Unsteady Continuous Adjoint Method Using Pod For Jet-Based Flow Control
An approximation method based on Proper Orthogonal Decomposition (POD) is used for the storage of the primal flow fields, needed during the solution of the unsteady adjoint equations in aerodynamic optimization problems. Here, without loss in generality, the presentation is restricted to flow control optimization; its extension to aerodynamic shape optimization is straightforward. In this paper, the use of the POD, as an alternative to full storage or the check-pointing technique to support
... ique to support unsteady adjoint methods, is demonstrated. POD approximates the time-evolution of the flow variables at each grid node, instead of repetitively re-computing them during the solution of the unsteady adjoint equations, while marching backwards in time. The solutions obtained with the POD method are compared to those reached by any accurate method (as such the check-pointing method is used), in terms of accuracy and overall simulation time. A parametric investigation of the POD implementation is carried out.Based on the approximately reconstructed primal fields for flow problem around a cylinder, a flow control optimization, using pulsating jets, is performed using the unsteady continuous adjoint method. The jet positions are fixed whereas their amplitudes are optimized aiming at minimal time-averaged drag.