A Preliminary Research On The Complex Geometric Modeling And Variance Reduction Techniques Of Monte Carlo Shielding Calculation

Monte Carlo method(MC method),as an essential method for particle transport calculations,comparing with deterministic method,it has advantages in flexible geometry respect and treating physical transportation with higher simulation accuracy.With the increasing demand for the accuracy of radiation field shielding calculation of nuclear facilities,Monte Carlo particle transport program has been more and more widely used in radiation field shielding calculation.The rapid development of computer technology also provides hardware support for the wide application of Monka programs.Based on the autonomic Monka program cosRMC and the practical requirements of Monte-Carlo method in engineering application,the geometric modeling and variance reduction in the calculation of fixed source shielding are studied in this paper.At the beginning,cosRMC is a Monte Carlo program for critical calculation of reactor.Therefore,its geometrie modeling function can be used to describe models that do not contain complex geometric surfaces,such as reactor cores and assemblies.However,there are still some limitations in the geometric modeling function of cosRMC for some complex geometric models in radiation shielding computation.In addition,for typical difficult problems in shielding calculation,such as those with deep penetration characteristics and small probabilistic detection problems,the common features of these problems are that the statistical results caused by the large attenuation of particle flux in the transport process are difficult to converge.It is necessary to use some variance reducing methods currently.The cosRMC program already has the primary cell importance,the weight window and so on to reduce variance,but still has the limitation to some problems which need to solve the global flux distribution.Therefore,this paper mainly focuses on the shortcomings of cosRMC in complex geometric modeling and the use of variance reduction methods in shielding computation.Firstly,the geometric modeling module of cosRMC program is extended to make it capable of describing some high-order surfaces such as torus,paraboloid,hyperboloid and so on.Then,the models are verified from simple to complex,and all the models are compared.The results show that the calculation results of the flux of the geometric model with higher order surfaces are in good agreement with the calculated values of the reference program under the same model and the same settings,thus proving the correctness of the function of the complex geometric modeling in this paper.Secondly,aiming at the limitation of weight window generator function in cosRMC program to solve the typical deep penetration problem in shielding calculation,a method of global variance reduction based on forward transport and flux statistics of Monte Carlo is implemented.A two-dimensional deep penetration calculation model of fixed-source is constructed.The global flux distribution of the whole model is calculated by using the global variance reducing method.The results show that,the global variance reducing method is used to iterate the whole model,and the zero-counting grid element ratio decreases rapidly,which verifies the effectiveness of the global minus variance method.Finally,in this paper,the CFETR three-dimensional model of China Fusion Engineering Experimental reactor is selected firstly,and the fine modeling is carried out by using cosRMC program,and the important neutronic parameters concerned in the design of fusion cladding,including tritium increment ratio,neutron wall load and nuclear thermal deposition were calculated and analyzed.The results show that the calculation results of the cosRMC program are in good agreement with the calculated values of the reference program and can meet the engineering design requirements of CFETR.In addition,the global flux distribution of CFETR one-dimensional model is calculated by using the global variance reducing method developed in this paper.The global flux distribution of the model is obtained by iteration,and the quality factor FOMGVR,which is used to evaluate the calculation efficiency,is increased 2.76 times.The applicability and accuracy of the complex geometric modeling and variance reduction function developed in this paper in practical problems are further proved by the testing and verification of CFETR three-dimensional and one-dimensional models.