A 2D/1D Neutron Transport Method with Improved Angular Coupling

computational neutron transport;2D/1D;MPACT;Nuclear Engineering and Radiological Sciences;Engineering;Nuclear Engineering & Radiological Sciences

Jarrett, MichaelStimpson, Shane ;

University of Michigan

关键词: computational neutron transport;    2D/1D;    MPACT;    Nuclear Engineering and Radiological Sciences;    Engineering;    Nuclear Engineering & Radiological Sciences;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/147498/jarremic_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Developing efficient and accurate three-dimensional (3D) neutron transport methods for nuclear reactor applications has long been a major objective for nuclear scientists in the field of reactor physics and radiation transport. Even with the large computers available today, exact 3D neutron transport methods are often too costly to be used for practical core design or safety analysis. Several methods have been developed that use various approximations to the neutron transport equation so that the calculations can be performed on commonly available computing platforms.One such method is the 2D/1D method, which decomposes 3D geometries into several 2D domains wherein 2D transport equations are solved. These 2D transport equations are coupled to one another through transverse, 1D, approximate transport solutions in the axial direction. The 2D/1D method is best suited for problems where the axial gradient of the solution is relatively weak, such as Light Water Reactor (LWR) problems. The 2D/1D method uses an accurate 2D transport solution to resolve the highly heterogeneous radial geometry, and treats the axial dimension with a lower-fidelity, more coarsely discretized solution, which is usually appropriate.Some of the typical assumptions made in many 2D/1D methods can negatively affect the accuracy of the solution in a non-negligible way. Two of the most significant are the isotropic approximations made to the transverse leakage (TL) and homogenized total cross section (XS) used to couple the 2D and 1D equations. In cases where the axial gradients are relatively strong, these assumptions are detrimental to the accuracy. The isotropic TL approximation was corrected in previous work. In this work, the XS is also allowed to be anisotropic. The results show that with both anisotropic TL and XS, the accuracy of 2D/1D is improved significantly.The 2D/1D methods with anisotropic TL and XS are significantly more expensive than the isotropic TL and XS method, which is the standard in the Michigan Parallel Characteristics Transport (MPACT) code. In this work, a 2D/1D method with polar angle parity is developed to significantly reduce the run time of the anisotropic TL and XS method while still significantly improving the accuracy compared to the isotropic TL and XS method.The theoretical accuracy limit of the 2D/1D methods are analyzed and compared to the 3D Simplified P3 (SP3) method. We find that the 2D/1D method with anisotropic TL preserves the 3D SP3 limit with only a few anisotropic TL moments, while the 2D/1D method with isotropic TL does not. As a result, the isotropic TL method is less accurate in problems where there are strong spatial gradients in the radial and axial dimensions.