SUMMARY

In loop-type molten salt reactors, gaseous fission products may become entrained in the primary loop. Additionally, sparging and cover gases, such as helium, may also become entrained. The fraction, by volume, of the loop occupied by gas is named the void fraction. Void fraction affects reactivity and thereby the power dynamics of the reactor. There currently do not exist methods of in-situ assay of void fraction durable to the high temperatures and corrosive effects of molten salt. However, indirect interrogation by activity analysis through gamma spectroscopy may be a viable solution to void fraction assay. An in-silico study of such a solution utilizing was executed. Uncertainty quantification was conducted through repeated MCNP simulations in order to estimate the accuracy and precision of such a system. A validation phase, with no variation in the composition of the salt, and an experimental phase, with variation in the composition of the salt, were executed for a variety of void fractions. The simulation system was strongly validated, with an absolute error in the reconstructed void fraction no greater than 0.01 and a 95% confidence interval no wider than 0.001. Error increased until a void fraction of near 0.4, and decreased thereafter. Uncertainty decreased monotonically with increasing void fraction. In the experimental phase, compared to the validation phase, uncertainty increased across the entire range of void fractions by between one to two orders of magnitude, depending on void fraction. Additionally, bias increased, by a factor ranging between 50 near void fractions of zero, to about 2 otherwise, due to fundamental limitations of the void fraction reconstruction algorithm. Error and uncertainty both decreased monotonically with increasing void fraction.