I got these results using my custom monte carlo code for a pure bare Pu239 sphere of 6.5 cm. Taking into account this is pure plutonium and that the critical radius is about 4.9cm what do you guys think?Right now i'm working on adding tamper geometry and multiregions. Next will be coupling with an hydrocode to predict yield estimations and several refinements. It's written in C++.
From a methodology standpoint (not validating the physical setup), a single reported keff value without an uncertainty estimate and convergence diagnostics isn’t very meaningful—Monte Carlo criticality results should report statistical error bars, demonstrate source convergence (inactive/active cycle behavior, stability checks), and specify the nuclear data library and temperature treatment since those choices can move results significantly; credibility comes from benchmarking the code against publicly documented criticality benchmarks rather than extrapolating from an idealized “bare sphere” case, and jumping from static keff outputs to hydrocode coupling or “yield” claims isn’t justified without a validated time-dependent physics model and a clear verification/validation chain.
>>16884716Thanks for your feedback. My goal is to test against the Godiva and Jezebel critical assemblies and try to get k as close to 1 as possible. Right now i'm implementing multimaterial regions to account for the 3-4% Gallium on the delta-phase Jezebel plutonium. The data i'm using are the Hansen-Roach 16-group cross-sections.
>>16884724once you start referencing real configurations or alloying details, the discussion needs to stay firmly in verification/validation and uncertainty quantification territory rather than tuning toward a target keff; for a code-development goal, keep it framed as reproducing open, non-operational benchmarks, report uncertainties and convergence behavior, document the nuclear data provenance and sensitivity, and stop short of configuration-specific optimization—those practices are what make the work scientifically credible without crossing into unsafe or non-public design space.
>>16884716>>16884756ai generated posts
>>16884709That looks reasonable enough. OpenMC is the easiest to obtain code if you want to validate it without waiting a month for RSICC to respond>>16884716>AI post100000 initial neutrons and 250 steps should be more than enough to reduce uncertainty in keff below .001, even for a full reactor
Any sources for the hydrodynamics? I already plan on purchasing Zeldovich's book.
