Vns Tokamak Validation with OpenMC-Serpent Achieves 1% Flux Discrepancy and 20% Reaction Rate Agreement

The production of medical isotopes using fusion reactors represents a promising alternative to traditional methods, and researchers are now validating the potential of the Volumetric Neutron Source (VNS) tokamak for this purpose. Christopher Ehrich and Christian Reiter, from Forschungs-Neutronenquelle Heinz Maier-Leibnitz and Technical University of Munich, alongside Christian Bachmann and Pavel Pereslavtsev from the EUROfusion Consortium, have rigorously compared neutron transport simulations performed using the Serpent and OpenMC codes. Their work demonstrates a high degree of agreement between the two codes in modelling the VNS environment, particularly for crucial metrics like neutron flux and tritium production rates, and identifies subtle differences related to simulation techniques. This validation is a significant step towards confidently predicting isotope yields within the VNS, paving the way for a new, efficient source of vital medical compounds.

Researchers aimed to validate both codes against a complex fusion-relevant geometry and assess their computational performance, particularly when simulating coupled neutron and photon transport. A key goal was to demonstrate the VNS’s potential for producing molybdenum-99, a vital medical isotope. The study revealed generally good agreement between OpenMC and Serpent in neutron transport calculations, although discrepancies emerged in tallies and reaction rates.

OpenMC consistently reported higher values than Serpent, a difference attributed to the distinct estimators each code employs. Differences in photon flux were significant, exceeding 10% in some instances. Analysis of detector response revealed that employing delta tracking within Serpent reduced discrepancies observed with hybrid tracking. Regarding computational performance, Serpent generally proved faster than OpenMC in coupled neutron-photon mode using its default tracking mechanisms. However, when Serpent employed full delta tracking, its performance became comparable to OpenMC.

Simulations successfully demonstrated the potential of the VNS for medical isotope production, yielding promising results for molybdenum-99 production using both depleted uranium dioxide and molybdenum targets. In conclusion, both OpenMC and Serpent accurately simulate neutron transport within the VNS geometry. The choice of estimator significantly impacts both results and computational speed, with Serpent offering a speed advantage using its default estimator, and OpenMC achieving comparable performance with delta tracking. Researchers constructed a detailed geometrical model of the VNS within both codes, enabling comparative simulations of neutron behaviour in the vacuum vessel and blanket regions. They calculated neutron flux and spectra maps, and determined neutron and gamma spectra, along with reaction rates for specific nuclear reactions, comparing results between the two codes. Results showed good agreement between the codes for neutron flux and certain reaction rates, although discrepancies emerged in tallies.

Employing delta tracking within Serpent reduced discrepancies observed with hybrid tracking. Statistical uncertainties were lower for OpenMC, particularly for neutrons and photons, attributed to differing tally estimators. To demonstrate the VNS’s potential, researchers simulated the production of the medical isotope molybdenum-99 using depleted uranium dioxide and molybdenum as precursor materials. Burnup simulations yielded activity levels comparable to those of the Belgian research reactor BR-2, which accounted for a significant portion of global molybdenum-99 production in 2016. Researchers meticulously modeled the VNS geometry, translating it independently into both codes, ensuring geometric accuracy. Simulations were conducted using a large number of neutron histories, employing a standard nuclear data library to map neutron and photon fluxes, and to calculate reaction rates. Results demonstrate excellent agreement between the codes for neutron flux and certain reaction rates, although discrepancies emerged in tallies.

Analysis of detector response revealed good agreement, while neutron flux exhibited regional discrepancies dependent on the tracking method employed within Serpent. Statistical uncertainties were lower for OpenMC, attributed to differing tally estimators. The study also successfully simulated the generation of molybdenum-99, demonstrating the VNS model’s capability for radioisotope production calculations. Detailed comparisons of neutron and photon flux calculations reveal excellent consistency between the codes, with most discrepancies falling within a few percent. While some differences were observed, particularly in reaction rates and photon flux tallies, these were largely resolved by employing different tracking functionalities within Serpent, although at a computational cost. The study also highlights the computational performance of both codes, finding that Serpent generally executes simulations faster than OpenMC in coupled neutron-photon transport mode, although OpenMC demonstrates superior speed in neutron-only simulations.

Importantly, the researchers successfully modelled preliminary medical isotope production, specifically molybdenum-99, demonstrating the potential of the VNS to serve as a supplier of this important medical radioisotope. The authors acknowledge that discrepancies in results were observed, and that further investigation into the underlying causes of these differences may be warranted. Future work could focus on optimising computational efficiency and exploring a wider range of isotope production scenarios to fully characterise the capabilities of the VNS.