In order to predict accurate requirements for design and optimization, coherent structures in the rotor wake need to be better understood. By clarifying characteristics of the coherent structures induced from boundaries, progress in optimization and design can be pursued. Computational costs can be infeasible for such problems with traditional techniques. A different path we are pursuing is development of new coupled methods for fluid-structure interaction, which can simulate the aerodynamics with different numerical schemes designed for near (e.g. compressible, shock-capturing) and away (e.g. incompressible, vorticity-velocity) from the body to expediently capture the wake. The vorticity transport equations with adaptive mesh refinement has been shown to be an efficient framework away from boundaries or in conjunction with some immersed boundary methods for these types of flows. Our schemes are designed to maintain vorticity persistence with relatively few grid cells.
The coupled framework can be efficiently used to investigate aeroelastic effects on wake physics and drive development of data-driven models towards forecasting flow evolution, while near-body schemes designed with efficient immersed boundary methods can be employed alone to capture the compressibility, dynamics and fluid-structure interactions in a variety of flows from turbo-machinery to cardiovascular flows.
Another aspect of this research is development of new algorithms for adaptive mesh refinement. We’ve developed a method based on compressive sampling and sparse sensing to identify the optimal locations in the flow. The algorithm works independent of numerical scheme. These locations tend to be collocated with dominant coherent structures such as wake meandering in a wind turbine wake or Kelvin-Helmholtz vortex street in turbulent jets.
