An Overview of Nuclear Criticality Safety Analyses Performed to Support Three Mile Island Unit 2 Defueling

Criticality safety and adherence to established k_eff criteria had to be demonstrated for the various defueling operations performed at Three Mile Island Unit 2. This included determination of adequate neutron poisoning for the reactor coolant system (RCS), design of the defueling canisters, the canister-handling devices, the storage racks, and the shipping cask. Tools, equipment, and support systems required for the defueling operation also had to adhere to criticality safety requirements. The k_eff criterion used to define the poison concentration for the RCS was ≤0.99. This criterion, coupled with an extremely conservative core model, led to a poison concentration that provided an appropriate margin of safety. To define the fixed poison requirements for the defueling canisters, a k_eff criterion of <0.95 was used for both single canisters and arrays in all credible configurations. For all design analyses, bounding assumptions were made. For each set of analyses performed, the evaluation of k_eff included an allowance to account for uncertainties in the calculated values. A criticality benchmark study was completed to determine appropriate computer code bias values for both the RCS and canister analyses. A lenticular model, which included the entire fuel inventory, was used to define the RCS boron concentration. Based on the analysis by Oak Ridge National Laboratory, a boron concentration of 4350 ppm was required to meet the k_eff criterion. Babcock and Wilcox Company performed the design analyses for the defueling canisters. Each type of canister was explicitly modeled and analyzed using the KENO code. The design requirements mandated that the diameter of the canisters be larger than critically safe dimensions; thus, fixed boron poisoning was utilized. The fixed poison used in the canisters was either Boral plates or stacked sintered boron carbide pellets. The quantity and location of the canister poison was determined based on k_eff limits and operational criteria for the canisters. Analyses were also performed to ensure that k_eff was ≤0.95 for canisters contained within the canister-handling devices.