通过模拟CFD离散化方案缓解伴随求解器的鲁棒性问题.pdfVIP

OPT-i An International Conference on Engineering and Applied Sciences Optimization M. Papadrakakis, M.G. Karlaftis, N.D. Lagaros (eds.) Kos Island, Greece, 4-6, June 2014 ALLEVIATING ADJOINT SOLVER ROBUSTNESS ISSUES VIA MIMETIC CFD DISCRETIZATION SCHEMES Mattia Oriani, Guillaume Pierrot Queen Mary, University of London 327 Mile End Road, E1 4NS London, UK e-mail: mattia.oriani@ ESI Group Parc d’Affaires SILIC, 99 Rue des Solets, BP 80112, 94513 Rungis cedex, France e-mail: guillaume.pierrot@ Keywords: Adjoint, Gradient-Based Optimization, Mimetic Finite Differences, Mixed Finite Volumes, Virtual Element Methods, Anisotropic Diffusion, Convection-Diffusion-Reaction Abstract. Adjoint methods allow to compute the gradient of a cost function at an expense that is independent of the number of design variables, thus representing an appealing tool to those who deal with large-scale gradient-based optimization processes. However, researchers in the field are currently facing difficulties related to poor convergence and high storage memory requirements. Both continuous and discrete adjoints require a fully converged solution to the primal prob- lem. It is argued that, specifically in Computational Fluid Dynamics, the presence of numerical ad