![]() |
![]() |
------------------------- Starting with Fluka2021.2 neutron transport can be done also using continuous (pointwise) cross sections. For 1H, 2H, 3He, 4He, and 12C fully correlated pointwise cross sections are built-in inside the code, and are automatically processed at the temperature(s) required by the problem. For many other isotopes fully correlated pointwise cross sections are available in external files (directory pwxs) generated out of evaluated data files. In all cases, all secondaries, including charged particles and recoils are explicitly generated. Binding effect (eg H in water, H in polyethylene, C in graphite...) are not yet implemented, all isotopes are treated as free gas in the thermal region. However by default, for those isotopes for which bound cross sections exist and are requested in the multi-group library, the pointwise treatment will be stopped at 3.059023 eV and the bound groupwise treatment used below that energy. This default behaviour can be overriden setting what(6) in LOW-NEUT, however this is strongly discouraged. The easiest way to activate pointwise cross sections is to issue a LOW-PWXS card like: All isotopes for all materials at 296 K, leaving FLUKA to decide which evaluations to pick up for each of them:*...+....1....+....2....+....3....+....4....+....5....+....6....+....7....+...LOW-PWXS 1.0 3.0 @LASTMAT The code will automatically split each element into the constituent isotopes and pick up the corresponding pointwise cross section for each isotope. Please look into the output file if you want to know the exact details about how the association isotope to pointwise cross section set is performed. The CPU penalty for using pointwise cross sections can be significant. The number of energy-cross section pairs can be very large (eg 96033 points for 56Fe total cross section, ENDF/B-VIIIR0, at 296 K), and the generation of secondaries much more complex with respect to multi-group cross sections. On top of these considerations, the transport of the generated charged products can in itself take up a considerable computer time.