Wide-Area Disturbance Detection: CERN Records the Iberian Blackout

1. PMUs at CERN record the blackout

On April 28^th, 2025, at 12:33 CET, a major blackout struck the Iberian Peninsula, disrupting the power grids of Spain and Portugal for several hours and briefly impacting parts of southwestern France. While CERN is geographically distant from the affected region, the wide-area impact of the event – propagating across the Continental European synchronous area – was clearly observable through its local monitoring infrastructure. The event was successfully captured by CERN’s recently deployed advanced grid monitoring system under the RF2.0 project. This system includes 24 Phasor Measurement Units (PMUs) strategically placed to monitor grid behavior in real time [¹, ²].

On May 9^th, 2025, an expert panel of ENTSO-E began an investigation of the root causes that led to the Iberian blackout [3]. According to ENTSO-E’s transparency platform [³], the impact of the blackout was recorded as an imbalance between generation and demand in the Spanish Peninsular System as shown in Fig. 1.

Figure 1: Total Load - Day Ahead / Actual of the Spanish Peninsular System [4]. — *Figure 1: Total Load – Day Ahead / Actual of the Spanish Peninsular System [*3].

At precisely 12:33:20 CET, the PMU at CERN’s 400 kV substation – CERN’s main connection point to France’s transmission grid (operated by RTE) – detected the typical signs of a blackout as part of the Continental European synchronous area: a) a sharp frequency drop from 50 Hz to a bit less than 49.85 Hz (Fig. 2) and b) a sharp voltage increase followed by rapid stabilization (Fig. 3).

*Figure 2: Power frequency and Rate of change of frequency (ROCOF) recordings from the 400 kV substation PMU at CERN between 12:20 and 13:30 on the 28^th of April 2025.*

*Figure 3: (a) Effect of voltage regulation from generators in the French HV system, (b) zoom on the recorded rapid voltage variation at the time of the blackout.*

2. Measurements preceding the incident

Interestingly, the PMUs at CERN also detected signs of potential inter-area oscillations at a frequency of about 0.2 Hz, a few minutes before the inception of the event which led to the blackout. These oscillations were visible in both frequency and voltage. Two distinct events were recorded:

between 12:03 and 12:08 CET and
between 12:19 and12:22 CET

as highlighted in red on Fig. 4. In addition, the second event also had an impact on the voltages, where oscillations of the same frequency (0.2 Hz) were recorded as shown in Fig. 5.

These recorded signals correspond exactly to the data from ENTSO-E’s preliminary information on the chronology of events that led to the blackout. It should be noted that with their swift reaction the French and the Spanish TSO’s were able to mitigate these issues as at the time of the blackout there were no oscillations and the systems were under normal operating conditions [⁴].

*Figure 4: Power frequency oscillation events recorded between 12:03 and 12:08 and between 12:19 and 12:22.*

*Figure 5: Voltage oscillations at the 400 kV substation matching the power frequency oscillations.*

3. Validating PMU Capabilities Through a Real-World Grid Event

The Iberian blackout offered a valuable real-world scenario to assess the capabilities of wide-area monitoring systems and especially PMUs. Their installation at CERN demonstrated several key advantages:

High-resolution detection of system anomalies including frequency drops, RoCoF spikes, and inter-area oscillations,
Excellent time-synchronization, enabling accurate correlation with ENTSO-E’s official event timeline,
Insightful early warning capabilities, particularly in identifying pre-fault oscillatory conditions.

Notably, CERN’s electrical network remained fully operational throughout the event, indicating a degree of robustness and ride-through capability in the face of external grid perturbations. This is particularly important for facilities such as particle accelerators, which rely on stringent power quality and continuity standards to avoid disruptions in sensitive experimental processes.

References