تجزیه و تحلیل حساسیت برای یک معیار نوترون 14 MeV با استفاده از روشهای مونت کارلو محاسباتی قطعی

A sensitivity analysis has been performed for a 14 MeV neutron benchmark on an iron assembly, typical for a fusion neutronic integral experiment. Probabilistic and deterministic computational methods have been used in the sensitivity calculations with the main objective to check and validate the novel Monte Carlo technique for calculating point detector sensitivities. Good agreement has been achieved between the Monte Carlo and the deterministic approaches for the individual calculated sensitivity profiles, the uncertainties and the neutron flux spectra. It is thus concluded that the Monte Carlo technique for calculating point detector sensitivities and related uncertainties as being implemented in MCSEN, a local version of the MCNP4A code with the capability to calculate point detector sensitivities, is well qualified for sensitivity and uncertainty analyses of integral experiments.

Sensitivity and uncertainty analysis is a powerful means to assess uncertainties of nuclear responses in neutron transport calculations and track down these uncertainties to specific nuclides, reaction cross-sections and energy ranges. When applied to the analysis of integral experiments, it enables the assessment of the calculational accuracy and provides information for improving the cross-section data evaluations. The differential operator method developed originally by Hall [1] is a suitable method to calculate sensitivities with the Monte Carlo technique. Based on this method, an algorithm has been developed at the Hebrew University of Jerusalem [2] for calculating point detector sensitivities which has been implemented into a local update to the Los Alamos Monte Carlo code MCNP4A [3]. The point detector method is best suited for analysing integral benchmark experiments when leakage spectra are measured by detectors that are located far from the irradiated material assembly. Experiments of this kind are in frequent use for testing and validating fusion nuclear data evaluations, as well as for assessing their uncertainty margins. In these experiments, material assemblies are irradiated with 14 MeV (d, t) neutrons and the neutron leakage flux spectrum is measured by means of experimental techniques such as the time-of-flight (TOF) method. This work is devoted to the sensitivity analysis of a 14 MeV neutron benchmark on an iron assembly typical for a fusion neutronic integral experiment. Probabilistic and deterministic computational methods have been used in the sensitivity calculations with the main objective to check and validate the novel Monte Carlo technique for calculating point detector sensitivities. There follows a short description of the benchmark problem, the computational procedures for both the deterministic and the Monte Carlo calculation of point detector sensitivities as well as a detailed presentation and discussion of the sensitivity results obtained in this analysis.

A detailed sensitivity analysis has been performed for a 14 MeV neutron benchmark on an iron assembly, typical for a fusion neutronic integral experiment. Probabilistic and deterministic computational methods have been used in the sensitivity calculations with the main objective to check and validate the novel Monte Carlo technique for calculating point detector sensitivities. Good agreement has been achieved between the Monte Carlo and the deterministic approaches. This applies for the individual sensitivity profiles and uncertainties and the neutron flux spectra as well. In addition, the calculated sensitivities were shown to be consistent with the results of direct transport calculations using cross-sections that were changed by a specified percentage for selected reactions in selected energy ranges. It is thus concluded that the Monte Carlo technique for calculating point detector sensitivities and related uncertainties as implemented in MCSEN, our local version of MCNP4A, is well qualified for sensitivity and uncertainty analyses of integral experiments. In addition to the well established deterministic approaches, an independent and complementary method is thus available.