An operational concern in natural-convection-cooled research reactors is pool-top 16N activity (PTNA). The conventional technique for reducing PTNA is to disperse the water plume rising above the core by a planar water jet and thus increase the transit time of 16N nuclei to the pool top. The extension in transit time is a function of pool dynamics under dispersion. Ideally, a sufficiently deep stagnant water layer is formed below the pool top to confine 16N activity to lower pool regions. The effects of changes in pool configuration and disperser design parameters on pool dynamics are not well known. These effects are important in determining the feasibility of a power upgrade without major facility modifications. Due to the complexity of pool geometry, pool dynamics under dispersion cannot be described by simple flow models. The COMMIX-1A code is used to simulate the pool dynamics of a typical natural-convection-cooled research reactor with plate-type elements as a function of pool configuration and disperser design parameters. The pool is partly described as continuum and partly as porous medium. All the major pool components are explicitly modeled. The differences between the shapes of some pool structures and computational cells are accounted for using the concept of directional surface permeability. The importance of local turbulence effects and cross-flow friction losses at the guide tubes above the core are also investigated. The results show the following:

  1. The conventional technique for reducing PTNA is effective in the power upgrade of natural-convection-cooled research reactors with large pools.
  2. If pool dimensions are small, a more feasible way to reduce PTNA is to place pool outlet suction points close to the top of the core.
  3. Containing the core outflow within a shroud increases the pool heat removal system inlet temperature, as well as reducing PTNA.