PAKS FUEL CLEANING CONTAINER - DESCRIPTION OF COOLING SYSTEM, & POSSIBLE EXPLANATION OF FUEL DAMAGE

The key design feature, cooling-wise, was free flow of cooling medium around the fuel assemblies. The flow was routinely provided by a pump. There were three pumps. The one used during the cleaning process provided a greater flow rate than the pump used during post-cleaning cooldown. The third pump was a backup for either of the other two. In the event of loss of pump flow, Framatome calculations indicate that the valve at the bottom of the tank, if opened, could cool the fuel by thermal convection.

If the fuel assemblies in the tank were positioned improperly, it is possible that the flow of coolant could have been disturbed, especially after the switch to lower flow pump when cleaning part of the process was complete. Industry newsletter Nucleonics Week reports that this possibility is one of the factors being investigated. One difference between the five successful runs of the new cleaning system prior to April 9, and the fateful sixth batch was that the first five sets of assemblies were removed from the tank pretty quickly after cleaning. The cleaning of the sixth batch was complete at about 4 pm on April 10. The switch to the lower capacity pump was done at that time. The lid to the tank remained closed until early on April 11. Framatome analysis suggests that the undercooled fuel caused the coolant at the top of the tank to boil, which created a steam bubble inside the closed tank. The bubble would have displaced coolant, and the fuel assemblies' cladding would have become embrittled from a temperature-related water-zirconium reaction. When the lid was opened, the relatively cold pool water would have wreaked havoc on the embrittled assemblies.

[Ref: Ann MacLachlan (McGraw Hill-Paris), "Paks-2 Unable To Restart Because It Lacks Damaged Fuel Replacement", Nucleonics Week, May 8, 2003, p. 1 (subscription required)]