SCONE Simulation Settings

These parameters control accuracy, cost, and what gets written to disk. Start conservative (small pop and few active cycles), prove geometry and materials, then scale up.

SCONE supports two primary simulation modes. Eigenvalue (k-eigenvalue) calculations determine multiplication factors and flux distributions in fissile systems, commonly used in reactor analysis. Fixed-source calculations model systems with known neutron sources, ideal for shielding or radiation transport problems.

Matches tutorial_config_testlib lines 5–15 — tutorial-sized pop/active/inactive; increase for production.

type eigenPhysicsPackage;
pop      10000;
active   50;
inactive 20;
XSdata   ce;
dataType ce;

collisionOperator { neutronCE { type neutronCEstd; } }
transportOperator { type transportOperatorDT; }

inactiveTally {}

Use type eigenPhysicsPackage; for criticality calculations or type fixedSourcePhysicsPackage; for fixed-source runs. The physics package type must be declared at the top of the input.

From tutorial_shielding_slab lines 6–13 (verified fixed-source header + operators).

type fixedSourcePhysicsPackage;
pop      10000;
cycles   100;
XSdata   ce;
dataType ce;

collisionOperator { neutronCE { type neutronCEstd; } }
transportOperator { type transportOperatorDT; }

Fixed-source keyword names can differ by SCONE version — confirm cycles / batch controls in the Input Manual.

The accuracy and efficiency of your simulation depend heavily on core control parameters. For eigenvalue calculations, pop sets particles per cycle, active sets the number of cycles used for statistics, and inactive sets cycles for source convergence before tallying begins.

pop      100000;   // particles per cycle (higher = better statistics)
active   300;     // active cycles (more = smaller k-eff uncertainty)
inactive 200;    // inactive cycles (more = better source convergence)

// For eigenvalue: total cycles = inactive + active
// Statistics improve with sqrt(pop * active)

For eigenvalue calculations, additional parameters control source convergence and iteration behavior. Tune pop, active, and inactive first; then adjust tally-map resolution as needed.

activeTally with 2D map — excerpt from tutorial_examples_hub_testlib lines 19–31 (pin-cell grid; full-core models use wider min/max and larger N).

activeTally {
  fissionRate {
    type collisionClerk;
    response (fission);
    fission { type macroResponse; MT -6; }
    map {
      type multiMap;
      maps (xax yax);
      xax { type spaceMap; axis x; grid lin; N 64; min -0.63; max 0.63; }
      yax { type spaceMap; axis y; grid lin; N 64; min -0.63; max 0.63; }
    }
  }
}

SCONE offers flexible output options. The outputFormat directive controls how tally results are written. Use asciiMATLAB (default) for MATLAB-compatible output or asciiJSON for structured JSON output.

// Output format: asciiMATLAB (default) or asciiJSON
outputFormat asciiJSON;

// Default is asciiMATLAB if outputFormat is omitted

The viz block configures visualization output. You can define bitmap images of the geometry (e.g., material map, flux) at specified planes. Each viz entry specifies the output file, view centre, width, axis, and resolution.

viz {
  bmpZ {
    type bmp;
    output imgZ;
    what material;
    centre (0.0 0.0 422.5);
    width (500.0 500.0);
    axis z;
    res (5000 5000);
  }
}

When configuring SCONE, start with conservative settings during testing. Begin with smaller particle populations and fewer cycles to quickly identify input errors. For source convergence, monitor cycle-wise k and any entropy or diagnostic tallies you add (the Input Manual defines a shannonEntropyClerk if you choose to use it). Full-core models usually need more inactive cycles than pin cells.

Use fixed random seeds during development to ensure reproducible results. Consider your output needs carefully. Remember that simulation settings involve tradeoffs between accuracy, performance, and resource usage.