Font Size: a A A

Three-Dimensional Hybrid Monte Carlo-Discrete Ordinate Method For Neutron Transport Calculation Based On Energy Region Division

Posted on:2020-08-25Degree:DoctorType:Dissertation
Country:ChinaCandidate:S X ZhuangFull Text:PDF
GTID:1360330572974799Subject:Nuclear science and engineering
Abstract/Summary:PDF Full Text Request
Monte Carlo(MC)and the deterministic methods are conventional methods for neutron transport simualtion.The MC method accurately resolves problems with minimal simplifications and assumptions,and could obtain accurate solution.But it requires long computation time and large memory.The deterministic method approximately solves the Boltzmann equation,which is efficient in engineering application,but the accuracy is poor especially in resonance region.With the development of advanced reactors,demand for simulation with both high efficiency and accuracy has been put forward,invoking the research of hybrid MC-deterministic method.In this work,based on the mature-developed discrete ordinate method(SN),a three-dimensional hybrid MC-Sn method based on energy region division was developed.The deterministic method is applied in the fast and thermal regions for fast solution,while the MC method is applied to simulate neutron transport in the resonance energy region for high accuracy.The mature-developed Sn method was used for its flexible applicability and higher efficiency in three-dimensional problems.And the key issues,including the division of energy regions,the processing of upscattering source,and the source sampling in MC simulation,were studied.The main work and innovations of this research include:(1)A hybrid MC-SN method based on transition energy region was developed to achieve efficient processing of upscattering source.For the processing of upscattering source between resonance and thermal energy regions,the source iteration method can not be applied in the hybrid MC-SN method,as iterated simulations of resonance energy region are required and thus the calculation is poor in efficiency.The calcualtion method based on transition energy region was proposed in this research.The transition energy region,where neutron could be upscattered into resonance energy region,was solved by MC method.Thus the iteration of upscattering source in outer iteration was eliminated and efficient hybrid MC-Sn calculation was achieved.(2)A sampling method of multi-mesh sources for MC simulation based on particle distribution was developed.In hybrid calculation,mesh sources varying largely in probability should be defined in MC simulation.In order to obtain reliable source distribution,large number of particles should be simulated,resulting in the high computational cost in resonance energy region.For this problem,the sampling method of multi-mesh sources based on particles distribution was proposed.Based on source probability,the source particles in each mesh are directly distributed and particle weights are adapted,guaranteeing the reliable source sampling in resonance energy region and ensuring the calculation efficiency and accuracy of hybrid MC-Sn method.Based on analysis of critical parameters,calculation cases,including shielding analysis of heavy concrete plug in ITER lower ports and IAEA-ADS benchmark,were simulated to analyze the calculation efficiency and accuracy of the hybrid method.All the results showed good agreement with the reference results.In the hybrid method,accurate point-wise cross section data were employed in resonance energy region,thus the calculation accuracy can be 3.94 for fixed-source case and 2.08 for criticality case to that of SN calculation.And the fast solutions in fast and thermal region result in higher calculation efficiency and smaller statistical fluctuation,so the calculation efficiency can be 26.59 for fixed-source case and 7.94 for criticality case to that of MC calculation.The reliability of hybrid method has been confirmed,and it could be the technical support for advanced physical design and analysis.
Keywords/Search Tags:Neutron transport, Hybrid MC-S_N method, Monte Carlo, Discrete ordinate
PDF Full Text Request
Related items