A comparison of specularly reflective boundary conditions and rotationally invariant formulations for Discrete Ordinate Methods in axisymmetric geometries

•Two boundary conditions on wedge geometries using a 3D conventional DOM solver are compared.•The first boundary condition is based on specular reflection.•The second boundary condition enforces rotational invariance.•Consistency and accuracy are demonstrated in one-dimensional axisymmetric radiation.•Accuracy for nongray non-uniform radiation is investigated using an axisymmetric jet flame.

In simulations of periodic or symmetric geometries, computational domains are reduced by imaginary boundaries that exploit the symmetry conditions. Two boundary conditions are proposed for Discrete Ordinate Methods to solve axisymmetric radiation problems. Firstly, a specularly reflective boundary condition similar to that is used in Photon Monte Carlo methods is developed for Discrete Ordinate Methods. Secondly, the rotational invariant formulation is revisited for axisymmetric wedge geometries. Correspondingly, a new rotationally invariant boundary condition specially designed for axisymmetric problems on wedge shape is proposed to enforce the rotational invariance properties possessed by the radiative transfer equation (RTE) but violated by three-dimensional conventional Discrete Ordinate Methods. Both boundary conditions have the advantage that the discretization and linear equation solution procedures of conventional three-dimensional DOM are not affected by changing to a reduced geometry. Consistency, accuracy and efficiency of the new boundary conditions are demonstrated by multiple numerical examples involving periodic symmetry and axisymmetry. A comparison between specularly reflective boundary conditions and the rotationally invariant formulation shows that the latter offers several advantages for wedge geometries. In other symmetry conditions, when the rotational invariant formulation is not applicable, specular reflective boundary conditions are still effective.