Discrete cones method in two-dimensional neutron transport computation

The discrete ordinates method (the S/sub N/ method) has been widely used to solve the two-dimensional neutron transport equation for decades. However, the S/sub N/ method cannot be accurate in some situations because of an anomaly, the ray effects due to representing particle directions by a finite number of discrete directions. Such situations arise when a source is localized in rarely scattering materials such as a void or a highly absorbing material. In this work the author proposes a new approximation to eliminate the ray effects. The method is called the discrete cones method and is abbreviated as the DC/sub N/ method. In this method, the unit spherical surface a direction vector covers is partitioned into subsurfaces, cones, and particles in a cone are simultaneously traced instead of only in discrete directions. The method uses the transfer and escape matrices obtained from an integral form of the neutron transport equation, and the equation is numerically solved using cell-by-cell iteration