ERROR ANALYSIS FOR THE FINITE-ELEMENT APPROXIMATION OF A RADIATIVE-TRANSFER MODEL

This note deals with the numerical solution of a weakly singular integ ral equation arising in radiative or neutron transfer physics. In cont rast to the common discrete ordinates method, the approach considered here is designed for physical applications where the major interest is in computing the scalar flux or only a limited number of flow intensi ties in arbitrary directions rather than the whole radiative field. Ap art from a global convergence result, we also derive interior a priori and a posteriori error estimates particularly tempting for astrophysi cal applications where the computational boundary is mostly artificial .