Refinements of a treat model for transient analysis
In order to prepare for the TREAT reactor being taken off standby, the Gateway for Advanced Modeling and Simulation Lab has used a set of KENO models to simulate certain TREAT temperature- limited transients using the Transient Reactor excursion simulator (T-ReX) code, a hybrid deterministic-stochastic code. These transients are from the M8CAL experiment series, a set of experiments done to calibrate the reactor for the planned M8 series of fuel tests. Recently, the model has been updated in media definitions and geometry. Models were examined and calculations were performed to low uncertainty values to find the k{sub eff} and reactivity insertions. Alterations to modeled values of shutdown and initial transient rod insertion depth were made to create models which had experimental values of reactivity insertion, pre-transient k{sub eff}, or both. It was found that two of the altered media inputs, fuel and Zircaloy-3 cladding, had significant effect on the k{sub eff}. Surprisingly, the Zircaloy-3 composition had a larger effect on k{sub eff}. The reactivity insertions in this model's 2855, 2856 and 2857 temperature-limited transients were found to be above experimental values by a factor of 5.5% to 7%, and that pre-transient k{sub eff} was too high, at 1.002. It was found that increasing shutdown rod insertion by 2 to 2.5 cm decreased pre-transient k{sub eff} to 1.0, and that increasing shutdown rod insertion by 3 to 5 cm and decreasing initial transient rod insertion by 1 to 2 cm gave perfect pre-transient k{sub eff} and total reactivity insertion. (authors)