations. This reversal reduced system voltage, but static reactive power devices are less flexible in controlling voltage than the control provided by generators and other dynamic reactive devices. Whereas generators and other devices can dynamically regulate voltage, meaning they can incrementally control the reactive power injected or absorbed, switched static devices like series capacitors and reactors can only change voltage in an all-or-nothing stepwise fashion.
About 10 minutes before the blackout the grid operator also directed two gas combined cycle generators to start up to help dynamically manage voltage. However, the combined cycle owners reported they would require 1.5 hours in one case, and 2-2.5 hours in the other case, to come online. That startup process began, but was halted by the blackout ten minutes later. (Red Electrica at 6-7) In contrast to the many minutes to hours required to start up and synchronize most types of conventional generators so they can begin providing reliability services, one of the advantages of battery storage resources is that they can start up and begin controlling voltage or frequency almost instantly. Moreover, many modern IBRs can be configured so that their power electronics can regulate voltage without even generating real power. The Spanish grid operator does not currently allow for these resources to provide voltage control, however.
About one minute prior to the blackout, voltages “begin to increase throughout the entire transmission network in an almost linear manner, going, for example, in SE Olmedilla from 413 kV to 428 kV in 57 seconds or in SE Arroyo de San Serván 400 kV from 411 kV to 424 kV in the same time.” (government report at 39)
This voltage increase coincides with demand increasing by 525 MW, apparently due to a combination of distributed generation disconnecting to protect itself from rapidly increasing voltage and the “electrotechnical effect that generates an increase in demand with the rise in voltage.” (government report at 88) Because power consumed is voltage times current, increasing voltage proportionally increases power consumption. The concern with distributed generators remaining online or “riding through” voltage or frequency disturbances on the grid has been addressed in the U.S. through standards like the 2018 update to the Institute of Electrical and Electronics Engineers (IEEE) 1547 standard. The government report also blames “Insufficient absorption of reactive energy by the generation that controls dynamic voltage (large synchronous generators such as nuclear or combined cycles” (89) for failing to halt the increase in voltage.
During the minute before the blackout, some utility-scale generators were disconnected as voltage increased. Voltages were about 10% above normal when these initial disconnections occurred, about 239-241 kV at substations that are supposed to operate at 220kV. While there is some uncertainty regarding precisely at what voltage some disconnections occurred, it appears that voltages were within the bounds where generators should have remained online.
The two initial large disconnections of generation, of 355 MW and 582 MW, occurred when two transformers serving multiple renewable generators tripped, apparently because the taps on the low-voltage side of the transformers failed to respond quickly enough to the rapid increases in grid voltage. The Red Electrica report highlights that two other transformers had tripped for the same reason during the period of rapidly increasing voltage just after 11:00 that morning, which is shown in the voltage chart above. Red Electrica notes that this earlier event “suggests that the owner of the connecting facilities did not adapt the transformer taps quickly enough to the voltage increase. This situation that occurred in the transformers of these two substations, although it did not have a major impact, is important because it is considered a prelude to what could have occurred later in other facilities.”
Once the transformers interconnecting the 355 MW and 582 MW of generation disconnected in the minute leading up to the blackout, any type of generation connected via those transformers would have gone offline, so this does not appear to implicate the ride-through performance of those generators themselves. The government report states that for both events “Once the 220 kV bar switch is disconnected, the entire downstream network was left islanded and the collector substations were recording undervoltage or overfrequency (when an island of generation is left without load to serve)…” (90, 94)
The voltage increase caused by these initial generator trips subsequently caused voltage to approach or exceed the level at which generator disconnections are permissible, causing a cascading effect. As the Red Electrica report explains “With each generation disconnection, the system voltage increased, causing additional generation disconnections.” (12) The government report explains this “becomes a “point of no return” by starting a “chain reaction”… In other words, once this phase has begun, the way of containing the system would have been a sufficient absorption of reactive energy to reduce the voltage faster than the “chain reaction” tended to raise it…” (97-98)
The initial loss of generation due to transformer disconnections that began the cascading effect does not indicate a similar risk for the U.S. power system.
First, the connection of multiple unaffiliated generators through a single transformer is not typically used in the United States. In the limited instances where such configurations do exist, NERC mitigates the associated risks through a combination of performance, protection, and system modeling standards that address these shared facilities. The government report explains that this arrangement is common in Spain, though: “For economic and environmental efficiency, to take advantage of electrical corridors and to minimize impact and costs, private substation networks have been developed and “Christmas tree” shaped collector lines, from which they sometimes hang several dozen plants belonging to different owners who, through agreements private, built and/or use common export infrastructure.” (113)
Second, more effective voltage control would have counteracted the voltage increases from the initial loss of generation, preventing the cascading effect of additional generator trips. As noted above and below, the U.S. power system uses a more diverse pool of resources to control voltage, including renewable and battery resources that provide fast and accurate voltage control. Red Electrica documents that during the voltage increase following these trips, “Most conventional generation with dynamic voltage control does not absorb the reactive power required under the Operating Procedures, particularly a group that was for voltage control in the southern zone, a group located in Extremadura, and a group that was for voltage control in the central zone.”(12)
Finally, strict generator ride-through requirements under FERC Order 2023, IEEE 2800-2022, and North American Electric Reliability Corporation (NERC) standards ensure IBRs remain online for voltage disturbances of this magnitude in the U.S. The existing NERC standard PRC-024 specifies generator protective relays cannot trip generators of all types for voltages 10% above normal, and cannot trip for voltages in the 10-20% above normal range unless they persist for dozens of electrical cycles. The PRC-029 standard NERC has filed at FERC requires IBRs to remain online for one second for voltages between 10-20% above normal, and remain online indefinitely for voltages up to 10% above normal.
The cascading generator outages in Spain caused frequency to decline, which initiated the disconnection of the transmission ties with the rest of Europe and Morocco, successfully preventing the system collapse from spreading to those areas. The frequency drop also initiated automatic load-shedding, i.e. controlled rolling blackouts of a subset of customers, which Red Electrica correctly notes is “a universal mechanism that helps restore frequency in scenarios of severe generation-demand imbalances…” in an attempt to avoid an uncontrolled blackout of the entire system. However, Red Electrica raises the interesting point that load shedding also “raises voltage, since disconnecting demand naturally increases system voltage. This worsens the existing voltage control problem in the system…” Under-frequency load shedding is valuable because of its fast automatic response, but it is not helpful for high voltage. This reinforces the value of IBRs’ ability to provide fast dynamic voltage regulation.
Both reports also put to rest the preliminary but erroneous speculation from many parties, including some in the U.S., that the event was caused by the high penetration of renewable generation causing inadequate inertia to maintain system frequency. The government report documents that inertia levels were more than adequate, as “the system operator indicates that before the incident the system had inertia levels of 2.3 seconds (excluding contributions through the interconnections), which is a value higher than the 2 seconds recommended by the European Network of Transmission System Operators for Electricity (ENTSOE)…” (107) Red Electrica also explains that “the system’s inertia in this incident is irrelevant because the system was already doomed by the massive loss of generation.” (12)
What are the reports’ recommended solutions?
As noted above, a primary recommendation of both reports is to use renewable and battery resources to regulate voltage going forward. Spanish grid operating procedures currently allow only conventional synchronous generators to control voltage, even though renewable and battery resources can quickly and accurately regulate voltage. Spain’s practice differs from requirements in the U.S., where for nearly a decade FERC Order 827 has required newly installed IBRs to regulate voltage.
Wind, solar, and battery resources interconnect to the grid via sophisticated power electronics that allow extremely fast and accurate regulation of voltage. For example, as illustrated by the actual power system data presented in the 2008 chart below, wind turbines can significantly improve power system voltage stability. This is indicated by the fact that power system voltage is much better regulated when wind turbine generators (WTGs) are online than when they are not.
IBR power electronics can even provide voltage regulation when the plant is not producing real power. For example, solar plants can be configured to regulate voltage at night by pulling a small amount of power from the grid to operate the power electronics to inject or consume reactive power. Batteries also have extremely fast voltage and frequency regulation capability, and the government report recommends “increased strength and flexibility of the electrical system” by “boosting electrical storage.” (153-154)
The government report explains that “for years renewables already have the technological capacity to operate on command, although the regulations do not yet require or permit it.” (66) A footnote further explains “it is estimated that, in the peninsular electricity system, 19 GW of installed photovoltaic power (almost half of the total) and 5 GW of wind power (16% of the total) currently comply with the European grid code and, therefore, have the capacity to regulate voltage.” (Id.)
The Spanish government report notes that allowing renewable resources to regulate voltage will improve reliability, and save money and reduce emissions. (150) Currently during some periods of low demand and high renewable output, like April 28, Spanish thermal generators are being operated solely to regulate voltage, even though they are uneconomic and their real power is not needed to serve load. As the government report explains:
When a generator is scheduled for voltage control, it is typically scheduled at the “technical minimum,” that is, the minimum generation required to be coupled, since the system does not “need” that energy; rather, it is the result of having the generator coupled for voltage control. To ensure a balance between generation and demand, this “excess” generation from generators connected via voltage control limits other generation that would have been matched on the market, increasing generation spills. (150, footnote 18)
The government report proposes that voltage service be provided by “resources selected based on lowest cost.” (149) It continues by noting that “This change will mean that facilities distributed throughout the country may contribute to voltage control, thus reinforcing the tools available with wide territorial distribution. Being an open service, reinforces technological neutrality allowing it to achieve the desired objectives at the lowest cost to consumers even savings are expected compared to the starting point.” (Id.) The report also recommends that “penalties are established for non-compliance with the voltage control obligations that remain applicable to the synchronous groups.” (Id.)
As noted above, the third primary recommendation of the government report is “continue promoting the increase in interconnections with the European system,” (151) as “the peninsular electrical system has a low level of interconnection with the European continent, with barely 3% of installed capacity, far from the 15% target set in European regulation.” (98) The report notes greater interconnection would have dampened the oscillations that preceded the blackout. It also explains that “Interconnections can play a relevant role in the event of an incident, providing instantaneous inertia and primary frequency regulation in the form of active power,” and that following the Spanish blackout the “interconnections were essential for the rapid replacement of the supply.” (99)
Conclusion
The April 28 Iberian blackout highlights how reliability risks can arise from inadequate operational practices and underperformance by conventional generation. Reports from the Spanish government and Red Electrica found that the primary cause was the failure of synchronous generators to manage voltage, not the presence of renewable resources. In fact, renewable energy and battery storage were identified as essential to improving system stability and preventing future events. The government report also recommends expanding Spain’s transmission ties with neighboring power systems to improve stability and reduce reliability risks.
The U.S. is less vulnerable to similar failures due to differences in system design and regulatory oversight. In the U.S., new renewable and battery resources have been required to regulate voltage for nearly a decade.
Spain’s experience reinforces the importance of modernizing voltage control practices, expanding the use of battery storage, and applying consistent reliability standards to all resources. The reports confirm that renewable energy did not cause the blackout, and in fact is a critical part of the solution.