Data-driven acceleration of thermal radiation transfer calculations with the dynamic mode decomposition and a sequential singular value decomposition

We present a method for accelerating discrete ordinates radiative transfer calculations transfer. Our works with nonlinear positivity fixes, in contrast to most acceleration schemes. The is based on the dynamic mode decomposition (DMD) and using sequence of rank-one updates compute singular value needed DMD. Using sequential allows us automatically determine number solution vectors include DMD acceleration. results slab geometry standard temperature linearization. Compared positive source iteration, our demonstrate that reduces transport sweeps required solve problem by factor about 3 diffusive Marshak wave problem, several thousand cooling where effective scattering ratio approaches unity, 20 improvement realistic, multimaterial radiating shock problem.