Analysis Of Benchmark Experiments Research Articles

Calculation and experimental analysis of benchmark experiments with a fast neutron spectrum and models of sodium and lead cooled fast reactors using different evaluated nuclear data libraries

The paper presents the results of a comparative analysis of criticality calculations using a Monte-Carlo code with the BNAB-93 and BNAB-RF neutron group constants, as well as with evaluated neutron data files from the Russian ROSFOND evaluated nuclear data library and other evaluated nuclear data libraries (ENDF, JEFF, JENDL) from different years. A set of integral experiments on BFS critical assemblies carried out in different years at the Institute of Physics and Power Engineering (60 different critical configurations) was analyzed. The considered integral experiments are included in the database of evaluated experimental neutronic data used to justify the neutronic performance of sodium and lead cooled fast reactors, to verify codes and nuclear data as well as to estimate uncertainties in neutronic parameters due to the nuclear data uncertainties. It has been shown that the ROSFOND evaluated nuclear data library is a library that minimizes the calculation and experimental discrepancies for the considered set of integral experiments. The paper also presents the results of criticality calculations for models of sodium and lead cooled fast reactors based on different evaluated neutron data libraries and provides estimates for the uncertainty in criticality associated with nuclear data.

Impact of uncertainties in the uranium 235 cross section resonance structure on characteristics measured in the BFS-79 critical assemblies

The report presents the results of an analysis of benchmark experiments form the international ICSBEP Handbook (HEU-MET-INTER-005) carried out at the the SSC RF – IPPE in cooperation with the Idaho National Laboratory (INL, USA) applicable to the verification of calculations of a wide range of tasks related to safe storage of vitrified radioactive waste. Experiments on the BFS assemblies make it possible to perform a large series of studies needed for neutron data refinement, including measurements of reactivity effects which allow testing the neutron cross section resonance structure. This series of studies is considered as a sample joint analysis framework for differential and integral experiments required to correct nuclea data files of the ROSFOND evaluated neutron data library. Thus, it is shown that despite the wide range of available experimental data, in so far as it relates to the resonance region refinement, the experiments on reactivity measurement make it possible to more subtly reflect the resonance structure peculiarities in addition to the time-of-flight measurement method.

(Psycho-)analysis of benchmark experiments: A formal framework for investigating the relationship between data sets and learning algorithms

It is common knowledge that the performance of different learning algorithms depends on certain characteristics of the data—such as dimensionality, linear separability or sample size. However, formally investigating this relationship in an objective and reproducible way is not trivial. A new formal framework for describing the relationship between data set characteristics and the performance of different learning algorithms is proposed. The framework combines the advantages of benchmark experiments with the formal description of data set characteristics by means of statistical and information-theoretic measures and with the recursive partitioning of Bradley–Terry models for comparing the algorithms’ performances. The formal aspects of each component are introduced and illustrated by means of an artificial example. Its real-world usage is demonstrated with an application example consisting of thirteen widely-used data sets and six common learning algorithms. The Appendix provides information on the implementation and the usage of the framework within the R language.

The Design and Analysis of Benchmark Experiments

The assessment of the performance of learners by means of benchmark experiments is an established exercise. In practice, benchmark studies are a tool to compare the performance of several competing algorithms for a certain learning problem. Cross-validation or resampling techniques are commonly used to derive point estimates of the performances which are compared to identify algorithms with good properties. For several benchmarking problems, test procedures taking the variability of those point estimates into account have been suggested. Most of the recently proposed inference procedures are based on special variance estimators for the cross-validated performance. We introduce a theoretical framework for inference problems in benchmark experiments and show that standard statistical test procedures can be used to test for differences in the performances. The theory is based on well-defined distributions of performance measures which can be compared with established tests. To demonstrate the usefulness in practice, the theoretical results are applied to regression and classification benchmark studies based on artificial and real world data.

Cross-section adjustment in the analysis of benchmark experiments

An iteration procedure is given for the solution of the exact non-linear least-squares equation for cross-section adjustment. The solution is made feasible by properly weighting all detectors used in the Benchmark experiment and combining them to one ‘global’ detector. The adjoint fluxes and sensitivity coefficients must be calculated only for this one global detector so that in each iteration cycle the code ANISN and SWANLAKE have to be run only once. Simultaneously an expression was found for the error-covariance matrix of the adjusted cross-section set.