SERPENT Guide
Physics Settings
Configuring neutron transport physics, energy cutoffs, thermal scattering treatments, and cross-section processing for accurate simulations
Physics Models
Serpent's physics settings control thermal scattering, resonance absorption, Doppler broadening, and nuclear reaction channels. The defaults work well for most reactor physics applications; configuration uses set cards.
Basic Physics Configuration
% Standard reactor physics settings
set title "PWR Pin Cell Calculation"
% Transport parameters
set bc 1 % Boundary conditions (1=black, 2=reflective, 3=periodic)
set pop 10000 500 100 % neutrons/cycle, active cycles, inactive cycles
set pcc 0 % Disable predictor-corrector calculation for burnupEnergy Cutoffs and Physics Options
Energy cutoffs define the energy range over which particles are tracked. Serpent determines its energy bounds from the loaded ACE cross-section data — there is no explicit user card to set them. The default low-energy bound is approximately 1×10⁻¹¹ MeV (10⁻⁵ eV), which covers the full thermal range for LWR applications. Shielding calculations may tolerate a narrower range since dose-contributing neutrons behind thick shields are typically well above thermal energies.
Neutron Energy Cutoffs
% Energy cutoffs are inherited from the ACE cross-section data.
% The default low-energy cutoff is ~1E-11 MeV (~1E-5 eV),
% which covers the full thermal range for LWR applications.
%
% For time-dependent or kinetics calculations, a time cutoff
% can be applied via the tcut card in the source definition.
%
% No explicit 'set ecut' card exists in Serpent 2; energy
% bounds are determined by the loaded nuclear data libraries.Time cutoffs terminate particles exceeding a specified flight time (relevant for time-dependent calculations). Weight cutoffs control statistical weight through Russian roulette (killing low-weight particles with a survival probability preserving expected weight) and splitting (dividing high-weight particles into multiple copies). These variance reduction techniques improve efficiency in deep-penetration problems where weights span many orders of magnitude.
Weight Cutoffs and Variance Reduction
% Russian roulette and splitting parameters
set roulette 0.001 0.5 % W0=weight cutoff, P=survival probability
% Weight window meshes for deep-penetration problems
% are defined using the wwgen (generation) and wwin (use) cards.
% Example: generate a weight window mesh over a detector region
% wwgen det_name 1 mesh_type NX NY NZ
% XMIN XMAX YMIN YMAX ZMIN ZMAXThermal Neutron Physics
Thermal neutron scattering treatment is critical in light water reactors where moderator temperature strongly influences the thermal spectrum. Serpent provides thermal scattering law data (S(alpha,beta) tables), temperature-dependent cross sections via target motion sampling (TMS), and unresolved resonance probability table sampling. For PWR and BWR calculations, enabling TMS, unresolved resonance treatment, and Doppler broadening is essential for accurate temperature coefficients, spectral indices, and integral reaction rates.
Thermal Scattering Configuration
% Unresolved resonance probability table sampling
set ures 1 % 0=off, 1=on (recommended for reactor calculations)
% Thermal scattering law (S(alpha,beta)) via therm card
% Syntax: therm NAME LIB1 [LIB2]
% If two libraries given, Serpent interpolates between them.
therm lwtr lwj3.11t % Light water at 300 K
therm lwtr lwj3.15t lwj3.17t % Interpolation between 500 K and 600 K
% Link to material with the 'moder' keyword:
mat water -0.7 moder lwtr 1001
1001.06c 2.0
8016.06c 1.0
% Doppler broadening via TMS: assign temperature on the mat card
mat fuel -10.4 tmp 900 % 900 K for Doppler broadening
92235.09c -0.032
92238.09c -0.838
8016.09c -0.130
% Or use TMS with temperature limits for on-the-fly broadening
mat fuel2 -10.4 tmp 900 tft 300 1500
92235.09c -0.032
92238.09c -0.838
8016.09c -0.130Serpent handles both resolved and unresolved resonance regions. Probability table sampling accounts for statistical fluctuations of resonance parameters in the unresolved region. The resolved region uses direct interpolation of pointwise cross-section data with optional Doppler broadening for target nuclear thermal motion.
Resonance Treatment Settings
% Resonance absorption treatment
set ures 1 % Unresolved resonance probability tables
% The unresolved resonance energy range is determined
% automatically from the ACE cross-section libraries.
% Resolved resonance handling uses direct interpolation
% of pointwise data with Doppler broadening via TMS.Advanced Physics Features
Serpent supports coupled neutron-photon transport for shielding analysis, dose rate calculations, and gamma heating. When enabled, it tracks photons from neutron capture, inelastic scattering, and fission, modeling their interactions (Compton scattering, photoelectric absorption, pair production).
Coupled Neutron-Photon Transport
% Enable coupled neutron-photon transport
set ngamma 1 % 1=on, 0=off (neutron-induced photon production)
% Photon interaction data must be available in the
% cross-section library. Serpent models Compton scattering,
% photoelectric absorption, pair production, and Rayleigh
% scattering when photon transport data is loaded.
%
% Photon energy deposition mode
set edepmode 1 % 0=no photon heating, 1=local photon energy deposition
% Photon detectors score gamma flux and dose rates
% using standard det cards with photon-compatible responses.Serpent tracks delayed neutron precursors and samples delayed neutron emission from the appropriate energy spectra, which is essential for reactor kinetics and transient analysis. The effective delayed neutron fraction (beta-effective) is calculated directly from the Monte Carlo simulation, providing a more accurate value than the simple ratio of delayed to total neutron production.
Delayed Neutron Physics
% Delayed neutron modeling
set delnu 1 % Include delayed neutrons (default in Serpent)
% Serpent automatically uses delayed neutron data from the
% cross-section libraries. The effective delayed neutron
% fraction (beta-effective) is reported in the standard
% output for criticality calculations.The cross-section library path is set with set acelib, and OpenMP threads for parallel execution are controlled by the -omp command-line option. For large problems with many materials and nuclides, parallelization and library setup significantly affect performance.
Cross-Section Processing and Parallelization
% Cross-section library path
set acelib "/path/to/sss_endfb7u.xsdata"
% OpenMP threads: use command-line flag -omp N (e.g., sss2 -omp 4 input)Performance Optimization
Serpent's memory requirements depend on geometric complexity, number of materials and nuclides, population size, and tally count. set pcc 0 disables the predictor-corrector burnup calculation (faster, less accurate). set lost sets the maximum particles lost to geometry errors before termination.
Memory Management Settings
% Burnup and convergence settings
set pcc 0 % Disable predictor-corrector calculation for burnup
set lost 100 % Lost particle limit before terminationFor production results, 20,000–50,000 neutrons per cycle with 1,000–2,000 active cycles typically achieves uncertainties below 10 pcm on k-effective and 1–2% on local reaction rates. Monitor Shannon entropy to confirm fission source convergence before active cycles begin. Tightening the lost-particle limit identifies modeling errors early.
Convergence and Statistics
% Statistical convergence criteria
set pop 20000 400 100 % Particles, active cycles, inactive cycles
set entr 20 20 1 % Shannon entropy mesh (20x20x1 spatial bins)
% Optional: tighten lost-particle limit
set lost 50Application-Specific Settings
PWR/BWR reactor physics calculations typically use reflective boundary conditions, TMS Doppler broadening, unresolved resonance probability tables, and a thermal-appropriate energy cutoff. Assembly-level calculations may additionally generate homogenized few-group cross sections for downstream nodal diffusion codes.
Reactor Physics Configuration
% Standard PWR/BWR reactor physics
set title "PWR Core Physics Analysis"
set acelib "/path/to/sss_endfb7u.xsdata"
set bc 2 % Reflective boundaries (core symmetry)
set ures 1 % Unresolved resonance probability tables
set pop 20000 400 100 % Good statistics
set power 3400E6 % Normalize to thermal power (W)
% Group constant generation for nodal codes
set gcu 1 2 3 % Generate group constants for universes 1, 2, 3Shielding calculations require coupled neutron-photon transport with variance reduction. Higher energy cutoffs are often acceptable since dose-contributing radiation behind shielding typically has energies well above thermal. Weight windows, Russian roulette, and splitting are essential for adequate statistics in deep-penetration regions.
Shielding Calculation Configuration
% Shielding analysis
set title "Spent Fuel Cask Shielding"
set acelib "/path/to/sss_endfb7u.xsdata"
set bc 1 % Black (vacuum) boundaries
set pop 50000 500 50 % Higher population for shielding
set ngamma 1 % Enable photon transport for dose
% Variance reduction: Russian roulette and weight windows
set roulette 1E-5 0.5 % Weight cutoff 1E-5, survival prob 0.5
% Generate importance map over the shielding region
% wwgen dose_det 1 1 50 50 50
% XMIN XMAX YMIN YMAX ZMIN ZMAXActivation calculations determine radioactive isotope production rates in structural materials, coolant, and other non-fuel components. They require decay data libraries and appropriate burnup sequence configuration. Delayed neutron tracking is typically unnecessary for pure activation problems.
Activation Calculation Configuration
% Neutron activation analysis
set title "Reactor Vessel Activation"
set acelib "/path/to/sss_endfb7u.xsdata"
set bc 1 % Black boundaries
set pop 10000 200 50 % Population parameters
% Activation-specific settings
set pcc 0 % Disable predictor-corrector calculation
set powdens 40.0 % Power density for normalization (W/gHM)
% Burnup and decay data libraries are specified via
% set declib and set nfylib cards for activation analysis.