Uncertainty quantification propagation in FNG HCPB-tritium breeding module using OpenMC-SANDY workflow

Nuclear data (ND) uncertainty quantification (UQ) supports the establishment of safety margins for fusion neutronic designs. This study applies an open-source, stochastic UQ workflow — coupling the continuous-energy Monte Carlo transport code OpenMC with the SANDY nuclear-data sampling toolkit — to the Frascati Neutron Generator (FNG) Helium-Cooled Pebble Bed (HCPB) Tritium Breeding Module (TBM) mock-up benchmark. ENDF/B-VIII.0 and JEFF-3.3 serve as the transport base libraries, with SANDY generating 400 random samples from its covariance evaluation to propagate within-library ND uncertainty to the Tritium Production Rate ( T P R ) in the L i 2 C O 3 detector pellets and to activation reaction rates at increasing shielding depth. A prior sensitivity screening identified the 9 ⁢ B e ⁡ ( 𝑛 , t o t a l ) and ( 𝑛 , 2 ⁢ 𝑛 ) channels and 6 ⁢ L i ⁡ ( 𝑛 , 𝛼 ) ⁢ t as the leading contributors. 9 ⁢ B e cross-section uncertainty dominates the computational variance. The ND contribution to calculated T P R is approximately 4% ( 2 ⁢ 𝜎 coverage), with per-run Monte Carlo statistical error held below 0.5%. Measured TPR carries an experimental uncertainty of approximately 7.8% ( 2 ⁢ 𝜎 ). Rather than summing these into a single figure, calculation and measurement are compared as C/E ratios per pellet stack, which show reasonable agreement with the uncertainty band. A spatial gradient in ND uncertainty is observed, the calculated spread widening from 2 – 3 % at the front stack to 3 – 5 % at the rear. Calculated activation and TPR profiles show reasonable agreement with experiment, with the exception of a consistent underprediction of low-energy monitors, most notably 1 9 7 ⁢ A u ⁡ ( 𝑛 , 𝛾 ) . It is attributed to an under-moderated calculated flux at depth, implicating the thermal-scattering ( 𝑆 ⁡ ( 𝛼 , 𝛽 ) ) and elastic-scattering data of the beryllium multiplier — rather than to the Au evaluation itself, and is identified for further investigation on that basis. Because the same effect implies that parasitic low-energy capture may be underestimated in the model, its impact on the tritium breeding ratio (TBR) is not assumed to be conservative; the directional effect is quantified directly rather than asserted.