Fire from a check valve when troublshooting

The event occurred at a nuclear power plant of the pressurised water type and affected the non-nuclear part of the plant. A valve on the hydrogen supply line to the generator cooling system failed at stem during a routine procedure aiming at adding hydrogen to the system. The location was inside the turbine hall. The released hydrogen ignited with some delay.

A team was preparing to add hydrogen to the main generator. A step in this procedure was the verification of the status of the inlet isolation valve which allowed hydrogen flow from the station to generator. This valve was a normally-open-valve, and it had to be verified that it was in the open position. A team member tried to move the handle in both directions, unsuccessfully. Assuming that the valve was stuck on its closed seat, the operator used a pipe wrench and proceeded to rotate the valve handle in the open direction. Suddenly, the bonnet, hand wheel, and valve internals assembly ejected from the valve body, releasing hydrogen under pressure to the atmosphere. The operator immediately exited the area and the hydrogen quickly ignited.

The reactor was working normally at 90% of the nominal power. In response to the fire, it was manually shutdown. The hydrogen supply was isolated, and fire extinguished with water in ten minutes. However, it re-flashed burning electrical cables. No injury and negligible property damage, beyond the loss of business caused by the reactor shutdown.

The INITIATING CAUSE was a check valve difficult to open and close. Wrong manual operation by a team member caused the damaging of the valve and its failure.
The operator first attempted to open the valve using both hands with no success, then unsuccessfully attempted to close it using both hands. Assuming the valve was stuck on its closed seat, by means of a pipe wrench, the operator succeeded to rotate the valve handle in the open direction approximately six turns. Because the valve was actually on its open seat, this action resulted in the valve bonnet unscrewing from the valve body.

The power plant operator concluded that the ROOT CAUSE of this event was a personnel error. THe operator was trained for the operation, but '...did not apply acquired skills and abilities to the task.' However, the valve had a critical role in the supply of hydrogen to the generator and was very probably considered part of the safety system. The facts that (A) its handle was not moving, and (B) it was lacking visual indications of the position, suggest a lack of operative supervision in guaranteeing correct inspection and maintenance and clarity of the instruction.

valve,
supply line,
cables

The reactor was working normally at 90% of the nominal power. The operation of adding hydrogen to the generator cooling system appears to be a routine work, probably aiming at compensating for the operative hydrogen loss of the coolant circuit.
There have been no previous similar events in which a fire resulted in a reactor trip at this plant.

According to the plant operator final report, operations personnel responded quickly and appropriately to the hydrogen fire by immediately isolating the hydrogen supply and extinguishing the fire.
The decision to shut down the rector was a precautionary, conservative decision. During the shutdown, all systems performed as designed during and after the plant trip. Main steam safety valves lifted as expected following the plant trip from 90 percent power and subsequently reseated, as designed.

A Notification of Unusual Event was declared after that the fire in a protected area had not been extinguished within 10 minutes.

Remediation training and requalification provided to the operator involved with this event.