Geometry & CAD · Particle transport
FusionFissionGeneral nuclear

DAGMC.

Direct Accelerated Geometry Monte Carlo

DAGMC is a C++ geometry layer that lets Monte Carlo transport codes track particles directly through faceted CAD solid models instead of hand-built constructive solid geometry, using bounding-box trees for fast ray tracing and an automatically generated implicit complement for unmodelled space. It has been coupled to OpenMC, MCNP, Geant4, FLUKA, Tripoli4 and Shift, and its unified workflow lets one tagged geometry carry materials and tallies into several of them.

Hub betaListedSources checked 2026-10-10
Engineering questions it helps answer
  • What flux, nuclear heating or dose does a geometrically complex blanket, shield or port assembly receive when modelled straight from its engineering CAD?
  • Can one CAD-derived model with one set of material and tally tags be run in several transport codes to cross-check results?
  • Where are particles lost in a CAD assembly, pointing to gaps, overlaps or faceting problems in the design model?
Physics scope
  • CAD-based geometry for Monte Carlo radiation transport (physics comes from the host code)
  • Ray tracing on faceted surfaces with oriented-bounding-box trees
  • Implicit complement for undefined space
  • Material, density, boundary-condition and tally metadata via Cubit group names
  • Unified workflow (UWUW) across host codes with PyNE material libraries

Limitations

Not a transport code: a host Monte Carlo code is always required, and DAG-MCNP additionally needs export-controlled MCNP source. Model preparation depends on Coreform Cubit for faceting and tagging, and the documented clean-up (imprint, merge, make watertight) is iterative; faceting introduces tolerance error and lost particles. The documentation site lags the code releases.

Fusion relevance

DAGMC grew out of the University of Wisconsin–Madison fusion neutronics effort and its foundational 2010 paper is about CAD-based fusion neutronics. It is the standard route to CAD-based neutronics for tokamak and stellarator components in OpenMC and MCNP, while remaining a general-purpose geometry layer used for fission and accelerator work too.

Upstream examples

Published by the project itself. These are the project's own tutorials, not Morton Labs runs.

Reference workflows on this site

No reference workflow documents this tool yet. Its evidence status is “Listed”: A reviewed catalog entry with upstream links.

People

Developed by University of Wisconsin–Madison, Computational Nuclear Engineering Research Group (CNERG), published under the Svalinn GitHub organisation. No individual profiles are published for this project yet.

Roles name a specific relationship with evidence. Authorship of a paper is not maintainership, and a listed person is not a Morton Labs partner or endorser.

Source collections

This entry was written by Morton Labs from public sources and last checked on 2026-10-10. Spotted an error, or want something changed or removed? Email zach@mortonlabs.ai with the page address.

Sources checked for this page (4)