(GRS-842) Analysis of the fixed-source Option and the weight-window Implementation in Serpent

In-house research Project 4722E03240

In this work, the capabilities of the code Serpent are evaluated, which are relevant for shielding calculations: fixed-source calculation mode, photon transport and application of variance reduction methods.

The fixed-source calculation mode is tested in two different aspects. One is the ability to control population of particle tracks in multiplying media, e. g. to model decay neutrons in an irradiated fuel assembly storage. Several methods available in Serpent are tested on a model that admits adjustment of the neutron multiplication. The efficiency of the methods and their impact on the calculated time-dependent neutron flux behaviour is evaluated. It is shown that the most general method is the dynamic external source simulation mode.

The other aspect of the fixed-source calculation mode is the application of weight-window variance reduction methods to control track populations in space and energy domains. Serpent provides means to apply a user-given weight-window parameter defined on a cylinder or rectangular spatial mesh and on an arbitrary energy grid, as well as means to generate such weight-window meshes globally or for particular responses. The iterative method to generate weight-window meshes is tested on two models that differ by the flux attenuation factor. Some recommendations for the choice of the mesh and the iteration parameters are formulated.

The neutron and photon transport physics modelling in Serpent is validated by comparing against experimental results and correspondent results of MCNP calculations. One of the shielding benchmark problems from the International Criticality Safety Benchmark Evaluation Project (ICSBEP) handbook is calculated with Serpent. The results show that Serpent performs well comparing to MCNP. The effect of some modelling options is investigated in detail.

The performed work shows that Serpent can be applied routinely for shielding calculations as a viable alternative for MCNP.