Deriving Linear and Adjoint Codes for CFD using Automatic Differentiation
نویسندگان
چکیده
Sensitivity analysis is a research area of increasing interest within CFD. Recent developments in Automatic Differentiation software (AD) make the discrete adjoint methods particularly attractive. We outline the theory of adjoint sensitivities and discuss how to use AD to derive adjoint versions from CFD codes. In particular we review the properties of the four major AD packages, we explain how to prepare the code for AD, how to treat Runge-Kutta timestepping, how to validate the derived code and how to optimise the CFD code such as to improve the performance of the derived code
منابع مشابه
Generating efficient derivative code with TAF: Adjoint and tangent linear Euler flow around an airfoil
FastOpt’s new automatic differentiation tool TAF is applied to the two-dimensional Navier-Stokes solver NSC2KE. For a configuration that simulates the Euler flow around a NACA airfoil, TAF has generated the tangent linear and adjoint models as well as the second derivative (Hessian) code. Owing to TAF’s capability of generating efficient adjoints of iterative solvers, the derivative code has a ...
متن کاملA Discrete Adjoint Version of an Unsteady Incompressible Solver for Openfoam Using Algorithmic Differentiation
The comparatively low computational costs of adjoint based gradient methods for optimization problems with a high number of degrees of freedom have allowed them to become one of the most interesting approaches in CFD shape optimization. At the core of such optimization techniques lies the computation of topology sensitivity maps. The two most common approaches for computing adjoint based sensit...
متن کاملAutomatic Differentiation and Sensitivity Analysis Methods for Computational Fluid Dynamics
In the final phase before submission to a journal, we have prepared an internal draft report for a limited distribution. This is not a final document. Updates and corrections are likely to be included prior to publication. In recognition of this, we would appreciate any comments or critics you might have on our research or this document. We welcome any opportunity to improve or clarify any part...
متن کاملBuilding the Tangent and Adjoint codes of the Ocean General Circulation Model OPA with the Automatic Differentiation tool TAPENADE
The ocean general circulation model OPA is developed by the LODYC team at Paris VI university. OPA has recently undergone a major rewriting, migrating to FORTRAN95, and its adjoint code needs to be rebuilt. For earlier versions, the adjoint of OPA was written by hand at a high development cost. We use the Automatic Differentiation tool TAPENADE to build mechanicaly the tangent and adjoint codes...
متن کاملAutomatic Differentiation of Advanced Cfd Codes for Multidisciplinary Design
Automated multidisciplinary design of aircraft and other ight vehicles requires the optimization of complex performance objectives with respect to a number of design parameters and constraints. The e ect of these independent design variables on the system performance criteria can be quanti ed in terms of sensitivity derivatives which must be calculated and propagated by the individual disciplin...
متن کامل