Discrete adjoint methodology for general multiphysics problems

Abstract This article presents a methodology whereby adjoint solutions for partitioned multiphysics problems can be computed efficiently, in way that is completely independent of the underlying physical sub-problems, associated numerical solution methods, and number type couplings between them. By applying reverse mode algorithmic differentiation to each discipline, by using specialized recording strategy, diagonal cross terms evaluated individually, thereby allowing different methods generic coupled problem (for example block-Jacobi or block-Gauss-Seidel). Based on an implementation open-source simulation design software SU2, we demonstrate how same algorithm applied shape sensitivity analysis heat exchanger (conjugate transfer), deforming wing (fluid–structure interaction), cooled turbine blade where both effects are simultaneously taken into account.