Akademik Veri Yönetim Sistemi
37th AIAA Fluid Dynamics Conference, Miami, FL, Amerika Birleşik Devletleri, 25 - 28 Haziran 2007, cilt.3, ss.2433-2451
The aim of this research program is flow state based feedback control of a circular cylinder wake flow at a Reynolds number of Re=20,000. CFD simulations were performed with an unstructured grid of 4 diameters span and periodic boundary conditions. In addition, wind tunnel experiments were performed to validate the Strouhal number, surface pressures, drag coefficient and velocity profiles in the cylinder wake. The simulation results were compared to both experiments in literature and the experimental results obtained in this study. POD was performed on the surface pressure data and Double Proper Orthogonal Decomposition (DPOD) was performed on the wake velocity field obtained from the CFD results. A low dimensional model was developed based on the DPOD mode amplitudes. For flow field observation, a four sensor configuration of surface mounted pressure sensors was heuristically selected based on the POD spatial Eigenfunctions of the surface pressure. Then an artificial neural network estimator was designed to estimate the mode amplitudes of the 6 dominant DPOD wake velocity modes from the sensor data. For the validation data set, the estimation errors are relatively small and remain bounded. Also, it appears that all 6 DPOD modes are observable at each of the nine spanwise planes that were analyzed.