Modeling the Molten Salt Reactor Experiment With the GOTHIC Code

Abstract

Abstract GOTHIC 8.3(QA) includes capabilities for modeling advanced, non-light water cooled reactors. Important capabilities introduced in GOTHIC 8.3(QA) include fluid property tables for various molten salts, an enhancement to the tracer tracking module to allow radioactive decay energy to be released locally in the carrier fluid and other improvements to the neutron kinetics module. With these new capabilities in place, GOTHIC is used to benchmark steady-state and transient conditions in the Molten Salt Reactor Experiment (MSRE), which operated at Oak Ridge National Laboratory from 1965 to 1969. In this experimental reactor, UF4 fuel was dissolved in molten fluoride salt, and criticality could be achieved only in the graphite moderated core. An air-cooled radiator transferred fission and decay heat to the environment. The design thermal output of the MSRE was 10 MWt, but the radiator design limited the output to 8 MWt. The original design parameters neglected the impact of decay heat on system temperatures. GOTHIC is used to benchmark system operating parameters at both the 10 MWt design condition and the 8 MWt operating condition, both with and without decay heat. The cases that include decay heat apply 7% of the nominal thermal output using the eleven decay heat precursors from ASB 9-2 as tracers. The results of the benchmark exhibit good agreement with design and operating data and demonstrate heat-up due to decay heat in the fuel salt outside the core. In the MSRE, delayed neutron precursors are not confined to the core because the fuel and fission products flow through the system. As a result, there are different values for (effective) delayed neutron fraction with and without flow, and the decay of delayed neutron precursors outside the core under full-flow conditions reduces reactivity by 0.212 % δk/k. Zero power physics testing included fuel salt pump start-up and coast-down transients with a control rod automatically moving to maintain criticality. The control rod motion calculated by GOTHIC is a reasonable match to measured data from these transients. Low power testing included a natural convection transient with no control rod motion such that reactor power was responding to heat load demand from the radiator. The reactor power and fuel salt and coolant salt temperatures calculated by GOTHIC exhibit good agreement with measured data.