Fission Stories #178
On March 22, 1975, a worker using a lit candle to check for air leaks in the room directly below the control room for the Unit 1 and Unit 2 reactors at the Browns Ferry Nuclear Plant near Athens, Alabama accidentally ignited highly flammable material that had been used to seal openings in the wall where metal trays filled with electrical cables passed through. These cables connected switches, gauges, indicators, alarms, and other devices in the control room above to equipment throughout the plant. The fire burned for over six hours. As insulation burned away, exposed cables touched each other or the metal trays and shorted out. The extensive cable damage disabled all of the emergency core cooling systems for the Unit 1 reactor and most of these systems on Unit 2. Only heroic and ad hoc actions by workers prevented two core meltdowns.
In response to that near-miss, the NRC adopted broader and more stringent fire protection regulations. The goal was to lessen the chances that a single fire disables vital safety systems and their backups as happened at Browns Ferry. A common method for reaching this goal is to separate the electrical cables for a primary system from the electrical cables from its backup. Separation can be achieved by physically routing the cables far apart. Alternatively, one or both sets of cables can be wrapped in a fire retardant material such that the fire gets extinguished before having time to damage all the cables.
A recent event at the Quad Cities nuclear plant in Illinois falls into the truth-is-stranger-than-fiction category. It also shows that cables need to be protected from self-inflicted fire damage.
Fire at Quad Cities
Workers were restarting the Quad Cities Unit 2 reactor following a recent maintenance outage. With the reactor power level at about 8%, alarms sounded in the control room caused by smoke detectors in the turbine building. The control room operators dispatched the Station Fire Brigade Leader to the scene.
About ten minutes later, the fire brigade leader reported back to the control room operators from the scene that there was no fire or smoke, but there was steam leaking into the area. Uncertain as to the source of the steam leak, the control room operators made preparations to roll the main turbine up to speed and place the generator online. Apparently, they figured that routing the steam through the turbine might stop it from leaking into the turbine building.
When it became clear that conditions were getting worse instead of better, the control room operators manually scrammed (shut down) the reactor, tripped the main turbine, and closed the main steam isolation valves. Their first action caused all the control rods to fully insert into the reactor core within seconds, shutting it down. Their latter action stopped the flow of steam into the turbine building, terminating the steam leak.
Thirty-nine minutes after the initial fire alarms and about half an hour after getting an “all clear” report from the fire brigade leader at the scene, other workers in the turbine building informed the control room operators they were seeing electrical sparks. The operators re-sounded the fire alarm and dispatched another fire brigade leader to the scene. This time, the control room was notified about thick black smoke in the turbine building.
Fifty-three minutes after the initial fire alarm, the fire melted fusible links that automatically actuated the fire sprinkler system. The water sprayed from the fire nozzles put out the fire.
After the dust settled and the smoke cleared, an investigation team determined that a 42-year-old electrical cable that had been routed through a tight corner. Excessive bending stressed the insulation on the cable to the point of failure. When a steam leak developed in the turbine building, the high humidity and moisture it caused in the area allowed electricity to flow through the failed insulation and spark when it contacted metal conduit and cable trays. This sparking started the fire.
The investigation team also found that a contributing factor to the steam leak was an expansion joint that had been installed backwards during original construction of the plant decades ago.
Our Takeaway
Fires typically are not sneaky and stealthy. Yet this Quad Cities fire eluded one and perhaps two fire brigade leaders and burned until it automatically triggered the fire sprinkler system that finally put it.
The first fire brigade leader reported no signs of smoke or fire and attributed the smoke detector alarms to steam leaking from an unidentified location. The second fire brigade leader wandered about for over ten minutes looking for the sneaky fire that produced visible black smoke but remained hidden until it eventually burned enough equipment to automatically trigger the fire sprinklers that put it out (wherever it was).
More than forty reactors in the United States do not comply with the fire protection regulations adopted by the NRC following the Browns Ferry fire (including the three reactors at Browns Ferry). Instead, their owners are trying to achieve compliance with an alternative set of fire protection regulations adopted by the NRC in 2004.
These alternative regulations seek to manage the fire hazard by constructing parallel timelines. The first timeline tracks how long the fire burns and the equipment damaged by its flames and smoke. The second timeline tracks how long it takes for the fire to be detected and extinguished.
Sometimes, as in this Quad Cities case, the fire is extinguished by suppression systems that automatically actuate. But often the fire must be put out by workers using portable extinguishers or manually actuating installed suppression systems.
These fire analyses make assumptions about how long it takes responders to find fires and extinguish them. These analyses seldom assume that fire brigade leaders will roam about totally unable to find a fire, let alone put it out.
This is not to suggest that the responders at Quad Cities were inept or incompetent. They left areas of safety to enter areas where radioactive steam was leaking with reports of fire and sparks. They very likely did as good a job as possible under such challenging, high-stress conditions.
This is to suggest that fire analysts should properly account for this reality and not falsely assume that all fires will be readily found and fought. This is to further suggest that the NRC’s reviewers of these fire analyses must not accept overly optimistic “guesses” about fires always being extinguished promptly.
“Fission Stories” is a weekly feature by Dave Lochbaum. For more information on nuclear power safety, see the nuclear safety section of UCS’s website and our interactive map, the Nuclear Power Information Tracker.