Research On The High-efficiency Even-parity Discrete Ordinates Method Based On Finite Volume Method

The accuracy and efficiency of the physical calculation of the core will greatly affect the safety of reactors during the design and operation phase.With the development of reactors,the non-uniformity of cores is getting stronger and stronger,and accurate core calculation methods based on transport theory is widely concerned.Discrete ordinates method is the most widely used deterministic transport calculation method,but it is faced with computational storage problem and efficiency problem in large-scale engineering calculation.The high efficiency even-parity discrete ordinates method(HEPSN)is a simplified method of evenparity discrete ordinates method,which can effectively save the storage and improve the calculation efficiency.The study of HEPSN method is of engineering significance.The HEPSN equation is a second-order partial differential equation with elliptic properties.The finite volume method is widely used in the second-order equation.In this paper,the finite volume method for elliptic equations is introduced as the spatial variable processing method of HEPSN.The three-dimensional finite volume method HEPSN based on isotropic scattering is deduced in detail,and the processing of its boundary condition is described in detail.The finite volume method has a clear physical meaning,and its introduction can improve the computational accuracy of HEPSN.On the other hand,in order to further improve the computational efficiency of HEPSN,this paper studies the parallel strategy of HEPSN on the basis of serial HEPSN.The parallel method based on angle decomposition and spatial decomposition make HEPSN more suitable for large-scale core calculation.This paper improves the engineering application ability of HEPSN in this two aspects.In this paper,the critical calculation capability of HEPSN is tested by the Takeda benchmarks,which contain a small fast breeder reactor core model,an axial inhomogeneous fast neutron breeder reactor core model and a small light water reactor core model.The numerical results of the Keff and the regional average neutron flux show that the accuracy of the HEPSN method is between the first-order SN method and the diffusion theory.HEPSN significantly improve the computational efficiency compared with the traditional first-order SN method.For the larger core model,the results of finite volume method is closer to the reference value than the that of finite difference method.The Keff deviation is reduced to less than 270 pcm with the finite volume method.The finite volume method has a significant improvement in the accuracy of HEPSN.HEPSN does not suitable for high-leakage models.The test of parallel method shows that the parallel method based on the angle region decomposition can effectively improve the computational efficiency of HEPSN.In this paper,the HEPSN method is improved in the calculation accuracy and computational efficiency.