State estimation in turbulent channel flow from limited observations

Estimation of the initial state turbulent channel flow from limited data is investigated using an adjoint-variational approach. The are generated a reference direct numerical simulation (DNS) which sub-sampled at different spatiotemporal resolutions. When velocity 1/4096 resolution DNS, correlation coefficient between true and estimated exceeds 99 percent. robustness algorithm to observation noise demonstrated. In addition, impact density on estimation quality evaluated, threshold established for successful reconstruction. critical spanwise proportional Taylor microscale, characterizes domain dependence location. Due mean advection, either streamwise or temporal must satisfy criterion based microscale. A second configuration considered where comprised velocities in outer layer wall shear stresses only. near-wall statistics coherent structures, although not sampled, accurately reconstructed, possible because coupling motions. Finally, most challenging addressed only spatiotemporally resolved observed. remains accurate within viscous sublayer deteriorates significantly with distance wall. units, this trend nearly independent Reynolds number considered, indicative fundamental difficulty reconstructing wall-detached motions data.