Recurrent inverter failures at Kopli Solar Power Plant, Estonia

A peer-reviewed engineering investigation published in 2024 documents recurrent destructive inverter failures at Kopli Solar Power Plant, a utility-scale solar plant in Kopli, Estonia. Multiple inverter failure events were captured through field monitoring and waveform analysis. The investigators then carried out controlled laboratory experiments to reproduce the fault scenarios identified in the field, and combined both lines of evidence to identify the failure mechanism. The case is documented with explicit fault signatures, telemetry and reproducible laboratory validation, which is uncommon for solar inverter failures.

According to the published investigation, the failures originated on the DC side of the inverters, specifically negative DC input terminal short-circuit conditions. The investigation identified three contributing factors that allowed the initial faults to escalate to destruction of the units: (1) insufficient isolation relay responsiveness, in that the protective relays did not trip quickly enough to interrupt the fault current before damage propagated; (2) inadequate semiconductor transient ratings, in that the IGBT components within the inverters were not rated for the transient conditions developed during the fault; (3) ineffective internal insulation, which allowed sustained ground-fault arcing once the fault had initiated. Field waveforms were reported to match laboratory-induced fault signatures. The cascade described was a DC-side short circuit, escalating to a multi-phase short circuit, then a sustained ground-fault arc, then semiconductor breakdown and destruction of the inverter.

Lessons publicly drawn from the investigation:

1. treat DC-side protection in utility-scale solar inverters as an explicit design and verification requirement rather than assuming the inverter will protect itself;
2. specify isolation relay response times and test them against realistic fault scenarios rather than relying on data sheet values alone;
3. select semiconductor transient ratings against the transient profile a unit will see in service, including DC-side fault transients, which the investigation reported can exceed AC-side transients in severity;
4. design and assess internal insulation for sustained arc conditions, not only steady-state operating conditions;
5. where several inverters at the same site fail in a similar way, treat the pattern as a possible systemic issue and commission an engineering investigation rather than replacing units individually;
6. retain detailed waveform and telemetry data, as sites that do so are better placed to establish the root cause after an event.

ncbi.nlm.nih.gov/pmc/articles/PMC12197166