Advancing the agent methodology to include the higher order of neutron anisotropy with accelerated solutions

Nader Satvat, Purdue University


With the development of new core designs for generation IV reactors with their complexity and newer fuel designs, the need for consideration of neutron anisotropic scattering is becoming important for enchasing the economy and reliability of these designs. The theory and accurate modeling of neutron anisotropy is one of the most important problems of the transport solution to neutron Boltzmann equation. A number of methods based on careful theoretical developments, were established to numerically determine the effect of anisotropy; some of these methods are: the spherical harmonics method, the so-called function method (FN), the discrete ordinate method, and the Monte Carlo method. The AGENT methodology, based on the method of characteristics, currently the most accurate neutron transport method, represents the state-of-the-art advanced neutronics simulation tool available for 2D, 3D, and full core modeling. The higher order of anisotropic scattering (with no limitation of the number of expansion) is introduced into the AGENT code. An extensive analysis is performed to verify and validate this new model. It is shown that anisotropic scattering is important to be considered for complex geometries due to high angular dependence of neutron flux. The first principle in physics were used to explain the effects of anisotropic scattering (at the level on particle interactions), importance in including the higher moments in flux development for the core designs of high heterogonous structure promoting biased scattering (at the level of heterogeneous reactor assemblies in 2D and 3D). This inclusion of higher order of anisotropic scattering as expected increased the complexity of the mathematical model which in turn increased the computational time. An analysis of the computational time dependence on anisotropic scattering and the method of characteristics resolution parameters are analyzed with accurate predictions of scaling to larger geometries. Finally, an accelerated module was developed for speeding up the solution prediction for anisotropic method of characteristics. The accelerated module has the ability to predict criticality of a heterogeneous system very efficiently and accurately.^




Tatjana Jevremovic, Purdue University, Lefteri H. Tsoukalas, Purdue University.

Subject Area

Engineering, Nuclear

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server