Research On Multigroup Nuclear Data Adjustment And Accuracy Assessment Method For Fast Spectrum Reactor

With the improvement of calculation models and numerical methods,the inherent uncertainties of nuclear data become the most important uncertainty sources of fast spectrum reactor physics calculations.The design safety margin of fast spectrum reactor can be reduced by the numerical simulation results of high accuracy.Therefore,it is of great significance to improve the uncertainties of nuclear data.The accuracies of calculation results can be improved by the adjusted nuclear data using integral experiment data.An indicative quantitative evaluation of nuclear data improvement requirements can be provided by the performance of the target accuracy assessment based on the design target accuracy.This paper focuses on the methods of nuclear data adjustment and target accuracy assessment.Firstly,based on nuclear data adjustment method,a nuclear data adjustment code named GLAD is developed.The reliability of GLAD is vertified by the comparison of numerical results with reference results.Aiming at the problem nuclear data may appear false changes when there are unkown systematic errors in integral parameters used in the nucelar data adjustment,a nuclear data adjustment method removing systematic errors is proposed.The results show that the potential system errors of integral parameter have a great influence on nuclear data adjustment and the de-systematic error adjustment can effectively avoid unreasonable adjustment of nuclear data.Secondly,the target accuracy assessment problem is researched.The constrained optimization problem solver is developed in GLAD based on the differential evolution algorithm and the feasibility rule.In order to analyze the impact of cross-correlations among isotopes,nuclear reaction and energy range of nuclear data for the target accuracy assessment problem,three different models of target accuracy assessment are established and used in the analysis of the sodium-cooled fast reactor JAEA-FBR.The results show the cross-correlations of nuclear data have a strong impact on the target accuracy assessment problem.Therefore,it is necessary to involve the reasonable cross-correlations of nuclear data in the target accuracy assessment model.Finally,the target accuracy study has been performed for two sodium-cooled reactors,JAEA-FBR and ZPPR-9,including effective multiplication factors and sodium void reactivity coefficients.Meanwhile,the impact of the cost factors on target accuracy assessment problem is investigated.The results show this kind of nuclear data with large cost factors have significant relaxed requirements on uncertainties and the other nuclear data with small cost factors have strict requirements on uncertainties to meet the target parameter accuracy requirements.Therefore,the cost factors characterizing the improvement difficulties of nuclear data uncertainties should be reasonably set based on the practical engineering experience.Through the above research,based on the nuclear data adjustment method and novel method of solving target accuracy assessment problem,a nuclear data adjustment code named GLAD with independent intellectual property rights is developed.The studies of nuclear data adjustment and target accuracy assessment for fast spectrum reactor are carried out,which provide an important basis for the improvement of nuclear data.