Computational analyses were performed to quantify the inherent ππ’ scaling
distortion caused by similarity scaling between FLiBe and its surrogate fluid
Dowtherm-A for a randomly packed pebble bed akin to the MK-1 PB-FHR core.
Preliminary studies were conducted to determine the optimum modeling approach,
geometrical flow path design, minimum gap size between pebbles, mesh
specifications, and the number of required pebbles. Results showed using the
realistic approach with artificial spacing between pebbles with πΏπ = 0.005 mm, π
= 75, and β = 0.5 mm to be sufficient in producing an accurate ππ’ to the validation
data. Using similarity scaling techniques, the requisite flow and heating conditions
for various length scales (1:1, 0.75:1, and 0.5:1) using Dowtherm-A were
determined. Simulations were performed for FLiBe and Dowtherm-A at the three-
length scales for the investigated ranges of 500 < π
π < 2500 and 7.95 < ππ < 19.4.
The FLiBe numerical results showed similar behavior to the Wakao correlation,
which is currently the recommended ππ’ correlation for PB-FHR cores. The
collected data for FLiBe was used to develop a ππ’ correlation. The resulting
correlation, ππ’ = 0.633π
π0.6ππ0.5, showed good agreement with the numerical
data, producing only Β±5% error. The ππ’ results of the Dowtherm-A data at various
length scales displayed similar behavior despite having distinctly different flow and
heating conditions between each length scale. The collected Dowtherm-A data was
also used to develop a ππ’ correlation, independent of the length scale. The resulting
correlation, ππ’ = 0.268π
π0.6ππ0.745, showed good agreement with the numerical
data, showing Β±10% error. Comparison between the FLiBe and Dowtherm-A data
sets showed similar distortion values for all investigated length scales, showing -
14% at ππ = 19.4 and -30% at ππ = 7.95 for Dowtherm-A. A ππ’ distortion
correction factor (Ξ¨) was developed, resulting in Ξ¨ = 2.362ππβ0.245. Thus, if
Dowtherm-A were to be experimentally investigated for a PB-FHR core with
correctly reduced power and flow conditions, the ππ’ data results can be multiplied
by this distortion correction factor to achieve similar results for FLiBe. By
implementing the developed ππ’ distortion correction into the Dowtherm-A data
sets, roughly Β±2% error at ππ = 19.4 and Β±8% error at ππ = 7.95 for the FLiBe
numerical data results was observed.
Keywords: pebble bed, fluoride salt, similarity scaling, Nusselt number |