Skip to content

Operating experience: Davis-Besse vessel-head erosion (March 2002)

Davis-Besse Unit 1 (a Babcock & Wilcox PWR, but the lessons generalise across PWR designs) discovered during a refueling outage in March 2002 that boric-acid corrosion had eaten a football-sized hole through the reactor pressure vessel head, leaving only the stainless-steel cladding (~10 mm) between the reactor coolant inventory and a containment-bypass leak path.

The event itself was a finding, not an accident — the plant operated safely until shutdown. But the operating-experience value is enormous: the corrosion had been progressing for years, and the indicators that should have triggered investigation had been explained away repeatedly.

What was missed

  • Boron crystallisation on the vessel head was observed and documented multiple times across several outages. Each observation was attributed to small leaks from CRDM (control rod drive mechanism) flange seals — plausible-sounding but never independently confirmed.
  • Filters in the containment air-handling system were clogging with boron-bearing particulate more frequently than design predicted. This was tracked as a maintenance nuisance, not a systemic indication.
  • Containment dome radiation surveys showed elevated levels consistent with primary leakage. These were also normalised.

No single indicator was load-bearing. But the pattern — multiple weakly-suggestive indicators all pointing the same direction — was the signal that was missed.

Why this matters for EOP-era operator training

The Davis-Besse event is taught not as an EOP failure (no EOP applied) but as a cultural failure mode with direct relevance to EOP-era operator behaviour:

  • Normalisation of anomalous indications. When an unusual reading recurs, the bias is to develop an explanation for it rather than escalate it. This is the cultural force the "questioning attitude" principle is designed to counter — see safety-culture.md.
  • Multi-event pattern recognition. Individual operators see individual events. The pattern across events is visible only to whoever is tracking near-miss reports. Davis-Besse had diminishing near-miss reporting in the years before discovery.
  • Authority gradient effect on inspection findings. Maintenance staff who saw the boron buildup typically did not formally escalate it. The culture did not invite escalation.

Connection to EOPs

A LOCA originating from the eroded vessel head would have entered E-0 → E-1 like any other LOCA. The pre-LOCA cultural problem that allowed the corrosion to progress unchecked is not addressed by EOPs; it is addressed by the conduct-of-operations framework that surrounds them. See procedure-usage.md for the hierarchy.

Cross-reference

References

  • NRC LER 50-346/2002-002 (Davis-Besse Unit 1).
  • NRC NUREG/BR-0294 — Davis-Besse Reactor Vessel Head Degradation: Lessons Learned, Industry Response, and NRC Action Plan.
  • OECD-NEA CSNI/R(2003)10 — Operating Experience with Boric Acid Corrosion.