| No. | Year | Title | Venue |
|---|---|---|---|
| 01 | 2014 | Calculation of the Radial and Axial Flux and Power Distribution for a CANDU 6 Reactor with both the MCNP6 and Serpent CodesM. S. Hussein, H. W. Bonin, B. J. LewisBuilt full 3-D MCNP6 and Serpent models of the CANDU 6 (Gentilly-2) core with 37-element bundles and compared flux and power-density maps; Serpent matched MCNP6 and was the more efficient of the two. | PBNC 2014Conference · Vancouver |
| 02 | 2014 | Burnup Calculation of a CANDU6 Reactor Using the Serpent and MCNP6 CodesM. S. Hussein, H. W. Bonin, B. J. LewisValidated whole-core burnup in MCNP6 and Serpent against WIMS-AECL, the industry lattice code, with k-effective versus burnup in excellent agreement. Still cited in 2025 work on SMR spent fuel in CANDU. | PBNC 2014Conference · Vancouver |
| 03 | 2017 | Design of a Multispectrum CANDU Reactor for Burning Actinides: An Approach of Non-Proliferation and Nuclear Fuel RecyclingMohamed Salah HusseinSix full-core MSCR models in Serpent. The best uranium model destroyed 10.7 % of fissile isotopes in 1000 days; the best plutonium model 15.3 % over 2600 days. Either core alone is subcritical, so disturbing one shuts the reactor down. | PhD thesis, RMCThesis · Kingston, Ontario |
OpenMC, learned for the next comparison.
He completed the OECD Nuclear Energy Agency's Introduction to OpenMC in June 2026, adding the open-source Monte Carlo code to a decade of reactor models built and cross-checked in MCNP and Serpent.
What OpenMC is
OpenMC is an open-source Monte Carlo code for neutron and photon transport. It began at the Massachusetts Institute of Technology and is now developed in the open by a community that includes national laboratories and universities. It solves the same kind of problem as MCNP and Serpent: it follows particles one at a time through a three-dimensional model of a core and reports the multiplication factor, flux, reaction rates and power. Models are defined in XML input files, usually generated from Python through the code’s Python interface, and the code includes fuel depletion, so a model can follow how the fuel changes over time. Because anyone can read, run and modify it, it has become a common choice for work on small modular reactors and other advanced designs, and for teaching.
His record with it so far
His OpenMC record is short and stated plainly here: in June 2026 he completed the OECD Nuclear Energy Agency’s Introduction to OpenMC. His reason for taking it, in his own posts at the time, was that proficiency in more than one code is what makes verification and validation possible, and that he wanted OpenMC for future simulation and benchmarking work. The publications listed on this page were produced with MCNP and Serpent, not OpenMC; they are here because they show the kind of model and comparison he intends to repeat with it.
What carries over from MCNP and Serpent
The hard parts of Monte Carlo reactor work are not specific to one code. From 2013 to 2017 he built the same models in more than one code and compared them: coupled-core calculations on the Deuterium Critical Assembly in MCNP5 and Serpent; a full CANDU 6 core at Gentilly-2, with every 37-element bundle, in MCNP6.1 and Serpent; and whole-core burnup in both, checked against WIMS-AECL, the Canadian industry’s lattice code. Then he used Serpent for six full-core models of a new reactor concept in his doctoral thesis.
That work is mostly about decisions that any code needs made well: how much geometric detail the fuel needs, how to convert track-length tallies into power per bundle, how many particle histories a result needs before it is converged, how to normalise flux to reactor power, and what counts as agreement between two codes. OpenMC adds a third, independent code to those comparisons. A CANDU model checked in MCNP, Serpent and OpenMC gives three independent answers to the same question.
§1 Papers
The work this rests on.
§2 Related