Attila4MC-Cottonwood was developed with the goal of providing an efficient, intelligent, and highly automatic CADIS1 and FW-CADIS2 workflow, regardless of geometric complexity. Whether your model contains several or several thousand parts, Attila4MC-Cottonwood can dramatically shorten and automate your MCNP® analysis cycle for deep penetration calculations.
Attila4MC-Cottonwood directly reads the MCNP abaqus® mesh, the MCNP input file, and several user specified parameters. From this information, Attila4MC-Cottonwood will automatically generate an arbitrary mesh refinement (AMR) computational grid for the deterministic weight window generation (see Figure 1). Local refinement and coarsening are automatically performed based on calculated material half value layers, eliminating the need for users to manually define a Cartesian computational/weight window grid.
Figure 1 - CAD Model of Fusion Tokamak Inside a 200cm thick Concrete Building (left) and Automatically Generated Arbitrary Mesh Refinement (AMR) Attila4MC-Cottonwood Calculation Grid (right)
Similarly, Attila4MC-Cottonwood also provides an optional automatic weight window grid definition feature, which will perform local clustering based on material half value layers, eliminating the need for users to manually construct a weight window grid definition, frequently a trial-and-error process. Once the deterministic calculation is complete, weight windows are automatically mapped to the MCNP weight window grid definition, and source biasing parameters are optionally punched to the MCNP input deck.
Attila4MC-Cottonwood generated weight windows can increase solution speed by as much as 3 to 4 orders of magnitude over analog with minimal user effort. For the tokamak example shown, the figure-of-merit (FOM) for the outside air dose increases about 3 orders of magnitude, from 7.2e-3 to 7.5 for the analog and Attila4MC-cottonwood FW-CADIS cases, respectively. With these efficiency gains, detailed full-field Monte Carlo solutions can be obtained within hours on laptops or desktops.
Figure 2 - MCNP Calculated Neutron Dose Field Using Attila4MC-Cottonwood Generated FW-CADIS Weight Windows. Solution obtained in 35 minutes on a Dell Precision laptop
Reach out to us at sales@silverfirsoftware.com to see how Attila4MC-Cottonwood and Attila4MC can help streamline your calculation workflow, and to receive a no-obligation web demonstration on one of your real-world CAD assemblies.
Funding for the initial development of Attila4MC-Cottonwood was provided by Commonwealth Fusion Systems.
1 J Wagner, A Haghighat, “Automated Variance Reduction of Monte Carlo Shielding Calculations Using the Discrete Ordinates Adjoint Function”,
Nuclear Science and Technology, Vol 128(2), Feb 1998
2 J Wagner, D Pepolow, S Mosher, “FW-CADIS Method for Global and Regional Variance Reduction of Monte Carlo Radiation Transport Calculations”,
Nuclear Science and Technology, Vol 176(1), 2014