摘要: To deal with 3-D heterogeneous reactor problems avoiding direct 3-D transport calculations, 2-D/1-D methods have been developed in recent years. There are two kinds of 2-D/1-D methods: "fusion" and "hybrid." The 2-D/1-D fusion met... 展开 To deal with 3-D heterogeneous reactor problems avoiding direct 3-D transport calculations, 2-D/1-D methods have been developed in recent years. There are two kinds of 2-D/1-D methods: "fusion" and "hybrid." The 2-D/1-D fusion method employs a 2-D MOC calculation in the radial (i.e., x-y) direction and a 1-D S_n calculation in the axial (i.e., z) direction. The 2-D/1-D hybrid method employs a 2-D MOC calculation in the radial direction and 1-D diffusion (or SP_3) approximation in the axial direction. These methods give accurate solutions for several 3-D reactor problems. However, the radial heterogeneity affects the axial leakage and the diffusion approximation would be inaccurate in non-diffusive region of even a simple geometry. To confirm these facts, this paper presents a comparison of the fusion method and the hybrid method via numerical performance on a typical reactor problem. To concentrate and achieve the goal, a 2-D transport problem is considered by reducing one dimension in the radial direction. A test problem is an x-z cut view of the 3-D C5G7 extension benchmark problem. The results in this paper indicate that the diffusion approximation of axial leakage is not accurate for the test problem which has the axially piecewise homogeneous geometry. It causes significant errors in the multiplication factor and the flux distribution of the hybrid method. On the other hand, the fusion method gives very accurate results, because the fusion method does not use any approximations from the transport equation except for the discretization. 收起