Control and Estimation of Turbulent Shear Flows using Modal Analysis

Modal analysis methods can serve as powerful tools in the development of reduced-complexity models for fluid flows. This talk presents two applications of modal analysis involving control and estimation of turbulent shear flows. First, we consider the development of passive control techniques for wall-bounded flows. We show that the drag reducing effect of sharkskin-inspired riblet surfaces is reproduced by a limited number of high-gain modes obtained via resolvent analysis. This reduced representation enables optimization of riblet geometry and provides insight into the development of permeable materials for turbulence control. Next, we attempt parameter estimation and flow reconstruction for stratified wakes from limited and noisy measurements.  We create a library of flow features from prior simulation data using dynamic mode decomposition (DMD). We then use a stepwise regression technique to sequentially identify DMD modes that best represent the measurements and calibrate their amplitudes. The resulting sparse model enables flow reconstruction as well as estimation of Reynolds and Froude numbers for the object that created the wake. We demonstrate successful estimation of wake properties using this approach from both point and field measurements of velocity.